1 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
3 * config/arc/simdext.md (UNSPEC_ARC_SIMD_VLD32WH): Delete.
4 (UNSPEC_ARC_SIMD_VLD32WL): Likewise.
5 (vld32wh_insn, vld32wl_insn): Delete commented-out old
6 versions of these patterns.
8 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
10 * config/arc/arc.c (arc_conditional_register_usage):
11 Use ARC_FIRST_SIMD_VR_REG / ARC_LAST_SIMD_VR_REG.
12 Also set reg_alloc_order for DMA config regs.
14 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
15 Jeremy Bennett <jeremy.bennett@embecosm.com>
17 * doc/install.texi (--with-cpu): Mention ARC.
18 (arc-*-elf32): New paragraph.
19 (arc-linux-uclibc): Likewise.
20 * doc/md.texi (Machine Constraints): Add ARC part.
21 * doc/invoke.texi: (menu): Add ARC Options.
22 (Machine Dependent Options) <ARC Options>: Add synopsis.
23 (node ARC Options): Add.
24 * doc/extend.texi (long_call / short_call attribute): Add ARC.
25 (ARC Built-in Functions): New section defining
26 generic ARC built-in functions.
27 (ARC SIMD Built-in Functions): New section defining SIMD specific
29 (Declaring Attributes of Functions): Extended
30 description of short_call and long_call attributes for ARC and
33 2013-10-01 Saurabh Verma <saurabh.verma@codito.com>
34 Ramana Radhakrishnan <ramana.radhakrishnan@codito.com>
35 Joern Rennecke <joern.rennecke@embecosm.com>
36 Muhammad Khurram Riaz <khurram.riaz@arc.com>
37 Brendan Kehoe <brendan@zen.org>
38 Michael Eager <eager@eagercon.com>
39 Simon Cook <simon.cook@embecosm.com>
40 Jeremy Bennett <jeremy.bennett@embecosm.com>
42 * config/arc, common/config/arc: New directories.
44 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
45 Brendan Kehoe <brendan@zen.org>
46 Simon Cook <simon.cook@embecosm.com>
48 * config.gcc (arc*-*-elf*, arc*-*-linux-uclibc*): New configurations.
50 2013-10-01 Andrew MacLeod <amacleod@redhat.com>
52 * tree-ssa-live.h (coalesce_ssa_name): Move Prototype to...
53 * tree-ssa-coalesce.h: New. Move prototype to here.
54 * tree-outof-ssa.h: Include tree-ssa-coalesce.h.
55 * tree-ssa-coalesce.c: Include tree-outof-ssa.h.
56 (gimple_can_coalesce_p): Move to...
57 * gimple.c (gimple_can_coalesce_p): Here.
59 2013-10-01 Andrew MacLeod <amacleod@redhat.com>
61 * tree-into-ssa.c (enum need_phi_state): Relocate from tree-flow.h.
62 (dump_decl_set): Move to gimple.c.
63 * gimple.h: Don't include tree-ssa-operands.h.
64 (dump_decl_set): Add prototype.
65 (gimple_vuse_op, gimple_vdef_op, update_stmt, update_stmt_if_modified):
67 (phi_ssa_name_p, phi_nodes, phi_nodes_ptr, gimple_phi_arg_def,
68 gimple_phi_arg_def_ptr, gimple_phi_arg_edge, gimple_phi_arg_location,
69 gimple_phi_arg_location_from_edge, gimple_phi_arg_set_location,
70 gimple_phi_arg_has_location): Relocate from tree-flow-inline.h
71 * gimple.c (walk_stmt_load_store_ops): Use gimple_phi_arg_def rather
73 (dump_decl_set): Relocate here.
74 * gimple-ssa.h: New file.
75 (gimple_vuse_op, gimple_vdef_op, update_stmt, update_stmt_if_modified):
76 Relocate from gimple.h.
77 * tree-cfg.c (has_zero_uses_1, single_imm_use_1): Move to...
78 * tree-ssa-operands.c (swap_ssa_operands): Rename from
79 swap_tree_operands and remove non-ssa path.
80 (has_zero_uses_1, single_imm_use_1): Relocate from tree-cfg.c.
81 * tree-ssa-reassoc.c (linearize_expr_tree, repropagate_negates): Use
83 * tree-vect-loop.c (destroy_loop_vec_info, vect_is_slp_reduction,
84 vect_is_simple_reduction_1): Use swap_ssa_operands.
85 * tree-flow.h: Move various prototypes to tree-phinodes.h.
86 (enum need_phi_state): Move to tree-into-ssa.c.
87 (struct immediate_use_iterator_d, FOR_EACH_IMM_*,
88 BREAK_FROM_IMM_USE_STMT): Move to ssa-iterators.h.
89 (swap_tree_operands): Rename and move prototype to tree-ssa-operands.h.
90 * tree-flow-inline.h (delink_imm_use, link_imm_use_to_list,
91 link_imm_use, set_ssa_use_from_ptr, link_imm_use_stmt, relink_imm_use,
92 relink_imm_use_stmt, end_readonly_imm_use_p, first_readonly_imm_use,
93 next_readonly_imm_use, has_zero_uses, has_single_use, single_imm_use,
94 num_imm_uses): Move to ssa-iterators.h.
95 (get_use_from_ptr, get_def_from_ptr): Move to tree-ssa-operands.h
96 (gimple_phi_arg_imm_use_ptr, phi_arg_index_from_use): Move to
98 (op_iter_done, op_iter_next_def, op_iter_next_tree,
99 clear_and_done_ssa_iter, op_iter_init, op_iter_init_use,
100 op_iter_init_def, op_iter_init_tree, single_ssa_tree_operand,
101 single_ssa_use_operand, single_ssa_def_operand, zero_ssa_operands,
102 num_ssa_operands, delink_stmt_imm_use, single_phi_def,
103 op_iter_init_phiuse, op_iter_init_phidef, end_imm_use_stmt_p,
104 end_imm_use_stmt_traverse, move_use_after_head, link_use_stmts_after,
105 first_imm_use_stmt, next_imm_use_stmt, first_imm_use_on_stmt,
106 end_imm_use_on_stmt_p, next_imm_use_on_stmt): Move to ssa-iterators.h.
107 (gimple_phi_arg_def, gimple_phi_arg_def_ptr, gimple_phi_arg_edge,
108 gimple_phi_arg_location, gimple_phi_arg_location_from_edge,
109 gimple_phi_arg_set_location, gimple_phi_arg_has_location, phi_nodes,
110 phi_nodes_ptr, phi_ssa_name_p): Move to gimple.h.
111 (set_phi_nodes): Move to tree-phinodes.h.
112 * tree-ssa-operands.h (enum ssa_op_iter_type,
113 struct ssa_operand_iterator_d, SSA_OP*, FOR_EACH_SSA*, SINGLE_SSA*,
114 ZERO_SSA_OPERANDS, NUM_SSA_OPERANDS): Move to ssa-iterators.h.
115 (dump_decl_set): Remove prototype.
116 (get_use_from_ptr, get_def_from_ptr): Relocate from tree-flow.h.
117 * tree-phinodes.h: New file. Move some prototypes from tree-flow.h.
118 (set_phi_nodes): Relocate from tree-flow-inline.h.
119 (gimple_phi_arg_imm_use_ptr, phi_arg_index_from_use): Relocate from
121 * tree-ssa.h: Add tree-phinodes.h, gimple-ssa.h, ssa-iterators.h to
122 include list. Temporarily add gimple.h to include list.
123 * ssa-iterators.h: New file.
124 (struct immediate_use_iterator_d, FOR_EACH_IMM_*,
125 BREAK_FROM_IMM_USE_STMT): Relocate from tree-flow.h.
126 (enum ssa_op_iter_type, struct ssa_operand_iterator_d, SSA_OP*,
127 FOR_EACH_SSA*, SINGLE_SSA*, ZERO_SSA_OPERANDS, NUM_SSA_OPERANDS):
128 Relocate from tree-ssa-operands.h.
129 (delink_imm_use, link_imm_use_to_list, link_imm_use,
130 set_ssa_use_from_ptr, link_imm_use_stmt, relink_imm_use,
131 relink_imm_use_stmt, end_readonly_imm_use_p, first_readonly_imm_use,
132 next_readonly_imm_use, has_zero_uses, has_single_use, single_imm_use,
133 num_imm_uses, get_use_from_ptr, get_def_from_ptr,
134 phi_arg_index_from_use, op_iter_done, op_iter_next_def,
135 op_iter_next_tree, clear_and_done_ssa_iter, op_iter_init,
136 op_iter_init_use, op_iter_init_def, op_iter_init_tree,
137 single_ssa_tree_operand, single_ssa_use_operand, single_ssa_def_operand,
138 zero_ssa_operands, num_ssa_operands, delink_stmt_imm_use,
139 single_phi_def, op_iter_init_phiuse, op_iter_init_phidef,
140 end_imm_use_stmt_p, end_imm_use_stmt_traverse, move_use_after_head,
141 link_use_stmts_after, first_imm_use_stmt, next_imm_use_stmt,
142 first_imm_use_on_stmt, end_imm_use_on_stmt_p, next_imm_use_on_stmt):
143 Relocate from tree-flow-inline.h.
144 * tree-outof-ssa.h: Change _SSAEXPAND_H macro to GCC_TREE_OUTOF_SSA_H.
146 2013-10-01 Vidya Praveen <vidyapraveen@arm.com>
149 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l2<mode>_internal): Rename to ...
150 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal): ... this;
151 Insert '\t' to output template.
152 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_lo_internal): New.
153 (aarch64_saddl2<mode>, aarch64_uaddl2<mode>): Modify to call
154 gen_aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal() instead.
155 (aarch64_ssubl2<mode>, aarch64_usubl2<mode>): Ditto.
157 2013-10-01 Uros Bizjak <ubizjak@gmail.com>
159 * doc/install.texi (Host/target specific installation notes for GCC):
160 Put @anchor before @heading.
161 * doc/gcc.texi (titlepage): Use @uref and http:// prefix for website.
162 Use @email for email addresses.
164 2013-10-01 Jeff Law <law@redhat.com>
166 * tree-ssa-threadedge.c (thread_across_edge): Make path a pointer to
167 a vec. Only delete the path if we create one without successfully
168 registering a jump thread.
169 * tree-ssa-threadupdate.h (register_jump_thread): Pass in path vector
171 * tree-ssa-threadupdate.c (threaded_edges): Remove. No longer used
172 (paths): New vector of jump threading paths.
173 (THREAD_TARGET, THREAD_TARGET2): Remove accessor macros.
174 (THREAD_PATH): New accessor macro for the entire thread path.
175 (lookup_redirection_data): Get intermediate and final outgoing edge
176 from the thread path.
177 (create_edge_and_update_destination_phis): Copy the threading path.
178 (ssa_fix_duplicate_block_edges): Get edges and block types from the
180 (ssa_redirect_edges): Get edges and block types from the jump threading
181 path. Free the path vector.
182 (thread_block): Get edges from the jump threading path. Look at the
183 entire path to see if we thread to a loop exit. If we cancel a jump
184 thread request, then free the path vector.
185 (thread_single_edge): Get edges and block types from the jump threading
186 path. Free the path vector.
187 (thread_through_loop_header): Get edges and block types from the jump
188 threading path. Free the path vector.
189 (mark_threaded_blocks): Iterate over the vector of paths and store
190 the path on the appropriate edge. Get edges and block types from the
192 (mark_threaded_blocks): Get edges and block types from the jump
193 threading path. Free the path vector.
194 (thread_through_all_blocks): Use the vector of paths rather than
195 a vector of 3-edge sets.
196 (register_jump_thread): Accept pointer to a path vector rather
197 than the path vector itself. Store the path vector for later use.
200 2013-10-01 Jakub Jelinek <jakub@redhat.com>
201 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
204 * config/s390/s390.c (s390_split_branches): Modify check for table
206 (s390_chunkify_start): Rearrange table jump insn check in order to
207 deal with compare and branch insns correctly.
209 2013-10-01 Kugan Vivekanandarajah <kuganv@linaro.org>
213 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
214 * config/arm/arm.md (arm_ashldi3_1bit): define_insn into
215 define_insn_and_split.
216 (arm_ashrdi3_1bit,arm_lshrdi3_1bit): Likewise.
217 (shiftsi3_compare): New pattern.
219 * config/arm/unspecs.md (UNSPEC_RRX): New.
221 2013-10-01 Alan Modra <amodra@gmail.com>
223 * stmt.c (expand_asm_operands): Revert part of 2013-09-24 special
224 casing inout operands.
226 2013-10-01 Richard Biener <rguenther@suse.de>
228 PR tree-optimization/58553
229 * tree-loop-distribution.c (struct partition_s): Add niter member.
230 (classify_partition): Populate niter member for the partition
231 and properly identify whether the relevant store happens before
232 or after the loop exit.
233 (generate_memset_builtin): Use niter member from the partition.
234 (generate_memcpy_builtin): Likewise.
236 2013-09-30 Richard Sandiford <rdsandiford@googlemail.com>
238 * vec.h (vec_prefix, vec): Prefix member names with "m_".
239 * vec.c (vec_prefix::calculate_allocation): Update accordingly.
241 2013-09-30 Richard Sandiford <rdsandiford@googlemail.com>
243 * basic-block.h (edge_list): Prefix member names with "m_".
244 * context.h (context): Likewise.
245 * domwalk.h (dom_walker): Likewise.
246 * gengtype-state.c (s_expr_writer, state_writer): Likewise.
247 * graphite-sese-to-poly.c (sese_dom_walker): Likewise.
248 * hash-table.h (hash_table): Likewise.
249 * machmode.h (bit_field_mode_iterator): Likewise.
250 * pass_manager.h (pass_list): Likewise.
251 * tree-into-ssa.c (mark_def_dom_walker): Likewise.
252 * tree-pass.h (pass_data): Likewise.
253 * tree-ssa-dom.c (dom_opt_dom_walker): Likewise.
254 * tree-ssa-phiopt.c (nontrapping_dom_walker): Likewise,
255 * tree-ssa-uncprop.c (uncprop_dom_walker): Likewise.
256 * asan.c (pass_data_asan): Update accordingly.
257 * cfganal.c (control_dependences::find_control_dependence): Likewise.
258 (control_dependences::control_dependences): Likewise.
259 (control_dependences::~control_dependences): Likewise.
260 (control_dependences::~control_dependences): Likewise.
261 (control_dependences::get_edges_dependent_on): Likewise.
262 * cgraphbuild.c (pass_data_rebuild_cgraph_edges::clone): Likewise.
263 (pass_data_remove_cgraph_callee_edges::clone): Likewise.
264 * context.c (gcc::context::context): Likewise.
265 * cprop.c (pass_rtl_cprop::clone): Likewise.
266 * domwalk.c (dom_walker::walk): Likewise.
267 * ipa-inline-analysis.c (pass_inline_parameters::clone): Likewise.
268 * ipa-pure-const.c (pass_local_pure_const::clone): Likewise.
269 * mode-switching.c (pass_mode_switching::clone): Likewise.
270 * passes.c (opt_pass::opt_pass): Likewise.
271 (pass_manager::pass_manager): Likewise.
272 * predict.c (pass_strip_predict_hints::clone): Likewise.
273 * recog.c (pass_data pass_data_peephole2::clone): Likewise.
274 (pass_split_all_insns::clone): Likewise.
275 * stor-layout.c (bit_field_mode_iterator::bit_field_mode_iterator):
277 (bit_field_mode_iterator::next_mode): Likewise.
278 (bit_field_mode_iterator::prefer_smaller_modes): Likewise.
279 * tree-cfg.c (pass_split_crit_edges::clone): Likewise.
280 * tree-cfgcleanup.c (pass_merge_phi::clone): Likewise.
281 * tree-complex.c (pass_lower_complex::clone): Likewise.
282 * tree-eh.c (pass_cleanup_eh::clone): Likewise.
283 * tree-object-size.c (pass_object_sizes::clone): Likewise.
284 * tree-optimize.c (pass_fixup_cfg::clone): Likewise.
285 * tree-ssa-ccp.c (pass_data_ccp::clone): Likewise.
286 (pass_fold_builtins::clone): Likewise.
287 * tree-ssa-copy.c (pass_data_copy_prop::clone): Likewise.
288 * tree-ssa-copyrename.c (pass_rename_ssa_copies::clone): Likewise.
289 * tree-ssa-dce.c (pass_dce::clone, pass_dce_loop::clone): Likewise.
290 (pass_cd_dce::clone): Likewise.
291 * tree-ssa-dom.c (pass_dominator::clone): Likewise.
292 (pass_phi_only_cprop::clone): Likewise.
293 * tree-ssa-dse.c (pass_dse::clone): Likewise.
294 * tree-ssa-forwprop.c (pass_forwprop::clone): Likewise.
295 * tree-ssa-loop.c (pass_lim::clone): Likewise.
296 * tree-ssa-phiopt.c (pass_phiopt::clone): Likewise.
297 * tree-ssa-pre.c (pass_fre::clone): Likewise.
298 * tree-ssa-reassoc.c (pass_reassoc::clone): Likewise.
299 * tree-ssa-uninit.c (pass_late_warn_uninitialized::clone): Likewise.
300 * tree-tailcall.c (pass_tail_recursion::clone): Likewise.
301 * tree-vect-generic.c (pass_lower_vector_ssa::clone): Likewise.
302 * tree-vrp.c (pass_vrp::clone): Likewise.
303 * tsan.c (pass_tsan::clone): Likewise.
305 2013-09-30 Jakub Jelinek <jakub@redhat.com>
308 * fold-const.c (tree_unary_nonnegative_warnv_p): Use
309 INTEGRAL_TYPE_P (t) instead of TREE_CODE (t) == INTEGER_TYPE.
312 * fold-const.c (fold_ternary_loc): For A < 0 : <sign bit of A> : 0
313 optimization, punt if sign_bit_p looked through any zero extension.
315 2013-09-30 Teresa Johnson <tejohnson@google.com>
317 * tree-ssa-threadupdate.c (ssa_fix_duplicate_block_edges):
318 Update redirected out edge count in joiner case.
319 (ssa_redirect_edges): Common the joiner and non-joiner cases
320 so that joiner case gets profile updates.
322 2013-09-30 Richard Biener <rguenther@suse.de>
324 PR tree-optimization/58554
325 * tree-loop-distribution.c (classify_partition): Require
326 unconditionally executed stores for memcpy and memset recognition.
327 (tree_loop_distribution): Calculate dominance info.
329 2013-09-30 Venkataramanan Kumar <venkataramanan.kumar@linaro.org>
331 * config/aarch64/aarch64.h (MCOUNT_NAME): Define.
332 (NO_PROFILE_COUNTERS): Likewise.
333 (PROFILE_HOOK): Likewise.
334 (FUNCTION_PROFILER): Likewise.
335 * config/aarch64/aarch64.c (aarch64_function_profiler): Remove.
337 2013-09-30 Iain Sandoe <iain@codesourcery.com>
339 * config/rs6000/darwin.md (load_macho_picbase_si): Wrap machopic
340 calls and defines in TARGET_MACHO conditional.
341 (load_macho_picbase_di): Likewise.
342 (reload_macho_picbase): Likewise.
343 (reload_macho_picbase_si): Likewise.
344 (reload_macho_picbase_di): Likewise.
345 (nonlocal_goto_receiver): Likewise.
347 2013-09-30 Nick Clifton <nickc@redhat.com>
349 * config/msp430/msp430.c (msp430x_names): New array. Lists MCUs
350 that use the MSP430X ISA.
351 (msp430_option_override): Scan -mmcu command line option for any
352 MCU name that supports the MSP430X ISA.
353 * config/msp430/t-msp430 (MULTILIB_MATCHES): Add matches for known
354 -mmcu options which enable the MSP430X ISA.
356 2013-09-30 Richard Biener <rguenther@suse.de>
359 * tree-cfg.c (make_abnormal_goto_edges): Skip debug statements
360 before looking for setjmp-like calls.
362 2013-09-29 Iain Sandoe <iain@codesourcery.com>
365 * config/darwin-protos.h (machopic_get_function_picbase): New.
366 * config/darwin.c (machopic_get_function_picbase): New.
367 * config/rs6000/darwin.md (load_macho_picbase_si): Update picbase
368 label for a new func. (load_macho_picbase_di): Likewise.
369 (reload_macho_picbase): New expand.
370 (reload_macho_picbase_si): New insn.
371 (reload_macho_picbase_di): New insn.
372 (nonlocal_goto_receiver): New define and split.
373 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_RELD_MPIC.
374 (unspecv enum): Add UNSPECV_NLGR.
376 2013-09-29 Iain Sandoe <iain@codesourcery.com>
378 * config/rs6000/rs6000.c (rs6000_init_dwarf_reg_sizes_extra): Ensure
379 that altivec registers are correctly sized on Darwin.
381 2013-09-29 Iain Sandoe <iain@codesourcery.com>
383 * config/t-darwin (darwin.o, darwin-c.o, darwin-f.o,
384 darwin-driver.o): Use COMPILE and POSTCOMPILE.
385 * config/x-darwin (host-darwin.o): Likewise.
386 * config/i386/x-darwin (host-i386-darwin.o): Likewise.
387 * config/rs6000/x-darwin (host-ppc-darwin.o): Likewise.
388 * config/rs6000/x-darwin64 (host-ppc64-darwin.o): Likewise.
390 2013-09-29 Uros Bizjak <ubizjak@gmail.com>
392 * doc/invoke.texi: Fix usage of @tie{} command.
394 2013-09-29 Eric Botcazou <ebotcazou@adacore.com>
396 * config/sparc/sync.md: Add peephole for consecutive memory barriers.
398 2013-09-28 Jan Hubicka <jh@suse.cz>
400 * config/i386/x86-tune.def: Add documentation for each of the options;
403 2013-09-28 Jan Hubicka <jh@suse.cz>
405 * x86-tune.def (X86_TUNE_SSE_UNALIGNED_LOAD_OPTIMAL): Enable for
407 (X86_TUNE_SSE_UNALIGNED_STORE_OPTIMAL): Likewise.
408 (X86_TUNE_FOUR_JUMP_LIMIT): Drop for generic and buldozer.
409 (X86_TUNE_PAD_RETURNS): Drop for buldozer chips.
410 (X86_TUNE_AVOID_VECTOR_DECODE): Drop for generic.
411 (X86_TUNE_REASSOC_FP_TO_PARALLEL): Enable for generic.
413 2013-09-28 Richard Sandiford <rdsandiford@googlemail.com>
415 * alloc-pool.c, asan.c, auto-inc-dec.c, basic-block.h, bb-reorder.c,
416 bitmap.c, bitmap.h, bt-load.c, builtins.c, calls.c, cfgcleanup.c,
417 cfgexpand.c, cfghooks.c, cfgloop.c, cfgloopmanip.c, cfgrtl.c, cgraph.c,
418 cgraph.h, cgraphbuild.c, cgraphclones.c, cgraphunit.c, collect2.c,
419 combine-stack-adj.c, combine.c, compare-elim.c, context.c, context.h,
420 cprop.c, cse.c, cselib.c, dbxout.c, dce.c, defaults.h, df-core.c,
421 df-problems.c, df-scan.c, df.h, diagnostic.c, double-int.c, dse.c,
422 dumpfile.c, dwarf2asm.c, dwarf2cfi.c, dwarf2out.c, emit-rtl.c,
423 errors.c, except.c, expmed.c, expr.c, file-find.c, final.c,
424 fixed-value.c, fold-const.c, function.c, fwprop.c, gcc-ar.c, gcc.c,
425 gcov-io.c, gcov-io.h, gcov.c, gcse.c, genattr-common.c, genattr.c,
426 genattrtab.c, genautomata.c, genconfig.c, genemit.c, genextract.c,
427 genflags.c, gengenrtl.c, gengtype-state.c, gengtype.c, genmodes.c,
428 genopinit.c, genoutput.c, genpeep.c, genpreds.c, genrecog.c,
429 gensupport.c, ggc-common.c, ggc-page.c, gimple-fold.c, gimple-low.c,
430 gimple-pretty-print.c, gimple-ssa-strength-reduction.c, gimple.c,
431 gimple.h, godump.c, graphite-clast-to-gimple.c,
432 graphite-optimize-isl.c, graphite-poly.h, graphite-sese-to-poly.c,
433 graphite.c, haifa-sched.c, hash-table.c, hash-table.h, hwint.c,
434 hwint.h, ifcvt.c, incpath.c, init-regs.c, input.h, intl.c, intl.h,
435 ipa-cp.c, ipa-devirt.c, ipa-inline-analysis.c, ipa-inline.c,
436 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c,
437 ipa-utils.c, ipa.c, ira-build.c, ira.c, jump.c, loop-doloop.c,
438 loop-init.c, loop-invariant.c, loop-iv.c, lower-subreg.c, lto-cgraph.c,
439 lto-streamer-in.c, lto-streamer-out.c, lto-wrapper.c, mcf.c,
440 mode-switching.c, modulo-sched.c, omp-low.c, optabs.c, opts.c,
441 pass_manager.h, passes.c, plugin.c, postreload-gcse.c, postreload.c,
442 predict.c, prefix.c, pretty-print.c, print-rtl.c, print-tree.c,
443 profile.c, read-md.c, real.c, real.h, recog.c, ree.c, reg-stack.c,
444 regcprop.c, reginfo.c, regmove.c, regrename.c, regs.h, regstat.c,
445 reload1.c, reorg.c, rtl.c, rtl.h, rtlanal.c, sbitmap.c, sched-rgn.c,
446 sdbout.c, sel-sched-ir.c, sel-sched.c, sparseset.c, stack-ptr-mod.c,
447 statistics.c, stmt.c, stor-layout.c, store-motion.c, streamer-hooks.h,
448 system.h, target-hooks-macros.h, targhooks.c, targhooks.h, toplev.c,
449 tracer.c, trans-mem.c, tree-browser.c, tree-call-cdce.c, tree-cfg.c,
450 tree-cfgcleanup.c, tree-complex.c, tree-data-ref.c, tree-data-ref.h,
451 tree-eh.c, tree-emutls.c, tree-flow.h, tree-if-conv.c, tree-into-ssa.c,
452 tree-iterator.c, tree-loop-distribution.c, tree-mudflap.c,
453 tree-nested.c, tree-nomudflap.c, tree-nrv.c, tree-object-size.c,
454 tree-optimize.c, tree-pass.h, tree-pretty-print.c, tree-profile.c,
455 tree-scalar-evolution.c, tree-sra.c, tree-ssa-ccp.c,
456 tree-ssa-coalesce.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
457 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
458 tree-ssa-ifcombine.c, tree-ssa-live.c, tree-ssa-loop-ch.c,
459 tree-ssa-loop-im.c, tree-ssa-loop-ivopts.c, tree-ssa-loop-prefetch.c,
460 tree-ssa-loop.c, tree-ssa-math-opts.c, tree-ssa-operands.c,
461 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
462 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
463 tree-ssa-structalias.c, tree-ssa-threadedge.c, tree-ssa-threadupdate.c,
464 tree-ssa-uncprop.c, tree-ssa-uninit.c, tree-ssa.c, tree-ssanames.c,
465 tree-stdarg.c, tree-switch-conversion.c, tree-tailcall.c,
466 tree-vect-data-refs.c, tree-vect-generic.c, tree-vect-loop-manip.c,
467 tree-vect-stmts.c, tree-vectorizer.c, tree-vectorizer.h, tree-vrp.c,
468 tree.c, tree.h, tsan.c, tsystem.h, value-prof.c, var-tracking.c,
469 varasm.c, vec.h, vmsdbgout.c, vtable-verify.c, web.c: Add missing
470 whitespace before "(".
472 2013-09-28 Sandra Loosemore <sandra@codesourcery.com>
474 * expr.h (extract_bit_field): Remove packedp parameter.
475 * expmed.c (extract_fixed_bit_field): Remove packedp parameter
476 from forward declaration.
477 (store_split_bit_field): Remove packedp arg from calls to
478 extract_fixed_bit_field.
479 (extract_bit_field_1): Remove packedp parameter and packedp
480 argument from recursive calls and calls to extract_fixed_bit_field.
481 (extract_bit_field): Remove packedp parameter and corresponding
482 arg to extract_bit_field_1.
483 (extract_fixed_bit_field): Remove packedp parameter. Remove code
485 (extract_split_bit_field): Remove packedp arg from call to
486 extract_fixed_bit_field.
487 * expr.c (emit_group_load_1): Adjust calls to extract_bit_field.
488 (copy_blkmode_from_reg): Likewise.
489 (copy_blkmode_to_reg): Likewise.
490 (read_complex_part): Likewise.
491 (store_field): Likewise.
492 (expand_expr_real_1): Likewise.
493 * calls.c (store_unaligned_arguments_into_pseudos): Adjust call
494 to extract_bit_field.
495 * config/tilegx/tilegx.c (tilegx_expand_unaligned_load): Adjust
496 call to extract_bit_field.
497 * config/tilepro/tilepro.c (tilepro_expand_unaligned_load): Adjust
498 call to extract_bit_field.
499 * doc/invoke.texi (Code Gen Options): Remove mention of warnings
500 and special packedp behavior from -fstrict-volatile-bitfields
503 2013-09-27 Jan-Benedict Glaw <jbglaw@lug-owl.de>
505 * lra-eliminations.c (init_elim_table): Guard value_p.
507 2013-09-27 Michael Meissner <meissner@linux.vnet.ibm.com>
509 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): Allow
510 DFmode, DImode, and SFmode in the upper VSX registers based on the
511 -mupper-regs-{df,sf} flags. Fix wu constraint to be ALTIVEC_REGS
512 if -mpower8-vector. Combine -mvsx-timode handling with the rest
513 of the VSX register handling.
515 * config/rs6000/rs6000.md (f32_lv): Use %x0 for VSX regsters.
517 (zero_extendsidi2_lfiwzx): Add support for loading into the
518 Altivec registers with -mpower8-vector. Use wu/wv constraints to
519 only do VSX memory options on Altivec registers.
520 (extendsidi2_lfiwax): Likewise.
521 (extendsfdf2_fpr): Likewise.
522 (mov<mode>_hardfloat, SF/SD modes): Likewise.
523 (mov<mode>_hardfloat32, DF/DD modes): Likewise.
524 (mov<mode>_hardfloat64, DF/DD modes): Likewise.
525 (movdi_internal64): Likewise.
527 2013-09-27 Xinliang David Li <davidxl@google.com>
529 * opts.c (finish_options): Adjust parameters
530 according to vect cost model.
531 (common_handle_option): Set dynamic vect cost
533 targhooks.c (default_add_stmt_cost): Compute stmt cost
535 * tree-vect-loop.c (vect_estimate_min_profitable_iters):
537 * tree-vectorizer.h (unlimited_cost_model): New function.
538 * tree-vect-slp.c (vect_slp_analyze_bb_1): Use helper function.
539 * tree-vect-data-refs.c (vect_peeling_hash_insert): Use helper
541 (vect_enhance_data_refs_alignment): Ditto.
542 * flag-types.h: New enum.
543 * common/config/i386/i386-common.c (ix86_option_init_struct):
544 No need to initialize vect_cost_model flag.
545 * config/i386/i386.c (ix86_add_stmt_cost): Compute stmt cost
548 2013-09-27 Diego Novillo <dnovillo@google.com>
550 * gimple.h (enum ssa_mode): Remove.
552 2013-09-27 Paulo Matos <pmatos@broadcom.com>
555 * gcc.dg/pr58463.c: New test.
557 2013-09-27 Paulo Matos <pmatos@broadcom.com>
559 * cfgloop.h (number_of_loops): Fix typo in check for null.
561 2013-09-27 Jakub Jelinek <jakub@redhat.com>
564 * tree-cfg.c (move_sese_region_to_fn): Also move loops that
565 are children of outermost saved_cfun's loop, and set it up to
566 be moved to dest_cfun's outermost loop. Fix up num_nodes adjustments
567 if loop != loop0 and SESE region contains bbs that belong to loop0.
569 2013-09-27 Richard Sandiford <rdsandiford@googlemail.com>
571 * rtlanal.c (must_be_base_p, must_be_index_p): Delete.
572 (binary_scale_code_p, get_base_term, get_index_term): New functions.
573 (set_address_segment, set_address_base, set_address_index)
574 (set_address_disp): Accept the argument unconditionally.
575 (baseness): Remove must_be_base_p and must_be_index_p checks.
576 (decompose_normal_address): Classify as much as possible in the
579 2013-09-27 Richard Sandiford <rdsandiford@googlemail.com>
581 * cse.c (count_reg_usage): Handle INT_LIST.
582 * lra-eliminations.c (lra_eliminate_regs_1): Likewise.
583 * reginfo.c (reg_scan_mark_refs): Likewise.
584 * reload1.c (eliminate_regs_1): Likewise.
586 2013-09-27 Iain Sandoe <iain@codesourcery.com>
589 * rtlanal.c (lsb_bitfield_op_p): Make both parts of the comparison
592 2013-09-27 Richard Biener <rguenther@suse.de>
594 PR tree-optimization/58459
595 * tree-ssa-forwprop.c (forward_propagate_addr_expr): Remove
596 restriction not propagating into loops.
598 2013-09-26 Florian Weimer <fw@deneb.enyo.de>
600 * tree-ssa.h (walk_use_def_chains_fn, walk_use_def_chains): Delete.
601 * tree-ssa.c (walk_use_def_chains_1, walk_use_def_chains): Delete.
602 * doc/tree-ssa.texi (Walking use-def chains): Delete.
604 2013-09-26 Richard Biener <rguenther@suse.de>
606 * tree-into-ssa.c (rewrite_into_ssa): Make more SSA names to anonymous.
608 2013-09-26 Richard Biener <rguenther@suse.de>
610 * alias.h (component_uses_parent_alias_set): Rename to ...
611 (component_uses_parent_alias_set_from): ... this.
612 * alias.c (component_uses_parent_alias_set): Rename to ...
613 (component_uses_parent_alias_set_from): ... this and return
615 (reference_alias_ptr_type_1): Use the result from
616 component_uses_parent_alias_set_from instead of stripping
617 components one at a time.
618 * emit-rtl.c (set_mem_attributes_minus_bitpos): Adjust.
620 2013-09-26 Andrew MacLeod <amacleod@redhat.com>
622 * tree-ssa-live.h (find_replaceable_exprs, dump_replaceable_exprs):
623 Move prototypes to...
624 * tree-ssa-ter.h: New File. Move prototypes here.
625 * tree-flow.h (stmt_is_replaceable_p): Remove prototype.
626 * tree-outof-ssa.h: New. Rename ssaexpand.h, include tree-ssa-ter.h.
627 * tree-outof-ssa.c (ssa_is_replaceable_p): New. Refactor common bits
628 from is_replaceable_p.
629 * tree-ssa-ter.c (is_replaceable_p, stmt_is_replaceable_p): Delete.
630 (ter_is_replaceable_p): New. Use new refactored ssa_is_replaceable_p.
631 (process_replaceable): Use ter_is_replaceable_p.
632 (find_replaceable_in_bb): Use ter_is_replaceable_p.
633 * expr.c (stmt_is_replaceable_p): Relocate from tree-ssa-ter.c. Use
634 newly refactored ssa_is_replaceable_p.
635 * cfgexpand.c: Include tree-outof-ssa.h.
636 * ssaexpand.h: Delete.
638 2013-09-26 Andrew MacLeod <amacleod@redhat.com>
640 * gimple.c (gimple_replace_lhs): Move to tree-ssa.c and rename.
641 (struct count_ptr_d, count_ptr_derefs, count_uses_and_derefs): Move to
643 (create_gimple_tmp): Delete.
644 (get_expr_type, build_assign, build_type_cast): Move to...
645 * gimple-builder.c: New File.
646 (get_expr_type): Relocate from gimple.c.
647 (build_assign, build_type_cast): Change to only create ssanames.
648 * gimple.h: Move prototypes to...
649 * gimple-builder.h: New File. Here.
650 * tree-ssa.h: And here.
651 * tree-ssa.c (struct count_ptr_d, count_ptr_derefs,
652 count_uses_and_derefs): Relocate from gimple.c.
653 (gimple_replace_ssa_lhs): Renamed gimple_replace_ssa from gimple.c
654 * tree-ssa-reassoc.c (repropagate_negates): Use gimple_replace_ssa_lhs.
655 * tree-ssa-math-opts (execute_cse_reciprocals): Use
656 gimple_replace_ssa_lhs.
657 * asan.c: Include gimple-builder.h.
658 * Makefile.in: Add gimple-builder.o.
660 2013-09-26 Richard Biener <rguenther@suse.de>
662 * tree-ssa-live.c (var_map_base_init): Handle SSA names with
663 DECL_IGNORED_P base VAR_DECLs like anonymous SSA names.
664 (loe_visit_block): Use gcc_checking_assert.
665 * tree-ssa-coalesce.c (create_outofssa_var_map): Use
666 gimple_assign_ssa_name_copy_p.
667 (gimple_can_coalesce_p): Adjust according to the var_map_base_init
670 2013-09-26 David Edelsohn <dje.gcc@gmail.com>
672 * config/rs6000/t-rs6000 (rs6000.o): Remove.
673 (rs6000-c.o): Use COMPILE and POSTCOMPILE.
675 2013-09-26 Richard Biener <rguenther@suse.de>
677 PR tree-optimization/58539
678 * tree-vect-loop.c (vect_create_epilog_for_reduction): Honor
679 the fact that debug statements are not taking part in loop-closed
682 2013-09-26 Nick Clifton <nickc@redhat.com>
684 * config/msp430/msp430.c (msp430_expand_epilogue): Fix compile
685 time warning message.
686 (msp430_print_operand_raw): Delete unused letter parameter.
687 (TARGET_PRINT_OPERAND_ADDRESS): Define.
688 (msp430_print_operand_address): New function.
689 (msp430_print_operand): Move address printing code from here to
691 * config/msp430/msp430.md (movsipsi2): Add comment in generated
693 (zero_extendpsisi2): Likewise.
694 (extendpsisi2): New pattern.
695 (andneghi3): New pattern.
697 2013-09-26 Yvan Roux <yvan.roux@linaro.org>
699 * config/aarch64/aarch64.opt (mlra): New option.
700 * config/aarch64/aarch64.c (aarch64_lra_p): New function.
701 (TARGET_LRA_P): Define.
703 2013-09-26 Eric Botcazou <ebotcazou@adacore.com>
705 * expr.c (expand_assignment): Remove obsolete comment.
707 2013-09-25 Jeff Law <law@redhat.com>
709 * tree-flow.h (thread_through_all_blocks): Prototype moved into
710 tree-ssa-threadupdate.h.
711 (register_jump_thread): Similarly.
712 * tree-ssa-threadupdate.h: New header file.
713 * tree-ssa-dom.c: Include tree-ssa-threadupdate.h.
714 * tree-vrp.c: Likewise.
715 * tree-ssa-threadedge.c: Include tree-ssa-threadupdate.h.
716 (thread_around_empty_blocks): Change type of path vector argument to
717 an edge,type pair from just an edge. Initialize both elements when
718 appending to a jump threading path. Tweak references to elements
720 (thread_across_edge): Similarly. Release memory for the elements
722 * tree-ssa-threadupdate.c: Include tree-ssa-threadupdate.h.
723 (dump_jump_thread_path): New function broken out from
724 register_jump_thread.
725 (register_jump_thread): Use dump_jump_thread_path. Change type of
726 path vector entries. Search the path for NULL edges and dump
727 the path if one is found. Tweak the conversion of path to 3-edge
728 form to use the block copy type information embedded in the path.
730 2013-09-25 Yvan Roux <yvan.roux@linaro.org>
732 * lra.c (update_inc_notes): Remove all REG_DEAD and REG_UNUSED notes.
734 2013-09-25 Yvan Roux <yvan.roux@linaro.org>
735 Vladimir Makarov <vmakarov@redhat.com>
737 * rtlanal.c (lsb_bitfield_op_p): New predicate for bitfield operations
738 from the least significant bit.
739 (strip_address_mutations): Add bitfield operations handling.
740 (must_be_index_p): Add shifting and rotate operations handling.
741 (set_address_base): Use must_be_base_p predicate.
742 (set_address_index): Use must_be_index_p predicate.
744 2013-09-25 Alexander Ivchenko <alexander.ivchenko@intel.com>
745 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
746 Sergey Lega <sergey.s.lega@intel.com>
747 Anna Tikhonova <anna.tikhonova@intel.com>
748 Ilya Tocar <ilya.tocar@intel.com>
749 Andrey Turetskiy <andrey.turetskiy@intel.com>
750 Ilya Verbin <ilya.verbin@intel.com>
751 Kirill Yukhin <kirill.yukhin@intel.com>
752 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
754 * config/i386/i386.c (ix86_avx256_split_vector_move_misalign):
756 (ix86_expand_vector_move_misalign): Support new unaligned load and
757 stores and use new names.
758 (CODE_FOR_sse2_storedqu): Rename to ...
759 (CODE_FOR_sse2_storedquv16qi): ... this.
760 (CODE_FOR_sse2_loaddqu): Rename to ...
761 (CODE_FOR_sse2_loaddquv16qi): ... this.
762 (CODE_FOR_avx_loaddqu256): Rename to ...
763 (CODE_FOR_avx_loaddquv32qi): ... this.
764 (CODE_FOR_avx_storedqu256): Rename to ...
765 (CODE_FOR_avx_storedquv32qi): ... this.
766 * config/i386/i386.md (fpint_logic): New.
767 * config/i386/sse.md (VMOVE): Extend for AVX512.
771 (VI_UNALIGNED_LOADSTORE): Ditto.
772 (sse2_avx_avx512f): Ditto.
773 (sse2_avx2): Extend for AVX512.
774 (sse4_1_avx2): Ditto.
776 (sse): Extend for AVX512.
779 (avxsizesuffix): Ditto.
780 (sseintvecmode): Ditto.
782 (<sse>_loadu<ssemodesuffix><avxsizesuffix>): Ditto.
783 (<sse>_storeu<ssemodesuffix><avxsizesuffix>): Ditto.
784 (<sse2>_loaddqu<avxsizesuffix>): Extend for AVX512 and rename to ...
785 (<sse2_avx_avx512f>_loaddqu<mode>): ... this.
786 (<sse2>_storedqu<avxsizesuffix>): Extend for AVX512 and rename to ...
787 (<sse2_avx_avx512f>_storedqu<mode): ... this.
788 (<sse>_movnt<mode>): Replace constraint "x" with "v".
789 (STORENT_MODE): Extend for AVX512.
790 (*absneg<mode>2): Replace constraint "x" with "v".
791 (*mul<mode>3): Ditto.
792 (*ieee_smin<mode>3): Ditto.
793 (*ieee_smax<mode>3): Ditto.
794 (avx_cmp<mode>3): Replace VF with VF_128_256.
795 (*<sse>_maskcmp<mode>3_comm): Ditto.
796 (<sse>_maskcmp<mode>3): Ditto.
797 (<sse>_andnot<mode>3): Extend for AVX512.
798 (<code><mode>3, anylogic): Replace VF with VF_128_256.
799 (<code><mode>3, fpint_logic): New.
800 (*<code><mode>3): Extend for AVX512.
801 (avx512flogicsuff): New.
802 (avx512f_<logic><mode>): Ditto.
803 (<sse>_movmsk<ssemodesuffix><avxsizesuffix>): Replace VF with
805 (<sse4_1>_blend<ssemodesuffix><avxsizesuffix>): Ditto.
806 (<sse4_1>_blendv<ssemodesuffix><avxsizesuffix>): Ditto.
807 (<sse4_1>_dp<ssemodesuffix><avxsizesuffix>): Ditto.
808 (avx_vtest<ssemodesuffix><avxsizesuffix>): Ditto.
809 (<sse4_1>_round<ssemodesuffix><avxsizesuffix>): Ditto.
810 (xop_vpermil2<mode>3): Ditto.
811 (*avx_vpermilp<mode>): Extend for AVX512 and rename to ...
812 (*<sse2_avx_avx512f>_vpermilp<mode>): ... this.
813 (avx_vpermilvar<mode>3): Extend for AVX512 and rename to ...
814 (<sse2_avx_avx512f>_vpermilvar<mode>3): ... this.
816 2013-09-25 Tom Tromey <tromey@redhat.com>
818 * Makefile.in (PARTITION_H, LTO_SYMTAB_H, COMMON_TARGET_DEF_H)
819 (RTL_ERROR_H, TRANS_MEM_H, COVERAGE_H, DEMANGLE_H, ALIAS_H)
820 (SCHED_INT_H, SEL_SCHED_IR_H, SEL_SCHED_DUMP_H, VALTRACK_H, DDG_H)
821 (GGC_INTERNAL_H, DECNUM_H, BACKTRACE_H, MKDEPS_H, TREE_HASHER_H)
822 (TREE_SSA_LIVE_H, SSAEXPAND_H, DWARF2OUT_H, SCEV_H, OMEGA_H)
823 (TREE_DATA_REF_H, IRA_INT_H, LRA_INT_H, DBGCNT_H, DATA_STREAMER_H)
824 (GIMPLE_STREAMER_H, TREE_STREAMER_H, STREAMER_HOOKS_H)
825 (TREE_VECTORIZER_H, IPA_INLINE_H, GSTAB_H, LIBFUNCS_H)
826 (GRAPHITE_HTAB_H): Remove.
828 2013-09-25 Tom Tromey <tromey@redhat.com>
830 * config/mcore/t-mcore (CROSS_FLOAT_H): Remove.
832 2013-09-25 Tom Tromey <tromey@redhat.com>
834 * config/t-glibc (glibc-c.o): Use COMPILE and POSTCOMPILE.
836 2013-09-25 Tom Tromey <tromey@redhat.com>
838 * config/i386/t-i386 (i386.o): Remove.
839 (i386-c.o): Use COMPILE and POSTCOMPILE.
841 2013-09-25 Tom Tromey <tromey@redhat.com>
843 * Makefile.in ($(out_object_file)): Use COMPILE and POSTCOMPILE.
845 2013-09-25 Tom Tromey <tromey@redhat.com>
847 * Makefile.in (graph.o, sbitmap.o, sparseset.o, gcc-ar.o)
848 (gcc-ranlib.o, gcc-nm.o, collect2.o, collect2-aix.o, tlink.o)
849 (lto-wrapper.o, default-c.o, attribs.o, incpath.o, prefix.o)
850 (gcc.o, options.o, options-save.o, version.o, gtype-desc.o)
851 (trans-mem.o, ggc-common.o, ggc-page.o, ggc-none.o, stringpool.o)
852 (convert.o, double-int.o, lto-compress.o, data-streamer-in.o)
853 (data-streamer-out.o, data-streamer.o, gimple-streamer-in.o)
854 (gimple-streamer-out.o, tree-streamer.o, tree-streamer-in.o)
855 (tree-streamer-out.o, streamer-hooks.o, lto-cgraph.o)
856 (lto-streamer-in.o, lto-streamer-out.o, lto-section-in.o)
857 (lto-section-out.o, lto-opts.o, lto-streamer.o, langhooks.o)
858 (test-dump.o, tree.o, tree-dump.o, tree-inline.o, print-tree.o)
859 (stor-layout.o, asan.o, tsan.o, ubsan.o, tree-ssa-tail-merge.o)
860 (tree-ssa-structalias.o, tree-ssa-uninit.o, tree-ssa.o)
861 (tree-into-ssa.o, tree-ssa-ter.o, tree-ssa-coalesce.o)
862 (tree-outof-ssa.o, tree-ssa-dse.o, tree-ssa-forwprop.o)
863 (tree-ssa-phiprop.o, tree-ssa-ifcombine.o, tree-ssa-phiopt.o)
864 (tree-nrv.o, tree-ssa-copy.o, tree-ssa-propagate.o)
865 (tree-ssa-dom.o, tree-ssa-uncprop.o, tree-ssa-threadedge.o)
866 (tree-ssa-threadupdate.o, tree-ssanames.o, tree-phinodes.o)
867 (domwalk.o, tree-ssa-live.o, tree-ssa-copyrename.o)
868 (tree-ssa-pre.o, tree-ssa-sccvn.o)
869 (gimple-ssa-strength-reduction.o, tree-vrp.o, tree-cfg.o)
870 (tree-cfgcleanup.o, tree-tailcall.o, tree-ssa-sink.o)
871 (tree-nested.o, tree-if-conv.o, tree-iterator.o, tree-dfa.o)
872 (tree-ssa-operands.o, tree-eh.o, tree-ssa-loop.o)
873 (tree-ssa-loop-unswitch.o, tree-ssa-address.o)
874 (tree-ssa-loop-niter.o, tree-ssa-loop-ivcanon.o)
875 (tree-ssa-loop-ch.o, tree-ssa-loop-prefetch.o, tree-predcom.o)
876 (tree-ssa-loop-ivopts.o, tree-affine.o, tree-ssa-loop-manip.o)
877 (tree-ssa-loop-im.o, tree-ssa-math-opts.o, tree-ssa-alias.o)
878 (tree-ssa-reassoc.o, tree-optimize.o, gimplify.o)
879 (gimple-iterator.o, gimple-fold.o, gimple-low.o, omp-low.o)
880 (tree-browser.o, omega.o, tree-chrec.o, tree-scalar-evolution.o)
881 (tree-data-ref.o, sese.o, graphite.o, graphite-blocking.o)
882 (graphite-clast-to-gimple.o, graphite-dependences.o)
883 (graphite-interchange.o, graphite-poly.o)
884 (graphite-scop-detection.o, graphite-sese-to-poly.o)
885 (graphite-optimize-isl.o, tree-vect-loop.o)
886 (tree-vect-loop-manip.o, tree-vect-patterns.o, tree-vect-slp.o)
887 (tree-vect-stmts.o, tree-vect-data-refs.o, tree-vectorizer.o)
888 (vtable-verify.o, tree-loop-distribution.o, tree-parloops.o)
889 (tree-stdarg.o, tree-object-size.o, internal-fn.o, gimple.o)
890 (gimple-pretty-print.o, tree-mudflap.o, tree-nomudflap.o)
891 (tree-pretty-print.o, tree-diagnostic.o, fold-const.o)
892 (diagnostic.o, diagnostic-color.o, opts.o, opts-global.o)
893 (opts-common.o, targhooks.o, common/common-targhooks.o, input.o)
894 (toplev.o, hwint.o, passes.o, plugin.o, main.o, host-default.o)
895 (rtl-error.o, rtl.o, print-rtl.o, rtlanal.o, varasm.o, function.o)
896 (statistics.o, stmt.o, except.o, expr.o, dojump.o, builtins.o)
897 (calls.o, expmed.o, explow.o, optabs.o, dbxout.o, debug.o)
898 (sdbout.o, dwarf2out.o, dwarf2cfi.o, dwarf2asm.o, vmsdbgout.o)
899 (xcoffout.o, godump.o, emit-rtl.o, real.o, realmpfr.o, dfp.o)
900 (fixed-value.o, jump.o, simplify-rtx.o, symtab.o, cgraph.o)
901 (cgraphunit.o, cgraphclones.o, cgraphbuild.o, varpool.o, ipa.o)
902 (ipa-profile.o, ipa-devirt.o, ipa-prop.o, ipa-ref.o, ipa-cp.o)
903 (ipa-split.o, ipa-inline.o, ipa-inline-analysis.o)
904 (ipa-inline-transform.o, ipa-utils.o, ipa-reference.o)
905 (ipa-pure-const.o, coverage.o, cselib.o, cse.o, dce.o, dumpfile.o)
906 (dse.o, fwprop.o, web.o, ree.o, cprop.o, gcse.o, store-motion.o)
907 (resource.o, lcm.o, mode-switching.o, tree-ssa-dce.o)
908 (tree-call-cdce.o, tree-ssa-ccp.o, tree-ssa-strlen.o, tree-sra.o)
909 (tree-switch-conversion.o, tree-complex.o, tree-emutls.o)
910 (tree-vect-generic.o, df-core.o, df-problems.o, df-scan.o)
911 (regstat.o, valtrack.o, var-tracking.o, profile.o, mcf.o)
912 (tree-profile.o, value-prof.o, loop-doloop.o, alloc-pool.o)
913 (auto-inc-dec.o, cfg.o, cfghooks.o, cfgexpand.o, cfgrtl.o)
914 (cfganal.o, cfgbuild.o, cfgcleanup.o, cfgloop.o, cfgloopanal.o)
915 (graphds.o, loop-iv.o, loop-invariant.o, cfgloopmanip.o)
916 (loop-init.o, loop-unswitch.o, loop-unroll.o, dominance.o)
917 (et-forest.o, combine.o, reginfo.o, bitmap.o, vec.o, hash-table.o)
918 (reload.o, reload1.o, rtlhooks.o, postreload.o, postreload-gcse.o)
919 (caller-save.o, bt-load.o, reorg.o, alias.o, stack-ptr-mod.o)
920 (init-regs.o, ira-build.o, ira-costs.o, ira-conflicts.o)
921 (ira-color.o, ira-emit.o, ira-lives.o, ira.o, lra.o)
922 (lra-assigns.o, lra-coalesce.o, lra-constraints.o)
923 (lra-eliminations.o, lra-lives.o, lra-spills.o, regmove.o)
924 (combine-stack-adj.o, compare-elim.o, ddg.o, modulo-sched.o)
925 (haifa-sched.o, sched-deps.o, sched-rgn.o, sched-ebb.o)
926 (sched-vis.o, sel-sched.o, sel-sched-dump.o, sel-sched-ir.o)
927 (final.o, recog.o, reg-stack.o, sreal.o, predict.o, lists.o)
928 (bb-reorder.o, tracer.o, timevar.o, regcprop.o, regrename.o)
929 (ifcvt.o, params.o, pointer-set.o, hooks.o, pretty-print.o)
930 (errors.o, dbgcnt.o, lower-subreg.o, target-globals.o)
931 (hw-doloop.o, file-find.o, context.o, $(common_out_object_file))
932 (insn-attrtab.o, insn-automata.o, insn-dfatab.o, insn-emit.o)
933 (insn-enums.o, insn-extract.o, insn-latencytab.o, insn-modes.o)
934 (insn-opinit.o, insn-output.o, insn-peep.o, insn-preds.o)
935 (insn-recog.o, intl.o, cppbuiltin.o, cppdefault.o, gcov.o)
936 (gcov-dump.o): Remove.
937 (default-c.o): Use COMPILE and POSTCOMPILE.
938 (CFLAGS-gcc.o): New variable.
939 ($(common_out_object_file)): Use COMPILE and POSTCOMPILE.
941 2013-09-25 Tom Tromey <tromey@redhat.com>
943 * Makefile.in (c-family/cppspec.o, c-family/c-common.o)
944 (c-family/c-cppbuiltin.o, c-family/c-dump.o, c-family/c-format.o)
945 (c-family/c-gimplify.o, c-family/c-lex.o, c-family/c-omp.o)
946 (c-family/c-opts.o, c-family/c-pch.o, c-family/c-ppoutput.o)
947 (c-family/c-pragma.o, c-family/c-pretty-print.o)
948 (c-family/c-semantics.o, c-family/c-ada-spec.o)
949 (c-family/array-notation-common.o, c-family/stub-objc.o)
950 (c-family/c-ubsan.o): Remove.
952 2013-09-25 Tom Tromey <tromey@redhat.com>
954 * Makefile.in (C_TREE_H): Reference c/c-tree.h.
956 2013-09-25 Tom Tromey <tromey@redhat.com>
958 * Makefile.in (DRIVER_DEFINES): Use $(and), not shell code,
959 to add -DENABLE_SHARED_LIBGCC.
960 (gcc.o): Don't use subshell.
962 2013-09-25 Tom Tromey <tromey@redhat.com>
964 * Makefile.in (OUTPUT_OPTION): Define as "-o $@".
965 * configure.ac: Don't invoke AM_PROG_CC_C_O.
966 (NO_MINUS_C_MINUS_O, OUTPUT_OPTION): Don't subst.
967 * configure, config.in: Rebuild.
969 2013-09-25 Tom Tromey <tromey@redhat.com>
971 * Makefile.in (CCDEPMODE, DEPDIR, depcomp, COMPILE.base)
972 (COMPILE, POSTCOMPILE): New variables.
973 (.cc.o .c.o): Use COMPILE, POSTCOMPILE.
974 (DEPFILES): New variable.
976 * configure.ac: Add checks for dependency checking.
977 * configure, aclocal.m4: Regenerate.
979 2013-09-25 Tom Tromey <tromey@redhat.com>
981 * Makefile.in (ALL_HOST_BACKEND_OBJS): Add lto-wrapper.o.
982 ($(ALL_HOST_OBJS)): Move order-only dependency to end of file.
984 2013-09-25 Tom Tromey <tromey@redhat.com>
986 * Makefile.in (generated_files): Add options.h,
987 target-hooks-def.h, insn-opinit.h,
988 common/common-target-hooks-def.h, pass-instances.def,
989 c-family/c-target-hooks-def.h.
991 2013-09-25 Jeff Law <law@redhat.com>
993 * tree-ssa-threadedge.c (thread_across_edge): Use foo.last () rather
994 than foo[foo.length () - 1] to access last member in a vec.
995 * tree-ssa-threadupdate.c (register_jump_thread): Similarly.
997 2013-09-25 Richard Biener <rguenther@suse.de>
1000 * tree.c (iterative_hash_expr): Remove MEM_REF special handling.
1002 2013-09-25 Jan Hubicka <jh@suse.cz>
1004 * cgraph.c (cgraph_resolve_speculation): Use semantical equivalency
1007 2013-09-25 Marek Polacek <polacek@redhat.com>
1010 * ubsan.c (ubsan_type_descriptor): Handle IDENTIFIER_NODEs
1011 when determining the type name.
1013 2013-09-24 Oleg Endo <olegendo@gcc.gnu.org>
1015 * config/sh/sh.md: Fix formatting.
1017 2013-09-24 Xinliang David Li <davidxl@google.com>
1019 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Check
1020 max peel iterations parameter.
1021 * param.def: New parameter.
1022 * doc/invoke.texi: Document New parameter.
1024 2013-09-24 Christophe Lyon <christophe.lyon@linaro.org>
1026 * gimple-pretty-print.c: Various whitespace tweaks.
1027 * tree-core.h: Likewise.
1028 * tree-pretty-print.c: Likewise.
1029 * tree-ssa-alias.c: Likewise.
1030 * tree-ssa-copy.c: Likewise.
1031 * tree-ssanames.c: Likewise.
1032 * tree-ssanames.h: Likewise.
1033 * tree-vrp.c: Likewise.
1035 2013-09-24 Alan Modra <amodra@gmail.com>
1039 * stmt.c (expand_asm_operands): Call expand_expr with EXPAND_MEMORY
1040 for output operands that disallow regs. Don't use EXPAND_WRITE on
1043 2013-09-24 Richard Biener <rguenther@suse.de>
1046 * tree.c (reference_alias_ptr_type): Move ...
1047 * alias.c (reference_alias_ptr_type): ... here and implement
1048 in terms of the new reference_alias_ptr_type_1.
1049 (ref_all_alias_ptr_type_p): New helper.
1050 (get_deref_alias_set_1): Drop flag_strict_aliasing here,
1051 use ref_all_alias_ptr_type_p.
1052 (get_deref_alias_set): Add flag_strict_aliasing check here.
1053 (reference_alias_ptr_type_1): New function, split out from ...
1054 (get_alias_set): ... here.
1055 (alias_ptr_types_compatible_p): New function.
1056 * alias.h (reference_alias_ptr_type): Declare.
1057 (alias_ptr_types_compatible_p): Likewise.
1058 * tree.h (reference_alias_ptr_type): Remove.
1059 * fold-const.c (operand_equal_p): Use alias_ptr_types_compatible_p
1060 to compare MEM_REF alias types.
1062 2013-09-24 Richard Biener <rguenther@suse.de>
1064 * tree-vrp.c (vrp_finalize): Check for SSA name presence.
1066 2013-09-23 Michael Meissner <meissner@linux.vnet.ibm.com>
1068 * config/rs6000/rs6000.c (rs6000_vector_reload): Delete, combine
1069 reload helper function arrays into a single array reg_addr.
1070 (reload_fpr_gpr): Likewise.
1071 (reload_gpr_vsx): Likewise.
1072 (reload_vsx_gpr): Likewise.
1073 (struct rs6000_reg_addr): Likewise.
1074 (reg_addr): Likewise.
1075 (rs6000_debug_reg_global): Change rs6000_vector_reload,
1076 reload_fpr_gpr, reload_gpr_vsx, reload_vsx_gpr uses to reg_addr.
1077 (rs6000_init_hard_regno_mode_ok): Likewise.
1078 (rs6000_secondary_reload_direct_move): Likewise.
1079 (rs6000_secondary_reload): Likewise.
1081 * config/rs6000/rs6000.h (enum r6000_reg_class_enum): Add new
1082 constraints: wu, ww, and wy. Repurpose wv constraint added during
1083 power8 changes. Put wg constraint in alphabetical order.
1085 * config/rs6000/rs6000.opt (-mvsx-scalar-float): New debug switch
1086 for future work to add ISA 2.07 VSX single precision support.
1087 (-mvsx-scalar-double): Change default from -1 to 1, update
1088 documentation comment.
1089 (-mvsx-scalar-memory): Rename debug switch to -mupper-regs-df.
1090 (-mupper-regs-df): New debug switch to control whether DF values
1091 can go in the traditional Altivec registers.
1092 (-mupper-regs-sf): New debug switch to control whether SF values
1093 can go in the traditional Altivec registers.
1095 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print wu, ww,
1097 (rs6000_init_hard_regno_mode_ok): Use ssize_t instead of int for
1098 loop variables. Rename -mvsx-scalar-memory to -mupper-regs-df.
1099 Add new constraints, wu/ww/wy. Repurpose wv constraint.
1100 (rs6000_debug_legitimate_address_p): Print if we are running
1101 before, during, or after reload.
1102 (rs6000_secondary_reload): Add a comment.
1103 (rs6000_opt_masks): Add -mupper-regs-df, -mupper-regs-sf.
1105 * config/rs6000/constraints.md (wa constraint): Sort w<x>
1106 constraints. Update documentation string.
1107 (wd constraint): Likewise.
1108 (wf constraint): Likewise.
1109 (wg constraint): Likewise.
1110 (wn constraint): Likewise.
1111 (ws constraint): Likewise.
1112 (wt constraint): Likewise.
1113 (wx constraint): Likewise.
1114 (wz constraint): Likewise.
1115 (wu constraint): New constraint for ISA 2.07 SFmode scalar
1117 (ww constraint): Likewise.
1118 (wy constraint): Likewise.
1119 (wv constraint): Repurpose ISA 2.07 constraint that we not used in
1120 the previous submissions.
1121 * doc/md.texi (PowerPC and IBM RS6000): Likewise.
1123 2013-09-23 Richard Sandiford <rdsandiford@googlemail.com>
1125 * doc/rtl.texi (REG_NOTES): Say that int_list can also be used.
1126 (REG_BR_PROB): Say that the probability is stored in an int_list.
1127 * reg-notes.def: Update commentary to mention INT_LIST.
1128 * rtl.def (EXPR_LIST, INSN_LIST): Capitalize comments.
1129 (INT_LIST): New rtx.
1130 * rtl.h (add_int_reg_note, add_shallow_copy_of_reg_note): Declare.
1131 * rtlanal.c (int_reg_note_p): New function.
1132 (alloc_reg_note): Assert that the note does not have an int argument.
1133 (add_int_reg_note, add_shallow_copy_of_reg_note): New functions.
1134 * combine.c (distribute_notes): Use add_shallow_copy_of_rtx.
1135 * cse.c (cse_process_notes_1): Expect REG_EQUAL to be an EXPR_LIST
1136 rather than an INSN_LIST. Handle INT_LIST.
1137 * ifcvt.c (cond_exec_process_insns): Take the probability as an int
1138 rather than an rtx. Use gen_rtx_INT_LIST to create a REG_BR_PROB note.
1139 (cond_exec_process_if_block): Use XINT to extract REG_BR_PROB values.
1140 Manipulate them as ints rather than rtxes.
1141 * reg-stack.c (subst_asm_stack_regs): Only handle EXPR_LIST notes.
1142 * regmove.c (copy_src_to_dest): Likewise.
1143 * sched-vis.c (print_insn_with_notes): Handle INT_LIST.
1145 * config/i386/winnt.c (i386_pe_seh_unwind_emit): Sink pat assignment
1146 into the cases that need it.
1147 * config/arm/arm.c (arm_unwind_emit): Likewise.
1149 * asan.c (asan_clear_shadow): Use add_int_reg_note for REG_BR_PROB.
1150 * emit-rtl.c (try_split, emit_copy_of_insn_after): Likewise.
1151 * loop-doloop.c (add_test, doloop_modify): Likewise.
1152 * loop-unswitch.c (compare_and_jump_seq): Likewise.
1153 * optabs.c (emit_cmp_and_jump_insn_1): Likewise.
1154 * predict.c (combine_predictions_for_insn): Likewise.
1155 * print-rtl.c (print_rtx): Handle INT_LIST.
1156 * config/aarch64/aarch64.c (aarch64_emit_unlikely_jump): Likewise.
1157 * config/alpha/alpha.c (emit_unlikely_jump): Likewise.
1158 * config/arm/arm.c (emit_unlikely_jump): Likewise.
1159 * config/i386/i386.c (ix86_expand_split_stack_prologue): Likewise.
1160 (ix86_split_fp_branch, predict_jump): Likewise.
1161 * config/rs6000/rs6000.c (emit_unlikely_jump): Likewise.
1162 * config/sh/sh.c (expand_cbranchsi4): Likewise.
1163 * config/spu/spu.c (ea_load_store_inline): Likewise.
1165 * cfgbuild.c (compute_outgoing_frequencies): Use XINT to access the
1166 value of a REG_BR_PROB note.
1167 * cfgrtl.c (force_nonfallthru_and_redirect): Likewise.
1168 (update_br_prob_note, rtl_verify_edges, purge_dead_edges): Likewise.
1169 * emit-rtl.c (try_split): Likewise.
1170 * predict.c (br_prob_note_reliable_p): Likewise.
1171 (invert_br_probabilities, combine_predictions_for_insn): Likewise.
1172 * reorg.c (mostly_true_jump): Likewise.
1173 * config/bfin/bfin.c (cbranch_predicted_taken_p): Likewise.
1174 * config/frv/frv.c (frv_print_operand_jump_hint): Likewise.
1175 * config/i386/i386.c (ix86_print_operand): Likewise.
1176 * config/ia64/ia64.c (ia64_print_operand): Likewise.
1177 * config/mmix/mmix.c (mmix_print_operand): Likewise.
1178 * config/rs6000/rs6000.c (output_cbranch): Likewise.
1179 * config/s390/s390.c (s390_expand_tbegin): Likewise.
1180 * config/sh/sh.c (sh_print_operand, sh_adjust_cost): Likewise.
1181 * config/sparc/sparc.c (output_cbranch): Likewise.
1182 * config/spu/spu.c (get_branch_target): Likewise.
1183 * config/tilegx/tilegx.c (cbranch_predicted_p): Likewise.
1184 * config/tilepro/tilepro.c (cbranch_predicted_p): Likewise.
1186 2013-09-23 Jan Hubicka <jh@suse.cz>
1188 * ipa-cp.c (ipa_get_indirect_edge_target_1): Add sanity check
1190 * ipa-utils.h (possible_polymorphic_call_target_p): New function.
1191 * ipa-devirt.c (possible_polymorphic_call_target_p): Be tolerant
1193 * gimple-fold.c: Include ipa-utils.h and gimple-pretty-print.h
1194 (gimple_fold_call): Dump inconsistent devirtualizations; add
1195 sanity check for type based devirtualizations.
1196 * ipa-prop.c: Include ipa-utils.h
1197 (ipa_intraprocedural_devirtualization): Add sanity check.
1198 (try_make_edge_direct_virtual_call): Likewise.
1200 2013-09-23 Eric Botcazou <ebotcazou@adacore.com>
1202 * tree-ssa-ccp.c (insert_clobber_before_stack_restore): Recurse on copy
1203 assignment statements.
1205 2013-09-23 Kugan Vivekanandarajah <kuganv@linaro.org>
1207 * gimple-pretty-print.c (dump_ssaname_info): New function.
1208 (dump_gimple_phi): Call it.
1209 (pp_gimple_stmt_1): Likewise.
1210 * tree-core.h (tree_ssa_name): New union ssa_name_info_type field.
1211 (range_info_def): Declare.
1212 * tree-pretty-print.c (pp_double_int): New function.
1213 (dump_generic_node): Call it.
1214 * tree-pretty-print.h (pp_double_int): Declare.
1215 * tree-ssa-alias.c (dump_alias_info): Check pointer type.
1216 * tree-ssanames.h (range_info_def): New structure.
1217 (value_range_type): Move definition here.
1218 (set_range_info, value_range_type, duplicate_ssa_name_range_info):
1220 * tree-ssanames.c (make_ssa_name_fn): Check pointer type at
1222 (set_range_info): New function.
1223 (get_range_info): Likewise.
1224 (duplicate_ssa_name_range_info): Likewise.
1225 (duplicate_ssa_name_fn): Check pointer type and call
1226 duplicate_ssa_name_range_info.
1227 * tree-ssa-copy.c (fini_copy_prop): Likewise.
1228 * tree-vrp.c (value_range_type): Remove definition, now in
1230 (vrp_finalize): Call set_range_info to update value range of SSA_NAMEs.
1231 * tree.h (SSA_NAME_PTR_INFO): Macro changed to access via union.
1232 (SSA_NAME_RANGE_INFO): New macro.
1234 2013-09-23 Richard Biener <rguenther@suse.de>
1236 PR tree-optimization/58464
1237 * tree-ssa-pre.c (phi_trans_lookup): Remove.
1238 (phi_trans_add): Change to add conditionally on being not present.
1239 (phi_translate_1): Remove recursion detection here.
1240 (phi_translate): Pre-seed the cache with NULL to catch
1241 recursion here in a more generic way.
1242 (bitmap_find_leader): Adjust comment.
1243 (get_representative_for): Dump value-numbers.
1244 (create_expression_by_pieces): Likewise.
1245 (insert_into_preds_of_block): Likewise.
1247 2013-09-23 Christian Bruel <christian.bruel@st.com>
1250 * config/sh/sh.md (movsf_ie): Allow fpul_operand.
1251 * config/sh/predicate.md (arith_reg_operand): Disallow FPUL_REG.
1253 2013-09-23 James Greenhalgh <james.greenhalgh@arm.com>
1256 2013-09-20 Renlin Li <renlin.li@arm.com>
1258 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use
1260 (aarch64_expand_epilogue): Likewise.
1261 (aarch64_legitimize_reload_address): Likewise.
1263 2013-09-22 Eric Botcazou <ebotcazou@adacore.com>
1265 * gimplify.c (gimplify_asm_expr): Reset the TREE_CHAIN of clobbers to
1266 NULL_TREE before pushing them onto the vector. Likewise for labels.
1268 2013-09-21 Eric Botcazou <ebotcazou@adacore.com>
1270 * config/ia64/predicates.md (ia64_cbranch_operator): Accept unordered
1271 comparison operators when -fno-trapping-math is in effect.
1272 * config/ia64/ia64.c (ia64_expand_compare): Add support for unordered
1273 comparison operators in TFmode and assert that unsupported operators
1275 (ia64_print_operand): Likewise.
1277 2013-09-21 Jan Hubicka <jh@suse.cz>
1279 * x86-tune.def (partial_reg_stall): Disable for CoreI7 and newer.
1280 (sse_typeless_stores): Enable for core
1281 (sse_load0_by_pxor): Likewise.
1282 (four_jump_limit): Disable for core.
1283 (pad_returns): Likewise.
1284 (avoid_vector_decode): Likewise.
1285 (fuse_cmp_and_branch): Enable for cores.
1286 * i386.c (x86_accumulate_outgoing_args): Disable for cores.
1288 2013-09-20 John David Anglin <danglin@gcc.gnu.org>
1291 * config/pa/pa.c (pa_option_override): Disable auto increment and
1292 decrement instructions until reload is completed.
1294 * config/pa/pa-linux.h (TARGET_OS_CPP_BUILTINS): Define
1295 __GCC_HAVE_SYNC_COMPARE_AND_SWAP_1, __GCC_HAVE_SYNC_COMPARE_AND_SWAP_2,
1296 and __GCC_HAVE_SYNC_COMPARE_AND_SWAP_4.
1298 2013-09-20 DJ Delorie <dj@redhat.com>
1299 Nick Clifton <nickc@redhat.com>
1301 * config/rl78/rl78.c: Various whitespace and comment tweaks.
1302 (need_to_save): Save bank 0 on interrupts.
1303 (characterize_address): Strip far address wrappers.
1304 (rl78_as_legitimate_address): Likewise.
1305 (transcode_memory_rtx): Likewise.
1306 (rl78_peep_movhi_p): Disable this peephole after devirt.
1307 (rl78_propogate_register_origins): Forget all origins when a
1309 * config/rl78/rl78-virt.md: Various whitespace tweaks.
1310 * config/rl78/rl78-real.md: Various whitespace tweaks. Additional
1312 * config/rl78/rl78.md (sel_rb): Disable for G10 just in case.
1313 * config/rl78/rl78-expand.md (movqi): Check for subregs of consts.
1314 * config/rl78/rl78.h (LINK_SPEC): Pass -gc-sections unless
1316 * config/rl78/constraints.md: Various whitespace and paren tweaks.
1318 2013-09-20 John David Anglin <danglin@gcc.gnu.org>
1320 * config/pa/pa.md: In "scc" insn patterns, change output template to
1321 handle const0_rtx in reg_or_0_operand operands.
1323 2013-09-20 Martin Husemann <martin@NetBSD.org>
1326 * config/vax/vax.c (vax_output_int_move): Use D format specifier.
1327 * config/vax/vax.md (ashldi3, <unnamed>): Ditto.
1329 2013-09-20 Richard Biener <rguenther@suse.de>
1332 * tree-scalar-evolution.c (struct scev_info_str): Shrink by
1333 remembering SSA name version and block index.
1334 (new_scev_info_str): Adjust.
1335 (hash_scev_info): Likewise. Also hash the block index.
1336 (eq_scev_info): Adjust.
1337 (find_var_scev_info): Likewise.
1338 (struct instantiate_cache_entry): Remove.
1339 (struct instantiate_cache_type): Use a htab to map name, block
1341 (instantiate_cache_type::~instantiate_cache_type): Adjust.
1342 (get_instantiated_value_entry): Likewise.
1343 (hash_idx_scev_info, eq_idx_scev_info): New functions.
1344 (instantiate_scev_name): Adjust.
1346 2013-09-20 Jeff Law <law@redhat.com>
1348 * tree-ssa-dom.c (record_temporary_equivalences): Add comment.
1350 2013-09-20 Yufeng Zhang <yufeng.zhang@arm.com>
1352 * config/aarch64/aarch64-builtins.c (aarch64_simd_expand_args):
1353 Call aarch64_simd_expand_args to update op[argc].
1355 2013-09-20 Basile Starynkevitch <basile@starynkevitch.net>
1357 * plugin.c (parse_plugin_arg_opt): Accept equal sign inside
1360 2013-09-20 Basile Starynkevitch <basile@starynkevitch.net>
1362 * gengtype.c (file_rules): Added rule for *.cc files.
1363 (get_output_file_with_visibility): Give fatal message when no
1366 2013-09-20 Renlin Li <renlin.li@arm.com>
1368 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use plus_constant.
1369 (aarch64_expand_epilogue): Likewise.
1370 (aarch64_legitimize_reload_address): Likewise.
1372 2013-09-20 Bernd Edlinger <bernd.edlinger@hotmail.de>
1375 * expr.c (expand_assignment): Remove misalignp code path.
1377 2013-09-20 Marek Polacek <polacek@redhat.com>
1380 * ubsan.c (get_ubsan_type_info_for_type): Use TYPE_SIZE instead of
1381 TYPE_PRECISION. Add asserts.
1383 2013-09-20 Richard Biener <rguenther@suse.de>
1385 PR tree-optimization/58453
1386 * tree-loop-distribution.c (distribute_loop): Apply the cost
1387 model for -ftree-loop-distribute-patterns, too.
1389 2013-09-20 Richard Biener <rguenther@suse.de>
1392 * tree-chrec.h (build_polynomial_chrec): Use gcc_checking_assert,
1393 make type comparison less strict.
1395 2013-09-20 Alan Modra <amodra@gmail.com>
1397 * configure: Regenerate.
1398 * aclocal.m4: Regenerate.
1400 2013-09-20 Marek Polacek <polacek@redhat.com>
1403 * doc/extend.texi: Document that attribute used is meant to be used
1404 on variables with static storage duration.
1406 2013-09-19 Jakub Jelinek <jakub@redhat.com>
1408 PR tree-optimization/58472
1409 * tree-vect-stmts.c (vectorizable_store, vectorizable_load): For
1410 simd_lane_access set inv_p = false.
1411 * omp-low.c (lower_rec_input_clauses): Set TREE_NO_WARNING on
1412 the simduid magic VAR_DECL.
1414 2013-09-19 Jan Hubicka <jh@suse.cz>
1416 * i386.c (generic_memcpy, generic_memset): Fix 32bit template.
1418 2013-09-17 Jeff Law <law@redhat.com>
1420 * tree-ssa-dom.c (record_temporary_equivalences): New function
1421 split out of dom_opt_dom_walker::after_dom_children.
1422 (dom_opt_dom_walker::thread_across_edge): Move common code
1423 in here from dom_opt_dom_walker::after_dom_children.
1424 (dom_opt_dom_walker::after_dom_children): Corresponding simplifictions.
1426 2013-09-19 Jan Hubicka <jh@suse.cz>
1428 * i386.h (TARGET_GENERIC32, TARGET_GENERIC64): Remove.
1429 (TARGET_GENERIC): Use PROCESOR_GENERIC
1430 (enum processor_type): Unify generic32 and 64.
1431 * i386.md (cpu): Likewise.
1432 * x86-tune.def (use_leave): Enable for generic32.
1433 (avoid_vector_decode, slow_imul_imm32_mem, slow_imul_imm8): Likewise.
1434 * athlon.md: Change generic64 to generic in all occurences.
1435 * i386-c.c (ix86_target_macros_internal): Unify generic64 and 32.
1436 (ix86_target_macros_internal): Likewise.
1437 * driver-i386.c (host_detect_local_cpu): Likewise.
1438 * i386.c (generic64_memcpy, generic64_memset, generic64_cost): Rename
1440 (generic_memcpy, generic_memset, generic_cost): This one.
1441 (generic32_memcpy, generic32_memset, generic32_cost): Remove.
1442 (m_GENERIC32, m_GENERIC64): Remove.
1443 (m_GENERIC): Turn into one flag.
1444 (processor_target): Unify generic tunnings.
1445 (ix86_option_override_internal): Replace generic32/64 by generic.
1446 (ix86_issue_rate): Likewise.
1447 (ix86_adjust_cost): Likewise.
1449 2013-09-19 Jan Hubicka <jh@suse.cz>
1451 * cgraph.c (cgraph_create_edge_1): Avoid uninitialized read
1452 of speculative flag.
1454 2013-09-19 Jakub Jelinek <jakub@redhat.com>
1456 * omp-low.c (expand_omp_sections): Always pass len - 1 to
1457 GOMP_sections_start, even if !exit_reachable.
1459 2013-09-18 Vladimir Makarov <vmakarov@redhat.com>
1461 * lra-constraints.c (need_for_all_save_p): Use macro
1462 HARD_REGNO_CALL_PART_CLOBBERED.
1463 * lra-lives.c (check_pseudos_live_through_calls): Use the macro to
1464 set up pseudo conflict hard regs.
1466 2013-09-18 Michael Meissner <meissner@linux.vnet.ibm.com>
1469 * config/rs6000/paired.md (movmisalignv2sf): Fix to allow memory
1472 2013-09-18 Vladimir Makarov <vmakarov@redhat.com>
1474 PR rtl-optimization/58438
1475 * lra.c (lra): Clear lra_optional_reload_pseudos in upper loop.
1476 * lra-constraints.c (undo_optional_reloads): Keep optional reloads
1477 from previous subpasses.
1479 2013-09-18 Richard Earnshaw <rearnsha@arm.com>
1481 * arm.c (arm_get_frame_offsets): Validate architecture supports
1482 LDRD/STRD before accepting the tuning preference.
1483 (arm_expand_prologue): Likewise.
1484 (arm_expand_epilogue): Likewise.
1486 2013-09-18 Richard Biener <rguenther@suse.de>
1488 PR tree-optimization/58417
1489 * tree-chrec.c (chrec_fold_plus_1): Assert that we do not
1490 have chrecs with symbols defined in the loop as operands.
1491 (chrec_fold_multiply): Likewise.
1492 * tree-scalar-evolution.c (interpret_rhs_expr): Instantiate
1493 parameters before folding binary operations.
1494 (struct instantiate_cache_entry_hasher): Remove.
1495 (struct instantiate_cache_type): Use a pointer-map.
1496 (instantiate_cache_type::instantiate_cache_type): New function.
1497 (instantiate_cache_type::get): Likewise.
1498 (instantiate_cache_type::set): Likewise.
1499 (instantiate_cache_type::~instantiate_cache_type): Adjust.
1500 (get_instantiated_value_entry): Likewise.
1501 (global_cache): New global.
1502 (instantiate_scev_r, instantiate_scev_poly, instantiate_scev_binary,
1503 instantiate_array_ref, instantiate_scev_convert, instantiate_scev_3,
1504 instantiate_scev_2, instantiate_scev_1): Do not pass along cache.
1505 (instantiate_scev_name): Adjust.
1506 (instantiate_scev): Construct global instead of local cache.
1507 (resolve_mixers): Likewise.
1509 2013-09-18 Daniel Morris <danielm@ecoscentric.com>
1510 Paolo Carlini <paolo.carlini@oracle.com>
1513 * doc/implement-cxx.texi: Fix references to the C++ standards.
1515 2013-09-18 Jakub Jelinek <jakub@redhat.com>
1517 * omp-low.c (copy_var_decl): Copy DECL_ATTRIBUTES.
1518 * tree-vect-data-refs.c (vect_analyze_data_refs): For
1519 simd_lane_access drs, update also DR_ALIGNED_TO.
1521 2013-09-18 Marek Polacek <polacek@redhat.com>
1524 * doc/extend.texi: Document no_sanitize_undefined attribute.
1525 * builtins.c (fold_builtin_0): Don't sanitize function if it has the
1526 no_sanitize_undefined attribute.
1528 2013-09-18 Nick Clifton <nickc@redhat.com>
1530 * config/msp430/msp430.h (ASM_SPEC): Pass -md on to the assembler.
1531 (ASM_DECLARE_FUNCTION_NAME): Define.
1533 2013-09-17 Trevor Saunders <tsaunders@mozilla.com>
1535 * compare-elim.c (find_comparison_dom_walker): New class
1536 (find_comparisons_in_bb): Rename to
1537 find_comparison_dom_walker::before_dom_children
1538 (find_comparisons): Adjust
1539 * domwalk.c (walk_dominator_tree): Rename to dom_walker::walk, and
1541 (init_walk_dominator_tree, fini_walk_dominator_tree): Remove
1542 * domwalk.h (dom_walk_data): Convert it To a class dom_walker.
1543 (init_walk_dominator_tree): Remove declaration.
1544 (fini_walk_dominator_tree): Remove declaration.
1545 * fwprop.c (single_def_use_dom_walker): New class
1546 (single_def_use_enter_block): Convert to
1547 single_def_use_dom_walker::before_dom_children.
1548 (single_def_use_leave_block): Convert to
1549 single_def_use_dom_walker::after_dom_children.
1550 (build_single_def_use_links): Adjust.
1551 * gimple-ssa-strength-reduction.c (find_candidates_dom_walker): New
1553 (find_candidates_in_block): Convert to
1554 find_candidates_dom_walker::before_dom_children.
1555 (execute_strength_reduction): Adjust.
1556 * graphite-sese-to-poly.c (struct bsc, build_sese_conditions): Remove.
1557 (sese_dom_walker): New class.
1558 (sese_dom_walker::sese_dom_walker): New constructor.
1559 (sese_dom_walker::~sese_dom_walker): New destructor.
1560 (build_sese_conditions_before): Convert to
1561 sese_dom_walker::before_dom_children.
1562 (build_sese_conditions_after): Convert to
1563 sese_dom_walker::after_dom_children.
1564 (build_poly_scop): Adjust
1565 * tree-into-ssa.c (rewrite_dom_walker): New class
1566 (rewrite_enter_block): Convert to
1567 rewrite_dom_walker::before_dom_children.
1568 (rewrite_leave_block): Convert to
1569 rewrite_dom_walker::after_dom_children.
1570 (rewrite_update_dom_walker): New class.
1571 (rewrite_update_enter_block): Convert to
1572 rewrite_update_dom_walker::before_dom_children.
1573 (rewrite_update_leave_block): Convert to
1574 rewrite_update_dom_walker::after_dom_children.
1575 (rewrite_blocks, rewrite_into_ssa): Adjust.
1576 (mark_def_dom_walker): New class.
1577 (mark_def_dom_walker::mark_def_dom_walker): New constructor.
1578 (mark_def_dom_walker::~mark_def_dom_walker): New destructor.
1579 (mark_def_sites_blocks): Convert to
1580 mark_def_dom_walker::before_dom_children.
1581 (mark_def_site_blocks): Remove.
1582 * tree-ssa-dom.c (dom_opt_dom_walker): New class.
1583 (tree_ssa_dominator_optimize): Adjust.
1584 (dom_thread_across_edge): Convert to method
1585 dom_opt_dom_walker::thread_across_edge.
1586 (dom_opt_enter_block): Convert to member function
1587 dom_opt_dom_walker::before_dom_children.
1588 (dom_opt_leave_block): Convert to member function
1589 dom_opt_dom_walker::after_dom_children.
1590 * tree-ssa-dse.c (dse_dom_walker): New class.
1591 (dse_enter_block): Convert to member function
1592 dse_dom_walker::before_dom_children.
1593 (tree_ssa_dse): Adjust.
1594 * tree-ssa-loop-im.c (invariantness_dom_walker): New class.
1595 (determine_invariantness_stmt): Convert to method
1596 invariantness_dom_walker::before_dom_children.
1597 (determine_invariantness): Remove
1598 (move_computations_dom_walker): New class.
1599 (move_computations_stmt): Convert to method
1600 move_computations_dom_walker::before_dom_children.
1601 (move_computations, tree_ssa_lim): Adjust.
1602 * tree-ssa-phiopt.c (nontrapping_dom_walker): new class
1603 (nt_init_block): Make method
1604 notrappping_dom_walker::before_dom_children.
1605 (nt_fini_block): Make
1606 method nontrapping_dom_walker::after_dom_children.
1607 (get_non_trapping): Adjust.
1608 * tree-ssa-pre.c (eliminate_dom_walker): New class.
1609 (eliminate_bb): Make method eliminate_dom_walker::before_dom_children.
1610 (eliminate_leave_block): Make method.
1611 eliminate_dom_walker::after_dom_children.
1613 * tree-ssa-strlen.c (strlen_dom_walker): New class.
1614 (strlen_enter_block): Make method
1615 strlen_dom_walker::before_dom_children.
1616 (strlen_leave_block): Make
1617 method strlen_dom_walker::after_dom_children.
1618 (tree_ssa_strlen): Adjust.
1619 * tree-ssa-uncprop.c (uncprop_dom_walker): New class.
1620 (tree_ssa_uncprop): Adjust.
1621 (uncprop_leave_block): Make method
1622 uncprop_dom_walker::after_dom_children.
1623 (uncprop_leave_block): Make method
1624 uncprop_dom_walker::before_dom_children.
1626 2013-09-18 Bin Cheng <bin.cheng@arm.com>
1628 * config/arm/arm.c (thumb1_reorg): Search for flag setting insn before
1629 branch in same basic block. Check both src and dest of the move insn.
1631 2013-09-17 Nick Clifton <nickc@redhat.com>
1633 * config/rl78/rl78-real.md (bf): New pattern.
1635 * config/rl78/rl78.c (rl78_print_operand_1): Handle %B.
1636 (rl78_print_operand): Do not put a # before a %B.
1637 * config/rl78/rl78.opt: Tweak doc strings.
1639 2013-09-17 DJ Delorie <dj@redhat.com>
1641 * config/rl78/constraints.md (Wcv): Allow up to $r31.
1642 * config/rl78/rl78.c (rl78_asm_file_start: Likewise.
1643 (rl78_option_override): Likewise, if -mallregs.
1644 (is_virtual_register): Likewise.
1645 * config/rl78/rl78.h (reg_class): Extend VREGS to $r31.
1646 (REGNO_OK_FOR_BASE_P): Likewise.
1647 * config/rl78/rl78.opt (-mallregs): New.
1649 2013-09-17 Nick Clifton <nickc@redhat.com>
1651 * config/rl78/rl78.c (need_to_save): Change return type to bool.
1652 For interrupt functions: save all call clobbered registers if the
1653 interrupt handler is not a leaf function.
1654 (rl78_expand_prologue): Always recompute the frame information.
1655 For interrupt functions: only select bank 0 if one of the bank 0
1656 registers is going to be psuhed.
1658 2013-09-17 DJ Delorie <dj@redhat.com>
1660 * config/rl78/constraints.md: For each W* constraint, rename to C*
1661 and create a W* constraint that checks for an optional ES: prefix
1663 * config/rl78/rl78.md (UNS_ES_ADDR): New.
1664 (es_addr): New. Used to wrap far addresses.
1665 * config/rl78/rl78-protos.h (rl78_es_addr): New.
1666 (rl78_es_base): New.
1667 * config/rl78/rl78.c (rl78_as_legitimate_address): Accept "unspec"
1668 wrapped far addresses.
1669 (rl78_print_operand_1): Unwrap far addresses before processing.
1670 (rl78_lo16): Wrap far addresses in unspecs.
1671 (rl78_es_addr): New.
1672 (rl78_es_base): New.
1673 (insn_ok_now): Check for not-yet-wrapped far addresses.
1674 (transcode_memory_rtx): Properly re-wrap far addresses.
1676 2013-09-17 Sebastian Huber <sebastian.huber@embedded-brains.de>
1678 * config/sparc/t-rtems: Add leon3 multilibs.
1680 2013-09-17 Cong Hou <congh@google.com>
1682 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Fix a bug
1683 when checking the dot production pattern. The type of rhs operand
1684 of multiply is now checked correctly.
1686 2013-09-17 Jeff Law <law@redhat.com>
1688 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
1689 edge implied equivalences into successor phis.
1690 * tree-ssa-threadupdate.c (phi_args_equal_on_edges): Moved into
1691 here from tree-ssa-threadedge.c.
1692 (mark_threaded_blocks): When threading through a joiner, if both
1693 successors of the joiner's clone reach the same block, verify the
1694 PHI arguments are equal. If not, cancel the jump threading request.
1695 * tree-ssa-threadedge.c (phi_args_equal_on_edges): Moved into
1696 tree-ssa-threadupdate.c
1697 (thread_across_edge): Don't check PHI argument equality when
1698 threading through joiner block here.
1700 2013-09-17 Andrew MacLeod <amacleod@redhat.com>
1702 * tree-flow.h (ssa_undefined_value_p): Remove prototype.
1703 * tree-ssa.c (ssa_undefined_value_p): Move pass independent parts here.
1704 (warn_uninit, warn_uninitialized_vars,
1705 execute_early_warn_uninitialized, make_pass_early_warn_uninitialized):
1706 Move to tree-ssa-uninit.c.
1707 * tree-ssa-uninit.c (ssa_undefined_value_p): Move to tree-ssa.c
1708 (has_undefined_value_p): New. Pass dependant parts of
1709 ssa_undefined_value_p.
1710 (uninit_undefined_value_p): Use has_undefined_value_p.
1711 (warn_uninit, warn_uninitialized_vars,
1712 execute_early_warn_uninitialized, make_pass_early_warn_uninitialized):
1713 Move from tree-ssa.c.
1714 * tree-ssa.h: Adjust prototypes.
1716 2013-09-17 Jan Hubicka <jh@suse.cz>
1719 * cif-code.def (FUNCTION_NOT_OPTIMIZED): New CIF code.
1720 * ipa-inline.c (can_inline_edge_p): Do not downgrade
1721 FUNCTION_NOT_OPTIMIZED.
1722 * ipa-inline-analysis.c (compute_inline_parameters): Function
1723 not optimized is not inlinable unless it is alwaysinline.
1724 (inline_analyze_function): Force calls in not optimized
1725 function not inlinable.
1727 2013-09-17 Jan Hubicka <jh@suse.cz>
1730 * ipa-devirt.c (ipa_devirt): Be ready for symtab_nonoverwritable_alias
1732 * ipa.c (function_and_variable_visibility): Likewise.
1733 * ipa-profile.c (ipa_profile): Likewise.
1735 2013-09-17 Bernd Edlinger <bernd.edlinger@hotmail.de>
1738 * cgraph.c (cgraph_function_body_availability): Check for ifunc
1739 attribute, and don't inline the resolver in this case.
1741 2013-09-17 Teresa Johnson <tejohnson@google.com>
1743 * coverage.c (get_coverage_counts): Add missing newline.
1745 2013-09-17 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1747 PR tree-optimization/58088
1748 * fold-const.c (mask_with_trailing_zeros): New function.
1749 (fold_binary_loc): Make sure we don't recurse infinitely
1750 when the X in (X & C1) | C2 is a tree of the form (Y * K1) & K2.
1751 Use mask_with_trailing_zeros where appropriate.
1753 2013-09-17 Yuri Rumyantsev <ysrumyan@gmail.com>
1755 * config/i386/i386.c (distance_agu_use_in_bb) : Proper initialization
1756 of 'prev' var to get better distance estimation.
1758 2013-09-17 Eric Botcazou <ebotcazou@adacore.com>
1760 * tree-inline.h (struct copy_body_data): Add transform_parameter.
1761 * tree-inline.c (is_parameter_of): New predicate.
1762 (remap_gimple_op_r): Do not propagate TREE_THIS_NOTRAP on MEM_REF if
1763 a parameter has been remapped.
1764 (copy_tree_body_r): Likewise on INDIRECT_REF and MEM_REF.
1765 (optimize_inline_calls): Initialize transform_parameter.
1766 (copy_gimple_seq_and_replace_locals): Likewise.
1767 (tree_function_versioning): Likewise.
1768 (maybe_inline_call_in_expr): Likewise.
1770 2013-09-17 Nick Clifton <nickc@redhat.com>
1772 * config/msp430/msp430-protos.h: Add prototypes for new functions.
1773 * config/msp430/msp430.c (msp430_preserve_reg_p): Add support for
1775 (is_attr_func): New function.
1776 (msp430_is_interrupt_func): New function.
1777 (is_naked_func): New function.
1778 (is_reentrant_func): New function.
1779 (is_critical_func): New function.
1780 (msp430_start_function): Add annotations for function attributes.
1781 (msp430_attr): New function.
1782 (msp430_attribute_table): New.
1783 (msp430_function_section): New function.
1784 (TARGET_ASM_FUNCTION_SECTION): Define.
1785 (msp430_builtin): New enum.
1786 (msp430_init_builtins): New function.
1787 (msp430_builtin_devl): New function.
1788 (msp430_expand_builtin): New function.
1789 (TARGET_INIT_BUILTINS): Define.
1790 (TARGET_EXPAND_BUILTINS): Define.
1791 (TARGET_BUILTIN_DECL): Define.
1792 (msp430_expand_prologue): Add support for naked, interrupt,
1793 critical and reentrant functions.
1794 (msp430_expand_epilogue): Likewise.
1795 (msp430_print_operand): Handle 'O' character.
1796 * config/msp430/msp430.h (TARGET_CPU_CPP_BUILTINS): Define
1798 * config/msp430/msp430.md (unspec): Add UNS_DINT, UNS_EINT,
1799 UNS_PUSH_INTR, UNS_POP_INTR, UNS_BIC_SR, UNS_BIS_SR.
1800 (pushm): Use a 'n' rather than an 'i' constraint.
1801 (msp_return): Add generation of the interrupt return instruction.
1802 (disable_interrupts): New pattern.
1803 (enable_interrupts): New pattern.
1804 (push_intr_state): New pattern.
1805 (pop_intr_state): New pattern.
1806 (bic_SR): New pattern.
1807 (bis_SR): New pattern.
1808 * doc/extend.texi: Document MSP430 function attributes and builtin
1811 2013-09-17 Richard Biener <rguenther@suse.de>
1813 PR tree-optimization/58432
1814 * tree-loop-distribution.c (tree_loop_distribution): Also
1815 scan PHIs for outside loop uses and seed a partition from them.
1817 2013-09-17 Bin Cheng <bin.cheng@arm.com>
1819 * gimple-ssa-strength-reduction.c (backtrace_base_for_ref): New.
1820 (restructure_reference): Call backtrace_base_for_ref.
1822 2013-09-17 Alan Modra <amodra@gmail.com>
1825 * config/rs6000/driver-rs6000.c (elf_platform): Revert 2013-06-11
1828 2013-09-16 DJ Delorie <dj@redhat.com>
1830 * config/rl78/rl78.c (rl78_asm_file_start): Specify alternate
1831 vregs location for RL78/G10.
1832 (rl78_expand_prologue): Avoid SEL on G10.
1833 (rl78_expand_epilogue): Likewise.
1834 (rl78_peep_movhi_p): Can't move a constant to memory in HImode.
1835 * config/rl78/rl78.h (TARGET_CPU_CPP_BUILTINS): Define
1836 __RL78_G10__ when appropriate.
1837 (ASM_SPEC): Pass -mg10 along to the assembler.
1838 * config/rl78/rl78.md (sel_rb): Disable for G10.
1839 * config/rl78/rl78.opt: Add -mg10 option.
1840 * config/rl78/t-rl78: Add -mg10 multilib.
1842 2013-09-16 Xinliang David Li <davidxl@google.com>
1844 * tree-if-conv.c (main_tree_if_conversion): Check new flag.
1845 * omp-low.c (omp_max_vf): Ditto.
1846 (expand_omp_simd): Ditto.
1847 * tree-vectorizer.c (vectorize_loops): Ditto.
1848 (gate_vect_slp): Ditto.
1849 (gate_increase_alignment): Ditto.
1850 * tree-ssa-pre.c (inhibit_phi_insertion): Ditto.
1851 * tree-ssa-loop.c (gate_tree_vectorize): Ditto.
1852 (gate_tree_vectorize): Name change.
1853 (tree_vectorize): Ditto.
1854 (pass_vectorize::gate): Call new function.
1855 (pass_vectorize::execute): Ditto.
1856 * opts.c: O3 default setting change.
1857 (finish_options): Check new flag.
1858 * doc/invoke.texi: Document new flags.
1859 * common.opt: New flags.
1861 2013-09-16 Andreas Schwab <schwab@linux-m68k.org>
1863 * doc/tm.texi.in (Cond Exec Macros): Remove node.
1864 (Condition Code): Don't reference it.
1865 * doc/tm.texi: Regenerate.
1867 2013-09-16 Vladimir Makarov <vmakarov@redhat.com>
1870 * lra-constraints.c (undo_optional_reloads): Consider all optional
1871 reload even if it did not get a hard reg.
1873 2013-09-16 Teresa Johnson <tejohnson@google.com>
1875 * dumpfile.c (dump_loc): Remove newline emission.
1876 * tree-vect-data-refs.c (vect_lanes_optab_supported_p): Add newline
1877 emission to dump_printf_loc calls where missing.
1878 (vect_mark_for_runtime_alias_test): Ditto.
1879 (vect_analyze_data_ref_dependence): Ditto.
1880 (vect_analyze_data_ref_dependences): Ditto.
1881 (vect_slp_analyze_data_ref_dependence): Ditto.
1882 (vect_slp_analyze_data_ref_dependences): Ditto.
1883 (vect_compute_data_ref_alignment): Ditto.
1884 (vect_update_misalignment_for_peel): Ditto.
1885 (vect_verify_datarefs_alignment): Ditto.
1886 (vector_alignment_reachable_p): Ditto.
1887 (vect_get_data_access_cost): Ditto.
1888 (vect_enhance_data_refs_alignment): Ditto.
1889 (vect_find_same_alignment_drs): Ditto.
1890 (vect_analyze_data_refs_alignment): Ditto.
1891 (vect_analyze_group_access): Ditto.
1892 (vect_analyze_data_ref_access): Ditto.
1893 (vect_analyze_data_ref_accesses): Ditto.
1894 (vect_prune_runtime_alias_test_list): Ditto.
1895 (vect_analyze_data_refs): Ditto.
1896 (vect_create_addr_base_for_vector_ref): Ditto.
1897 (vect_create_data_ref_ptr): Ditto.
1898 (vect_grouped_store_supported): Ditto.
1899 (vect_grouped_load_supported): Ditto.
1900 * value-prof.c (check_counter): Ditto.
1901 (check_ic_target): Ditto.
1902 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Ditto.
1903 (vect_recog_widen_mult_pattern): Ditto.
1904 (vect_recog_widen_sum_pattern): Ditto.
1905 (vect_recog_over_widening_pattern): Ditto.
1906 (vect_recog_widen_shift_pattern): Ditto.
1907 (vect_recog_rotate_pattern): Ditto.
1908 (vect_recog_vector_vector_shift_pattern): Ditto.
1909 (vect_recog_divmod_pattern): Ditto.
1910 (vect_recog_mixed_size_cond_pattern): Ditto.
1911 (vect_recog_bool_pattern): Ditto.
1912 (vect_pattern_recog_1): Ditto.
1913 (vect_pattern_recog): Ditto.
1914 * tree-vect-loop.c (vect_determine_vectorization_factor): Ditto.
1915 (vect_is_simple_iv_evolution): Ditto.
1916 (vect_analyze_scalar_cycles_1): Ditto.
1917 (vect_get_loop_niters): Ditto.
1918 (vect_analyze_loop_1): Ditto.
1919 (vect_analyze_loop_form): Ditto.
1920 (vect_analyze_loop_operations): Ditto.
1921 (vect_analyze_loop_2): Ditto.
1922 (vect_analyze_loop): Ditto.
1923 (report_vect_op): Ditto.
1924 (vect_is_slp_reduction): Ditto.
1925 (vect_is_simple_reduction_1): Ditto.
1926 (vect_get_known_peeling_cost): Ditto.
1927 (vect_estimate_min_profitable_iters): Ditto.
1928 (vect_model_reduction_cost): Ditto.
1929 (vect_model_induction_cost): Ditto.
1930 (get_initial_def_for_induction): Ditto.
1931 (vect_create_epilog_for_reduction): Ditto.
1932 (vectorizable_reduction): Ditto.
1933 (vectorizable_induction): Ditto.
1934 (vectorizable_live_operation): Ditto.
1935 (vect_loop_kill_debug_uses): Ditto.
1936 (vect_transform_loop): Ditto.
1937 * tree-vect-stmts.c (vect_mark_relevant): Ditto.
1938 (vect_stmt_relevant_p): Ditto.
1939 (process_use): Ditto.
1940 (vect_mark_stmts_to_be_vectorized): Ditto.
1941 (vect_model_simple_cost): Ditto.
1942 (vect_model_promotion_demotion_cost): Ditto.
1943 (vect_model_store_cost): Ditto.
1944 (vect_get_store_cost): Ditto.
1945 (vect_model_load_cost): Ditto.
1946 (vect_get_load_cost): Ditto.
1947 (vect_init_vector_1): Ditto.
1948 (vect_get_vec_def_for_operand): Ditto.
1949 (vect_finish_stmt_generation): Ditto.
1950 (vectorizable_call): Ditto.
1951 (vectorizable_conversion): Ditto.
1952 (vectorizable_assignment): Ditto.
1953 (vectorizable_shift): Ditto.
1954 (vectorizable_operation): Ditto.
1955 (vectorizable_store): Ditto.
1956 (vectorizable_load): Ditto.
1957 (vectorizable_condition): Ditto.
1958 (vect_analyze_stmt): Ditto.
1959 (vect_transform_stmt): Ditto.
1960 (vect_is_simple_use): Ditto.
1961 * tree-vect-loop-manip.c (slpeel_make_loop_iterate_ntimes): Ditto.
1962 (vect_can_advance_ivs_p): Ditto.
1963 (vect_update_ivs_after_vectorizer): Ditto.
1964 (vect_do_peeling_for_loop_bound): Ditto.
1965 (vect_gen_niters_for_prolog_loop): Ditto.
1966 (vect_update_inits_of_drs): Ditto.
1967 (vect_create_cond_for_alias_checks): Ditto.
1968 * tree-vect-slp.c (vect_get_and_check_slp_defs): Ditto.
1969 (vect_build_slp_tree_1): Ditto.
1970 (vect_supported_load_permutation_p): Ditto.
1971 (vect_analyze_slp_instance): Ditto.
1972 (vect_analyze_slp): Ditto.
1973 (vect_make_slp_decision): Ditto.
1974 (vect_detect_hybrid_slp): Ditto.
1975 (vect_bb_vectorization_profitable_p): Ditto.
1976 (vect_slp_analyze_bb_1): Ditto.
1977 (vect_update_slp_costs_according_to_vf): Ditto.
1978 (vect_get_mask_element): Ditto.
1979 (vect_transform_slp_perm_load): Ditto.
1980 (vect_schedule_slp_instance): Ditto.
1981 (vect_schedule_slp): Ditto.
1982 (vect_slp_transform_bb): Ditto.
1983 * profile.c (read_profile_edge_counts): Ditto.
1984 (compute_branch_probabilities): Ditto.
1985 * coverage.c (get_coverage_counts): Ditto.
1987 2013-09-16 Diego Novillo <dnovillo@google.com>
1989 * tree-core.h: Add missing comment lines from refactoring of tree.h.
1991 2013-09-16 Jan Hubicka <jh@suse.cz>
1993 * gimple-fold.c (can_refer_decl_in_current_unit_p): Do not accept
1994 abstract functions; for static functions check the presence of body.
1996 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
1998 * config/aarch64/aarch64-simd-builtins.def (fma): New.
1999 * config/aarch64/aarch64-simd.md
2000 (aarch64_mla_elt<mode>): New.
2001 (aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
2002 (aarch64_mls_elt<mode>): Likewise.
2003 (aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
2004 (aarch64_fma4_elt<mode>): Likewise.
2005 (aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
2006 (aarch64_fma4_elt_to_128v2df): Likewise.
2007 (aarch64_fma4_elt_to_64df): Likewise.
2008 (fnma<mode>4): Likewise.
2009 (aarch64_fnma4_elt<mode>): Likewise.
2010 (aarch64_fnma4_elt_<vswap_width_name><mode>): Likewise.
2011 (aarch64_fnma4_elt_to_128v2df): Likewise.
2012 (aarch64_fnma4_elt_to_64df): Likewise.
2013 * config/aarch64/iterators.md (VDQSF): New.
2014 * config/aarch64/arm_neon.h
2015 (vfm<as><sdq>_lane<q>_f<32, 64>): Convert to C implementation.
2016 (vml<sa><q>_lane<q>_<fsu><16, 32, 64>): Likewise.
2018 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
2020 * config/aarch64/aarch64-simd.md (aarch64_mul3_elt<mode>): New.
2021 (aarch64_mul3_elt_<vswap_width_name><mode>): Likewise.
2022 (aarch64_mul3_elt_to_128df): Likewise.
2023 (aarch64_mul3_elt_to_64v2df): Likewise.
2024 * config/aarch64/iterators.md (VEL): Also handle DFmode.
2026 (VMUL_CHANGE_NLANES) Likewise.
2029 * config/aarch64/arm_neon.h
2030 (vmul<q>_lane<q>_<suf><16,32,64>): Convert to C implementation.
2032 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
2034 * config/aarch64/arm_neon.h
2035 (vcvtx_high_f32_f64): Fix parameters.
2037 2013-09-16 Jan-Benedict Glaw <jbglaw@lug-owl.de>
2038 Uros Bizjak <ubizjak@gmail.com>
2040 * config/alpha.c: Include tree-ssanames.h.
2042 2013-09-16 Richard Biener <rguenther@suse.de>
2044 * tree-loop-distribution.c (enum rdg_dep_type): Add control_dd.
2045 (dot_rdg_1): Handle control_dd.
2046 (create_edge_for_control_dependence): New function.
2047 (create_rdg_edges): Add control dependences if asked for.
2048 (build_rdg): Likewise.
2049 (generate_loops_for_partition): If there are not necessary
2050 control stmts remove all their dependencies.
2051 (collect_condition_stmts, rdg_flag_loop_exits): Remove.
2052 (distribute_loop): Pass on control dependences.
2053 (tree_loop_distribution): Compute control dependences and remove
2054 restriction on number of loop nodes.
2056 2013-09-16 Jakub Jelinek <jakub@redhat.com>
2058 * ipa-prop.c (ipa_compute_jump_functions_for_edge): Return early
2061 2013-09-16 Richard Sandiford <rdsandiford@googlemail.com>
2063 * cse.c (try_const_anchors): Punt on CC modes.
2065 2013-09-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
2067 * config/vax/constraints.md (T): Add missing CONSTANT_P check.
2069 2013-09-14 John David Anglin <danglin@gcc.gnu.org>
2072 * config/pa/pa.c (pa_expand_prologue): Change mode in gen_rtx_POST_INC
2075 2013-09-14 Iain Sandoe <iain@codesourcery.com>
2078 * config/darwin.c (darwin_objc2_section): Note if ObjC Metadata is
2080 (darwin_objc1_section): Likewise.
2081 (darwin_file_end): Emit Image Info section when required.
2083 2013-09-14 Jan Hubicka <jh@suse.cz>
2085 * tree-into-ssa.c (gate_into_ssa): New.
2086 (pass_data_build_ssa): Use it.
2087 * cgraph.h (expand_thunk): Update prototype.
2088 * cgraphunit.c (analyze_function): Expand thunks early.
2089 (expand_thunk): Fix DECL_CONTEXT of reust_decl;
2090 build proper cgraph; set in_ssa_p; clear bogus TREE_ASM_WRITTEN;
2091 set lowered flag; do not add new function.
2092 (assemble_thunks_and_aliases): Update.
2093 * tree-ssa.c (gate_init_datastructures): New gate.
2094 (pass_data_init_datastructures): Use it.
2096 2013-09-14 Iain Sandoe <iain@codesourcery.com>
2099 * config/i386/i386.c (ix86_function_arg_regno_p): Make Darwin use the
2100 xmm register set described in the psABI.
2102 2013-09-13 Evgeny Gavrin <e.gavrin@samsung.com>
2104 * dwarf2out.c (should_emit_struct_debug): Add check
2105 for type_decl variable is not NULL.
2107 2013-09-13 Jacek Caban <jacek@codeweavers.com>
2109 * config.gcc: Use new winnt-c.c target hooks
2110 * config/t-winnt: New file
2111 * config/winnt-c.c: New file
2112 * doc/tm.texi.in: Document new hook
2113 * doc/tm.texi: Regenerated
2115 2013-09-13 Jan Hubicka <jh@suse.cz>
2118 * ipa-inline.c (check_callers): New function.
2119 (check_caller_edge): Remove.
2120 (want_inline_function_to_all_callers_p): Also permit alises that are
2122 (inline_to_all_callers): Terminate the walk when devirtualization
2123 introduce new calls.
2125 2013-09-13 Jan Hubicka <jh@suse.cz>
2127 * ipa-inline-analysis.c (struct growth_data): Add node.
2128 (do_estimate_growth_1): Fix detection of recursion.
2130 2013-09-13 Jakub Jelinek <jakub@redhat.com>
2132 PR tree-optimization/58392
2133 * tree-cfg.c (move_sese_region_to_fn): Rename loop variable
2134 to avoid shadowing of outer loop variable. If
2135 saved_cfun->has_simduid_loops or saved_cfun->has_force_vect_loops,
2136 replace_by_duplicate_decl simduid of loops that have it set and
2137 set dest_cfun->has_simduid_loops and/or
2138 dest_cfun->has_force_vect_loops.
2139 * omp-low.c (build_outer_var_ref): Call maybe_lookup_decl_in_outer_ctx
2140 instead of maybe_lookup_decl.
2141 * tree-inline.c (copy_loops): Change blocks_to_copy argument to id.
2142 Use id->blocks_to_copy instead of blocks_to_copy. Adjust recursive
2143 call. Copy over force_vect and copy and remap simduid. Set
2144 cfun->has_simduid_loops and/or cfun->has_force_vect_loops.
2145 (copy_cfg_body): Remove blocks_to_copy argument. Use
2146 id->blocks_to_copy instead of blocks_to_copy. Adjust copy_loops
2147 caller. Don't set cfun->has_simduid_loops and/or
2148 cfun->has_force_vect_loops here.
2149 (copy_body): Remove blocks_to_copy argument. Adjust copy_cfg_body
2151 (expand_call_inline, tree_function_versioning): Adjust copy_body
2154 2013-09-13 Martin Jambor <mjambor@suse.cz>
2157 * ipa-prop.c (try_make_edge_direct_simple_call): Be less strict in
2158 the assert if the edge was a speculative one.
2160 2013-09-13 Richard Biener <rguenther@suse.de>
2162 * tree-data-ref.h (known_dependences_p): Move here ...
2163 * tree-loop-distribution.c (known_dependences_p): ... from here.
2164 (dump_rdg_component, debug_rdg_component): Remove.
2166 (generate_loops_for_partition): Use gimple_uid instead of
2167 relying on matching stmt visit order.
2168 (rdg_build_partitions): Take starting stmt vector.
2169 (ldist_gen): Merge into ...
2170 (distribute_loop): ... this function. Do not compute starting
2172 * tree-cfg.c (gimple_duplicate_bb): Copy UID for PHIs.
2174 2013-09-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2176 * config/arm/arm.md (arm_cmpsi_insn): Split rI alternative.
2177 Set type attribute correctly. Set predicable_short_it attribute.
2178 (cmpsi_shiftsi): Remove %? from output template.
2180 2013-09-13 Richard Biener <rguenther@suse.de>
2182 * tree-loop-distribution.c (struct rdg_component,
2183 rdg_defs_used_in_other_loops_p, free_rdg_components,
2184 rdg_build_components): Remove.
2185 (stmts_from_loop): Do not record virtual PHIs.
2186 (generate_loops_for_partition): Skip virtual PHIs.
2187 (build_rdg_partition_for_component): Rename to ...
2188 (build_rdg_partition_for_vertex): ... this and adjust.
2189 (rdg_build_partitions): Take a vector of starting vertices
2190 instead of components. Remove unnecessary leftover handling.
2191 (ldist_gen): Do not build components or record other stores.
2192 (distribute_loop): Do not distribute loops containing stmts
2195 2013-09-13 Christian Bruel <christian.bruel@st.com>
2198 * config/sh/sh.md (mov<mode>_reg_reg): Allow memory reloads.
2200 2013-09-13 Kai Tietz <ktietz@redhat.com>
2202 * config.gcc: Separate cases for mingw and cygwin targets,
2203 and add 64-bit cygwin target case.
2205 * config/i386/winnt-cxx.c (i386_pe_type_dllexport_p): Don't
2206 dll-export inline-functions.
2207 * config/i386/winnt.c (i386_pe_determine_dllexport_p): Likewise.
2210 2013-09-13 Jeff Law <law@redhat.com>
2214 2013-09-06 Jeff Law <law@redhat.com>
2216 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
2217 edge implied equivalences into successor phis.
2219 2013-09-12 DJ Delorie <dj@redhat.com>
2221 * config/rl78/rl78-virt.md: Change from | to \; for asm line
2224 2013-09-12 Brooks Moses <bmoses@google.com>
2227 * Makefile.in: Do not install driver binaries in $(target)/bin.
2229 2013-09-12 DJ Delorie <dj@redhat.com>
2231 * config/rl78/rl78.opt (mrelax): New.
2232 * config/rl78/rl78.h (ASM_SPEC): New, pass on -mrelax to gas.
2233 * config/rl78/rl78.h (LINK_SPEC): New, pass on -mrelax to ld.
2235 * config/rl78/rl78.c (rl78_expand_prologue): Use AX to copy
2237 (rl78_expand_epilogue): Likewise.
2239 2013-09-12 Vladimir Makarov <vmakarov@redhat.com>
2242 * lra-eliminations.c (remove_reg_equal_offset_note): New.
2243 (eliminate_regs_in_insn): Rewrite frame pointer to hard frame
2244 pointer elimination with using remove_reg_equal_offset_note.
2246 2013-09-12 DJ Delorie <dj@redhat.com>
2248 * config/msp430/: New port.
2249 * config.gcc (msp430): Added.
2250 * doc/invoke.texi: Document MSP430 options.
2251 * doc/install.texi: Document msp430-elf
2252 * doc/md.texi: Document msp430-elf
2253 * doc/contrib.texi: Document msp430-elf
2255 * cfgexpand.c (expand_debug_expr): Avoid sign-extending SImode to
2258 2013-09-12 Martin Jambor <mjambor@suse.cz>
2261 * ipa-prop.c (remove_described_reference): Give up if the edge in the
2262 reference descriptor is NULL.
2263 (ipa_edge_removal_hook): If owning a reference descriptor, set its
2266 2013-09-12 Andrew MacLeod <amacleod@redhat.com>
2268 * tree-flow.h (FREE_SSANAMES): Move to tree-ssanames.c
2269 (SSANAMES, MODIFIED_NORETURN_CALLS, DEFAULT_DEFS, ptr_info_def,
2270 num_ssa_names, ssa_name): Move to tree-ssanames.h + prototypes.
2271 * tree-flow-inline.h (make_ssa_name, copy_ssa_name, duplicate_ssa_name,
2272 make_temp_ssa_name): move to tree-ssanames.h
2273 * tree-ssa-alias.h: Move prototype.
2274 * tree-ssa.h: Include tree-ssanames.h.
2275 * tree-ssanames.c (FREE_SSANAMES): Move to here.
2276 * tree-ssanames.h: New. Move items from tree-flow*.h
2277 * Makefile.in (tree-ssanames.h): Add to tree-ssanames.o and GTFILES.
2279 2013-09-12 Richard Biener <rguenther@suse.de>
2281 PR tree-optimization/58404
2282 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Also
2283 propagate non-invariant addresses into dereferences wrapped
2284 in component references.
2286 2013-09-12 Richard Biener <rguenther@suse.de>
2288 PR tree-optimization/58402
2289 * passes.def: Move pass_late_warn_uninitialized later.
2291 2013-09-12 Andrew MacLeod <amacleod@redhat.com>
2293 * tree-ssa.h: New. Move content from tree-flow.h and
2295 * tree-flow.h (_edge_var_map, edge_var_map_vector): Move to tree-ssa.h.
2296 Move prototypes belonging to tree-ssa.c.
2297 * tree-flow-inline.h (redirect_edge_var_map_def,
2298 redirect_edge_var_map_result, redirect_edge_var_map_location): Move to
2300 * gimple.h: Adjust prototypes.
2301 * tree-ssa.c (useless_type_conversion_p, types_compatible_p): Move
2303 * gimple.c (useless_type_conversion_p, types_compatible_p): Here.
2304 * tree.h: Move prototype to tree-ssa.h.
2305 * gengtype.c (open_base_files): Replace tree-flow.h with tree-ssa.h.
2306 * Makefile.in: (TREE_SSA_H, TREE_FLOW_H): Adjust dependencies.
2307 * alias.c, asan.c, builtins.c, calls.c, cfgexpand.c, cfghooks.c,
2308 cfgloop.c, cfgloopmanip.c, cgraph.c, cgraphbuild.c, cgraphclones.c,
2309 cgraphunit.c, dse.c, except.c, expr.c, final.c, fold-const.c,
2310 ggc-page.c, gimple-fold.c, gimple-iterator.c, gimple-low.c,
2311 gimple-pretty-print.c, gimple-ssa-strength-reduction.c,
2312 gimple-streamer-in.c, gimple-streamer-out.c, gimple.c, gimplify.c,
2313 graphite-blocking.c, graphite-clast-to-gimple.c,
2314 graphite-dependences.c, graphite-interchange.c,
2315 graphite-optimize-isl.c, graphite-poly.c, graphite-scop-detection.c,
2316 graphite-sese-to-poly.c, graphite.c, ipa-cp.c, ipa-inline-analysis.c,
2317 ipa-inline-transform.c, ipa-inline.c, ipa-prop.c, ipa-pure-const.c,
2318 ipa-reference.c, ipa-split.c, ipa-utils.c,
2319 loop-init.c, lto-cgraph.c, lto-section-in.c, lto-section-out.c,
2320 lto-streamer-in.c, lto-streamer-out.c, lto-streamer.c, omp-low.c,
2321 passes.c, predict.c, print-tree.c, profile.c, sese.c, targhooks.c,
2322 tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c, tree-cfgcleanup.c,
2323 tree-chrec.c, tree-complex.c, tree-data-ref.c, tree-dfa.c, tree-eh.c,
2324 tree-emutls.c, tree-if-conv.c, tree-inline.c, tree-into-ssa.c,
2325 tree-loop-distribution.c, tree-mudflap.c, tree-nested.c, tree-nrv.c,
2326 tree-object-size.c, tree-optimize.c, tree-outof-ssa.c, tree-parloops.c,
2327 tree-phinodes.c, tree-predcom.c, tree-pretty-print.c, tree-profile.c,
2328 tree-scalar-evolution.c, tree-sra.c, tree-ssa*.c, tree-stdarg.c,
2329 tree-streamer-in.c, tree-switch-conversion.c, tree-tailcall.c,
2330 tree-vect-data-refs.c, tree-vect-generic.c, tree-vect-loop-manip.c,
2331 tree-vect-loop.c, tree-vect-patterns.c, tree-vect-slp.c,
2332 tree-vect-stmts.c, tree-vectorizer.c, tree-vrp.c, tsan.c,
2333 value-prof.c, var-tracking.c,
2334 varpool.c, vtable-verify.c: Replace tree-flow.h with tree-ssa.h
2336 2013-09-12 Richard Biener <rguenther@suse.de>
2338 PR tree-optimization/58396
2339 * tree-loop-distribution.c (create_rdg_edges): Free unused DDRs.
2340 (build_rdg): Take a loop-nest parameter, fix memleaks.
2341 (distribute_loop): Compute loop-nest here and pass it to build_rdg.
2343 2013-09-12 Yuri Rumyantsev <ysrumyan@gmail.com>
2345 * config/i386/x86-tune.def: Turn on X86_TUNE_AVOID_MEM_OPND_FOR_CMOVE
2348 2013-09-12 Cameron McInally <cameron.mcinally@nyu.edu>
2350 * doc/extend.texi: Fix errors in x86 FMA builtin naming.
2351 The FMA instruction names should have a 'v' prefix.
2353 2013-09-12 Richard Biener <rguenther@suse.de>
2355 * tree-loop-distribution.c (dot_rdg_1): Make graph prettier.
2356 (dot_rdg): Use popen instead of system in optional code.
2357 (remaining_stmts, upstream_mem_writes): Remove global bitmaps.
2358 (already_processed_vertex_p): Adjust.
2359 (has_anti_or_output_dependence, predecessor_has_mem_write,
2360 mark_nodes_having_upstream_mem_writes, has_upstream_mem_writes,
2361 rdg_flag_uses): Remove.
2362 (rdg_flag_vertex): Simplify.
2363 (rdg_flag_vertex_and_dependent): Rely on a correct RDG and
2365 (build_rdg_partition_for_component): Process the first vertex
2366 of a component only.
2367 (ldist_gen): Do not compute remaining_stmts or upstream_mem_writes.
2369 2013-09-12 Alan Modra <amodra@gmail.com>
2371 * config/rs6000/rs6000.c (toc_relative_expr_p): Use add_cint_operand.
2373 2013-09-11 DJ Delorie <dj@redhat.com>
2374 Nick Clifton <nickc@redhat.com>
2376 * config/rl78/predicates.md (rl78_cmp_operator_signed): New.
2377 (rl78_stack_based_mem): New.
2378 * config/rl78/constraints.md (Iv08): New.
2387 * config/rl78/rl78-expand.md (movqi): Reject more SUBREG operands.
2389 (movsi): Change from expand to insn-and-split.
2390 (ashrsi3): Clobber AX.
2394 * config/rl78/rl78.md (CC_REG): Fix.
2395 (addsi3): Allow memory and immediate operands.
2396 (addsi3_internal): Split into...
2397 (addsi3_internal_virt): ...new, and ...
2398 (addsi3_internal_real): ...new.
2400 (subsi3_internal_virt): New.
2401 (subsi3_internal_real): New.
2402 (mulsi3): Add memory operand.
2403 (mulsi3_rl78): Likewise.
2404 (mulsi3_g13): Likewise.
2405 * config/rl78/rl78-real.md (cbranchqi4_real_signed): New.
2406 (cbranchqi4_real): Add more constraint options.
2407 (cbranchhi4_real): Expand pattern.
2408 (cbranchhi4_real_signed): New.
2409 (cbranchhi4_real_inverted): New.
2410 (cbranchsi4_real_lt): New.
2411 (cbranchsi4_real_ge): New.
2412 (cbranchsi4_real_signed): New.
2413 (cbranchsi4_real): New.
2415 * config/rl78/rl78-virt.md (ashrsi3_virt): Add custom cases for
2417 (lshrsi3_virt): Likewise.
2418 (ashlsi3_virt): Likewise.
2419 (cbranchqi4_virt_signed): New.
2420 (cbranchhi4_virt_signed): New.
2421 (cbranchsi4_virt): New.
2422 * config/rl78/rl78.c: Whitespace fixes throughout.
2423 (move_elim_pass): New.
2424 (pass_data_rl78_move_elim): New.
2425 (pass_rl78_move_elim): New.
2426 (make_pass_rl78_move_elim): New.
2427 (rl78_devirt_info): Run devirt earlier.
2428 (rl78_move_elim_info): New.
2429 (rl78_asm_file_start): Register it.
2430 (rl78_split_movsi): New.
2431 (rl78_as_legitimate_address): Allow virtual base registers when
2433 (rl78_addr_space_convert): Remove spurious debug stuff.
2434 (rl78_print_operand_1): Add z,s,S,r,E modifiers.
2435 (rl78_print_operand): More cases for not printing '#'.
2436 (rl78_expand_compare): Remove most of the logic.
2437 (content_memory): New.
2438 (clear_content_memory): New.
2439 (get_content_index): New.
2440 (get_content_name): New.
2441 (display_content_memory): New.
2442 (update_content): New.
2443 (record_content): New.
2444 (already_contains): New.
2445 (insn_ok_now): Re-recog insns with virtual registers.
2446 (add_postponed_content_update): New.
2447 (process_postponed_content_update): New.
2448 (gen_and_emit_move): New.
2449 (transcode_memory_rtx): Record new location content.
2450 Use gen_and_emit_move.
2451 (force_into_acc): New.
2452 (move_to_acc): Use gen_and_emit_move.
2453 (move_from_acc): Likewise.
2454 (move_acc_to_reg): Likewise.
2455 (move_to_x): Likewise.
2456 (move_to_hl): Likewise.
2457 (move_to_de): Likewise.
2458 (rl78_alloc_physical_registers_op1): Record location content.
2459 (has_constraint): New.
2460 (rl78_alloc_physical_registers_op2): Record location content.
2462 (rl78_alloc_physical_registers_ro1): Likewise.
2463 (rl78_alloc_physical_registers_cmp): Likewise.
2464 (rl78_alloc_physical_registers_umul): Likewise.
2465 (rl78_alloc_address_registers_macax): New.
2466 (rl78_alloc_physical_registers): Initialize and set location
2467 content memory as needed.
2468 (rl78_reorg): Make sure split2 is called.
2469 (rl78_rtx_costs): New.
2471 2013-09-11 Richard Sandiford <rdsandiford@googlemail.com>
2473 * simplify-rtx.c (simplify_unary_operation_1): Use simplify_gen_binary
2474 for (not (neg ...)) and (neg (not ...)) cases.
2476 2013-09-11 Richard Biener <rguenther@suse.de>
2479 * passes.def: Split critical edges before late uninit warning passes.
2480 * tree-cfg.c (pass_split_crit_edges): Implement clone method.
2482 2013-09-11 Jakub Jelinek <jakub@redhat.com>
2484 PR tree-optimization/58385
2485 * fold-const.c (build_range_check): If both low and high are NULL,
2486 use omit_one_operand_loc to preserve exp side-effects.
2488 2013-09-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2490 * config/arm/arm.md (arm_shiftsi3): New alternative l/l/M.
2492 2013-09-11 Richard Biener <rguenther@suse.de>
2494 * tree-data-ref.c (dump_rdg_vertex, debug_rdg_vertex,
2495 dump_rdg_component, debug_rdg_component, dump_rdg, debug_rdg,
2496 dot_rdg_1, dot_rdg, rdg_vertex_for_stmt, create_rdg_edge_for_ddr,
2497 create_rdg_edges_for_scalar, create_rdg_edges, create_rdg_vertices,
2498 stmts_from_loop, known_dependences_p, build_empty_rdg,
2499 build_rdg, free_rdg, rdg_defs_used_in_other_loops_p): Move ...
2500 * tree-loop-distribution.c: ... here.
2501 * tree-data-ref.h (struct rdg_vertex, RDGV_STMT, RDGV_DATAREFS,
2502 RDGV_HAS_MEM_WRITE, RDGV_HAS_MEM_READS, RDG_STMT, RDG_DATAREFS,
2503 RDG_MEM_WRITE_STMT, RDG_MEM_READS_STMT, enum rdg_dep_type,
2504 struct rdg_edge, RDGE_TYPE, RDGE_LEVEL, RDGE_RELATION): Move ...
2505 * tree-loop-distribution.c: ... here.
2506 * tree-loop-distribution.c: Include gimple-pretty-print.h.
2507 (struct partition_s): Add loops member.
2508 (partition_alloc, partition_free, rdg_flag_uses, rdg_flag_vertex,
2509 rdg_flag_vertex_and_dependent, rdg_flag_loop_exits,
2510 build_rdg_partition_for_component, rdg_build_partitions): Adjust.
2512 2013-09-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
2513 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
2514 Sergey Lega <sergey.s.lega@intel.com>
2515 Anna Tikhonova <anna.tikhonova@intel.com>
2516 Ilya Tocar <ilya.tocar@intel.com>
2517 Andrey Turetskiy <andrey.turetskiy@intel.com>
2518 Ilya Verbin <ilya.verbin@intel.com>
2519 Kirill Yukhin <kirill.yukhin@intel.com>
2520 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
2522 * config/i386/constraints.md (k): New.
2524 * config/i386/i386.c (const regclass_map): Add new mask registers.
2525 (dbx_register_map): Ditto.
2526 (dbx64_register_map): Ditto.
2527 (svr4_dbx_register_map): Ditto.
2528 (ix86_conditional_register_usage): Squash mask registers if AVX512F is
2530 (ix86_preferred_reload_class): Disable constants for mask registers.
2531 (ix86_secondary_reload): Do spill of mask register using 32-bit insn.
2532 (ix86_hard_regno_mode_ok): Support new mask registers.
2533 (x86_order_regs_for_local_alloc): Ditto.
2534 * config/i386/i386.h (FIRST_PSEUDO_REGISTER): Update.
2535 (FIXED_REGISTERS): Add new mask registers.
2536 (CALL_USED_REGISTERS): Ditto.
2537 (REG_ALLOC_ORDER): Ditto.
2538 (VALID_MASK_REG_MODE): New.
2539 (FIRST_MASK_REG): Ditto.
2540 (LAST_MASK_REG): Ditto.
2541 (reg_class): Add MASK_EVEX_REGS, MASK_REGS.
2542 (MAYBE_MASK_CLASS_P): New.
2543 (REG_CLASS_NAMES): Add MASK_EVEX_REGS, MASK_REGS.
2544 (REG_CLASS_CONTENTS): Ditto.
2545 (MASK_REGNO_P): New.
2546 (ANY_MASK_REG_P): Ditto.
2547 (HI_REGISTER_NAMES): Add new mask registers.
2548 * config/i386/i386.md (MASK0_REG, MASK1_REG, MASK2_REG, MASK3_REG,
2549 MASK4_REG, MASK5_REG, MASK6_REG, MASK7_REG): Constants for new
2551 (attribute "type"): Add mskmov, msklog.
2552 (attribute "length_immediate"): Support them.
2553 (attribute "memory"): Ditto.
2554 (attribute "prefix_0f"): Ditto.
2555 (*movhi_internal): Support new mask registers.
2556 (*movqi_internal): Ditto.
2557 (define_split): Split out clobber pattern is a logic
2558 insn on mask registers.
2559 (*k<logic><mode>): New.
2560 (*andhi_1): Extend to support mask regs.
2561 (*andqi_1): Extend to support mask regs.
2563 (define_split): Split and-not to and and not if operands
2565 (*<code><mode>_1): Separate HI mode to new pattern...
2566 (*<code>hi_1): This.
2567 (*<code>qi_1): Extend to support mask regs.
2569 (kortestzhi): Ditto.
2570 (kortestchi): Ditto.
2572 (*one_cmpl<mode>2_1): Remove HImode and handle it...
2573 (*one_cmplhi2_1): ...Here, now with mask registers support.
2574 (*one_cmplqi2_1): Support new mask registers.
2575 (HI/QImode arithmetics splitter): Don't split if mask registers
2577 (HI/QImode not splitter): Ditto.
2578 * config/i386/predicated.md (mask_reg_operand): New.
2579 (general_reg_operand): Ditto.
2581 2013-09-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
2583 * doc/invoke.texi: Document fxsr, xsave and xsaveopt options.
2584 * doc/extend.texi: Document fxsr, xsave and xsaveopt builtins.
2586 2013-09-10 Jeff Law <law@redhat.com>
2588 PR tree-optimization/58380
2589 * tree-ssa-threadupdate.c (thread_block): Recognize another case
2590 of threading through a buried loop header.
2592 * tree-ssa-threadedge.c (thread_around_empty_blocks): Correct
2593 return value for single successor case.
2595 2013-09-10 Jan Hubicka <jh@suse.cz>
2597 * ipa-devirt.c (ipa_devirt): Enable with LTO.
2599 2013-09-10 Richard Earnshaw <rearnsha@arm.com>
2602 * arm/vfp.md (combine_vcvt_f32_<FCVTI32typename>): Fix pattern to
2603 support conditional execution.
2604 (combine_vcvt_f64_<FCVTI32typename>): Likewise.
2606 2013-09-10 Vladimir Makarov <vmakarov@redhat.com>
2608 * lra.c (lra): Clear lra_optional_reload_pseudos before every
2610 * lra-constraints.c (curr_insn_transform): Switch on optional reloads.
2611 Check destination too to check move insn.
2612 (undo_optional_reloads): Add check that the original peudo did not
2613 changed its allocation and the optional reload was inherited on last
2614 inheritance pass. Break loop after deciding to keep optional reload.
2615 (lra_undo_inheritance): Add check that inherited pseudo still in
2618 2013-09-10 James Greenhalgh <james.greenhalgh@arm.com>
2620 * config/aarch64/aarch64.md (generic_sched): New.
2621 * config/aarch64/aarch64-generic.md (load): Make conditional
2622 on generic_sched attribute.
2623 (nonload): Likewise.
2625 2013-09-10 Jan Hubicka <jh@suse.cz>
2627 * lto-cgraph.c: Include ipa-utils.h.
2628 (compute_ltrans_boundary): Also add possible targets into the boundary.
2630 2013-09-10 Jan Hubicka <jh@suse.cz>
2632 * gimple-fold.c (gimple_get_virt_method_for_binfo): Pass real
2633 VAR_DECL of vtable rather than full expression.
2635 2013-09-10 Jan Hubicka <jh@suse.cz>
2636 Paolo Carlini <paolo.carlini@oracle.com>
2638 * cgraphunit.c (analyze_functions): Save input_location, set it
2639 to UNKNOWN_LOCATION and restore it at the end.
2641 2013-09-10 Martin Jambor <mjambor@suse.cz>
2643 * ipa-cp.c (propagate_constants_topo): Do not ignore SCC
2644 represented by a thunk.
2646 2013-09-10 Jeff Law <law@redhat.com>
2648 PR tree-optimization/58343
2649 * tree-ssa-threadupdate.c (thread_block): Identify and disable
2650 jump threading requests through loop headers buried in the middle
2651 of a jump threading path.
2653 * tree-ssa-threadedge.c (thread_around_empty_blocks): Fix thinko
2654 in return value/type.
2656 2013-09-10 Jakub Jelinek <jakub@redhat.com>
2658 PR rtl-optimization/58365
2659 * cfgcleanup.c (merge_memattrs): Also clear MEM_READONLY_P
2660 resp. MEM_NOTRAP_P if they differ, or set MEM_VOLATILE_P if
2663 2013-09-10 Richard Biener <rguenther@suse.de>
2665 * tree-data-ref.h (build_rdg): Drop all parameters but loop.
2666 * tree-data-ref.c (create_rdg_vertices): Collect all data
2667 references, signal failure to the caller, use data-ref API.
2668 (build_rdg): Compute data references only once. Maintain lifetime
2669 of data references and data dependences from within RDG.
2670 (free_rdg): Free dependence relations.
2671 * tree-loop-distribution.c (rdg_flag_uses): Drop weird code
2672 inventing extra dependences.
2673 (distribute_loop): Update for RDG API changes.
2675 2013-09-10 Kai Tietz <ktietz@redhat.com>
2677 * doc/invoke.texi (fms-extensions): Document changed
2678 behavior for ms-abi targets.
2679 * config/i386/i386.c (ix86_option_override_internal):
2680 Set default value of option -fms-extension for ms-abi targets.
2682 2013-09-10 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
2684 * config/i386/i386.c (ix86_expand_movmem): Fix epilogue generation.
2686 2013-09-10 Alan Modra <amodra@gmail.com>
2689 * config/rs6000/rs6000.md (bswapdi2_64bit): Disable for volatile mems.
2691 2013-09-10 Alan Modra <amodra@gmail.com>
2693 * config/rs6000/predicates.md (add_cint_operand): New.
2694 (reg_or_add_cint_operand, small_toc_ref): Use add_cint_operand.
2695 * config/rs6000/rs6000.md (largetoc_high_plus): Restrict offset
2696 using add_cint_operand.
2697 (largetoc_high_plus_aix): Likewise.
2699 2013-09-09 Jakub Jelinek <jakub@redhat.com>
2701 PR tree-optimization/58364
2702 * tree-ssa-reassoc.c (init_range_entry): For BIT_NOT_EXPR on
2703 BOOLEAN_TYPE, only invert in_p and continue with arg0 if
2704 the current range can't be an unconditional true or false.
2706 2013-09-09 James Greenhalgh <james.greenhalgh@arm.com>
2708 * config/aarch64/arm_neon.h (vrsqrte_f64): Fix parameter type.
2710 2013-09-09 Uros Bizjak <ubizjak@gmail.com>
2712 * ipa-prop.c (ipa_modify_call_arguments): Initialize deref_align.
2714 2013-09-09 Paolo Carlini <paolo.carlini@oracle.com>
2717 * doc/invoke.texi (-Wdelete-incomplete): Document it.
2719 2013-09-09 Ian Bolton <ian.bolton@arm.com>
2721 * config/aarch64/aarch64.c (aarch64_preferred_reload_class): Return
2722 NO_REGS for immediate that can't be moved directly into FP_REGS.
2724 2013-09-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2726 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return CC_SWP for
2727 comparison with negated operand.
2728 * config/aarch64/aarch64.md (compare_neg<mode>): Match canonical
2731 2013-09-09 Richard Biener <rguenther@suse.de>
2734 * cfgloopmanip.c (fix_bb_placements): When fixing the placement
2735 of a subloop record all its block as affecting loop-closed SSA form.
2737 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
2739 * expmed.c (lshift_value): Take an unsigned HOST_WIDE_INT instead
2740 of an rtx/bitpos pair.
2741 (store_fixed_bit_field): Update accordingly.
2743 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
2745 * asan.c (asan_emit_stack_protection): Use gen_int_mode instead of
2747 * builtins.c (expand_errno_check): Likewise.
2748 * dwarf2cfi.c (init_return_column_size): Likewise.
2749 * except.c (sjlj_mark_call_sites): Likewise.
2750 * expr.c (move_by_pieces_1, store_by_pieces_2): Likewise.
2751 * lra-constraints.c (emit_inc): Likewise.
2752 * ree.c (combine_set_extension): Likewise.
2753 * regmove.c (fixup_match_2): Likewise.
2754 * reload1.c (inc_for_reload): Likewise.
2756 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
2758 * combine.c (simplify_set, expand_field_assignment, extract_left_shift)
2759 (force_to_mode, simplify_shift_const_1, simplify_comparison):
2760 Use gen_int_mode with the mode of the associated simplify_* call.
2761 * explow.c (probe_stack_range, anti_adjust_stack_and_probe): Likewise.
2762 * expmed.c (expand_shift_1): Likewise.
2763 * function.c (instantiate_virtual_regs_in_insn): Likewise.
2764 * loop-iv.c (iv_number_of_iterations): Likewise.
2765 * loop-unroll.c (unroll_loop_runtime_iterations): Likewise.
2766 * simplify-rtx.c (simplify_binary_operation_1): Likewise.
2768 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
2770 * asan.c (asan_clear_shadow): Use gen_int_mode with the mode
2771 of the associated expand_* call.
2772 (asan_emit_stack_protection): Likewise.
2773 * builtins.c (round_trampoline_addr): Likewise.
2774 * explow.c (allocate_dynamic_stack_space, probe_stack_range): Likewise.
2775 * expmed.c (expand_smod_pow2, expand_sdiv_pow2, expand_divmod)
2776 (emit_store_flag): Likewise.
2777 * expr.c (emit_move_resolve_push, push_block, emit_single_push_insn_1)
2778 (emit_push_insn, optimize_bitfield_assignment_op, expand_expr_real_1):
2780 * function.c (instantiate_virtual_regs_in_insn): Likewise.
2781 * ifcvt.c (noce_try_store_flag_constants): Likewise.
2782 * loop-unroll.c (unroll_loop_runtime_iterations): Likewise.
2783 * modulo-sched.c (generate_prolog_epilog): Likewise.
2784 * optabs.c (expand_binop, widen_leading, expand_doubleword_clz)
2785 (expand_ctz, expand_ffs, expand_unop): Likewise.
2787 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
2789 * alias.c (addr_side_effect_eval): Use gen_int_mode with the mode
2790 of the associated gen_rtx_* call.
2791 * caller-save.c (init_caller_save): Likewise.
2792 * combine.c (find_split_point, make_extraction): Likewise.
2793 (make_compound_operation): Likewise.
2794 * dwarf2out.c (mem_loc_descriptor): Likewise.
2795 * explow.c (plus_constant, probe_stack_range): Likewise.
2796 * expmed.c (expand_mult_const): Likewise.
2797 * expr.c (emit_single_push_insn_1, do_tablejump): Likewise.
2798 * reload1.c (init_reload): Likewise.
2799 * valtrack.c (cleanup_auto_inc_dec): Likewise.
2800 * var-tracking.c (adjust_mems): Likewise.
2801 * modulo-sched.c (sms_schedule): Likewise, but use gen_rtx_GT
2802 rather than gen_rtx_fmt_ee.
2804 2013-09-09 Jan Hubicka <jh@suse.cz>
2807 * value-prof.c (gimple_ic): Copy also abnormal edges.
2809 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
2811 * asan.c (asan_shadow_cst): Use gen_int_mode.
2813 2013-09-08 Jan Hubicka <jh@suse.cz>
2815 * ipa-profile.c: Add toplevel comment.
2816 (ipa_propagate_frequency_1): Be more conservative when profile is read.
2817 (contains_hot_call_p): New function.
2818 (ipa_propagate_frequency): Set frequencies based on counts when
2820 * predict.c (compute_function_frequency): Use PROFILE_READ gueard for
2821 profile; do not tamper with profile after inlining if it is read.
2823 2013-09-08 Jan Hubicka <jh@suse.cz>
2825 * ipa-prop.c (try_make_edge_direct_simple_call): Do not special case
2828 2013-09-08 Jan Hubicka <jh@suse.cz>
2830 * ipa.c (walk_polymorphic_call_targets): Fix redirection before IPA
2833 2013-09-08 Jeff Law <law@redhat.com>
2836 * tree-ssa-threadedge.c (thread_across_edge): Fix initialization
2839 2013-09-08 Andi Kleen <ak@linux.intel.com>
2841 * tree-inline.c (estimate_num_insns): Limit asm cost to 1000.
2843 2013-09-08 Jan Hubicka <jh@suse.cz>
2845 * ipa.c (walk_polymorphic_call_targets): Fix inliner summary update.
2847 2013-09-08 Richard Sandiford <rdsandiford@googlemail.com>
2849 * ira.c (update_equiv_regs): Only call set_paradoxical_subreg
2850 for non-debug insns.
2851 * lra.c (new_insn_reg): Take the containing insn as a parameter.
2852 Only modify lra_reg_info[].biggest_mode if it's non-debug insn.
2853 (collect_non_operand_hard_regs, add_regs_to_insn_regno_info): Update
2856 2013-09-08 Jan Hubicka <jh@suse.cz>
2858 * cgraphunit.c (walk_polymorphic_call_targets): Permit 0 possible
2859 targets and devirtualize to BUILT_IN_UNREACHABLE.
2860 * timevar.def (TV_IPA_UNREACHABLE): New timevar.
2861 * ipa.c (walk_polymorphic_call_targets): New function.
2862 (symtab_remove_unreachable_nodes): Use it; do not keep all virtual
2863 functions; use the new timevar.
2864 * ipa-devirt.c (maybe_record_node): Do not insert static nodes that
2865 was removed from the program.
2866 (record_binfo): If BINFO corresponds to an anonymous namespace, we may
2867 not consider it in the walk when its vtable is dead.
2868 (possible_polymorphic_call_targets_1): Pass anonymous flag to
2870 (devirt_variable_node_removal_hook): New function.
2871 (possible_polymorphic_call_targets): Also register
2872 devirt_variable_node_removal_hook.
2873 (ipa_devirt): Do not do non-speculative devirtualization.
2874 (gate_ipa_devirt): One execute if devirtualizing speculatively.
2876 2013-09-08 Jan Hubicka <jh@suse.cz>
2878 * cgraph.h (varpool_node_hook, varpool_node_hook_list,
2879 varpool_add_node_removal_hook, varpool_add_variable_insertion_hook,
2880 varpool_remove_variable_insertion_hook): Declare.
2881 * varpool.c (varpool_node_hook_list): New structure.
2882 (first_varpool_node_removal_hook,
2883 first_varpool_variable_insertion_hook): New variables.
2884 (varpool_add_node_removal_hook, varpool_remove_node_removal_hook,
2885 varpool_call_node_removal_hooks, varpool_add_variable_insertion_hook,
2886 varpool_remove_variable_insertion_hook,
2887 varpool_call_variable_insertion_hooks): New functions.
2888 (varpool_remove_node): Use it.
2890 2013-09-08 Paolo Carlini <paolo.carlini@oracle.com>
2893 * diagnostic.c (diagnostic_build_prefix): When s.file is
2894 "<built-in>" don't output line and column numbers.
2896 2013-09-06 Jan Hubicka <jh@suse.cz>
2898 * cgraphunit.c (expand_thunk): Get body before touching arguments.
2899 * lto-streamer-out.c: Stream thunks, too.
2900 * lto-streamer-in.c (input_function): Pop cfun here
2901 (lto_read_body): Instead of here.
2903 2013-09-06 Caroline Tice <cmtice@google.com>
2905 * doc/install.texi: Add documentation for the --enable-vtable-verify
2906 and the --disable-libvtv configure options.
2908 2013-09-06 Jeff Law <law@redhat.com>
2910 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
2911 edge implied equivalences into successor phis.
2913 2013-09-06 Joern Rennecke <joern.rennecke@embecosm.com>
2915 * resource.c (mark_referenced_resources): Handle COND_EXEC.
2917 2013-09-06 Claudiu Zissulescu <claziss@synopsys.com>
2919 * resource.c (mark_target_live_regs): Compute resources taking
2920 into account if a call is predicated or not.
2922 2013-09-06 Eric Botcazou <ebotcazou@adacore.com>
2924 * toplev.c (output_stack_usage): Be prepared for suffixes created by
2925 the compiler in the function names.
2927 2013-09-06 Jan Hubicka <jh@suse.cz>
2930 * ipa-inline.c (has_caller_p): New function.
2931 (want_inline_function_to_all_callers_p): Use it.
2932 (sum_callers, inline_to_all_callers): Break out from ...
2933 (ipa_inline): ... here.
2935 2013-09-06 Jan Hubicka <jh@suse.cz>
2937 * config/i386/i386.c (ix86_hard_regno_mode_ok): AVX modes are valid
2938 only when AVX is enabled.
2940 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
2942 * config/aarch64/aarch64.md
2943 (*movtf_aarch64): Use neon_<ls>dm_2 as type where v8type
2944 is fpsimd_<load/store>2.
2945 (load_pair<mode>): Likewise.
2946 (store_pair<mode>): Likewise.
2948 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
2950 * config/arm/types.md (type): Add "mrs" type.
2951 * config/aarch64/aarch64.md
2952 (aarch64_load_tp_hard): Make type "mrs".
2954 (load_tp_hard): Make type "mrs".
2955 * config/arm/cortex-a15.md: Update with new attributes.
2956 * config/arm/cortex-a5.md: Update with new attributes.
2957 * config/arm/cortex-a53.md: Update with new attributes.
2958 * config/arm/cortex-a7.md: Update with new attributes.
2959 * config/arm/cortex-a8.md: Update with new attributes.
2960 * config/arm/cortex-a9.md: Update with new attributes.
2961 * config/arm/cortex-m4.md: Update with new attributes.
2962 * config/arm/cortex-r4.md: Update with new attributes.
2963 * config/arm/fa526.md: Update with new attributes.
2964 * config/arm/fa606te.md: Update with new attributes.
2965 * config/arm/fa626te.md: Update with new attributes.
2966 * config/arm/fa726te.md: Update with new attributes.
2968 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
2970 * config/aarch64/aarch64.md
2971 (*movti_aarch64): Use "multiple" for type where v8type is "move2".
2972 (*movtf_aarch64): Likewise.
2974 (thumb1_movdi_insn): Use "multiple" for type where more than one
2975 instruction is used for a move.
2976 (*arm32_movhf): Likewise.
2977 (*thumb_movdf_insn): Likewise.
2979 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
2981 * config/arm/types.md (type): Rename fcpys to fmov.
2983 (*arm_movsi_vfp): Rename type fcpys as fmov.
2984 (*thumb2_movsi_vfp): Likewise
2985 (*movhf_vfp_neon): Likewise
2986 (*movhf_vfp): Likewise
2987 (*movsf_vfp): Likewise
2988 (*thumb2_movsf_vfp): Likewise
2989 (*movsfcc_vfp): Likewise
2990 (*thumb2_movsfcc_vfp): Likewise
2991 * config/aarch64/aarch64-simd.md
2992 (move_lo_quad_<mode>): Replace type mov_reg with fmovs.
2993 * config/aarch64/aarch64.md
2994 (*movsi_aarch64): Replace type mov_reg with fmovs.
2995 (*movdi_aarch64): Likewise
2996 (*movsf_aarch64): Likewise
2997 (*movdf_aarch64): Likewise
2999 (cortexa7_older_only): Rename TYPE_FCPYS to TYPE_FMOV.
3000 * config/arm/iwmmxt.md
3001 (*iwmmxt_movsi_insn): Rename type fcpys as fmov.
3002 * config/arm/arm1020e.md: Update with new attributes.
3003 * config/arm/cortex-a15-neon.md: Update with new attributes.
3004 * config/arm/cortex-a5.md: Update with new attributes.
3005 * config/arm/cortex-a53.md: Update with new attributes.
3006 * config/arm/cortex-a7.md: Update with new attributes.
3007 * config/arm/cortex-a8-neon.md: Update with new attributes.
3008 * config/arm/cortex-a9.md: Update with new attributes.
3009 * config/arm/cortex-m4-fpu.md: Update with new attributes.
3010 * config/arm/cortex-r4f.md: Update with new attributes.
3011 * config/arm/marvell-pj4.md: Update with new attributes.
3012 * config/arm/vfp11.md: Update with new attributes.
3014 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
3016 * config/aarch64/aarch64.md
3017 (*madd<mode>): Fix type attribute.
3018 (*maddsi_uxtw): Likewise.
3019 (*msub<mode>): Likewise.
3020 (*msubsi_uxtw): Likewise.
3021 (<su_optab>maddsidi4): Likewise.
3022 (<su_optab>msubsidi4): Likewise.
3024 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
3026 * config/arm/types.md: Split fdiv<sd> as fsqrt<sd>, fdiv<sd>.
3027 * config/arm/arm.md (core_cycles): Remove fdiv.
3028 * config/arm/vfp.md:
3029 (*sqrtsf2_vfp): Update for attribute changes.
3030 (*sqrtdf2_vfp): Likewise.
3031 * config/aarch64/aarch64.md:
3032 (sqrt<mode>2): Update for attribute changes.
3033 * config/arm/arm1020e.md: Update with new attributes.
3034 * config/arm/cortex-a15-neon.md: Update with new attributes.
3035 * config/arm/cortex-a5.md: Update with new attributes.
3036 * config/arm/cortex-a53.md: Update with new attributes.
3037 * config/arm/cortex-a7.md: Update with new attributes.
3038 * config/arm/cortex-a8-neon.md: Update with new attributes.
3039 * config/arm/cortex-a9.md: Update with new attributes.
3040 * config/arm/cortex-m4-fpu.md: Update with new attributes.
3041 * config/arm/cortex-r4f.md: Update with new attributes.
3042 * config/arm/marvell-pj4.md: Update with new attributes.
3043 * config/arm/vfp11.md: Update with new attributes.
3045 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
3047 * config/arm/types.md
3048 (type): Split f_cvt as f_cvt, f_cvtf2i, f_cvti2f.
3049 * config/aarch64/aarch64.md
3050 (l<fcvt_pattern><su_optab><GPF:mode><GPI:mode>2): Update with
3052 (fix_trunc<GPF:mode><GPI:mode>2): Likewise.
3053 (fixuns_trunc<GPF:mode><GPI:mode>2): Likewise.
3054 (float<GPI:mode><GPF:mode>2): Likewise.
3056 (*truncsisf2_vfp): Update with new attributes.
3057 (*truncsidf2_vfp): Likewise.
3058 (fixuns_truncsfsi2): Likewise.
3059 (fixuns_truncdfsi2): Likewise.
3060 (*floatsisf2_vfp): Likewise.
3061 (*floatsidf2_vfp): Likewise.
3062 (floatunssisf2): Likewise.
3063 (floatunssidf2): Likewise.
3064 (*combine_vcvt_f32_<FCVTI32typename>): Likewise.
3065 (*combine_vcvt_f64_<FCVTI32typename>): Likewise.
3066 * config/arm/arm1020e.md: Update with new attributes.
3067 * config/arm/cortex-a15-neon.md: Update with new attributes.
3068 * config/arm/cortex-a5.md: Update with new attributes.
3069 * config/arm/cortex-a53.md: Update with new attributes.
3070 * config/arm/cortex-a7.md: Update with new attributes.
3071 * config/arm/cortex-a8-neon.md: Update with new attributes.
3072 * config/arm/cortex-a9.md: Update with new attributes.
3073 * config/arm/cortex-m4-fpu.md: Update with new attributes.
3074 * config/arm/cortex-r4f.md: Update with new attributes.
3075 * config/arm/marvell-pj4.md: Update with new attributes.
3076 * config/arm/vfp11.md: Update with new attributes.
3078 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
3080 * config/aarch64/arm_neon.h
3081 (vqtbl<1,2,3,4><q>_s8): Fix control vector parameter type.
3082 (vqtbx<1,2,3,4><q>_s8): Likewise.
3084 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
3086 * config/arm/types.md: Add "no_insn", "multiple" and "untyped" types.
3087 * config/arm/arm-fixed.md: Add type attribute to all insn patterns.
3088 (add<mode>3): Add type attribute.
3089 (add<mode>3): Likewise.
3090 (usadd<mode>3): Likewise.
3091 (ssadd<mode>3): Likewise.
3092 (sub<mode>3): Likewise.
3093 (sub<mode>3): Likewise.
3094 (ussub<mode>3): Likewise.
3095 (sssub<mode>3): Likewise.
3096 (ssmulsa3): Likewise.
3097 (usmulusa3): Likewise.
3098 (arm_usatsihi): Likewise.
3100 (*movdi_vfp): Add types for all instructions.
3101 (*movdi_vfp_cortexa8): Likewise.
3102 (*movhf_vfp_neon): Likewise.
3103 (*movhf_vfp): Likewise.
3104 (*movdf_vfp): Likewise.
3105 (*thumb2_movdf_vfp): Likewise.
3106 (*thumb2_movdfcc_vfp): Likewise.
3107 * config/arm/arm.md: Add type attribute to all insn patterns.
3108 (*thumb1_adddi3): Add type attribute.
3109 (*arm_adddi3): Likewise.
3110 (*adddi_sesidi_di): Likewise.
3111 (*adddi_zesidi_di): Likewise.
3112 (*thumb1_addsi3): Likewise.
3113 (addsi3_compare0): Likewise.
3114 (*addsi3_compare0_scratch): Likewise.
3115 (*compare_negsi_si): Likewise.
3116 (cmpsi2_addneg): Likewise.
3117 (*addsi3_carryin_<optab>): Likewise.
3118 (*addsi3_carryin_alt2_<optab>): Likewise.
3119 (*addsi3_carryin_clobercc_<optab>): Likewise.
3120 (*subsi3_carryin): Likewise.
3121 (*subsi3_carryin_const): Likewise.
3122 (*subsi3_carryin_compare): Likewise.
3123 (*subsi3_carryin_compare_const): Likewise.
3124 (*arm_subdi3): Likewise.
3125 (*thumb_subdi3): Likewise.
3126 (*subdi_di_zesidi): Likewise.
3127 (*subdi_di_sesidi): Likewise.
3128 (*subdi_zesidi_di): Likewise.
3129 (*subdi_sesidi_di): Likewise.
3130 (*subdi_zesidi_ze): Likewise.
3131 (thumb1_subsi3_insn): Likewise.
3132 (*arm_subsi3_insn): Likewise.
3133 (*anddi3_insn): Likewise.
3134 (*anddi_zesidi_di): Likewise.
3135 (*anddi_sesdi_di): Likewise.
3136 (*ne_zeroextracts): Likewise.
3137 (*ne_zeroextracts): Likewise.
3138 (*ite_ne_zeroextr): Likewise.
3139 (*ite_ne_zeroextr): Likewise.
3140 (*anddi_notdi_di): Likewise.
3141 (*anddi_notzesidi): Likewise.
3142 (*anddi_notsesidi): Likewise.
3143 (andsi_notsi_si): Likewise.
3144 (thumb1_bicsi3): Likewise.
3145 (*iordi3_insn): Likewise.
3146 (*iordi_zesidi_di): Likewise.
3147 (*iordi_sesidi_di): Likewise.
3148 (*thumb1_iorsi3_insn): Likewise.
3149 (*xordi3_insn): Likewise.
3150 (*xordi_zesidi_di): Likewise.
3151 (*xordi_sesidi_di): Likewise.
3152 (*arm_xorsi3): Likewise.
3153 (*andsi_iorsi3_no): Likewise.
3154 (*smax_0): Likewise.
3155 (*smax_m1): Likewise.
3156 (*arm_smax_insn): Likewise.
3157 (*smin_0): Likewise.
3158 (*arm_smin_insn): Likewise.
3159 (*arm_umaxsi3): Likewise.
3160 (*arm_uminsi3): Likewise.
3161 (*minmax_arithsi): Likewise.
3162 (*minmax_arithsi_): Likewise.
3163 (*satsi_<SAT:code>): Likewise.
3164 (arm_ashldi3_1bit): Likewise.
3165 (arm_ashrdi3_1bit): Likewise.
3166 (arm_lshrdi3_1bit): Likewise.
3167 (*arm_negdi2): Likewise.
3168 (*thumb1_negdi2): Likewise.
3169 (*arm_negsi2): Likewise.
3170 (*thumb1_negsi2): Likewise.
3171 (*negdi_extendsid): Likewise.
3172 (*negdi_zero_extend): Likewise.
3173 (*arm_abssi2): Likewise.
3174 (*thumb1_abssi2): Likewise.
3175 (*arm_neg_abssi2): Likewise.
3176 (*thumb1_neg_abss): Likewise.
3177 (one_cmpldi2): Likewise.
3178 (extend<mode>di2): Likewise.
3179 (*compareqi_eq0): Likewise.
3180 (*arm_extendhisi2addsi): Likewise.
3181 (*arm_movdi): Likewise.
3182 (*thumb1_movdi_insn): Likewise.
3183 (*arm_movt): Likewise.
3184 (*thumb1_movsi_insn): Likewise.
3185 (pic_add_dot_plus_four): Likewise.
3186 (pic_add_dot_plus_eight): Likewise.
3187 (tls_load_dot_plus_eight): Likewise.
3188 (*thumb1_movhi_insn): Likewise.
3189 (*thumb1_movsf_insn): Likewise.
3190 (*movdf_soft_insn): Likewise.
3191 (*thumb_movdf_insn): Likewise.
3192 (cbranchsi4_insn): Likewise.
3193 (cbranchsi4_scratch): Likewise.
3194 (*negated_cbranchsi4): Likewise.
3195 (*tbit_cbranch): Likewise.
3196 (*tlobits_cbranch): Likewise.
3197 (*tstsi3_cbranch): Likewise.
3198 (*cbranchne_decr1): Likewise.
3199 (*addsi3_cbranch): Likewise.
3200 (*addsi3_cbranch_scratch): Likewise.
3201 (*arm_cmpdi_insn): Likewise.
3202 (*arm_cmpdi_unsig): Likewise.
3203 (*arm_cmpdi_zero): Likewise.
3204 (*thumb_cmpdi_zero): Likewise.
3205 (*deleted_compare): Likewise.
3206 (*mov_scc): Likewise.
3207 (*mov_negscc): Likewise.
3208 (*mov_notscc): Likewise.
3209 (*cstoresi_eq0_thumb1_insn): Likewise.
3210 (cstoresi_nltu_thumb1): Likewise.
3211 (cstoresi_ltu_thu): Likewise.
3212 (thumb1_addsi3_addgeu): Likewise.
3213 (*arm_jump): Likewise.
3214 (*thumb_jump): Likewise.
3215 (*check_arch2): Likewise.
3216 (arm_casesi_internal): Likewise.
3217 (thumb1_casesi_dispatch): Likewise.
3218 (*arm_indirect_jump): Likewise.
3219 (*thumb1_indirect_jump): Likewise.
3221 (*and_scc): Likewise.
3222 (*ior_scc): Likewise.
3223 (*compare_scc): Likewise.
3224 (*cond_move): Likewise.
3225 (*cond_arith): Likewise.
3226 (*cond_sub): Likewise.
3227 (*cmp_ite0): Likewise.
3228 (*cmp_ite1): Likewise.
3229 (*cmp_and): Likewise.
3230 (*cmp_ior): Likewise.
3231 (*ior_scc_scc): Likewise.
3232 (*ior_scc_scc_cmp): Likewise.
3233 (*and_scc_scc): Likewise.
3234 (*and_scc_scc_cmp): Likewise.
3235 (*and_scc_scc_nod): Likewise.
3236 (*negscc): Likewise.
3237 (movcond_addsi): Likewise.
3238 (movcond): Likewise.
3239 (*ifcompare_plus_move): Likewise.
3240 (*if_plus_move): Likewise.
3241 (*ifcompare_move_plus): Likewise.
3242 (*if_move_plus): Likewise.
3243 (*ifcompare_arith_arith): Likewise.
3244 (*if_arith_arith): Likewise.
3245 (*ifcompare_arith_move): Likewise.
3246 (*if_arith_move): Likewise.
3247 (*ifcompare_move_arith): Likewise.
3248 (*if_move_arith): Likewise.
3249 (*ifcompare_move_not): Likewise.
3250 (*if_move_not): Likewise.
3251 (*ifcompare_not_move): Likewise.
3252 (*if_not_move): Likewise.
3253 (*ifcompare_shift_move): Likewise.
3254 (*if_shift_move): Likewise.
3255 (*ifcompare_move_shift): Likewise.
3256 (*if_move_shift): Likewise.
3257 (*ifcompare_shift_shift): Likewise.
3258 (*ifcompare_not_arith): Likewise.
3259 (*ifcompare_arith_not): Likewise.
3260 (*if_arith_not): Likewise.
3261 (*ifcompare_neg_move): Likewise.
3262 (*if_neg_move): Likewise.
3263 (*ifcompare_move_neg): Likewise.
3264 (*if_move_neg): Likewise.
3265 (prologue_thumb1_interwork): Likewise.
3266 (*cond_move_not): Likewise.
3267 (*sign_extract_onebit): Likewise.
3268 (*not_signextract_onebit): Likewise.
3269 (stack_tie): Likewise.
3270 (align_4): Likewise.
3271 (align_8): Likewise.
3272 (consttable_end): Likewise.
3273 (consttable_1): Likewise.
3274 (consttable_2): Likewise.
3275 (consttable_4): Likewise.
3276 (consttable_8): Likewise.
3277 (consttable_16): Likewise.
3278 (*thumb1_tablejump): Likewise.
3279 (prefetch): Likewise.
3280 (force_register_use): Likewise.
3281 (thumb_eh_return): Likewise.
3282 (load_tp_hard): Likewise.
3283 (load_tp_soft): Likewise.
3284 (tlscall): Likewise.
3285 (*arm_movtas_ze): Likewise.
3286 (*arm_rev): Likewise.
3287 (*arm_revsh): Likewise.
3288 (*arm_rev16): Likewise.
3289 * config/arm/thumb2.md
3290 (*thumb2_smaxsi3): Likewise.
3291 (*thumb2_sminsi3): Likewise.
3292 (*thumb32_umaxsi3): Likewise.
3293 (*thumb2_uminsi3): Likewise.
3294 (*thumb2_negdi2): Likewise.
3295 (*thumb2_abssi2): Likewise.
3296 (*thumb2_neg_abss): Likewise.
3297 (*thumb2_movsi_insn): Likewise.
3298 (tls_load_dot_plus_four): Likewise.
3299 (*thumb2_movhi_insn): Likewise.
3300 (*thumb2_mov_scc): Likewise.
3301 (*thumb2_mov_negs): Likewise.
3302 (*thumb2_mov_negs): Likewise.
3303 (*thumb2_mov_nots): Likewise.
3304 (*thumb2_mov_nots): Likewise.
3305 (*thumb2_movsicc_): Likewise.
3306 (*thumb2_movsfcc_soft_insn): Likewise.
3307 (*thumb2_indirect_jump): Likewise.
3308 (*thumb2_and_scc): Likewise.
3309 (*thumb2_ior_scc): Likewise.
3310 (*thumb2_ior_scc_strict_it): Likewise.
3311 (*thumb2_cond_move): Likewise.
3312 (*thumb2_cond_arith): Likewise.
3313 (*thumb2_cond_ari): Likewise.
3314 (*thumb2_cond_sub): Likewise.
3315 (*thumb2_negscc): Likewise.
3316 (*thumb2_movcond): Likewise.
3317 (thumb2_casesi_internal): Likewise.
3318 (thumb2_casesi_internal_pic): Likewise.
3319 (*thumb2_alusi3_short): Likewise.
3320 (*thumb2_mov<mode>_shortim): Likewise.
3321 (*thumb2_addsi_short): Likewise.
3322 (*thumb2_subsi_short): Likewise.
3323 (thumb2_addsi3_compare0): Likewise.
3324 (*thumb2_cbz): Likewise.
3325 (*thumb2_cbnz): Likewise.
3326 (*thumb2_one_cmplsi2_short): Likewise.
3327 (*thumb2_negsi2_short): Likewise.
3328 (*orsi_notsi_si): Likewise.
3329 * config/arm/arm1020e.md: Update with new attributes.
3330 * config/arm/arm1026ejs.md: Update with new attributes.
3331 * config/arm/arm1136jfs.md: Update with new attributes.
3332 * config/arm/arm926ejs.md: Update with new attributes.
3333 * config/arm/cortex-a15.md: Update with new attributes.
3334 * config/arm/cortex-a5.md: Update with new attributes.
3335 * config/arm/cortex-a53.md: Update with new attributes.
3336 * config/arm/cortex-a7.md: Update with new attributes.
3337 * config/arm/cortex-a8.md: Update with new attributes.
3338 * config/arm/cortex-a9.md: Update with new attributes.
3339 * config/arm/cortex-m4.md: Update with new attributes.
3340 * config/arm/cortex-r4.md: Update with new attributes.
3341 * config/arm/fa526.md: Update with new attributes.
3342 * config/arm/fa606te.md: Update with new attributes.
3343 * config/arm/fa626te.md: Update with new attributes.
3344 * config/arm/fa726te.md: Update with new attributes.
3346 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
3348 * config/aarch64/aarch64-simd.md
3349 (aarch64_sqdml<SBINQOPS:as>l_n<mode>_internal): Use
3350 <vwx> iterator to ensure correct register choice.
3351 (aarch64_sqdml<SBINQOPS:as>l2_n<mode>_internal): Likewise.
3352 (aarch64_sqdmull_n<mode>): Likewise.
3353 (aarch64_sqdmull2_n<mode>_internal): Likewise.
3354 * config/aarch64/arm_neon.h
3355 (vml<as><q>_lane<q>_<su>16): Use 'x' constraint for element vector.
3356 (vml<as><q>_n_<su>16): Likewise.
3357 (vml<as>l_high_lane<q>_<su>16): Likewise.
3358 (vml<as>l_high_n_<su>16): Likewise.
3359 (vml<as>l_lane<q>_<su>16): Likewise.
3360 (vml<as>l_n_<su>16): Likewise.
3361 (vmul<q>_lane<q>_<su>16): Likewise.
3362 (vmul<q>_n_<su>16): Likewise.
3363 (vmull_lane<q>_<su>16): Likewise.
3364 (vmull_n_<su>16): Likewise.
3365 (vmull_high_lane<q>_<su>16): Likewise.
3366 (vmull_high_n_<su>16): Likewise.
3367 (vqrdmulh<q>_n_s16): Likewise.
3369 2013-09-06 Tejas Belagod <tejas.belagod@arm.com>
3371 * config/aarch64/arm_neon.h: Fix all vdup<bhsd_lane<q> intrinsics to
3372 have the correct lane parameter.
3374 2013-09-06 Richard Biener <rguenther@suse.de>
3376 * cfganal.c (control_dependences::~control_dependences):
3377 Properly free all of the vector.
3379 2013-09-06 Kirill Yukhin <kirill.yukhin@intel.com>
3382 * config/i386/i386.c (ix86_conditional_register_usage):
3383 Proper initialize extended SSE registers.
3385 2013-09-06 Jan Hubicka <jh@suse.cz>
3387 PR tree-optimization/58311
3388 * ipa-devirt.c (gate_ipa_devirt): Only execute when optimizing.
3390 2013-09-06 Jan Hubicka <jh@suse.cz>
3392 * Makefile.in (tree-sra.o): Update dependencies.
3393 * tree-sra.c: Include ipa-utils.h
3394 (scan_function): Use recursive_call_p.
3395 (has_caller_p): New function.
3396 (cgraph_for_node_and_aliases): Count also callers of aliases.
3398 2013-09-06 Jan Hubicka <jh@suse.cz>
3401 * cgraph.h (symtab_semantically_equivalent_p): Declare.
3402 * tree-tailcall.c: Include ipa-utils.h.
3403 (find_tail_calls): Use it.
3404 * ipa-pure-const.c (check_call): Likewise.
3405 * ipa-utils.c (recursive_call_p): New function.
3406 * ipa-utils.h (recursive_call_p): Dclare.
3407 * symtab.c (symtab_nonoverwritable_alias): Fix formatting.
3408 (symtab_semantically_equivalent_p): New function.
3409 * Makefile.in (tree-tailcall.o): Update dependencies.
3411 2013-09-06 Eric Botcazou <ebotcazou@adacore.com>
3413 * ipa-split.c (split_function): Set DECL_NO_INLINE_WARNING_P on the
3416 2013-09-06 Richard Biener <rguenther@suse.de>
3418 * lto-streamer.h (lto_global_var_decls): Remove.
3419 * Makefile.in (OBJS): Remove lto-symtab.o.
3420 (lto-symtab.o): Remove.
3421 (GTFILES): Remove lto-symtab.c
3422 * lto-symtab.c: Move to lto/
3424 2013-09-06 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
3426 * config/s390/s390.md (UNSPEC_FPINT_FLOOR, UNSPEC_FPINT_BTRUNC)
3427 (UNSPEC_FPINT_ROUND, UNSPEC_FPINT_CEIL, UNSPEC_FPINT_NEARBYINT)
3428 (UNSPEC_FPINT_RINT): New constant definitions.
3429 (FPINT, fpint_name, fpint_roundingmode): New integer iterator
3430 definition with 2 attributes.
3431 ("<FPINT:fpint_name><BFP:mode>2", "rint<BFP:mode>2")
3432 ("<FPINT:fpint_name><DFP:mode>2", "rint<DFP:mode>2"): New pattern
3435 2013-09-06 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
3437 * config/s390/s390.md: Add "bcr_flush" value to mnemonic attribute.
3438 ("mem_thread_fence_1"): Use bcr 14,0 for z196 and later.
3439 Set the mnemonic attribute to "bcr_flush". Set the "z196prop"
3440 attribute to "z196_alone".
3441 * config/s390/2827.md: Add "bcr_flush" to "ooo_groupalone" and
3444 2013-09-06 Richard Biener <rguenther@suse.de>
3446 * basic-block.h (class control_dependences): New.
3447 * tree-ssa-dce.c (control_dependence_map): Remove.
3449 (EXECUTE_IF_CONTROL_DEPENDENT): Remove.
3450 (set_control_dependence_map_bit, clear_control_dependence_bitmap,
3451 find_pdom, find_control_dependence, find_all_control_dependences):
3453 (mark_control_dependent_edges_necessary,
3454 find_obviously_necessary_stmts, propagate_necessity, tree_dce_init,
3455 tree_dce_done, perform_tree_ssa_dce): Adjust.
3456 * cfganal.c (set_control_dependence_map_bit,
3457 clear_control_dependence_bitmap, find_pdom, find_control_dependence,
3458 find_all_control_dependences): Move from tree-ssa-dce.c and
3459 implement as methods of control_dependences class.
3460 (control_dependences::control_dependences): New.
3461 (control_dependences::~control_dependences): Likewise.
3462 (control_dependences::get_edges_dependent_on): Likewise.
3463 (control_dependences::get_edge): Likewise.
3465 2013-09-04 Jan Hubicka <jh@suse.cz>
3467 * tree.c (types_same_for_odr): Drop overactive check.
3468 * ipa-devirt.c (hash_type_name): Likewise.
3470 2013-09-04 Jan Hubicka <jh@suse.cz>
3472 * cgraphunit.c (walk_polymorphic_call_targets): Break out from ...
3473 (analyze_functions): ... here.
3475 2013-09-04 Jan Hubicka <jh@suse.cz>
3478 * cgraphunit.c (analyze_functions): Clear AUX fields
3479 after processing; initialize assembler name has.
3481 2013-09-05 Jeff Law <law@redhat.com>
3483 * tree-ssa-threadedge.c (thread_around_empty_blocks): Renamed
3484 from thread_around_empty_block. Record threading path into PATH.
3485 Recurse if threading through the initial block is successful.
3486 (thread_across_edge): Corresponding changes to slightly simplify.
3488 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
3490 * config/aarch64/aarch64.md
3491 (type): Remove frecpe, frecps, frecpx.
3492 (aarch64_frecp<FRECP:frecp_suffix><mode>): Move to aarch64-simd.md,
3493 fix to be a TARGET_SIMD instruction.
3494 (aarch64_frecps): Remove.
3495 * config/aarch64/aarch64-simd.md
3496 (aarch64_frecp<FRECP:frecp_suffix><mode>): New, moved from aarch64.md
3497 (aarch64_frecps<mode>): Handle all float/vector of float modes.
3499 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
3500 Sofiane Naci <sofiane.naci@arm.com>
3502 * config/arm/types.md (define_attr "type"): Expand "arlo_imm"
3503 into "adr", "alu_imm", "alus_imm", "logic_imm", "logics_imm".
3504 Expand "arlo_reg" into "adc_reg", "adc_imm", "adcs_reg", "adcs_imm",
3505 "alu_ext", "alu_reg", "alus_ext", "alus_reg", "bfm", "csel",
3506 "logic_reg", "logics_reg", "rev". Expand "arlo_shift" into
3507 "alu_shift_imm", "alus_shift_imm", "logic_shift_imm",
3508 "logics_shift_imm". Expand "arlo_shift_reg" into "alu_shift_reg",
3509 "alus_shift_reg", "logic_shift_reg", "logics_shift_reg". Expand "clz"
3510 into "clz, "rbit". Rename "shift" to "shift_imm".
3511 * config/arm/arm.md (define_attr "core_cycles"): Update for attribute
3512 changes. Update for attribute changes all occurrences of arlo_* and
3514 * config/arm/arm-fixed.md: Update for attribute changes
3515 all occurrences of arlo_* types.
3516 * config/arm/thumb2.md: Update for attribute changes all occurrences
3518 * config/arm/arm.c (xscale_sched_adjust_cost): (rtx insn, rtx
3519 (cortexa7_older_only): Likewise.
3520 (cortexa7_younger): Likewise.
3521 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
3522 (1020alu_shift_op): Likewise.
3523 (1020alu_shift_reg_op): Likewise.
3524 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
3525 (alu_shift_op): Likewise.
3526 (alu_shift_reg_op): Likewise.
3527 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
3528 (11_alu_shift_op): Likewise.
3529 (11_alu_shift_reg_op): Likewise.
3530 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
3531 (9_alu_shift_reg_op): Likewise.
3532 * config/arm/cortex-a15.md (cortex_a15_alu): Update for
3534 (cortex_a15_alu_shift): Likewise.
3535 (cortex_a15_alu_shift_reg): Likewise.
3536 * config/arm/cortex-a5.md (cortex_a5_alu): Update for
3538 (cortex_a5_alu_shift): Likewise.
3539 * config/arm/cortex-a53.md (cortex_a53_alu): Update for
3541 (cortex_a53_alu_shift): Likewise.
3542 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for
3544 (cortex_a7_alu_reg): Likewise.
3545 (cortex_a7_alu_shift): Likewise.
3546 * config/arm/cortex-a8.md (cortex_a8_alu): Update for
3548 (cortex_a8_alu_shift): Likewise.
3549 (cortex_a8_alu_shift_reg): Likewise.
3550 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
3551 (cortex_a9_dp_shift): Likewise.
3552 * config/arm/cortex-m4.md (cortex_m4_alu): Update for
3554 * config/arm/cortex-r4.md
3555 (cortex_r4_alu): Update for attribute changes.
3556 (cortex_r4_mov): Likewise.
3557 (cortex_r4_alu_shift_reg): Likewise.
3558 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
3559 (526_alu_shift_op): Likewise.
3560 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
3561 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
3562 (626te_alu_shift_op): Likewise.
3563 * config/arm/fa726te.md (726te_alu_op): Update for attribute changes.
3564 (726te_alu_shift_op): Likewise.
3565 (726te_alu_shift_reg_op): Likewise.
3566 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
3567 (mp626_alu_shift_op): Likewise.
3568 * config/arm/marvell-pj4.md (pj4_alu): Update for attribute changes.
3569 (pj4_alu_conds): Likewise.
3570 (pj4_shift): Likewise.
3571 (pj4_shift_conds): Likewise.
3572 (pj4_alu_shift): Likewise.
3573 (pj4_alu_shift_conds): Likewise.
3574 * config/aarch64/aarch64.md: Update for attribute change
3575 all occurrences of arlo_* and shift* types.
3577 2013-09-05 Mike Stump <mikestump@comcast.net>
3579 * tree.h: Move documentation for tree_function_decl to tree-core.h
3580 with the declaration.
3582 2013-09-05 Peter Bergner <bergner@vnet.ibm.com>
3585 * reginfo.c (choose_hard_reg_mode): Scan through all mode classes
3586 looking for widest mode.
3588 2013-09-05 Eric Botcazou <ebotcazou@adacore.com>
3590 * config.gcc (*-*-vxworks*): Do not override an existing extra_objs.
3592 2013-09-05 Richard Biener <rguenther@suse.de>
3594 PR tree-optimization/58137
3595 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size):
3596 Do not create vectors of pointers.
3597 * tree-vect-loop.c (get_initial_def_for_induction): Use proper
3598 types for the components of the vector initializer.
3599 * tree-cfg.c (verify_gimple_assign_binary): Remove special-casing
3600 allowing pointer vectors with PLUS_EXPR/MINUS_EXPR.
3602 2013-09-05 Martin Jambor <mjambor@suse.cz>
3604 * ipa-prop.c (remove_described_reference): Accept missing references,
3605 return false if that hppens, otherwise return true.
3606 (cgraph_node_for_jfunc): New function.
3607 (try_decrement_rdesc_refcount): Likewise.
3608 (try_make_edge_direct_simple_call): Use them.
3609 (ipa_edge_removal_hook): Remove references from rdescs.
3610 (ipa_edge_duplication_hook): Clone rdescs and their references
3611 when the new edge has the same caller as the old one.
3612 * cgraph.c (cgraph_resolve_speculation): Remove speculative
3613 reference before removing any edges.
3615 2013-09-05 Richard Earnshaw <rearnsha@arm.com>
3617 * arm.c (thumb2_emit_strd_push): Rewrite to use pre-decrement on
3619 * thumb2.md (thumb2_storewb_parisi): New pattern.
3621 2013-09-05 Yufeng Zhang <yufeng.zhang@arm.com>
3623 * config/aarch64/aarch64-option-extensions.def: Add
3624 AARCH64_OPT_EXTENSION of 'crc'.
3625 * config/aarch64/aarch64.h (AARCH64_FL_CRC): New define.
3626 (AARCH64_ISA_CRC): Ditto.
3627 * doc/invoke.texi (-march and -mcpu feature modifiers): Add
3628 description of the CRC extension.
3630 2013-09-05 Alexander Ivchenko <alexander.ivchenko@intel.com>
3632 * config/rs6000/linux64.h: Define OPTION_BIONIC and OPTION_UCLIBC.
3633 * config/rs6000/linux.h: Ditto.
3634 * alpha/linux.h: Ditto.
3635 * config/bfin/uclinux.h: Define TARGET_LIBC_HAS_FUNCTION as
3636 no_c99_libc_has_function.
3637 * config/c6x/uclinux-elf.h: Ditto.
3638 * config/lm32/uclinux-elf.h: Ditto.
3639 * config/m68k/uclinux.h: Ditto.
3640 * config/moxie/uclinux.h: Ditto.
3641 * config.gcc (bfin*-linux-uclibc*): Add t-linux-android to tmake_file.
3642 (crisv32-*-linux*, cris-*-linux*): Ditto.
3643 * config/bfin/bfin.c: Include "tm_p.h".
3645 2013-09-05 Richard Biener <rguenther@suse.de>
3647 * tree-vect-loop.c (vect_analyze_loop_operations): Properly
3648 check for a definition without a basic-block.
3650 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
3651 Sofiane Naci <sofiane.naci@arm.com>
3653 * config/aarch64/aarch64.md
3654 (*movti_aarch64): Rename r_2_f and f_2_r.
3655 (*movsf_aarch64): Likewise.
3656 (*movdf_aarch64): Likewise.
3657 (*movtf_aarch64): Likewise.
3658 (aarch64_movdi_<mode>low): Likewise.
3659 (aarch64_movdi_<mode>high): Likewise.
3660 (aarch64_mov<mode>high_di): Likewise.
3661 (aarch64_mov<mode>low_di): Likewise.
3662 (aarch64_movtilow_tilow): Likewise.
3663 * config/arm/arm.md (attribute "neon_type"): Delete. Move attribute
3664 values to config/arm/types.md
3665 (attribute "conds"): Update for attribute change.
3666 (anddi3_insn): Likewise.
3667 (iordi3_insn): Likewise.
3668 (xordi3_insn): Likewise.
3669 (one_cmpldi2): Likewise.
3670 * config/arm/types.md (type): Add Neon types.
3671 * config/arm/neon.md (neon_mov<mode>): Remove "neon_type" attribute,
3672 use "type" attribute.
3673 (movmisalign<mode>_neon_store): Likewise.
3674 (movmisalign<mode>_neon_load): Likewise.
3675 (vec_set<mode>_internal): Likewise.
3676 (vec_setv2di_internal): Likewise.
3677 (vec_extract<mode>): Likewise.
3678 (vec_extractv2di): Likewise.
3679 (add<mode>3_neon): Likewise.
3680 (adddi3_neon): Likewise.
3681 (sub<mode>3_neon): Likewise.
3682 (subdi3_neon): Likewise.
3683 (mul<mode>3_neon): Likewise.
3684 (mul<mode>3add<mode>_neon): Likewise.
3685 (mul<mode>3neg<mode>add<mode>_neon): Likewise.
3686 (fma<VCVTF:mode>4)): Likewise.
3687 (fma<VCVTF:mode>4_intrinsic): Likewise.
3688 (fmsub<VCVTF:mode>4)): Likewise.
3689 (fmsub<VCVTF:mode>4_intrinsic): Likewise.
3690 (neon_vrint<NEON_VRINT:nvrint_variant><VCVTF:mode>): Likewise.
3691 (ior<mode>3): Likewise.
3692 (and<mode>3): Likewise.
3693 (anddi3_neon): Likewise.
3694 (orn<mode>3_neon): Likewise.
3695 (orndi3_neon): Likewise.
3696 (bic<mode>3_neon): Likewise.
3697 (bicdi3_neon): Likewise.
3698 (xor<mode>3): Likewise.
3699 (one_cmpl<mode>2): Likewise.
3700 (abs<mode>2): Likewise.
3701 (neg<mode>2): Likewise.
3702 (umin<mode>3_neon): Likewise.
3703 (umax<mode>3_neon): Likewise.
3704 (smin<mode>3_neon): Likewise.
3705 (smax<mode>3_neon): Likewise.
3706 (vashl<mode>3): Likewise.
3707 (vashr<mode>3_imm): Likewise.
3708 (vlshr<mode>3_imm): Likewise.
3709 (ashl<mode>3_signed): Likewise.
3710 (ashl<mode>3_unsigned): Likewise.
3711 (neon_load_count): Likewise.
3712 (ashldi3_neon_noclobber): Likewise.
3713 (signed_shift_di3_neon): Likewise.
3714 (unsigned_shift_di3_neon): Likewise.
3715 (ashrdi3_neon_imm_noclobber): Likewise.
3716 (lshrdi3_neon_imm_noclobber): Likewise.
3717 (widen_ssum<mode>3): Likewise.
3718 (widen_usum<mode>3): Likewise.
3719 (quad_halves_<code>v4si): Likewise.
3720 (quad_halves_<code>v4sf): Likewise.
3721 (quad_halves_<code>v8hi): Likewise.
3722 (quad_halves_<code>v16qi): Likewise.
3723 (reduc_splus_v2di): Likewise.
3724 (neon_vpadd_internal<mode>): Likewise.
3725 (neon_vpsmin<mode>): Likewise.
3726 (neon_vpsmax<mode>): Likewise.
3727 (neon_vpumin<mode>): Likewise.
3728 (neon_vpumax<mode>): Likewise.
3729 (ss_add<mode>_neon): Likewise.
3730 (us_add<mode>_neon): Likewise.
3731 (ss_sub<mode>_neon): Likewise.
3732 (us_sub<mode>_neon): Likewise.
3733 (neon_vadd<mode>_unspec): Likewise.
3734 (neon_vaddl<mode>): Likewise.
3735 (neon_vaddw<mode>): Likewise.
3736 (neon_vhadd<mode>): Likewise.
3737 (neon_vqadd<mode>): Likewise.
3738 (neon_vaddhn<mode>): Likewise.
3739 (neon_vmul<mode>): Likewise.
3740 (neon_vmla<mode>): Likewise.
3741 (neon_vmlal<mode>): Likewise.
3742 (neon_vmls<mode>): Likewise.
3743 (neon_vmlsl<mode>): Likewise.
3744 (neon_vqdmulh<mode>): Likewise.
3745 (neon_vqdmlal<mode>): Likewise.
3746 (neon_vqdmlsl<mode>): Likewise.
3747 (neon_vmull<mode>): Likewise.
3748 (neon_vqdmull<mode>): Likewise.
3749 (neon_vsub<mode>_unspec): Likewise.
3750 (neon_vsubl<mode>): Likewise.
3751 (neon_vsubw<mode>): Likewise.
3752 (neon_vqsub<mode>): Likewise.
3753 (neon_vhsub<mode>): Likewise.
3754 (neon_vsubhn<mode>): Likewise.
3755 (neon_vceq<mode>): Likewise.
3756 (neon_vcge<mode>): Likewise.
3757 (neon_vcgeu<mode>): Likewise.
3758 (neon_vcgt<mode>): Likewise.
3759 (neon_vcgtu<mode>): Likewise.
3760 (neon_vcle<mode>): Likewise.
3761 (neon_vclt<mode>): Likewise.
3762 (neon_vcage<mode>): Likewise.
3763 (neon_vcagt<mode>): Likewise.
3764 (neon_vtst<mode>): Likewise.
3765 (neon_vabd<mode>): Likewise.
3766 (neon_vabdl<mode>): Likewise.
3767 (neon_vaba<mode>): Likewise.
3768 (neon_vabal<mode>): Likewise.
3769 (neon_vmax<mode>): Likewise.
3770 (neon_vmin<mode>): Likewise.
3771 (neon_vpaddl<mode>): Likewise.
3772 (neon_vpadal<mode>): Likewise.
3773 (neon_vpmax<mode>): Likewise.
3774 (neon_vpmin<mode>): Likewise.
3775 (neon_vrecps<mode>): Likewise.
3776 (neon_vrsqrts<mode>): Likewise.
3777 (neon_vqabs<mode>): Likewise.
3778 (neon_vqneg<mode>): Likewise.
3779 (neon_vcls<mode>): Likewise.
3780 (clz<mode>2): Likewise.
3781 (popcount<mode>2): Likewise.
3782 (neon_vrecpe): Likewise.
3783 (neon_vrsqrte): Likewise.
3784 (neon_vget_lane<mode>_sext_internal): Likewise.
3785 (neon_vget_lane<mode>_zext_internal): Likewise.
3786 (neon_vdup_n<mode>): Likewise.
3787 (neon_vdup_nv2di): Likewise.
3788 (neon_vdpu_lane<mode>_internal): Likewise.
3789 (neon_vswp<mode>): Likewise.
3790 (float<mode><V_cvtto>2): Likewise.
3791 (floatuns<mode><V_cvtto>2): Likewise.
3792 (fix_trunc<mode><V_cvtto>)2): Likewise
3793 (fixuns_trunc<mode><V_cvtto)2): Likewise.
3794 (neon_vcvt<mode>): Likewise.
3795 (neon_vcvtv4sfv4hf): Likewise.
3796 (neon_vcvtv4hfv4sf): Likewise.
3797 (neon_vcvt_n<mode>): Likewise.
3798 (neon_vmovn<mode>): Likewise.
3799 (neon_vqmovn<mode>): Likewise.
3800 (neon_vqmovun<mode>): Likewise.
3801 (neon_vmovl<mode>): Likewise.
3802 (neon_vmul_lane<mode>): Likewise.
3803 (neon_vmull_lane<mode>): Likewise.
3804 (neon_vqdmull_lane<mode>): Likewise.
3805 (neon_vqdmulh_lane<mode>): Likewise.
3806 (neon_vmla_lane<mode>): Likewise.
3807 (neon_vmlal_lane<mode>): Likewise.
3808 (neon_vqdmlal_lane<mode>): Likewise.
3809 (neon_vmls_lane<mode>): Likewise.
3810 (neon_vmlsl_lane<mode>): Likewise.
3811 (neon_vqdmlsl_lane<mode>): Likewise.
3812 (neon_vext<mode>): Likewise.
3813 (neon_vrev64<mode>): Likewise.
3814 (neon_vrev32<mode>): Likewise.
3815 (neon_vrev16<mode>): Likewise.
3816 (neon_vbsl<mode>_internal): Likewise.
3817 (neon_vshl<mode>): Likewise.
3818 (neon_vqshl<mode>): Likewise.
3819 (neon_vshr_n<mode>): Likewise.
3820 (neon_vshrn_n<mode>): Likewise.
3821 (neon_vqshrn_n<mode>): Likewise.
3822 (neon_vqshrun_n<mode>): Likewise.
3823 (neon_vshl_n<mode>): Likewise.
3824 (neon_vqshl_n<mode>): Likewise.
3825 (neon_vqshlu_n<mode>): Likewise.
3826 (neon_vshll_n<mode>): Likewise.
3827 (neon_vsra_n<mode>): Likewise.
3828 (neon_vsri_n<mode>): Likewise.
3829 (neon_vsli_n<mode>): Likewise.
3830 (neon_vtbl1v8qi): Likewise.
3831 (neon_vtbl2v8qi): Likewise.
3832 (neon_vtbl3v8qi): Likewise.
3833 (neon_vtbl4v8qi): Likewise.
3834 (neon_vtbx1v8qi): Likewise.
3835 (neon_vtbx2v8qi): Likewise.
3836 (neon_vtbx3v8qi): Likewise.
3837 (neon_vtbx4v8qi): Likewise.
3838 (neon_vtrn<mode>_internal): Likewise.
3839 (neon_vzip<mode>_internal): Likewise.
3840 (neon_vuzp<mode>_internal): Likewise.
3841 (neon_vld1<mode>): Likewise.
3842 (neon_vld1_lane<mode>): Likewise.
3843 (neon_vld1_dup<mode>): Likewise.
3844 (neon_vld1_dupv2di): Likewise.
3845 (neon_vst1<mode>): Likewise.
3846 (neon_vst1_lane<mode>): Likewise.
3847 (neon_vld2<mode>): Likewise.
3848 (neon_vld2_lane<mode>): Likewise.
3849 (neon_vld2_dup<mode>): Likewise.
3850 (neon_vst2<mode>): Likewise.
3851 (neon_vst2_lane<mode>): Likewise.
3852 (neon_vld3<mode>): Likewise.
3853 (neon_vld3qa<mode>): Likewise.
3854 (neon_vld3qb<mode>): Likewise.
3855 (neon_vld3_lane<mode>): Likewise.
3856 (neon_vld3_dup<mode>): Likewise.
3857 (neon_vst3<mode>): Likewise.
3858 (neon_vst3qa<mode>): Likewise.
3859 (neon_vst3qb<mode>): Likewise.
3860 (neon_vst3_lane<mode>): Likewise.
3861 (neon_vld4<mode>): Likewise.
3862 (neon_vld4qa<mode>): Likewise.
3863 (neon_vld4qb<mode>): Likewise.
3864 (neon_vld4_lane<mode>): Likewise.
3865 (neon_vld4_dup<mode>): Likewise.
3866 (neon_vst4<mode>): Likewise.
3867 (neon_vst4qa<mode>): Likewise.
3868 (neon_vst4qb<mode>): Likewise.
3869 (neon_vst4_lane<mode>): Likewise.
3870 (neon_vec_unpack<US>_lo_<mode>): Likewise.
3871 (neon_vec_unpack<US>_hi_<mode>): Likewise.
3872 (neon_vec_<US>mult_lo_<mode>): Likewise.
3873 (neon_vec_<US>mult_hi_<mode>): Likewise.
3874 (neon_vec_<US>shiftl_<mode>): Likewise.
3875 (neon_unpack<US>_<mode>): Likewise.
3876 (neon_vec_<US>mult_<mode>): Likewise.
3877 (vec_pack_trunc_<mode>): Likewise.
3878 (neon_vec_pack_trunk_<mode>): Likewise.
3879 (neon_vabd<mode>_2): Likewise.
3880 (neon_vabd<mode>_3): Likewise.
3881 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
3882 (thumb2_movsi_vfp): Likewise.
3883 (movdi_vfp): Likewise.
3884 (movdi_vfp_cortexa8): Likewise.
3885 (movhf_vfp_neon): Likewise.
3886 (movhf_vfp): Likewiwse.
3887 (movsf_vfp): Likewiwse.
3888 (thumb2_movsf_vfp): Likewiwse.
3889 (movdf_vfp): Likewise.
3890 (thumb2_movdf_vfp): Likewise.
3891 (movsfcc_vfp): Likewise.
3892 (thumb2_movsfcc_vfp): Likewise.
3893 (movdfcc_vfp): Likewise.
3894 (thumb2_movdfcc_vfp): Likewise.
3895 * config/arm/arm.c (cortexa7_older_only): Update for attribute change.
3896 * config/arm/arm1020e.md (v10_c2v): Update for attribute change.
3897 (v10_v2c): Likewise.
3898 * config/arm/cortex-a15-neon.md (cortex_a15_neon_int_1): Update for
3900 (cortex_a15_neon_int_2): Likewise.
3901 (cortex_a15_neon_int_3): Likewise.
3902 (cortex_a15_neon_int_4): Likewise.
3903 (cortex_a15_neon_int_5): Likewise.
3904 (cortex_a15_neon_vqneg_vqabs): Likewise.
3905 (cortex_a15_neon_vmov): Likewise.
3906 (cortex_a15_neon_vaba): Likewise.
3907 (cortex_a15_neon_vaba_qqq): Likewise.
3908 (cortex_a15_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
3909 (cortex_a15_neon_mul_qqq_8_16_32_ddd_32): Likewise.
3910 (cortex_a15_neon_mul_qdd_64_32_long_qqd_16_ddd_32_\
3911 scalar_64_32_long_scalar): Likewise.
3912 (cortex_a15_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
3913 (cortex_a15_neon_mla_qqq_8_16): Likewise.
3914 (cortex_a15_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
3915 lotype_qdd_64_32_long): Likewise.
3916 (cortex_a15_neon_mla_qqq_32_qqd_32_scalar): Likewise.
3917 (cortex_a15_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
3918 (cortex_a15_neon_mul_qqd_32_scalar): Likewise.
3919 (cortex_a15_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
3920 (cortex_a15_neon_shift_1): Likewise.
3921 (cortex_a15_neon_shift_2): Likewise.
3922 (cortex_a15_neon_shift_3): Likewise.
3923 (cortex_a15_neon_vshl_ddd): Likewise.
3924 (cortex_a15_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
3925 (cortex_a15_neon_vsra_vrsra): Likewise.
3926 (cortex_a15_neon_fp_vadd_ddd_vabs_dd): Likewise.
3927 (cortex_a15_neon_fp_vadd_qqq_vabs_qq): Likewise.
3928 (cortex_a15_neon_fp_vmul_ddd): Likewise.
3929 (cortex_a15_neon_fp_vmul_qqd): Likewise.
3930 (cortex_a15_neon_fp_vmla_ddd): Likewise.
3931 (cortex_a15_neon_fp_vmla_qqq): Likewise.
3932 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
3933 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
3934 (cortex_a15_neon_fp_vrecps_vrsqrts_ddd): Likewise.
3935 (cortex_a15_neon_fp_vrecps_vrsqrts_qqq): Likewise.
3936 (cortex_a15_neon_bp_simple): Likewise.
3937 (cortex_a15_neon_bp_2cycle): Likewise.
3938 (cortex_a15_neon_bp_3cycle): Likewise.
3939 (cortex_a15_neon_vld1_1_2_regs): Likewise.
3940 (cortex_a15_neon_vld1_3_4_regs): Likewise.
3941 (cortex_a15_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
3942 (cortex_a15_neon_vld2_4_regs): Likewise.
3943 (cortex_a15_neon_vld3_vld4): Likewise.
3944 (cortex_a15_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
3945 (cortex_a15_neon_vst1_3_4_regs): Likewise.
3946 (cortex_a15_neon_vst2_4_regs_vst3_vst4): Likewise.
3947 (cortex_a15_neon_vst3_vst4): Likewise.
3948 (cortex_a15_neon_vld1_vld2_lane): Likewise.
3949 (cortex_a15_neon_vld3_vld4_lane" 10
3950 (cortex_a15_neon_vst1_vst2_lane): Likewise.
3951 (cortex_a15_neon_vst3_vst4_lane): Likewise.
3952 (cortex_a15_neon_vld3_vld4_all_lanes): Likewise.
3953 (cortex_a15_neon_ldm_2): Likewise.0
3954 (cortex_a15_neon_stm_2): Likewise.
3955 (cortex_a15_neon_mcr): Likewise.
3956 (cortex_a15_neon_mcr_2_mcrr): Likewise.
3957 (cortex_a15_neon_mrc): Likewise.
3958 (cortex_a15_neon_mrrc): Likewise.
3959 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
3961 (cortex_a15_alu_shift): Likewise.
3962 (cortex_a15_alu_shift_reg): Likewise.
3963 (cortex_a15_mult32): Likewise.
3964 (cortex_a15_mult64): Likewise.
3965 (cortex_a15_block): Likewise.
3966 (cortex_a15_branch): Likewise.
3967 (cortex_a15_load1): Likewise.
3968 (cortex_a15_load3): Likewise.
3969 (cortex_a15_store1): Likewise.
3970 (cortex_a15_store3): Likewise.
3971 (cortex_a15_call): Likewise.
3972 * config/arm/cortex-a5.md (cortex_a5_r2f): Update for attribute change.
3973 (cortex_a5_f2r): Likewise.
3974 * config/arm/cortex-a53.md (cortex_a53_r2f): Update for attribute
3976 (cortex_a53_f2r): Likewise.
3977 * config/arm/cortex-a7.md
3978 (cortex_a7_branch): Update for attribute change.
3979 (cortex_a7_call): Likewise.
3980 (cortex_a7_alu_imm): Likewise.
3981 (cortex_a7_alu_reg): Likewise.
3982 (cortex_a7_alu_shift): Likewise.
3983 (cortex_a7_mul): Likewise.
3984 (cortex_a7_load1): Likewise.
3985 (cortex_a7_store1): Likewise.
3986 (cortex_a7_load2): Likewise.
3987 (cortex_a7_store2): Likewise.
3988 (cortex_a7_load3): Likewise.
3989 (cortex_a7_store3): Likewise.
3990 (cortex_a7_load4): Likewise.
3991 (cortex_a7_store4): Likewise.
3992 (cortex_a7_fpalu): Likewise.
3993 (cortex_a7_fconst): Likewise.
3994 (cortex_a7_fpmuls): Likewise.
3995 (cortex_a7_neon_mul): Likewise.
3996 (cortex_a7_fpmacs): Likewise.
3997 (cortex_a7_neon_mla: Likewise.
3998 (cortex_a7_fpmuld: Likewise.
3999 (cortex_a7_fpmacd: Likewise.
4000 (cortex_a7_fpfmad: Likewise.
4001 (cortex_a7_fdivs: Likewise.
4002 (cortex_a7_fdivd: Likewise.
4003 (cortex_a7_r2f: Likewise.
4004 (cortex_a7_f2r: Likewise.
4005 (cortex_a7_f_flags: Likewise.
4006 (cortex_a7_f_loads: Likewise.
4007 (cortex_a7_f_loadd: Likewise.
4008 (cortex_a7_f_stores: Likewise.
4009 (cortex_a7_f_stored: Likewise.
4010 (cortex_a7_neon): Likewise.
4011 * config/arm/cortex-a8-neon.md
4012 (cortex_a8_neon_mrc): Update for attribute change.
4013 (cortex_a8_neon_mrrc): Likewise.
4014 (cortex_a8_neon_int_1): Likewise.
4015 (cortex_a8_neon_int_2): Likewise.
4016 (cortex_a8_neon_int_3): Likewise.
4017 (cortex_a8_neon_int_4): Likewise.
4018 (cortex_a8_neon_int_5): Likewise.
4019 (cortex_a8_neon_vqneg_vqabs): Likewise.
4020 (cortex_a8_neon_vmov): Likewise.
4021 (cortex_a8_neon_vaba): Likewise.
4022 (cortex_a8_neon_vaba_qqq): Likewise.
4023 (cortex_a8_neon_vsma): Likewise.
4024 (cortex_a8_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
4025 (cortex_a8_neon_mul_qqq_8_16_32_ddd_32): Likewise.
4026 (cortex_a8_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_\
4027 long_scalar): Likewise.
4028 (cortex_a8_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
4029 (cortex_a8_neon_mla_qqq_8_16): Likewise.
4030 (cortex_a8_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
4031 long_scalar_qdd_64_32_long): Likewise.
4032 (cortex_a8_neon_mla_qqq_32_qqd_32_scalar): Likewise.
4033 (cortex_a8_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
4034 (cortex_a8_neon_mul_qqd_32_scalar): Likewise.
4035 (cortex_a8_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
4036 (cortex_a8_neon_shift_1): Likewise.
4037 (cortex_a8_neon_shift_2): Likewise.
4038 (cortex_a8_neon_shift_3): Likewise.
4039 (cortex_a8_neon_vshl_ddd): Likewise.
4040 (cortex_a8_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
4041 (cortex_a8_neon_vsra_vrsra): Likewise.
4042 (cortex_a8_neon_fp_vadd_ddd_vabs_dd): Likewise.
4043 (cortex_a8_neon_fp_vadd_qqq_vabs_qq): Likewise.
4044 (cortex_a8_neon_fp_vsum): Likewise.
4045 (cortex_a8_neon_fp_vmul_ddd): Likewise.
4046 (cortex_a8_neon_fp_vmul_qqd): Likewise.
4047 (cortex_a8_neon_fp_vmla_ddd): Likewise.
4048 (cortex_a8_neon_fp_vmla_qqq): Likewise.
4049 (cortex_a8_neon_fp_vmla_ddd_scalar): Likewise.
4050 (cortex_a8_neon_fp_vmla_qqq_scalar): Likewise.
4051 (cortex_a8_neon_fp_vrecps_vrsqrts_ddd): Likewise.
4052 (cortex_a8_neon_fp_vrecps_vrsqrts_qqq): Likewise.
4053 (cortex_a8_neon_bp_simple): Likewise.
4054 (cortex_a8_neon_bp_2cycle): Likewise.
4055 (cortex_a8_neon_bp_3cycle): Likewise.
4056 (cortex_a8_neon_ldr): Likewise.
4057 (cortex_a8_neon_str): Likewise.
4058 (cortex_a8_neon_vld1_1_2_regs): Likewise.
4059 (cortex_a8_neon_vld1_3_4_regs): Likewise.
4060 (cortex_a8_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
4061 (cortex_a8_neon_vld2_4_regs): Likewise.
4062 (cortex_a8_neon_vld3_vld4): Likewise.
4063 (cortex_a8_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
4064 (cortex_a8_neon_vst1_3_4_regs): Likewise.
4065 (cortex_a8_neon_vst2_4_regs_vst3_vst4): Likewise.
4066 (cortex_a8_neon_vst3_vst4): Likewise.
4067 (cortex_a8_neon_vld1_vld2_lane): Likewise.
4068 (cortex_a8_neon_vld3_vld4_lane): Likewise.
4069 (cortex_a8_neon_vst1_vst2_lane): Likewise.
4070 (cortex_a8_neon_vst3_vst4_lane): Likewise.
4071 (cortex_a8_neon_vld3_vld4_all_lanes): Likewise.
4072 (cortex_a8_neon_mcr): Likewise.
4073 (cortex_a8_neon_mcr_2_mcrr): Likewise.
4074 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
4075 * config/arm/cortex-a9-neon.md (ca9_neon_mrc): Update for attribute
4077 (ca9_neon_mrrc): Likewise.
4078 (cortex_a9_neon_int_1): Likewise.
4079 (cortex_a9_neon_int_2): Likewise.
4080 (cortex_a9_neon_int_3): Likewise.
4081 (cortex_a9_neon_int_4): Likewise.
4082 (cortex_a9_neon_int_5): Likewise.
4083 (cortex_a9_neon_vqneg_vqabs): Likewise.
4084 (cortex_a9_neon_vmov): Likewise.
4085 (cortex_a9_neon_vaba): Likewise.
4086 (cortex_a9_neon_vaba_qqq): Likewise.
4087 (cortex_a9_neon_vsma): Likewise.
4088 (cortex_a9_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
4089 (cortex_a9_neon_mul_qqq_8_16_32_ddd_32): Likewise.
4090 (cortex_a9_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_\
4091 long_scalar): Likewise.
4092 (cortex_a9_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
4093 (cortex_a9_neon_mla_qqq_8_16): Likewise.
4094 (cortex_a9_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
4095 long_scalar_qdd_64_32_long): Likewise.
4096 (cortex_a9_neon_mla_qqq_32_qqd_32_scalar): Likewise.
4097 (cortex_a9_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
4098 (cortex_a9_neon_mul_qqd_32_scalar): Likewise.
4099 (cortex_a9_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
4100 (cortex_a9_neon_shift_1): Likewise.
4101 (cortex_a9_neon_shift_2): Likewise.
4102 (cortex_a9_neon_shift_3): Likewise.
4103 (cortex_a9_neon_vshl_ddd): Likewise.
4104 (cortex_a9_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
4105 (cortex_a9_neon_vsra_vrsra): Likewise.
4106 (cortex_a9_neon_fp_vadd_ddd_vabs_dd): Likewise.
4107 (cortex_a9_neon_fp_vadd_qqq_vabs_qq): Likewise.
4108 (cortex_a9_neon_fp_vsum): Likewise.
4109 (cortex_a9_neon_fp_vmul_ddd): Likewise.
4110 (cortex_a9_neon_fp_vmul_qqd): Likewise.
4111 (cortex_a9_neon_fp_vmla_ddd): Likewise.
4112 (cortex_a9_neon_fp_vmla_qqq): Likewise.
4113 (cortex_a9_neon_fp_vmla_ddd_scalar): Likewise.
4114 (cortex_a9_neon_fp_vmla_qqq_scalar): Likewise.
4115 (cortex_a9_neon_fp_vrecps_vrsqrts_ddd): Likewise.
4116 (cortex_a9_neon_fp_vrecps_vrsqrts_qqq): Likewise.
4117 (cortex_a9_neon_bp_simple): Likewise.
4118 (cortex_a9_neon_bp_2cycle): Likewise.
4119 (cortex_a9_neon_bp_3cycle): Likewise.
4120 (cortex_a9_neon_ldr): Likewise.
4121 (cortex_a9_neon_str): Likewise.
4122 (cortex_a9_neon_vld1_1_2_regs): Likewise.
4123 (cortex_a9_neon_vld1_3_4_regs): Likewise.
4124 (cortex_a9_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
4125 (cortex_a9_neon_vld2_4_regs): Likewise.
4126 (cortex_a9_neon_vld3_vld4): Likewise.
4127 (cortex_a9_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
4128 (cortex_a9_neon_vst1_3_4_regs): Likewise.
4129 (cortex_a9_neon_vst2_4_regs_vst3_vst4): Likewise.
4130 (cortex_a9_neon_vst3_vst4): Likewise.
4131 (cortex_a9_neon_vld1_vld2_lane): Likewise.
4132 (cortex_a9_neon_vld3_vld4_lane): Likewise.
4133 (cortex_a9_neon_vst1_vst2_lane): Likewise.
4134 (cortex_a9_neon_vst3_vst4_lane): Likewise.
4135 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
4136 (cortex_a9_neon_mcr): Likewise.
4137 (cortex_a9_neon_mcr_2_mcrr): Likewise.
4138 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
4139 (cortex_a9_fps): Likewise.
4140 * config/arm/cortex-m4-fpu.md (cortex_m4_vmov_2): Update for attribute
4142 (cortex_m4_fmuls): Likewise.
4143 * config/arm/cortex-r4f.md (cortex_r4_mcr): Update for attribute
4145 (cortex_r4_mrc): Likewise.
4146 * config/arm/iterators.md: Update comment referring to neon_type.
4147 * config/arm/iwmmxt.md (iwmmxt_arm_movdi): Update for attribute change.
4148 (iwmmxt_movsi_insn): Likewise.
4149 * config/arm/marvell-pj4.md (pj4_vfp_to_core): Update for
4151 (pj4_core_to_vfp): Likewise.
4152 * config/arm/neon-schedgen.ml (emit_insn_reservations): Update for
4154 * config/arm/vfp11.md (vfp_fload): Update for attribute change.
4155 (vfp_fstore): Likewise.
4156 * doc/md.texi: Change references to neon_type to refer to type.
4158 2013-09-04 Dodji Seketeli <dodji@redhat.com>
4160 * tree.h (DECL_BUILT_IN): Fix typo in comment.
4162 2013-09-04 David Edelsohn <dje.gcc@gmail.com>
4164 * config/rs6000/rs6000.h (ASM_OUTPUT_DEF_FROM_DECLS): Only emit
4167 2013-09-04 Easwaran Raman <eraman@google.com>
4170 * tree-ssa-reassoc.c (get_stmt_uid_with_default): New function,
4171 (build_and_add_sum): Use it.
4172 (appears_later_in_bb): Simplify code.
4174 2013-09-04 Teresa Johnson <tejohnson@google.com>
4176 * dumpfile.c (dump_finish): Don't close stderr/stdout.
4178 2013-09-04 James Greenhalgh <james.greenhalgh@arm.com>
4180 * config/aarch64/arm_neon.h (vaddvq_<su>64): Fix return types.
4182 2013-09-04 Jan Hubicka <jh@suse.cz>
4184 * Makefile.in (ipa-devirt.o): Add dependency on diagnostic.h
4185 * ipa-devirt.c: Include diganostic.h
4186 (odr_type_d): Add types and types_set.
4187 (hash_type_name): Work for types with vtables during LTO.
4188 (odr_hasher::remove): Fix comment; destroy types_set.
4189 (add_type_duplicate): New function,
4190 (get_odr_type): Use it.
4191 (dump_type_inheritance_graph): Dump type duplicates.
4192 * ipa.c (symtab_remove_unreachable_nodes): Build type inheritance
4194 * tree.c (types_same_for_odr): Give exact answers on types with
4197 2013-09-04 Dodji Seketeli <dodji@redhat.com>
4199 * tree.h (DECL_BUILT_IN, DECL_IS_BUILTIN): Add more comments
4200 explaining their differences.
4202 2013-09-04 Sandeep Kumar Singh<Sandeep.Singh2@kpitcummins.com>
4204 * config/rx/rx.h: Add option -mcpu for target variants RX100 and RX200.
4206 2013-09-03 Jeff Law <law@redhat.com>
4208 * tree-ssa-threadedge.c (thread_across_edge): Record entire path
4209 when not threading through a joiner block. Pass joiner/no joiner
4210 state to register_jump_thread.
4211 * tree-ssa-threadupdate.c (register_jump_thread): Get joiner/no joiner
4212 state from argument rather than implying on path length.
4213 Dump the entire jump thread path into debugging dump.
4214 * tree-flow.h (register_jump_thread): Update prototype.
4216 2013-08-29 Xinliang David Li <davidxl@google.com>
4218 * tree-vect-data-refs.c (vect_compute_data_ref_alignment):
4219 Remove a trivial gcc_assert.
4221 2013-08-29 Xinliang David Li <davidxl@google.com>
4223 * tree-vect-slp.c (destroy_bb_vec_info): Data ref cleanup.
4224 * tree-vect-loop.c (destroy_bb_vec_info): Ditto.
4225 * tree-vect-data-refs.c (vect_compute_data_ref_alignment):
4226 Delay base decl alignment adjustment.
4227 * tree-vectorizer.c (vect_destroy_datarefs): New function.
4228 * tree-vectorizer.h: New data structure.
4229 (set_dr_misalignment): New function.
4230 (dr_misalignment): Ditto.
4231 * tree-vect-stmts.c (vectorizable_store): Ensure alignment.
4232 (vectorizable_load): Ditto.
4233 (ensure_base_align): New function.
4234 (vectorize_loops): Add dbg_cnt support.
4235 (execute_vect_slp): Ditto.
4236 * dbgcnt.def: New debug counter.
4237 * Makefile: New dependency.
4239 2013-09-03 Meador Inge <meadori@codesourcery.com>
4243 2013-08-30 Meador Inge <meadori@codesourcery.com>
4245 * tree-vrp.c (check_array_ref): Bail out on zero-length arrays.
4247 2013-09-03 David Edelsohn <dje.gcc@gmail.com>
4249 * config/rs6000/rs6000.h (ASM_OUTPUT_DEF_FROM_DECLS): Emit lglobl for
4250 function descriptor.
4252 2013-09-03 Richard Biener <rguenther@suse.de>
4254 * tree-affine.c (add_elt_to_tree): Fix association issue,
4255 avoid useless converts and make sure to always return a
4256 properly typed result.
4258 2013-09-03 Richard Biener <rguenther@suse.de>
4261 * fold-const.c (negate_expr_p): Fix division case.
4262 (negate_expr): Likewise.
4264 2013-09-03 Richard Biener <rguenther@suse.de>
4267 * tree-streamer-out.c: Include tm.h.
4268 * Makefile.in (tree-streamer-out.o): Depend on $(TM_H).
4270 2013-09-03 Jan Hubicka <jh@suse.cz>
4272 * tree-profile.c (tree_profiling): Cleanup CFG when done.
4274 2013-09-03 Alan Modra <amodra@gmail.com>
4276 * config.gcc (powerpc*-*-linux*): Add support for little-endian
4277 multilibs to big-endian target and vice versa.
4278 * config/rs6000/t-linux64: Use := assignment on all vars.
4279 (MULTILIB_EXTRA_OPTS): Remove fPIC.
4280 (MULTILIB_OSDIRNAMES): Specify using mapping from multilib_options.
4281 * config/rs6000/t-linux64le: New file.
4282 * config/rs6000/t-linux64bele: New file.
4283 * config/rs6000/t-linux64lebe: New file.
4285 2013-09-02 Jan Hubicka <jh@suse.cz>
4287 * ipa-inline-transform.c (inline_transform): Do not
4288 optimize_inline_calls when not optimizing.
4290 2013-09-02 Jan Hubicka <jh@suse.cz>
4292 * lto-symtab.c (lto_symtab_merge_symbols): Add comments; merge
4293 duplicated nodes for assembler names.
4294 * symtab.c (symtab_unregister_node): Do not attempt to unlink
4295 hard registers from assembler name hash.
4297 2013-09-02 Jan Hubicka <jh@suse.cz>
4299 * ipa-split.c (execute_split_functions): Split externally visible
4300 functions called once.
4302 2013-09-02 Martin Jambor <mjambor@suse.cz>
4305 * ipa-prop.c (ipa_edge_duplication_hook): Always put new rdesc to the
4306 linked list. When finding the correct duplicate, also consider also
4307 the caller in additon to its inlined_to node.
4309 2013-09-02 James Greenhalgh <james.greenhalgh@arm.com>
4311 * config/aarch64/aarch64-simd-builtins.def
4312 (dup_lane_scalar): Remove.
4313 * config/aarch64/aarch64-simd.md
4314 (aarch64_simd_dup): Add 'w->w' alternative.
4315 (aarch64_dup_lane<mode>): Allow for VALL.
4316 (aarch64_dup_lane_scalar<mode>): Remove.
4317 (aarch64_dup_lane_<vswap_width_name><mode>): New.
4318 (aarch64_get_lane_signed<mode>): Add w->w altenative.
4319 (aarch64_get_lane_unsigned<mode>): Likewise.
4320 (aarch64_get_lane<mode>): Likewise.
4321 * config/aarch64/aarch64.c (aarch64_evpc_dup): New.
4322 (aarch64_expand_vec_perm_const_1): Use aarch64_evpc_dup.
4323 * config/aarch64/iterators.md (VSWAP_WIDTH): New.
4324 (VCON): Change container of V2SF.
4325 (vswap_width_name): Likewise.
4326 * config/aarch64/arm_neon.h
4327 (__aarch64_vdup_lane_any): New.
4328 (__aarch64_vdup<q>_lane<q>_<fpsu><8,16,32,64>): Likewise.
4329 (vdup<q>_n_<psuf><8,16,32,64>): Convert to C implementation.
4330 (vdup<q>_lane<q>_<fpsu><8,16,32,64>): Likewise.
4332 2013-09-02 Eric Botcazou <ebotcazou@adacore.com>
4335 * expr.c (emit_move_complex): Do not move complex FP values as parts if
4336 the source or the destination is a single hard register.
4338 2013-09-02 Richard Biener <rguenther@suse.de>
4341 * tree-scalar-evolution.c (instantiate_scev_name): Allow
4344 2013-09-02 Richard Biener <rguenther@suse.de>
4346 * tree-affine.c (add_elt_to_tree): Avoid converting all pointer
4347 arithmetic to sizetype.
4349 2013-09-02 Bin Cheng <bin.cheng@arm.com>
4351 * tree-ssa-loop-ivopts.c (set_autoinc_for_original_candidates):
4352 Find auto-increment use both before and after candidate.
4354 2013-09-02 Marek Polacek <polacek@redhat.com>
4356 * Makefile.in (ubsan.o): Add $(TM_P_H) dependency.
4358 2013-09-01 Jan Hubicka <jh@suse.cz>
4360 * Makefile.in: Add ipa-profile.o
4361 (ipa.o, ipa-devrit.o, ipa-inline-analysis.o): Adjust dependencies.
4362 * cgraph.c (struct cgraph_propagate_frequency_data,
4363 cgraph_propagate_frequency_1, cgraph_propagate_frequency): Move to
4364 ipa-profile.c; replace cgraph_ by ipa_ prefix.
4365 * cgraph.h (cgraph_propagate_frequency): Remove.
4366 * ipa-inline-analysis.c: Include ipa-utils.h;
4367 drop duplicated cfgloop.h.
4368 (inline_update_callee_summaries): Update.
4369 * ipa-profile.c: New file.
4370 * ipa-utils.h (ipa_propagate_frequency): Declare.
4371 * ipa.c: Do not include pointer-set.h, hash-table.h, lto-streamer.h,
4372 data-streamer.h, value-prof.h.
4373 (symtab_remove_unreachable_nodes): Update profile.
4374 (struct histogram_entry, histogram, histogram_pool, histogram_hash,
4375 account_time_size, cmp_counts, dump_histogram,
4376 ipa_profile_generate_summary, ipa_profile_write_summary,
4377 ipa_profile_read_summary, ipa_profile, gate_ipa_profile,
4378 pass_data_ipa_profile, pass_ipa_profile, make_pass_ipa_profile):
4379 Move to ipa-profile.c.
4381 2013-09-01 John David Anglin <danglin@gcc.gnu.org>
4383 * config/pa/pa.md: Allow "const 0" operand 1 in "scc" insns.
4385 2013-09-01 Jan Hubicka <jh@suse.cz>
4387 * common.opt (fdevirtualize-speculatively): New function.
4388 * invoke.texi (fdevirtualize-speculatively): Document.
4389 * ipa-devirt.c: Include ipa-inline.h
4390 (likely_target_p): New function.
4391 (ipa_devirt): New function.
4392 (gate_ipa_devirt): New function.
4393 (pass_data_ipa_devirt): New static var.
4394 (pass_ipa_devirt): Likewise.
4395 (make_pass_ipa_devirt): New function.
4396 * opts.c (default_options): Add OPT_fdevirtualize_speculatively.
4397 (common_handle_option): Disable devirtualization when
4398 value range profiling is available.
4399 * passes.def (pass_ipa_devirt): Add.
4400 * timever.def (TV_IPA_DEVIRT): New timevar.
4401 * tree-pass.h (make_pass_ipa_devirt):
4403 2013-09-01 Iain Sandoe <iain@codesourcery.com>
4405 * config/darwin.h (LINK_COMMAND_SPEC_A): Revise sanitizer specs to
4406 include sanitize(undefined).
4408 2013-08-31 Diego Novillo <dnovillo@google.com>
4410 * Makefile.in (TREE_CORE_H): Define.
4412 (GTFILES): Add tree-core.h.
4413 * builtins.c (built_in_class_names): Use BUILT_IN_LAST to
4415 * tree-core.h: New file.
4416 Move all data structures, enum, typedefs, global
4417 declarations and constants from ...
4420 2013-08-31 Jan Hubicka <jh@suse.cz>
4422 * bulitins.c (expand_builtin): Do not early exit for gcov
4423 instrumented functions.
4425 2013-08-31 Marek Polacek <polacek@redhat.com>
4427 * ubsan.c: Include tm_p.h.
4429 2013-08-31 Jan Hubicka <jh@suse.cz>
4431 * gimple-streamer-in.c (input_gimple_stmt): Silence parameter unused
4434 * cgraph.c (cgraph_get_body): Update call of lto_input_function_body.
4435 * gimple-streamer-in.c (input_gimple_stmt): Move sanity check to ...
4436 * tree-cfg.c (verify_gimple_label): ... here.
4437 * ipa-utils.c: Include lto-streamer.h, ipa-inline.h
4438 (ipa_merge_profiles): New function.
4439 * lto-streamer-in.c (lto_read_body): Take node instead of fn_decl.
4440 (lto_input_function_body): Likewise.
4441 * ipa-utils.h (ipa_merge_profiles): Declare.
4442 * lto-streamer.h (lto_input_function_body): Update prototype.
4443 (emit_label_in_global_context_p): Remove.
4444 * lto-symtab.c: Include ipa-utils.h
4445 (lto_cgraph_replace_node): Use ipa_merge_profiles.
4447 2013-08-31 Jan Hubicka <jh@suse.cz>
4449 * cgraph.c (cgraph_speculative_call_info): Fix ref lookup
4451 2013-08-31 Jan Hubicka <jh@suse.cz>
4453 * basic-block.h (apply_scale): Make scale parmeter gcov_type.
4455 2013-08-31 Uros Bizjak <ubizjak@gmail.com>
4457 * config/alpha/alpha.c (alpha_emit_conditional_move): Update
4458 "cmp" RTX before signed_comparison_operator check to account
4461 2013-08-30 Jan Hubicka <jh@suse.cz>
4463 * ipa-prop.c (ipa_set_jf_known_type): Check that we add only records.
4464 (detect_type_change_1): Rename to ...
4465 (detect_type_change): ... this one; early return on non-polymorphic
4467 (detect_type_change_ssa): Add comp_type parameter; update
4468 use of detect_type_change.
4469 (compute_complex_assign_jump_func): Add param_type parameter;
4470 update use of detect_type_change_ssa.
4471 (compute_complex_ancestor_jump_func): Likewise.
4472 (ipa_get_callee_param_type): New function.
4473 (ipa_compute_jump_functions_for_edge): Compute parameter type;
4474 update calls to the jump function computation functions.
4476 2013-08-30 Teresa Johnson <tejohnson@google.com>
4477 Steven Bosscher <steven@gcc.gnu.org>
4479 * cfgrtl.c (fixup_new_cold_bb): New routine.
4480 (commit_edge_insertions): Invoke fixup_partitions.
4481 (find_partition_fixes): New routine.
4482 (fixup_partitions): Ditto.
4483 (verify_hot_cold_block_grouping): Update comments.
4484 (rtl_verify_edges): Invoke find_partition_fixes.
4485 (rtl_verify_bb_pointers): Update comments.
4486 (rtl_verify_bb_layout): Ditto.
4487 * basic-block.h (probably_never_executed_edge_p): Declare.
4488 (fixup_partitions): Ditto.
4489 * cfgcleanup.c (try_optimize_cfg): Invoke fixup_partitions.
4490 * bb-reorder.c (sanitize_hot_paths): New function.
4491 (find_rarely_executed_basic_blocks_and_crossing_edges): Invoke
4493 * predict.c (probably_never_executed_edge_p): New routine.
4494 * cfg.c (check_bb_profile): Add partition insanity warnings.
4496 2013-08-30 Meador Inge <meadori@codesourcery.com>
4498 * tree-vrp.c (check_array_ref): Bail out on zero-length arrays.
4500 2013-08-30 Marek Polacek <polacek@redhat.com>
4502 * Makefile.in (ubsan.o): Add.
4503 (c-family/c-ubsan.o): Add.
4504 (builtins.o): Add ubsan.h dependency.
4505 * ubsan.h: New file.
4506 * ubsan.c: New file.
4507 * common.opt: Add -fsanitize=undefined option.
4508 (flag_sanitize): Add variable.
4509 (fsanitize=): Add option. Add Driver.
4510 (fsanitize=thread): Remove option.
4511 (fsanitize=address): Likewise.
4512 (static-libubsan): New option.
4513 * doc/invoke.texi: Document the new flag and -static-libubsan.
4514 * sanitizer.def (DEF_SANITIZER_BUILTIN): Define.
4515 (BUILT_IN_UBSAN_HANDLE_BUILTIN_UNREACHABLE): Define.
4516 * builtin-attrs.def (ATTR_COLD): Define.
4517 (ATTR_COLD_NOTHROW_LEAF_LIST): Define.
4518 * builtins.def (BUILT_IN_UBSAN_HANDLE_DIVREM_OVERFLOW,
4519 BUILT_IN_UBSAN_HANDLE_SHIFT_OUT_OF_BOUNDS): Define.
4520 * flag-types.h (sanitize_code): New enum.
4521 * opts.c (common_handle_option): Parse command line arguments
4522 of -fsanitize=. Add -fsanitize=unreachable option.
4523 * varasm.c (get_variable_section): Adjust.
4524 (assemble_noswitch_variable): Likewise.
4525 (assemble_variable): Likewise.
4526 (output_constant_def_contents): Likewise.
4527 (categorize_decl_for_section): Likewise.
4528 (place_block_symbol): Likewise.
4529 (output_object_block): Likewise.
4530 * builtins.def: Likewise.
4531 * toplev.c (compile_file): Likewise.
4532 (process_options): Likewise.
4533 * cppbuiltin.c: Likewise.
4534 * tsan.c (tsan_pass): Likewise.
4535 (tsan_gate): Likewise.
4536 (tsan_gate_O0): Likewise.
4537 * cfgexpand.c (partition_stack_vars): Likewise.
4538 (expand_stack_vars): Likewise.
4539 (defer_stack_allocation): Likewise.
4540 (expand_used_vars): Likewise.
4541 * cfgcleanup.c (old_insns_match_p): Likewise.
4542 * asan.c (asan_finish_file): Likewise.
4543 (asan_instrument): Likewise.
4544 (gate_asan): Likewise.
4545 (initialize_sanitizer_builtins): Build BT_FN_VOID_PTR_PTR_PTR.
4546 (ATTR_COLD_NOTHROW_LEAF_LIST): Define.
4547 (asan_global_struct): Use pointer_sized_int_node instead
4548 calling build_nonstandard_integer_type.
4549 (initialize_sanitizer_builtins): Likewise.
4550 (asan_finish_file): Likewise.
4551 * gcc.c: Document %{%:function(args):X}.
4552 (static_spec_functions): Add sanitize.
4553 (handle_spec_function): Add retval_nonnull argument and if non-NULL,
4554 store funcval != NULL there.
4555 (do_spec_1): Adjust handle_spec_function caller.
4556 (handle_braces): Allow %:function(args) as condition.
4557 (sanitize_spec_function): New function.
4558 (ADD_STATIC_LIBUBSAN_LIBS): Define.
4559 (LIBUBSAN_SPEC): Likewise.
4560 (LIBUBSAN_EARLY_SPEC): Likewise.
4561 (SANITIZER_SPEC): Handle libubsan.
4562 (SANITIZER_EARLY_SPEC): Likewise.
4563 * config/darwin.h (LINK_COMMAND_SPEC_A): Use %:sanitize(address)
4564 instead of fsanitize=address.
4565 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Use %:sanitize(address)
4566 instead of fsanitize=address*.
4567 * builtins.c: Include ubsan.h.
4568 (fold_builtin_0): Instrument __builtin_unreachable.
4569 * config/rs6000/rs6000.h (FRAME_GROWS_DOWNWARD): Use flag_sanitize
4570 instead of flag_asan.
4571 * tree.h (enum tree_index): Add TI_POINTER_SIZED_TYPE.
4572 (pointer_sized_int_node): Define.
4573 * tree.c (build_common_tree_nodes): Initialize pointer_sized_int_node.
4575 2013-08-30 Mike Stump <mikestump@comcast.net>
4577 * doc/install.texi (Prerequisites): Note regression in Tcl 8.6
4580 2013-08-29 Jan Hubicka <jh@suse.cz>
4582 * cgraph.c (cgraph_function_body_availability): Handle weakref
4584 * passes.def: Remove pass_fixup_cfg.
4585 * ipa-inline.c (ipa_inline): When not optimizing, do not inline;
4586 track when we need to remove functions.
4587 (gate_ipa_inline): Execute inlining always; add comment why.
4588 (pass_data_ipa_inline): Remove TODO_remove_functions.
4589 * ipa-inline-analysis.c (inline_generate_summary): When not optimizing
4590 do not produce summaries.
4591 * symtab.c (change_decl_assembler_name): Handle renaming of weakrefs.
4592 (symtab_nonoverwritable_alias): Assert we are not called on weakref.
4593 * varpool.c (cgraph_variable_initializer_availability): Fix weakrefs,
4594 constant pool and vtable.
4596 2013-08-30 Tejas Belagod <tejas.belagod@arm.com>
4598 * config/aarch64/arm_neon.h (__AARCH64_UINT64_C, __AARCH64_INT64_C):
4599 New arm_neon.h's internal macros to specify 64-bit constants.
4600 Avoid using stdint.h's macros.
4602 2013-08-30 Joern Rennecke <joern.rennecke@embecosm.com>
4604 * recog.c (verify_changes): Verify that changes[i].old is non-zero
4605 before applying REG_P.
4607 2013-08-30 Jakub Jelinek <jakub@redhat.com>
4609 PR tree-optimization/58277
4610 * tree-ssa-strlen.c (strlen_enter_block): If do_invalidate gave up
4611 after seeing too many stmts with vdef in between dombb and current
4612 bb, invalidate everything.
4614 2013-08-30 Richard Biener <rguenther@suse.de>
4616 * fold-const.c (fold_single_bit_test): Fix overflow test.
4618 2013-08-30 Eric Botcazou <ebotcazou@adacore.com>
4620 * function.c (assign_parm_setup_reg): For a parameter passed by pointer
4621 and which can live in a register, always retrieve the value on entry.
4622 * var-tracking.c (add_stores): Treat the copy on entry for a parameter
4623 passed by invisible reference specially.
4624 (emit_notes_in_bb) <MO_VAL_USE>: Emit notes before the instruction.
4625 (vt_add_function_parameter): Correctly deal with a parameter passed by
4626 invisible reference.
4628 2013-08-30 Jan Hubicka <jh@suse.cz>
4630 * tree.c (set_call_expr_flags): Fix handling of TM_PURE.
4632 2013-08-30 Richard Biener <rguenther@suse.de>
4634 PR tree-optimization/58228
4635 * tree-vect-data-refs.c (vect_analyze_data_ref_access): Do not
4636 allow invariant loads in nested loop vectorization.
4638 2013-08-30 Richard Biener <rguenther@suse.de>
4640 PR tree-optimization/58223
4641 * tree-loop-distribution.c (has_anti_dependence): Rename to ...
4642 (has_anti_or_output_dependence): ... this and adjust to also
4643 look for output dependences.
4644 (mark_nodes_having_upstream_mem_writes): Adjust.
4645 (rdg_flag_uses): Likewise.
4647 2013-08-30 Richard Biener <rguenther@suse.de>
4649 PR tree-optimization/58010
4650 * tree-vect-loop.c (vect_create_epilog_for_reduction): Remove
4651 assert that we have a loop-closed PHI.
4653 2013-08-29 Jan Hubicka <jh@suse.cz>
4655 * lto-symtab.c (lto_cgraph_replace_node): Free decl_in_state.
4656 * cgraph.c (cgraph_release_function_body): Free decl_in_state.
4657 * lto-section-in.c (lto_free_function_in_decl_state): New function.
4658 (lto_free_function_in_decl_state_for_node): New function.
4660 2013-08-29 Xinliang David Li <davidxl@google.com>
4662 * loop-unroll.c (report_unroll_peel): Minor message change.
4663 * tree-vect-loop-manip.c (vect_do_peeling_for_alignment):
4664 Emit alignment peeling message with default -fopt-info.
4665 (vect_loop_versioning): Emit loop version info message.
4666 * tree-vectorizer.c (vectorize_loops): Minor message change.
4667 (execute_vect_slp): Ditto.
4669 2013-08-29 Eric Botcazou <ebotcazou@adacore.com>
4671 * cgraphclones.c (cgraph_create_virtual_clone): Compute the DECL_NAME
4672 of the clone from the DECL_NAME of the original function.
4674 2013-08-29 Oleg Endo <olegendo@gcc.gnu.org>
4676 * passes.c (register_pass): Add overload.
4677 * tree-pass.h (register_pass): Forward declare it. Add comment.
4679 2013-08-29 Jan Hubicka <jh@suse.cz>
4681 * lto-streamer-out.c (hash_tree): Stream DECL_FINAL_P,
4682 DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P and TYPE_FINAL_P.
4683 * lto-streamer-in.c (unpack_ts_decl_with_vis_value_fields): Stream
4684 DECL_FINAL_P, DECL_CXX_CONSTRUCTOR_P and DECL_CXX_DESTRUCTOR_P.
4685 (unpack_ts_type_common_value_fields): Stream TYPE_FINAL_P.
4686 * tree-streamer-out.c (pack_ts_decl_with_vis_value_fields):
4687 Add DECL_FINAL_P, DECL_CXX_CONSTRUCTOR_P and DECL_CXX_DESTRUCTOR_P.
4688 (pack_ts_type_common_value_fields): Add TYPE_FINAL_P.
4690 2013-08-29 Teresa Johnson <tejohnson@google.com>
4692 * dumpfile.c (dump_loc): Output column number.
4693 * dumpfile.h (OPTGROUP_OTHER): Add and enable under OPTGROUP_ALL.
4694 * doc/invoke.texi: Document optall -fopt-info flag.
4695 * profile.c (read_profile_edge_counts): Use new dump framework.
4696 (compute_branch_probabilities): Ditto.
4697 * passes.c (pass_manager::register_one_dump_file): Use OPTGROUP_OTHER
4698 when pass not in any opt group.
4699 * pass_manager.h (pass_manager::get_pass_profile): New method.
4700 * value-prof.c (check_counter): Use new dump framework.
4701 (check_ic_target): Ditto.
4702 * coverage.c (get_coverage_counts): Ditto.
4703 (coverage_init): Setup new dump framework.
4705 2013-08-29 Richard Biener <rguenther@suse.de>
4707 PR tree-optimization/58246
4708 * tree-ssa-dce.c (mark_aliased_reaching_defs_necessary_1): Properly
4709 handle the dominance check inside a basic-block.
4711 2013-08-29 Richard Biener <rguenther@suse.de>
4714 * tree-ssa-copy.c (may_propagate_copy): Allow propagating
4715 of default defs that appear in abnormal PHI nodes.
4717 2013-08-29 Richard Biener <rguenther@suse.de>
4719 PR tree-optimization/57685
4720 * tree-vrp.c (register_edge_assert_for_1): Recurse only for
4721 single-use operands to avoid exponential complexity.
4723 2013-08-28 Dehao Chen <dehao@google.com>
4725 * ipa-inline.c (edge_badness): Fix integer underflow.
4727 2013-08-28 Uros Bizjak <ubizjak@gmail.com>
4729 * gtm-builtins.def (_ITM_free): Declare leaf.
4731 2013-08-28 Jakub Jelinek <jakub@redhat.com>
4734 * config/i386/i386.md (*tls_global_dynamic_64_largepic): New insn.
4735 (*tls_local_dynamic_base_64_largepic): Likewise.
4736 (tls_global_dynamic_64_<mode>, tls_local_dynamic_base_64_<mode>):
4737 Remove predicate from call operand.
4738 * config/i386/i386.c (ix86_tls_get_addr): For -mcmodel=large -fpic
4739 return sum of pic_offset_table_rtx and UNSPEC_PLTOFF of the symbol.
4741 2013-08-28 Jeff Law <law@redhat.com>
4743 * tree-ssa-threadedge.c (thread_around_empty_block): Remove
4744 checks for the number of predecessors and successors allowed.
4745 * tree-ssa-threadupdate.c (mark_threaded_blocks): Ignore requests
4746 which require copying a joiner block if there is a request which
4747 is a subpath that requires no joiner block copying.
4749 2013-08-28 Jan Hubicka <jh@suse.cz>
4751 * lto-streamer-out.c (DFS_write_tree_body): Drop
4752 BINFO_INHERITANCE_CHAIN, BINFO_SUBVTT_INDEX and BINFO_VPTR_INDEX.
4753 (hash_tree): Do not hash DECL_DEFER_OUTPUT, BINFO_INHERITANCE_CHAIN,
4754 BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX, DECL_IN_TEXT_SECTION.
4755 * tree-streamer-in.c (unpack_ts_decl_common_value_fields):
4756 Do not read DECL_ERROR_ISSUED.
4757 (unpack_ts_decl_with_vis_value_fields): Do not read
4759 (lto_input_ts_binfo_tree_pointers): Do not read
4760 BINFO_INHERITANCE_CHAIN, BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX
4761 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Do not
4762 write DECL_ERROR_ISSUED..
4763 (pack_ts_decl_with_vis_value_fields): Do not write
4765 (write_ts_binfo_tree_pointers): Do not read BINFO_INHERITANCE_CHAIN,
4766 BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX.
4767 * print-tree.c (print_node): Do not print DECL_ERROR_ISSUED.
4768 * tree.h (tree_decl_common): Update comment.
4769 (DECL_ERROR_ISSUED): Remove.
4771 2013-08-28 Jakub Jelinek <jakub@redhat.com>
4774 * omp-low.c (copy_var_decl): Copy over TREE_NO_WARNING flag.
4776 2013-08-28 Jan Hubicka <jh@suse.cz>
4778 * builtins.def (free): Declare leaf.
4780 2013-08-27 David Malcolm <dmalcolm@redhat.com>
4783 * configure.ac (gdbinit.in): Add import of gcc/gdbhooks.py.
4784 * configure: Regenerate.
4786 2013-08-27 Martin Jambor <mjambor@suse.cz>
4788 * ipa-prop.h (ipa_pass_through_data): New field type_preserved.
4789 (ipa_ancestor_jf_data): Likewise.
4790 (ipa_get_jf_pass_through_agg_preserved): Fix comment typo.
4791 (ipa_get_jf_pass_through_type_preserved): New function.
4792 (ipa_get_jf_ancestor_agg_preserved): Fix comment typo.
4793 (ipa_get_jf_ancestor_type_preserved): New function.
4794 * ipa-cp.c (ipa_get_jf_pass_through_result): Honor type_preserved flag.
4795 (ipa_get_jf_ancestor_result): Likewise.
4796 (propagate_vals_accross_pass_through): Use
4797 ipa_get_jf_pass_through_result to do all the value mappings.
4798 * ipa-prop.c (ipa_print_node_jump_functions_for_edge): Dump the
4799 type_preserved flag.
4800 (ipa_set_jf_cst_copy): New function.
4801 (ipa_set_jf_simple_pass_through): Set the type_preserved flag.
4802 (ipa_set_jf_arith_pass_through): Likewise.
4803 (ipa_set_ancestor_jf): Likewise.
4804 (compute_complex_assign_jump_func): Set type_preserved instead of
4806 (ipa_compute_jump_functions_for_edge): Likewise.
4807 (combine_known_type_and_ancestor_jfs): Honor type_preserved.
4808 (update_jump_functions_after_inlining): Update type_preserved.
4809 Explicitely create jump functions when combining one with pass_through.
4810 (ipa_write_jump_function): Stream the type_preserved flags.
4811 (ipa_read_jump_function): Likewise.
4813 2013-08-27 Jakub Jelinek <jakub@redhat.com>
4814 Aldy Hernandez <aldyh@redhat.com>
4816 * Makefile.in (omp-low.o): Depend on $(TARGET_H).
4817 * cfgloop.h (struct loop): Add safelen, force_vect, simduid.
4818 * function.h (struct function): Add has_force_vect_loops and
4820 * gimple-pretty-print.c (dump_gimple_omp_for): Handle GF_OMP_FOR_KIND*.
4821 * gimple.c (gimple_build_omp_critical): Add KIND argument and
4823 * gimple.def: Update CLAUSES comments.
4824 * gimple.h (enum gf_mask): Add GF_OMP_FOR_KIND_{FOR,SIMD}.
4825 (gimple_build_omp_for): Add argument to prototype.
4826 (gimple_omp_for_kind): New.
4827 (gimple_omp_for_set_kind): New.
4828 * gimplify.c (enum gimplify_omp_var_data): Add GOVD_LINEAR to
4829 GOVD_DATA_SHARE_CLASS.
4830 (enum omp_region_type): Add ORT_SIMD.
4831 (gimple_add_tmp_var): Handle ORT_SIMD.
4832 (gimplify_var_or_parm_decl): Same.
4833 (is_gimple_stmt): Same.
4834 (omp_firstprivatize_variable): Same.
4835 (omp_add_variable): Only use splay_tree_insert if lookup failed.
4836 (omp_notice_variable): Handle ORT_SIMD.
4837 (omp_is_private): Add SIMD argument and handle it as well as ORT_SIMD.
4838 (omp_check_private): Handle ORT_SIMD.
4839 (gimplify_scan_omp_clauses): Handle OMP_CLAUSE_LINEAR and
4841 (gimplify_adjust_omp_clauses_1): Handle GOVD_LINEAR.
4842 Handle OMP_CLAUSE_LASTPRIVATE.
4843 (gimplify_adjust_omp_clauses): Handle OMP_CLAUSE_LINEAR and
4845 (gimplify_omp_for): Handle OMP_SIMD and OMP_CLAUSE_LINEAR.
4846 (gimplify_expr): Handle OMP_SIMD.
4847 * internal-fn.c (expand_GOMP_SIMD_LANE): New.
4848 (expand_GOMP_SIMD_VF): New.
4849 (expand_GOMP_SIMD_LAST_LANE): New.
4850 * internal-fn.def (GOMP_SIMD_LANE): New.
4851 (GOMP_SIMD_VF): New.
4852 (GOMP_SIMD_LAST_LANE): New.
4853 * omp-low.c: Include target.h.
4854 (extract_omp_for_data): Handle OMP_SIMD, OMP_CLAUSE_LINEAR,
4856 (check_omp_nesting_restrictions): Same.
4858 (lower_rec_simd_input_clauses): New.
4859 (lower_rec_input_clauses): Handle OMP_SIMD, GF_OMP_FOR_KIND_SIMD,
4861 (lower_lastprivate_clauses): Handle OMP_CLAUSE_LINEAR,
4862 GF_OMP_FOR_KIND_SIMD, OMP_SIMD.
4863 (expand_omp_build_assign): New.
4864 (expand_omp_for_init_counts): New.
4865 (expand_omp_for_init_vars): New.
4866 (extract_omp_for_update_vars): New.
4867 (expand_omp_for_generic): Use expand_omp_for_{init,update}_vars
4868 and rewrite accordingly.
4869 (expand_omp_simd): New.
4870 (expand_omp_for): Use expand_omp_simd.
4871 (lower_omp_for_lastprivate): Unshare vinit when appropriate.
4872 (lower_omp_for): Do not lower the body.
4873 * tree-data-ref (get_references_in_stmt): Allow IFN_GOMP_SIMD_LANE
4875 * tree-flow.h (find_omp_clause): Remove prototype.
4876 * tree-if-conv.c (main_tree_if_conversion): Run if doing if conversion,
4877 forcing vectorization of the loop, or if flag_tree_vectorize.
4878 (gate_tree_if_conversion): Similarly.
4879 * tree-inline.c (remap_gimple_stmt): Pass for kind argument to
4880 gimple_build_omp_for.
4881 (copy_cfg_body): set has_force_vect_loops and has_simduid_loops.
4882 * tree-parloops (create_parallel_loop): Pass kind argument to
4883 gimple_build_omp_for.
4884 * tree-pretty-print.c (dump_omp_clause): Add cases for
4885 OMP_CLAUSE_UNIFORM, OMP_CLAUSE_LINEAR, OMP_CLAUSE_SAFELEN,
4886 OMP_CLAUSE__SIMDUID_.
4887 (dump_generic_node): Handle OMP_SIMD.
4888 * tree-ssa-ccp.c (likely_value): Handle IFN_GOMP_SIMD*.
4889 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely_1): Do not
4890 unroll OMP_SIMD loops here.
4891 * tree-ssa-loop.c (gate_tree_vectorize): Run if has_force_vect_loops.
4892 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Handle
4894 (vect_analyze_data_refs): Handle simd loops.
4895 * tree-vect-loop.c (vectorizable_live_operation): Handle
4897 * tree-vect-stmts.c (vectorizable_call): Handle IFN_GOMP_SIMD_LANE.
4898 (vectorizable_store): Handle STMT_VINFO_SIMD_LANE_ACCESS_P.
4899 (vectorizable_load): Same.
4900 * tree-vectorizer.c: Include hash-table.h and tree-ssa-propagate.h.
4901 (struct simduid_to_vf): New.
4902 (simduid_to_vf::hash): New.
4903 (simduid_to-vf::equal): New.
4904 (struct simd_array_to_simduid): New.
4905 (simd_array_to_simduid::hash): New.
4906 (simd_array_to_simduid::equal): New.
4907 (adjust_simduid_builtins): New.
4908 (struct note_simd_array_uses_struct): New.
4909 (note_simd_array_uses_cb): New.
4910 (note_simd_array_uses): New.
4911 (vectorize_loops): Handle simd hints and adjust simd builtins
4913 * tree-vectorizer.h (struct _stmt_vec_info): Add
4914 simd_lane_access_p field.
4915 (STMT_VINFO_SIMD_LANE_ACCESS_P): New macro.
4916 * tree.c (omp_clause_num_ops): Add entries for OMP_CLAUSE_LINEAR,
4917 OMP_CLAUSE_SAFELEN, OMP_CLAUSE__SIMDUID_, OMP_CLAUSE_UNIFORM.
4918 (omp_clause_code_name): Same.
4919 (walk_tree_1): Handle OMP_CLAUSE_UNIFORM, OMP_CLAUSE_SAFELEN,
4920 OMP_CLAUSE__SIMDUID_, OMP_CLAUSE_LINEAR.
4921 * tree.def (OMP_SIMD): New entry.
4922 * tree.h (enum omp_clause_code): Add entries for OMP_CLAUSE_LINEAR,
4923 OMP_CLAUSE_UNIFORM, OMP_CLAUSE_SAFELEN, OMP_CLAUSE__SIMDUID_.
4924 (OMP_CLAUSE_DECL): Adjust range for new clauses.
4925 (OMP_CLAUSE_LINEAR_NO_COPYIN): New.
4926 (OMP_CLAUSE_LINEAR_NO_COPYOUT): New.
4927 (OMP_CLAUSE_LINEAR_STEP): New.
4928 (OMP_CLAUSE_SAFELEN_EXPR): New.
4929 (OMP_CLAUSE__SIMDUID__DECL): New.
4930 (find_omp_clause): New prototype.
4932 2013-08-27 H.J. Lu <hongjiu.lu@intel.com>
4934 * config/i386/driver-i386.c (host_detect_local_cpu): Update
4935 Haswell processor detection.
4937 2013-08-27 Christian Widmer <shadow@umbrox.de>
4940 * config/i386/driver-i386.c (host_detect_local_cpu): Add detection
4941 of Ivy Bridge and Haswell processors. Assume core-avx2 for unknown
4942 AVX2 capable processors.
4944 2013-08-27 Tejas Belagod <tejas.belagod@arm.com>
4946 * config/aarch64/arm_neon.h: Replace all inline asm implementations
4947 of vget_low_* with implementations in terms of other intrinsics.
4949 2013-08-27 Marc Glisse <marc.glisse@inria.fr>
4952 * doc/extend.texi (__builtin_isinf_sign): Restrict the return
4953 values to -1, 0 and 1.
4955 2013-08-27 Vidya Praveen <vidyapraveen@arm.com>
4957 * config/aarch64/aarch64.md (unspec): Add UNSPEC_SISD_SSHL,
4958 UNSPEC_SISD_USHL, UNSPEC_USHL_2S, UNSPEC_SSHL_2S, UNSPEC_SISD_NEG.
4959 (<optab><mode>3_insn): Remove.
4960 (aarch64_ashl_sisd_or_int_<mode>3): New Pattern.
4961 (aarch64_lshr_sisd_or_int_<mode>3): Likewise.
4962 (aarch64_ashr_sisd_or_int_<mode>3): Likewise.
4963 (define_split for aarch64_lshr_sisd_or_int_di3): Likewise.
4964 (define_split for aarch64_lshr_sisd_or_int_si3): Likewise.
4965 (define_split for aarch64_ashr_sisd_or_int_di3): Likewise.
4966 (define_split for aarch64_ashr_sisd_or_int_si3): Likewise.
4967 (aarch64_sisd_ushl, aarch64_sisd_sshl): Likewise.
4968 (aarch64_ushl_2s, aarch64_sshl_2s, aarch64_sisd_neg_qi): Likewise.
4969 (ror<mode>3_insn): Likewise.
4970 * config/aarch64/predicates.md (aarch64_simd_register): New.
4972 2013-08-27 Richard Biener <rguenther@suse.de>
4974 PR tree-optimization/57521
4975 * tree-if-conv.c (if_convertible_bb_p): Verify that at least
4976 one edge is non-critical.
4977 (find_phi_replacement_condition): Make sure to use a non-critical
4978 edge. Cleanup and remove old bug workarounds.
4979 (bb_postdominates_preds): Remove.
4980 (if_convertible_loop_p_1): Do not compute post-dominators.
4981 (combine_blocks): Do not free post-dominators.
4982 (main_tree_if_conversion): Likewise.
4983 (pass_data_if_conversion): Add TODO_verify_ssa.
4985 2013-08-27 DJ Delorie <dj@redhat.com>
4987 * config/i386/djgpp.h (ASM_DECLARE_FUNCTION_NAME): New.
4989 2013-08-27 Yufeng Zhang <yufeng.zhang@arm.com>
4991 * function.c (assign_parm_find_data_types): Set passed_mode and
4992 nominal_mode to the TYPE_MODE of nominal_type for the built
4993 pointer type in case of the struct-pass-by-reference.
4995 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
4997 * config/avr/avr-stdint.h (INT16_TYPE): Change default to "int".
4998 (UINT16_TYPE): Change default to "unsigned int".
5000 * config/avr/avr.opt (mfract-convert-truncate): New option.
5001 * config/avr/avr.c (avr_out_fract): Unless TARGET_FRACT_CONV_TRUNC
5002 is set, round negative fractional integers according to n1169
5003 when converting to integer types.
5005 2013-08-26 Jan Hubicka <jh@suse.cz>
5007 * cgraph.c (cgraph_propagate_frequency): Do not assume that virtual
5008 methods can not be called indirectly when their address is not taken.
5010 2013-08-26 Jan Hubicka <jh@suse.cz>
5012 * gimple-fold.c (gimple_get_virt_method_for_binfo): Use
5015 2013-08-26 Jan Hubicka <jh@suse.cz>
5017 * ipa.c (comdat_can_be_unshared_p_1): C++ constructors and destructors
5020 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
5022 * reload.c (find_valid_class): Allow classes that do not include
5023 FIRST_PSEUDO_REGISTER - 1.
5025 2013-08-26 Jan Hubicka <jh@suse.cz>
5027 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Fix formatting;
5028 fix edge count/frequency when speculation failed; fix type check
5029 for the direct call.
5031 2013-08-26 Jan Hubicka <jh@suse.cz>
5033 * ipa-prop.c (ipa_print_node_params): Do not ICE during WPA.
5035 2013-08-26 Jan Hubicka <jh@suse.cz>
5037 * ipa-inline-transform.c (inline_transform): Be ready for basic block
5038 to be changed by edge redirection.
5040 2013-08-26 Jan Hubicka <jh@suse.cz>
5042 * cgraph.c (cgraph_speculative_call_info): Fix parameter order and
5043 formating; add sanity check.
5044 (cgraph_resolve_speculation): Add FIXME about scaling profiles.
5045 (cgraph_redirect_edge_call_stmt_to_callee): Fix ICE in debug dump.
5046 * ipa-inline.c (heap_edge_removal_hook): Reset node growth cache.
5047 (resolve_noninline_speculation): Update callee keys, too.
5049 2013-08-26 Jan Hubicka <jh@suse.cz>
5051 * tree.h (tree_decl_with_vis): Add cxx_constructor, cxx_destructor.
5052 (DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P): New macros.
5054 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
5056 * config/i386/i386.c (x86_64_elf_select_section): Put ATTRIBUTE_UNUSED
5059 2013-08-26 Uros Bizjak <ubizjak@gmail.com>
5061 * config/i386/i386.c (ix86_debug_options): Remove prototype.
5062 (x86_64_elf_select_section): Ditto.
5063 (ix86_handle_tm_regparm_attribute): Remove ATTRIBUTE_UNUSED on used
5065 (ix86_pass_by_reference): Ditto.
5066 (output_set_got): Ditto.
5067 (ix86_unary_operator_ok): Ditto.
5068 (ix86_expand_builtin): Ditto.
5070 2013-08-23 Jan Hubicka <jh@suse.cz>
5072 * cgraph.c (cgraph_turn_edge_to_speculative): Fix debug output.
5074 2013-08-23 Jan Hubicka <jh@suse.cz>
5076 * tree.h (TYPE_FINAL_P, DECL_FINAL_P): New macros.
5077 (tree_decl_with_vis): Add FINAL field.
5079 2013-08-23 Jeff Law <law@redhat.com>
5081 * tree-ssa-pre.c (do_regular_insertion): Include the expression in
5082 the debugging dump when the expression is fully redundant.
5084 2013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
5086 * diagnostic.c (diagnostic_set_caret_max_width): Use pp_buffer.
5087 * gimple-pretty-print.c (gimple_dump_bb_buff): Likewise.
5088 * pretty-print.c (pp_formatted_text_data): Likewise.
5089 (pp_write_text_to_stream): Likewise.
5090 (pp_write_text_as_dot_label_to_stream): Likewise.
5091 (pp_append_r): Likewise.
5092 (pp_format): Likewise.
5093 (pp_flush): Likewise.
5094 (pp_clear_output_area): Likewise.
5095 (pp_append_text): Likewise.
5096 (pp_formatted_text): Likewise.
5097 (pp_remaining_character_count_for_line): Likewise.
5098 (pp_newline): Likewise.
5099 (pp_character): Likewise.
5100 (output_buffer::~output_buffer): Define.
5101 (pretty_printer::~pretty_printer): Destruct output buffer.
5102 * pretty-print.h (output_buffer::~output_buffer): Declare.
5103 (pretty_printer::~pretty_printer): Declare virtual.
5105 2013-08-24 Marc Glisse <marc.glisse@inria.fr>
5108 * hwint.h (HOST_WIDE_INT_UC, HOST_WIDE_INT_1U, HOST_WIDE_INT_M1,
5109 HOST_WIDE_INT_M1U): New macros.
5110 * fold-const.c (sign_bit_p, build_range_check, fold_unary_loc,
5111 fold_binary_loc, fold_ternary_loc): Use the new macros. Use an
5112 unsigned -1 for lshift.
5113 * cse.c (cse_insn): Likewise.
5114 * double-int.c (rshift_double, lshift_double): Likewise.
5115 * builtins.c (fold_builtin_bitop): Likewise.
5116 * combine.c (force_to_mode): Likewise.
5117 * tree.c (integer_pow2p, tree_log2, tree_floor_log2): Likewise.
5118 * simplify-rtx.c (simplify_const_unary_operation,
5119 simplify_const_binary_operation): Likewise.
5120 * tree-stdarg.c (va_list_counter_bump, va_list_ptr_read,
5121 check_va_list_escapes): Likewise.
5122 * rtlanal.c (nonzero_bits1): Likewise.
5123 * expmed.c (expand_smod_pow2): Likewise.
5124 * tree-ssa-structalias.c (UNKNOWN_OFFSET): Use HOST_WIDE_INT_MIN.
5126 2013-08-23 Jan Hubicka <jh@suse.cz>
5128 * cgraph.c (cgraph_turn_edge_to_speculative): Mark target node
5129 as having address taken.
5131 2013-08-23 Jan Hubicka <jh@suse.cz>
5133 * ipa-utils.h (method_class_type): Declare.
5134 * ipa-devirt.c (method_class_type): Export.
5136 * cgraphunit.c (analyze_functions): Do basic devirtualization;
5137 do not walk base classes of anonymous types.
5139 2013-08-23 Kaz Kojima <kkojima@gcc.gnu.org>
5141 PR rtl-optimization/58220
5143 * final.c (reemit_insn_block_notes): Use NEXT_INSN to
5144 handle SEQUENCE insns properly.
5146 2013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
5148 * pretty-print.h (pp_newline_and_flush): Declare. Remove macro
5150 (pp_newline_and_indent): Likewise.
5151 (pp_separate_with): Likewise.
5152 * pretty-print.c (pp_newline_and_flush): Define.
5153 (pp_newline_and_indent): Likewise.
5154 (pp_separate_with): Likewise.
5156 2013-08-23 Jakub Jelinek <jakub@redhat.com>
5159 * config/i386/x86-64.h (TARGET_SECTION_TYPE_FLAGS): Define.
5160 * config/i386/i386.c (x86_64_elf_section_type_flags): New function.
5162 2013-08-23 Kirill Yukhin <kirill.yukhin@intel.com>
5164 * config/i386/predicates.md (ext_sse_reg_operand): New.
5165 * config/i386/i386.md (*movti_internal): Use
5166 predicate to determine if EVEX is needed.
5167 (*movsi_internal): Ditto.
5168 (*movdf_internal): Ditto.
5169 (*movsf_internal): Ditto.
5170 * config/i386/mmx.md (*mov<mode>_internal): Ditto.
5172 2013-08-23 Jakub Jelinek <jakub@redhat.com>
5174 PR tree-optimization/58209
5175 * tree-tailcall.c (process_assignment): Handle POINTER_PLUS_EXPR.
5176 (find_tail_calls): Give up for pointer result types if m is non-NULL.
5177 (adjust_return_value_with_ops): For PLUS_EXPR and pointer result type
5178 emit POINTER_PLUS_EXPR.
5179 (create_tailcall_accumulator): For pointer result type accumulate in
5182 2013-08-22 Paolo Carlini <paolo.carlini@oracle.com>
5184 * configure.ac: Add backslashes missing from the last change.
5185 * configure: Regenerate.
5187 2013-08-22 Jan Hubicka <jh@susue.cz>
5189 * ipa.c (function_and_variable_visibility): First remember function
5190 was global and then make it local.
5192 2013-08-22 Julian Brown <julian@codesourcery.com>
5194 * configure.ac: Add aarch64 to list of arches which use "nop" in
5196 * configure: Regenerate.
5198 2013-08-22 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5200 * config/s390/linux.h (TARGET_LIBC_HAS_FUNCTION): Define as
5201 gnu_libc_has_function.
5202 * config/s390/tpf.h: Likewise.
5204 2013-08-22 Jan Hubicka <jh@susue.cz>
5206 * timevar.c (validate_phases): Add cast.
5208 2013-08-22 Jan Hubicka <jh@susue.cz>
5210 * timevar.c (validate_phases): Use size_t for memory.
5211 * timevar.h (struct timevar_time_def): Use size_t for ggc_mem.
5213 2013-08-22 Gabriel Dos Reis <gdr@integrable-solutions.net>
5215 * pretty-print.h (output_buffer::output_buffer): Declare.
5216 (pretty_printer::pretty_printer): Likewise.
5217 (pp_construct): Remove.
5218 * pretty-print.c (output_buffer::output_buffer): Define.
5219 (pretty_printer::pretty_printer): Rename from pp_construct. Simplify.
5220 * gimple-pretty-print.c (print_gimple_stmt): Do not call pp_construct.
5221 (print_gimple_expr): Likewise.
5222 (print_gimple_seq): Likewise.
5223 (gimple_dump_bb): Likewise.
5224 * sched-vis.c (dump_value_slim): Likewise.
5225 (dump_insn_slim): Likewise.
5226 (dump_rtl_slim): Likewise.
5227 (str_pattern_slim): Likewise.
5228 * tree-mudflap.c (mf_varname_tree): Likewise.
5229 * graph.c (print_graph_cfg): Likewise.
5230 (start_graph_dump): Likewise.
5231 * tree-pretty-print.c (maybe_init_pretty_print): Likewise. Use
5233 * diagnostic.c (diagnostic_initialize): Simplify early diagnostic
5234 pretty printer initialization.
5235 * coretypes.h (diagnostic_context): Remove superflous type alias
5237 (pretty_printer): Likewise. Declare directly as a class.
5238 (pretty_print_info): Remove declaration as class.
5239 * asan.c (asan_emit_stack_protection): Remove call to pp_construct
5240 and pp_clear_output_area.
5241 (asan_add_global): Likewise.
5243 2013-08-22 Jan Hubicka <jh@suse.cz>
5245 * cgraphunit.c (analyze_functions) Use update_type_inheritance_graph.
5246 * ipa-utils.h (update_type_inheritance_graph): Declare.
5247 (possible_polymorphic_call_target_p): Declare.
5248 (possible_polymorphic_call_target_p): New.
5249 * ipa-devirt.c: Update toplevel comments.
5250 (cached_polymorphic_call_targets): Move up.
5251 (odr_type_d): Move ID down.
5252 (polymorphic_type_binfo_p): Update comment.
5253 (odr_hasher::remove): Likewise;
5254 (get_odr_type): Set anonymous_namespace.
5255 (dump_odr_type): Dump it.
5256 (dump_type_inheritance_graph): Do not ICE when there are no ODR types.
5257 (maybe_record_node): Record node in cached_polymorphic_call_targets.
5258 (record_binfo): Add comment.
5259 (free_polymorphic_call_targets_hash): Do not ICE when cache is not
5261 (devirt_node_removal_hook): Do not iCE when cache is freed.
5262 (possible_polymorphic_call_target_p): New predicate.
5263 (update_type_inheritance_graph): New function.
5265 2013-08-22 Alexander Ivchenko <alexander.ivchenko@intel.com>
5266 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
5267 Sergey Lega <sergey.s.lega@intel.com>
5268 Anna Tikhonova <anna.tikhonova@intel.com>
5269 Ilya Tocar <ilya.tocar@intel.com>
5270 Andrey Turetskiy <andrey.turetskiy@intel.com>
5271 Ilya Verbin <ilya.verbin@intel.com>
5272 Kirill Yukhin <kirill.yukhin@intel.com>
5273 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
5275 * common/config/i386/i386-common.c (OPTION_MASK_ISA_AVX512F_SET): New.
5276 (OPTION_MASK_ISA_AVX512CD_SET): Ditto.
5277 (OPTION_MASK_ISA_AVX512PF_SET): Ditto.
5278 (OPTION_MASK_ISA_AVX512ER_SET): Ditto.
5279 (OPTION_MASK_ISA_AVX2_UNSET): Update.
5280 (OPTION_MASK_ISA_AVX512F_UNSET): New.
5281 (OPTION_MASK_ISA_AVX512CD_UNSET): Ditto.
5282 (OPTION_MASK_ISA_AVX512PF_UNSET): Ditto.
5283 (OPTION_MASK_ISA_AVX512ER_UNSET): Ditto.
5284 (ix86_handle_option): Handle OPT_mavx512f, OPT_mavx512cd,
5285 OPT_mavx512pf, OPT_mavx512er cases.
5286 * config/i386/constraints.md (v): New constraint.
5287 (Yi, Yj): Replace SSE_REGS with ALL_SSE_REGS.
5288 * config/i386/cpuid.h (bit_AVX512F, bit_AVX512PF, bit_AVX512ER)
5289 (bit_AVX512CD): New.
5290 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
5291 AVX512F, AVX512ER, AVX512PF, AVX512CD features.
5292 * config/i386/i386-c.c (ix86_target_macros_internal):
5293 Conditionally define __AVX512F__, __AVX512ER__, __AVX512CD__,
5295 * config/i386/i386-modes.def (VECTOR_MODES (INT, 128))
5296 (VECTOR_MODES (FLOAT, 128), INT_MODE (XI, 64)): New modes.
5297 * config/i386/i386.c (regclass_map, dbx_register_map)
5298 (dbx64_register_map, svr4_dbx_register_map): Add new SSE registers.
5299 (gate_insert_vzeroupper): Disable vzeroupper for TARGET_AVX512F.
5300 (ix86_target_string): Define -mavx512f, -mavx512er, -mavx512cd,
5302 (ix86_option_override_internal): Define PTA_AVX512F, PTA_AVX512ER,
5303 PTA_AVX512PF, PTA_AVX512CD. Handle -mavx512f, -mavx512er, -mavx512cd,
5304 -mavx512pf options. Fix formatting.
5305 (ix86_conditional_register_usage): Squash EXT_REX_SSE_REGs for 32-bit
5306 targets. Squash EVEX_SSE_REGS if AVX512F is disabled.
5307 (ix86_valid_target_attribute_inner_p): Handle -mavx512f, -mavx512er,
5308 -mavx512cd, -mavx512pf options.
5309 (standard_sse_constant_opcode): Add vpternlogd for 512-bit modes.
5310 (print_reg, ix86_print_operand): Handle 'g' to output 512-bit operands.
5311 (ix86_preferred_output_reload_class): Replace SSE_REGS with
5313 (ix86_hard_regno_mode_ok): Support 512-bit registers.
5314 (ix86_set_reg_reg_cost): Ditto.
5315 (x86_order_regs_for_local_alloc): Ditto.
5316 (MAX_VECT_LEN): Extend to 64-byte.
5317 (ix86_spill_class): Replace SSE_REGS with ALL_SSE_REGS.
5318 * config/i386/i386.h (TARGET_AVX512F, TARGET_AVX512PF)
5319 (TARGET_AVX512ER, TARGET_AVX512CD): New.
5320 (BIGGEST_ALIGNMENT): Extend to 512-bits.
5321 (FIRST_PSEUDO_REGISTER, FIXED_REGISTERS): Add new registers.
5322 (CALL_USED_REGISTERS, REG_ALLOC_ORDER): Likewise.
5323 (VALID_AVX512F_SCALAR_MODE, VALID_AVX512F_REG_MODE): New.
5324 (SSE_REG_MODE_P): Support new modes.
5325 (FIRST_MMX_REG, FIRST_REX_INT_REG, FIRST_REX_SSE_REG): Add comments.
5326 (FIRST_EXT_REX_SSE_REG, LAST_EXT_REX_SSE_REG): New.
5327 (reg_class, REG_CLASS_NAMES): Add EVEX_SSE_REGS, ALL_SSE_REGS.
5328 (SSE_CLASS_P, MAYBE_SSE_CLASS_P): Replace SSE_REGS with ALL_SSE_REGS.
5329 (REG_CLASS_CONTENTS): Add new registers.
5330 (SSE_REGNO_P, SSE_REGNO, HARD_REGNO_RENAME_OK): Support new registers.
5331 (EXT_REX_SSE_REGNO_P): New.
5332 (HI_REGISTER_NAMES): Add new registers.
5333 * config/i386/i386.md: Define constants for new registers.
5334 (mode): Add new 512-bit modes.
5335 (prefix): Support evex prefix.
5336 (isa): Support avx512f, noavx512f, fma_avx512f.
5337 (ssemodesuffix): Add new 512-bit modes.
5339 (*movxi_internal_avx512f): Ditto.
5340 (*movdi_internal): Replace constraint "x" with the new constraint "v".
5342 (*movsi_internal): Likewise.
5343 (*movdf_internal): Likewise.
5344 (*movsf_internal): Likewise.
5345 (*fop_<mode>_comm_sse): Replace constraint "x" with new constraint "v".
5346 (<code><mode>3): Likewise.
5347 * config/i386/i386.opt (mavx512f, mavx512pf, mavx512er, mavx512cd):
5349 * config/i386/mmx.md (*mov<mode>_internal): Replace constraint "x"
5350 with the new constraint "v".
5351 * config/i386/sse.md (*mov<mode>_internal): Support new registers and
5353 (<sse>_loadu<ssemodesuffix><avxsizesuffix>): Replace constraint "x"
5354 with the new constraint "v".
5355 (<sse2>_loaddqu<avxsizesuffix>): Likewise.
5356 (<sse2>_storedqu<avxsizesuffix>): Likewise.
5357 (*<plusminus_insn><mode>3): Likewise.
5358 (<sse>_vm<plusminus_insn><mode>3): Likewise.
5359 (*mul<mode>3): Likewise.
5360 (<sse>_vmmul<mode>3): Likewise.
5361 (<sse>_div<mode>3): Likewise.
5362 (<sse>_vmdiv<mode>3): Likewise.
5363 (<sse>_sqrt<mode>2): Likewise.
5364 (<sse>_vmsqrt<mode>2): Likewise.
5365 (*<code><mode>3_finite): Likewise.
5366 (*<code><mode>3) <smaxmin>: Likewise.
5367 (<sse>_vm<code><mode>3): Likewise.
5368 (*<code><mode>3) <any_logic>: Likewise.
5369 (*fma_fmadd_<mode>): Likewise.
5370 (*fma_fmsub_<mode>): Likewise.
5371 (*fma_fnmadd_<mode>): Likewise.
5372 (*fma_fnmsub_<mode>): Likewise.
5373 (*fma_fmaddsub_<mode>): Likewise.
5374 (*fma_fmsubadd_<mode>): Likewise.
5375 (*fmai_fmadd_<mode>): Likewise.
5376 (*fmai_fmsub_<mode>): Likewise.
5377 (*fmai_fnmadd_<mode>): Likewise.
5378 (*fmai_fnmsub_<mode>): Likewise.
5379 (sse_cvtsi2ss): Likewise.
5380 (sse_cvtsi2ssq): Likewise.
5381 (sse_cvtss2si): Likewise.
5382 (sse_cvtss2si_2): Likewise.
5383 (sse_cvtss2siq): Likewise.
5384 (sse_cvtss2siq_2): Likewise.
5385 (sse_cvttss2si): Likewise.
5386 (sse_cvtss2siq_2): Likewise.
5387 (float<sseintvecmodelower><mode>2): Likewise.
5388 (sse2_cvtsd2si_2): Likewise.
5389 (sse2_cvtsd2siq_2): Likewise.
5390 (*<plusminus_insn><mode>3): Likewise.
5391 (*<sse2_avx2>_<plusminus_insn><mode>3): Likewise.
5392 (*<sse4_1_avx2>_mul<mode>3): Likewise.
5393 (ashr<mode>3): Likewise.
5394 (<shift_insn><mode>3): Likewise.
5395 (avx2_<code><mode>3): Likewise.
5396 (*avx2_<code><mode>3): Likewise.
5397 (*andnot<mode>3): Likewise.
5398 (*<code><mode>3) <any_logic>: Likewise.
5399 (abs<mode>2): Likewise.
5400 (avx2_permvar<mode>): Likewise.
5401 (avx2_perm<mode>_1): Likewise.
5402 (*avx_vpermilp<mode>): Likewise.
5403 (avx_vpermilvar<mode>3): Likewise.
5404 (avx2_ashrv<mode>): Likewise.
5405 (avx2_<shift_insn>v<mode>): Likewise.
5406 * doc/invoke.texi: Document -mavx512f, -mavx512pf, -mavx512er,
5408 * doc/rtl.texi: Document XImode.
5410 2013-08-21 Jeff Law <law@redhat.com>
5412 * tree-flow.h (register_jump_thread): Pass vector of edges
5413 instead of each important edge.
5414 * tree-ssa-threadedge.c (thread_across_edge): Build the jump
5415 thread path into a vector and pass that to register_jump_thread.
5416 * tree-ssa-threadupdate.c (register_jump_thread): Conver the
5417 passed in edge vector to the current 3-edge form.
5420 2013-08-20 Alexey Makhalov <makhaloff@gmail.com>
5422 * dce.c (fini_dce): Call df_analyze again just in case
5423 delete_unmarked_insns removed anything.
5425 2013-08-21 Joern Rennecke <joern.rennecke@embecosm.com>
5427 * reload.h (struct reg_equivs): Rename to ..
5428 (struct reg_equivs_s): .. this.
5430 2013-08-20 Martin Liska <marxin.liska@gmail.com>
5432 * ipa.c (ipa_profile_read_summary): Fix buffer overflow.
5434 2013-08-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
5436 * config/sol2-10.h (TARGET_LIBC_HAS_FUNCTION): Don't nest comment.
5438 2013-08-21 Jeff Law <law@redhat.com>
5440 * tree-vrp.c (simplify_stmt_for_jump_threading): Try to
5441 simplify assignments too. If the RHS collapses to a singleton
5442 range, then return the value for the range.
5444 2013-08-21 Kirill Yukhin <kirill.yukhin@intel.com>
5446 * config/i386/sse.md (V16): Rename to...
5448 (mov<mode>): Update iterator name.
5449 (*mov<mode>_internal): Ditto.
5450 (push<mode>1): Ditto.
5451 (movmisalign<mode>): Ditto.
5453 2013-08-20 Jan Hubicka <jh@suse.cz>
5456 * cgraph.c (cgraph_add_edge_to_call_site_hash): Overwrite hash
5457 entry for direct edges.
5458 (cgraph_turn_edge_to_speculative): Fix setting of can_throw_external.
5460 2013-08-20 David Malcolm <dmalcolm@redhat.com>
5462 Revert my last two changes, r201865 and r201864:
5465 2013-08-20 David Malcolm <dmalcolm@redhat.com>
5467 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
5468 instances can own GC refs.
5470 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
5471 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
5472 (gcc::context::gt_pch_nx): Likewise.
5473 (gcc::context::gt_pch_nx): Likewise.
5474 * ggc.h (gt_ggc_mx <T>): New.
5475 (gt_pch_nx_with_op <T>): New.
5476 (gt_pch_nx <T>): New.
5477 * passes.c (opt_pass::gt_ggc_mx): New.
5478 (opt_pass::gt_pch_nx): New.
5479 (opt_pass::gt_pch_nx_with_op): New.
5480 (pass_manager::gt_ggc_mx): New.
5481 (pass_manager::gt_pch_nx): New.
5482 (pass_manager::gt_pch_nx_with_op): New.
5483 (pass_manager::operator new): Use
5484 ggc_internal_cleared_alloc_stat rather than xcalloc.
5485 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
5486 (pass_manager::gt_ggc_mx): New.
5487 (pass_manager::gt_pch_nx): New.
5488 (pass_manager::gt_pch_nx_with_op): New.
5489 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
5490 (opt_pass::operator new): New.
5491 (opt_pass::gt_ggc_mx): New.
5492 (opt_pass::gt_pch_nx): New.
5493 (opt_pass::gt_pch_nx_with_op): New.
5496 2013-08-20 David Malcolm <dmalcolm@redhat.com>
5498 * Makefile.in (GTFILES): Add context.h.
5499 * context.c (gcc::context::operator new): New.
5500 (gcc::context::gt_ggc_mx): New.
5501 (gcc::context::gt_pch_nx): New.
5502 (gcc::context::gt_pch_nx): New.
5503 * context.h (gcc::context): Add GTY((user)) marking.
5504 (gcc::context::operator new): New.
5505 (gcc::context::gt_ggc_mx): New.
5506 (gcc::context::gt_pch_nx): New.
5507 (gcc::context::gt_pch_nx): New.
5508 (g): Add GTY marking.
5509 (gt_ggc_mx (gcc::context *)): New.
5510 (gt_pch_nx (gcc::context *)): New.
5511 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
5512 void *cookie)): New.
5513 * gengtype.c (open_base_files) <ifiles>: Add context.h.
5515 2013-08-20 Alexey Makhalov <makhaloff@gmail.com>
5517 * dce.c (fini_dce): Call df_analyze again just in case
5518 delete_unmarked_insns removed anything.
5520 2013-08-20 Teresa Johnson <tejohnson@google.com>
5522 PR rtl-optimizations/57451
5523 * final.c (reemit_insn_block_notes): Prevent lexical blocks
5524 from crossing split section boundaries.
5526 2013-08-20 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
5528 * config/arm/linux-elf.h (MULTILIB_DEFAULTS): Remove definition.
5529 * config/arm/t-linux-eabi (MULTILIB_OPTIONS): Document association
5530 with MULTLIB_DEFAULTS.
5532 2013-08-20 Nick Clifton <nickc@redhat.com>
5534 * target.def (narrow_volatile_bitfield): Note that the default
5535 value is false, not !TARGET_STRICT_ALIGN.
5536 * doc/tm.texi: Regenerate.
5538 2013-08-20 Pavel Chupin <pavel.v.chupin@intel.com>
5540 Fix LIB_SPEC for systems without libpthread.
5542 * config/gnu-user.h: Introduce GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC.
5543 * config/arm/linux-eabi.h: Use GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC
5545 * config/i386/linux-common.h: Likewise.
5546 * config/mips/linux-common.h: Likewise.
5548 2013-08-20 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
5550 * tree-ssa-ccp.c (get_default_value): Remove redundant condition
5553 2013-08-20 David Malcolm <dmalcolm@redhat.com>
5555 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
5556 instances can own GC refs.
5558 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
5559 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
5560 (gcc::context::gt_pch_nx): Likewise.
5561 (gcc::context::gt_pch_nx): Likewise.
5562 * ggc.h (gt_ggc_mx <T>): New.
5563 (gt_pch_nx_with_op <T>): New.
5564 (gt_pch_nx <T>): New.
5565 * passes.c (opt_pass::gt_ggc_mx): New.
5566 (opt_pass::gt_pch_nx): New.
5567 (opt_pass::gt_pch_nx_with_op): New.
5568 (pass_manager::gt_ggc_mx): New.
5569 (pass_manager::gt_pch_nx): New.
5570 (pass_manager::gt_pch_nx_with_op): New.
5571 (pass_manager::operator new): Use
5572 ggc_internal_cleared_alloc_stat rather than xcalloc.
5573 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
5574 (pass_manager::gt_ggc_mx): New.
5575 (pass_manager::gt_pch_nx): New.
5576 (pass_manager::gt_pch_nx_with_op): New.
5577 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
5578 (opt_pass::operator new): New.
5579 (opt_pass::gt_ggc_mx): New.
5580 (opt_pass::gt_pch_nx): New.
5581 (opt_pass::gt_pch_nx_with_op): New.
5583 2013-08-20 David Malcolm <dmalcolm@redhat.com>
5585 * Makefile.in (GTFILES): Add context.h.
5586 * context.c (gcc::context::operator new): New.
5587 (gcc::context::gt_ggc_mx): New.
5588 (gcc::context::gt_pch_nx): New.
5589 (gcc::context::gt_pch_nx): New.
5590 * context.h (gcc::context): Add GTY((user)) marking.
5591 (gcc::context::operator new): New.
5592 (gcc::context::gt_ggc_mx): New.
5593 (gcc::context::gt_pch_nx): New.
5594 (gcc::context::gt_pch_nx): New.
5595 (g): Add GTY marking.
5596 (gt_ggc_mx (gcc::context *)): New.
5597 (gt_pch_nx (gcc::context *)): New.
5598 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
5599 void *cookie)): New.
5600 * gengtype.c (open_base_files) <ifiles>: Add context.h.
5602 2013-08-20 Alan Modra <amodra@gmail.com>
5605 * config/rs6000/rs6000.c (rs6000_emit_prologue): Correct ool_adjust.
5606 (rs6000_emit_epilogue): Likewise.
5608 2013-08-19 Dehao Chen <dehao@google.com>
5610 * value-prof.c (gimple_ic): Fix the bug of adding EH edge.
5612 2013-08-19 Peter Bergner <bergner@vnet.ibm.com>
5613 Jakub Jelinek <jakub@redhat.com>
5615 * builtins.def (BUILT_IN_FABSD32): New DFP ABS builtin.
5616 (BUILT_IN_FABSD64): Likewise.
5617 (BUILT_IN_FABSD128): Likewise.
5618 * builtins.c (expand_builtin): Add support for new DFP ABS builtins.
5619 (fold_builtin_1): Likewise.
5620 * config/rs6000/dfp.md (*negtd2_fpr): Handle non-overlapping
5621 destination and source operands.
5622 (*abstd2_fpr): Likewise.
5623 (*nabstd2_fpr): Likewise.
5625 2013-08-19 Richard Sandiford <rdsandiford@googlemail.com>
5627 * config/mips/mips.c (mips_adjust_insn_length): Add checks for
5630 2013-08-19 Aldy Hernandez <aldyh@redhat.com>
5632 * doc/invoke.texi (-fcilkplus): Clarify that implementation is
5635 2013-08-19 Alexander Ivchenko <alexander.ivchenko@intel.com>
5637 * target.def (TARGET_LIBC_HAS_FUNCTION): New target hook.
5638 * builtins.c (default_libc_has_function): New.
5639 (gnu_libc_has_function): Ditto.
5640 (no_c99_libc_has_function): Ditto.
5641 (expand_builtin_cexpi): Using new target hook TARGET_LIBC_HAS_FUNCTION
5642 instead of TARGET_HAS_SINCOS and TARGET_C99_FUNCTIONS.
5643 (fold_builtin_sincos): Likewise.
5644 (fold_builtin_cexp): Likewise.
5645 * builtins.def (DEF_C94_BUILTIN): Likewise.
5646 (DEF_C99_BUILTIN): Likewise.
5647 (DEF_C99_C90RES_BUILTIN): Likewise.
5648 (DEF_C99_COMPL_BUILTIN): New define. Change all complex c99 builtin
5649 definitions to using this define.
5650 * config/darwin-protos.h (darwin_libc_has_function): New.
5651 * config/darwin.c (darwin_libc_has_function): Ditto.
5652 * config/alpha/linux.h: Remove TARGET_C99_FUNCTIONS and
5653 TARGET_HAS_SINCOS. Redefine TARGET_LIBC_HAS_FUNCTION.
5654 * config/darwin.h: Ditto.
5655 * config/elfos.h: Ditto.
5656 * config/freebsd.h: Ditto.
5657 * config/i386/cygming.h: Ditto.
5658 * config/i386/djgpp.h: Ditto.
5659 * config/i386/i386-interix.h: Ditto.
5660 * config/microblaze/microblaze.h: Ditto.
5661 * config/mmix/mmix.h: Ditto.
5662 * config/gnu-user.h: Ditto.
5663 * config/ia64/hpux.h: Ditto.
5664 * config/pa/pa-hpux.h: Ditto.
5665 * config/pdp11/pdp11.h: Ditto.
5666 * config/picochip/picochip.h: Ditto.
5667 * config/linux.h: Ditto.
5668 * config/netbsd.h: Ditto.
5669 * config/openbsd.h: Ditto.
5670 * config/rs6000/aix43.h: Ditto.
5671 * config/rs6000/aix51.h: Ditto.
5672 * config/rs6000/aix52.h: Ditto.
5673 * config/rs6000/aix53.h: Ditto.
5674 * config/rs6000/aix61.h: Ditto.
5675 * config/rs6000/darwin.h: Ditto.
5676 * config/rs6000/linux.h: Ditto.
5677 * config/rs6000/linux64.h: Ditto.
5678 * config/s390/tpf.h: Ditto.
5679 * config/sol2-10.h: Ditto.
5680 * config/sol2.h: Ditto.
5681 * config/vms/vms.h: Ditto.
5682 * config/vxworks.h: Ditto.
5683 * config/linux-android.c (linux_android_libc_has_function):
5684 New linux-specific implementation of TARGET_LIBC_HAS_FUNCTION.
5685 * config/linux-protos.h (linux_android_libc_has_function):
5687 * config/i386/i386.c (ix86_libc_has_function): New.
5688 * config/i386/i386-protos.h
5689 (ix86_libc_has_function): New declaration.
5690 * config/i386/i386.md
5691 ("isinfxf2"): Change condition for TARGET_LIBC_HAS_FUNCTION.
5692 ("isinf<mode>2): Likewise.
5693 * convert.c (convert_to_integer): Using new target hook
5694 TARGET_LIBC_HAS_FUNCTION istead of TARGET_HAS_SINCOS and
5695 TARGET_C99_FUNCTIONS.
5696 * fortran/f95-lang.c (gfc_init_builtin_functions): Ditto.
5697 * tree-ssa-math-opts.c (execute_cse_sincos): Ditto.
5698 * coretypes.h (function_class): New enum for different
5699 classes of functions.
5700 * defaults.h: Remove TARGET_C99_FUNCTIONS and TARGET_HAS_SINCOS.
5701 * doc/tm.texi.in (TARGET_C99_FUNCTIONS): Remove documentation.
5702 (TARGET_HAS_SINCOS): Likewise.
5703 (TARGET_LIBC_HAS_FUNCTION): New.
5704 * doc/tm.texi: Regenerated.
5705 * targhooks.h (default_libc_has_function): New declaration.
5706 (no_c99_libc_has_function): Ditto.
5707 (gnu_libc_has_function): Ditto.
5708 * system.h: Add the poisoning of TARGET_C99_FUNCTIONS
5709 and TARGET_HAS_SINCOS.
5711 2013-08-18 Jan Hubicka <jh@suse.cz>
5713 * Makeifle-in (ipa-devirt.o): New.
5714 (GTFILES): Add ipa-utils.h and ipa-devirt.c
5715 * cgraphunit.c (decide_is_symbol_needed): Do not care about virtuals.
5716 (analyze_functions): Look into possible targets of polymorphic call.
5717 * dumpfile.c (dump_files): Add type-inheritance dump.
5718 * dumpfile.h (TDI_inheritance): New.
5719 * ipa-devirt.c: New file.
5720 * ipa-utils.h (odr_type_d): Forward declare.
5721 (odr_type): New type.
5722 (build_type_inheritance_graph): Declare.
5723 (possible_polymorphic_call_targets): Declare and introduce inline
5724 variant when only edge is pased.
5725 (dump_possible_polymorphic_call_targets): Likewise.
5726 * timevar.def (TV_IPA_INHERITANCE, TV_IPA_VIRTUAL_CALL): New.
5727 * tree.c (type_in_anonymous_namespace_p): Break out from ...
5728 (types_same_for_odr): ... here.
5729 * tree.h (type_in_anonymous_namespace_p): Declare.
5731 2013-08-18 Jakub Jelinek <jakub@redhat.com>
5733 PR tree-optimization/58006
5734 * tree-parloops.c (take_address_of): Don't ICE if get_name
5736 (eliminate_local_variables_stmt): Remove clobber stmts.
5738 2013-08-18 Eric Botcazou <ebotcazou@adacore.com>
5740 * cgraphunit.c (handle_alias_pairs): Reset the alias flag after the
5741 error message is issued for an alias to undefined symbol.
5743 2013-08-18 Jan Hubicka <jh@suse.cz>
5745 * cgraph.c (cgraph_create_indirect_edge): Discover
5746 polymorphic calls and record basic info into indirect_info.
5747 * gimple-fold.c (gimple_fold_call): When doing BINFO based
5748 devirtualization, ignore objc function calls.
5749 * ipa-cp.c (initialize_node_lattices): Be ready for polymorphic
5750 call with no parm index info.
5751 * ipa-prop.c (ipa_analyze_call_uses): Likewise.
5752 * tree.c (virtual_method_call_p): New function.
5753 * tree.h (virtual_method_call_p): Declare.
5755 2013-08-16 Jan Hubicka <jh@suse.cz>
5758 * tree.c (obj_type_ref_class): Do not ICE on non-method calls.
5760 2013-08-16 David Edelsohn <dje.gcc@gmail.com>
5762 * config/rs6000/rs6000.md (rs6000_get_timebase_ppc32): Add length
5765 2013-08-16 David Malcolm <dmalcolm@redhat.com>
5767 * gengtype.c (type_for_name): Add special-case support for
5768 locating types within the "gcc::" namespace.
5769 (open_base_files): Emit a "using namespace gcc" directive.
5771 2013-08-16 Michael Meissner <meissner@linux.vnet.ibm.com>
5774 * config/rs6000/predicates.md (fusion_gpr_mem_load): Allow the
5775 memory rtx to contain ZERO_EXTEND and SIGN_EXTEND.
5777 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): Pass operands
5778 array instead of each individual operand as a separate argument.
5779 (emit_fusion_gpr_load): Likewise.
5780 (expand_fusion_gpr_load): Add new function declaration.
5782 * config/rs6000/rs6000.c (fusion_gpr_load_p): Change the calling
5783 signature to have the operands passed as an array, instead of as
5784 separate arguments. Allow ZERO_EXTEND to be in the memory
5785 address, and also SIGN_EXTEND if -mpower8-fusion-sign. Do not
5786 depend on the register live/dead flags when peepholes are run.
5787 (expand_fusion_gpr_load): New function to be called from the
5788 peephole2 pass, to change the register that addis sets to be the
5790 (emit_fusion_gpr_load): Change the calling signature to have the
5791 operands passed as an array, instead of as separate arguments.
5792 Allow ZERO_EXTEND to be in the memory address, and also
5793 SIGN_EXTEND if -mpower8-fusion-sign.
5795 * config/rs6000/rs6000.md (UNSPEC_FUSION_GPR): Delete unused
5797 (power8 fusion peephole/peephole2): Rework the fusion peepholes to
5798 adjust the register addis loads up in the peephole2 pass. Do not
5799 depend on the register live/dead state when the peephole pass is done.
5801 2013-08-16 David Malcolm <dmalcolm@redhat.com>
5803 * gengtype.c (create_user_defined_type): Ensure that the kind
5804 is set to TYPE_USER_STRUCT, fixing a bug seen when an incomplete
5805 declaration is seen before the GTY((user)) marking.
5807 2013-08-16 Bernd Edlinger <bernd.edlinger@hotmail.de>
5810 * config/i386/i386.c (make_resolver_func): Set DECL_UNINLINABLE.
5812 2013-08-16 Jan Hubicka <jh@suse.cz>
5814 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Add new
5815 arugment expected_type.
5816 (gimple_fold_call): Use it.
5817 * gimple.h (gimple_extract_devirt_binfo_from_cst): Update prototype.
5818 * ipa-cp.c (ipa_get_indirect_edge_target_1): Update.
5819 * ipa-prop.c (ipa_analyze_virtual_call_uses): Use obj_type_ref_class.
5820 (try_make_edge_direct_virtual_call): Likewise.
5821 * tree.c (obj_type_ref_class): New.
5822 * tree.h (obj_type_ref_class): Use it.
5824 2013-08-16 Gabriel Dos Reis <gdr@integrable-solutions.net>
5826 * sched-vis.c (rtl_slim_pp_initialized): Remove.
5827 (rtl_slim_pp): Likewise.
5828 (init_rtl_slim_pretty_print): Likewise.
5829 (dump_value_slim): Don't call it. Use local pretty printer.
5830 (dump_insn_slim): Likewise.
5831 (dump_rtl_slim): Likewise.
5832 (str_pattern_slim): Likewise.
5833 * tree-mudflap.c (mf_varname_tree): Use local pretty printer.
5836 2013-08-16 Jakub Jelinek <jakub@redhat.com>
5838 PR tree-optimization/58164
5839 * gimple.c (walk_stmt_load_store_addr_ops): For visit_addr
5840 walk gimple_goto_dest of GIMPLE_GOTO.
5842 PR tree-optimization/58165
5843 * tree-call-cdce.c (shrink_wrap_one_built_in_call): If
5844 bi_call must be the last stmt in a bb, don't split_block, instead
5845 use fallthru edge from it and give up if there is none.
5846 Release conds vector when returning early.
5848 2013-08-14 Xinliang David Li <davidxl@google.com>
5850 * config/i386/i386.c (ix86_option_override_internal):
5851 Remove unused variable and field.
5853 2013-08-14 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5856 * doc/invoke.texi: Add documentation of mcompat-align-parm option.
5857 * config/rs6000/rs6000.opt: Add mcompat-align-parm option.
5858 * config/rs6000/rs6000.c (rs6000_function_arg_boundary): For AIX
5859 and Linux, correct BLKmode alignment when 128-bit alignment is
5860 required and compatibility flag is not set.
5861 (rs6000_gimplify_va_arg): For AIX and Linux, honor specified alignment
5862 for zero-size arguments when compatibility flag is not set.
5864 2013-08-14 Jakub Jelinek <jakub@redhat.com>
5866 PR tree-optimization/58145
5867 * tree-sra.c (build_ref_for_offset): If prev_base has
5868 TREE_THIS_VOLATILE or TREE_SIDE_EFFECTS, propagate it to MEM_REF.
5870 2013-08-14 Xinliang David Li <davidxl@google.com>
5872 * config/i386/i386.c (ix86_option_override_internal):
5873 Fix uninitialized variable error.
5875 2013-08-14 Xinliang David Li <davidxl@google.com>
5877 * config/i386/i386.opt: Define two new options.
5878 * config/i386/x86-tune.def: Add arch selector field in macros.
5879 * config/i386/i386.h: Adjust macro definition.
5880 * config/i386/i386.c (ix86_option_override_internal):
5882 (parse_mtune_ctrl_str): New function.
5883 (set_ix86_tune_features): New function.
5884 (ix86_function_specific_restore): Call the new helper function.
5886 2013-08-14 Andrey Belevantsev <abel@ispras.ru>
5888 PR rtl-optimization/57662
5889 * sel-sched.c (code_motion_process_successors): When the current insn
5890 is removed after the recursive traversal, break from the loop.
5891 Add comments and debug printouts.
5893 2013-08-14 Jakub Jelinek <jakub@redhat.com>
5894 Alexandre Oliva <aoliva@redhat.com>
5897 * config/i386/i386.c (ix86_delegitimize_address): For CM_MEDIUM_PIC
5898 and CM_LARGE_PIC ix86_cmodel fall thru into the -m32 code, handle
5899 there also UNSPEC_PLTOFF.
5901 2013-08-14 Marek Polacek <polacek@redhat.com>
5903 * ipa-inline-analysis.c (add_clause): Avoid shifting integer
5904 NUM_CONDITIONS bit positions.
5906 2013-08-13 Cary Coutant <ccoutant@google.com>
5908 * dwarf2out.c (CHECKSUM_BLOCK): New macro.
5909 (attr_checksum): Hash vector contents instead of pointer.
5910 (attr_checksum_ordered): Likewise.
5912 2013-08-13 Uros Bizjak <ubizjak@gmail.com>
5914 * config/i386/sse.md (*sse2_maskmovdqu): Emit addr32 prefix
5915 when Pmode != word_mode. Add length_address attribute.
5916 (sse3_monitor_<mode>): Merge from sse3_monitor and
5917 sse3_monitor64_<mode> insn patterns. Emit addr32 prefix when
5918 Pmode != word_mode. Update insn length attribute.
5919 * config/i386/i386.c (ix86_option_override_internal): Update
5920 ix86_gen_monitor selection for merged sse3_monitor insn.
5922 2013-08-13 Julian Brown <julian@codesourcery.com>
5924 * config/rs6000/rs6000.c (rs6000_legitimize_reload_address): Don't
5925 perform invalid legitimization on greater-than-word-size modes for
5928 2013-08-13 Vladimir Makarov <vmakarov@redhat.com>
5930 * ira.c (setup_class_translate_array): Use aclass instead of cl
5931 for classes not fully covered by allocno classes.
5933 2013-08-13 Jakub Jelinek <jakub@redhat.com>
5935 PR tree-optimization/57661
5936 * tree-inline.h (struct copy_body_data): Add blocks_to_copy field.
5937 * tree-inline.c (tree_function_versioning): Initialize it.
5938 (remap_gimple_stmt): Return GIMPLE_NOP for MEM_REF lhs clobber stmts
5939 if id->blocks_to_copy and MEM_REF's SSA_NAME is defined in a block
5940 that is not being copied.
5943 * asan.c (instrument_strlen_call): Fix typo in comment.
5944 Use char * type even for the lhs of POINTER_PLUS_EXPR.
5946 2013-08-13 Steve Ellcey <sellcey@mips.com>
5948 * config/mips/mips.md (prefetch): Use lw instead of ld on
5949 loongson in 32bit mode.
5951 2013-08-13 Nick Clifton <nickc@redhat.com>
5953 * config.gcc: (avr-linux): Allow for tmake_file not being empty.
5955 2013-08-13 Jan Hubicka <jh@suse.cz>
5957 * cgraph.c (cgraph_turn_edge_to_speculative): Return newly
5958 introduced edge; fix typo in sanity check.
5959 (cgraph_resolve_speculation): Export; improve diagnostic.
5960 (cgraph_redirect_edge_call_stmt_to_callee): Better diagnostic; cancel
5961 speculation at type mismatch.
5962 * cgraph.h (cgraph_turn_edge_to_speculative): Update.
5963 (cgraph_resolve_speculation): Declare.
5964 (symtab_can_be_discarded): New function.
5965 * value-prof.c (gimple_ic_transform): Remove actual transform code.
5966 * ipa-inline-transform.c (speculation_removed): New global var.
5967 (clone_inlined_nodes): See if speculation can be removed.
5968 (inline_call): If speculations was removed, we growths may not match.
5969 * ipa-inline.c (can_inline_edge_p): Add DISREGARD_LIMITS parameter.
5970 (speculation_useful_p): New function.
5971 (resolve_noninline_speculation): New function.
5972 (inline_small_functions): Resolve useless speculations.
5973 * ipa-inline.h (speculation_useful_p): Declare
5974 * ipa.c (can_replace_by_local_alias): Simplify.
5975 (ipa_profile): Produce speculative calls in non-lto, too;
5976 add simple cost model; produce local aliases.
5978 2013-08-13 David Malcolm <dmalcolm@redhat.com>
5980 * config/i386/t-i386 (i386.o): Rename stray PIPELINE_H to
5983 2013-08-12 Paolo Carlini <paolo.carlini@oracle.com>
5985 * config/i386/i386.c (ix86_function_versions): Use error + inform.
5987 2013-08-12 Uros Bizjak <ubizjak@gmail.com>
5989 * config/i386/i386.md (floatunssi<mode>2 expand): Use MODEF mode
5990 iterator instead of X87MODEF.
5992 2013-08-12 Perez Read <netfirewall@gmail.com>
5995 * config/i386/i386.md (*movabs<mode>_1): Add <ptrsize> PTR before
5996 operand 0 for intel asm alternative.
5997 (*movabs<mode>_2): Ditto for operand 1.
5999 2013-08-12 James Greenhalgh <james.greenhalgh@arm.com>
6001 * config/aarch64/arm_none.h
6002 (vdup<bhsd>_lane_<su><8,16,32,64>): Fix macro call.
6004 2013-08-12 Nick Clifton <nickc@redhat.com>
6006 * config.gcc (m32r-linux): Allow for tmake_file not being empty.
6008 2013-08-12 Yuri Rumyantsev <ysrumyan@gmail.com>
6010 * config/i386/i386.md (floatunssi<mode>2 expand): Add new
6011 expand for QI/HImode operand to produce more effictive code for
6012 unsigned char(short) --> float(double) conversion.
6014 2013-08-12 Alexander Monakov <amonakov@ispras.ru>
6016 * doc/invoke.texi: Mention that -ftls-model does not force the final
6019 2013-08-12 Marek Polacek <polacek@redhat.com>
6020 Marc Glisse <marc.glisse@inria.fr>
6022 PR tree-optimization/57980
6023 * tree-tailcall.c (process_assignment): Call build_minus_one_cst
6024 when creating -1 constant.
6026 2013-08-10 Jan Hubicka <jh@suse.cz>
6028 Workaround binutils PR14342.
6029 * tree-profile.c (init_ic_make_global_vars): Add LTO path.
6030 (gimple_init_edge_profiler): Likewise.
6031 (gimple_gen_ic_func_profiler): Likewise.
6033 2013-08-09 Jan Hubicka <jh@suse.cz>
6035 * cgraph.c (cgraph_create_edge_1): Clear speculative flag.
6037 2013-08-09 Xinliang David Li <davidxl@google.com>
6039 * config/i386/stringop.def: New file.
6040 * config/i386/stringop.opt: New file.
6041 * config/i386/i386-opts.h: Include stringopt.def.
6042 * config/i386/i386.opt: Include stringopt.opt.
6043 * config/i386/i386.c (ix86_option_override_internal):
6044 Override default size based stringop inline strategies with options.
6045 * config/i386/i386.c (ix86_parse_stringop_strategy_string):
6048 2013-08-09 Jan Hubicka <jh@suse.cz>
6050 * ipa-ref.c (ipa_clear_stmts_in_references): Clear lto_stmt_uid, too.
6052 2013-08-09 Jan Hubicka <jh@suse.cz>
6054 * cgraph.c (cgraph_resolve_speculation): Cut frequency to
6056 (dump_cgraph_node): Dump profile-id.
6057 * cgraph.h (cgraph_indirect_call_info): Add common_target_id
6058 and common_target_probability.
6059 * lto-cgraph.c (lto_output_edge): Stream common targets.
6060 (lto_output_node): Stream profile ids.
6061 (input_node): Stream profile ids.
6062 (input_edge): Stream common targets.
6063 * lto-streamer-in.c (fixup_call_stmt_edges_1): Fix formatting.
6064 * ipa.c: Include value-prof.h
6065 (ipa_profile_generate_summary): Turn indirect call statement histograms
6066 into common targets.
6067 (ipa_profile): Turn common targets into speculative edges.
6069 2013-08-09 Jan Hubicka <jh@suse.cz>
6071 * cgraph.h (cgraph_node): Add profile_id.
6072 * value-prof.c (cgraph_node_map): Turn into pointer_map.
6073 (init_node_map): Rewrite to handle hashes increas of incremental IDs.
6074 (del_node_map): Update.
6075 (find_func_by_funcdef_no): Replace by ...
6076 (find_func_by_profile_id): ... this one.
6077 (gimple_ic_transform): Do not remove useful histograms when
6078 speculation is not done; dump info when indirect call removal
6080 * value-prof.h (find_func_by_profile_id, gimple_ic): Declare.
6081 * gcov-io.h (__gcov_indirect_call_profiler): Replace by ...
6082 (__gcov_indirect_call_profiler_v2): .. this one.
6083 * profile.h (init_node_map): Update.
6084 * coverage.c (coverage_compute_profile_id): New function.
6085 * coverage.h (coverage_compute_profile_id): Declare.
6086 * tree-profile.c (init_ic_make_global_vars): Make
6087 __gcov_indirect_call_callee and __gcov_indirect_call_counters global.
6088 (gimple_init_edge_profiler): Update prototype of
6089 __gcov_indirect_call_profiler.
6090 (gimple_gen_ic_func_profiler): Simplify.
6091 (tree_profiling): Use init_node_map
6093 2013-08-09 Jan Hubicka <jh@suse.cz>
6095 * cgraphbuild.c (cgraph_rebuild_references): Rebuild only
6096 non-speculative refs.
6097 * cgraph.c (cgraph_update_edge_in_call_site_hash): New function.
6098 (cgraph_add_edge_to_call_site_hash): Deal with speculative calls.
6099 (cgraph_set_call_stmt): Likewise.
6100 (cgraph_create_edge_1): Fix release checking compilatoin;
6102 (cgraph_free_edge): Free indirect info.
6103 (cgraph_turn_edge_to_speculative): New function.
6104 (cgraph_speculative_call_info): New function.
6105 (cgraph_make_edge_direct): Return direct edge; handle speculation.
6106 (cgraph_redirect_edge_call_stmt_to_callee): Expand speculative edges.
6107 (dump_cgraph_node): Dump speculation.
6108 (verify_edge_count_and_frequency): Accept speculative edges.
6109 (verify_edge_corresponds_to_fndecl): Handle partitioned cgraph.
6110 (verify_cgraph_node): Handle speculation.
6111 * cgraph.h (cgraph_edge): Add SPECULATIVE flag.
6112 (cgraph_set_call_stmt): Update prototype.
6113 (cgraph_make_edge_direct): Update prototype.
6114 (cgraph_speculative_call_info): Declare.
6115 * ipa-cp.c (ipcp_discover_new_direct_edges): Be ready for edge
6116 to change; update call of ipa_find_references.
6117 * ipa-ref.c (ipa_record_reference): Fix return value; clear
6118 lto_stmt_uid and speculative flags.
6119 (ipa_dump_references): Dump speculation.
6120 (ipa_clone_references): Clone speculative flag.
6121 (ipa_clone_referring): Likewise.
6122 (ipa_clone_ref): New function.
6123 (ipa_find_reference): Look into lto_stmt_uids
6124 (ipa_clear_stmts_in_references): Do not clear speculative calls.
6125 * ipa-ref.h (ipa_ref): Add lto_stmt_uid and speculative flags.
6126 (ipa_find_reference): Update declaration.
6127 (ipa_clone_ref): Declare.
6128 * lto-cgraph.c (lto_output_edge): Make lto_stmt_uids start from 0;
6129 stream speculative flag.
6130 (lto_output_ref): Stream statements uids and speculation.
6131 (input_ref): Likewise.
6132 (input_edge): Stream speuclation.
6133 * cgraphclones.c (cgraph_clone_edge): Clone speculation.
6134 (cgraph_set_call_stmt_including_clones): Handle speculation.
6135 * ipa-inline.c (heap_edge_removal_hook): New function.
6136 (inline_small_functions): Register it.
6137 * lto-streamer-in.c (fixup_call_stmt_edges_1): Bounds checking;
6138 also initialize refs.
6139 * ipa-prop.c (ipa_make_edge_direct_to_target): Be ready for
6141 (try_make_edge_direct_simple_call): Likewise.
6142 (try_make_edge_direct_simple_call): Likewise.
6143 (update_indirect_edges_after_inlining): Likewise.
6144 (remove_described_reference): Look proper lto_stmt_uid.
6145 (propagate_controlled_uses): Likewise.
6146 (propagate_controlled_uses): Liekwise.
6147 * tree-inline.c (copy_bb): Copy speculative edges.
6148 (redirect_all_calls): New function.
6149 (copy_cfg_body): Do redirection after loop info is updated.
6150 (delete_unreachable_blocks_update_callgraph): Updadte speculation.
6152 2013-08-09 Jan Hubicka <jh@suse.cz>
6154 * lto-streamer-out.c (output_function): Renumber PHIs.
6155 * lto-streamer-in.c (input_function): Likewise.
6157 2013-08-09 James Greenhalgh <james.greenhalgh@arm.com>
6159 * config/aarch64/aarch64-simd-builtins.def (get_lane_signed): Remove.
6160 (get_lane_unsigned): Likewise.
6161 (dup_lane_scalar): Likewise.
6162 (get_lane): enable for VALL.
6163 * config/aarch64/aarch64-simd.md
6164 (aarch64_dup_lane_scalar<mode>): Remove.
6165 (aarch64_get_lane_signed<mode>): Likewise.
6166 (aarch64_get_lane_unsigned<mode>): Likewise.
6167 (aarch64_get_lane_extend<GPI:mode><VDQQH:mode>): New.
6168 (aarch64_get_lane_zero_extendsi<mode>): Likewise.
6169 (aarch64_get_lane<mode>): Enable for all vector modes.
6170 (aarch64_get_lanedi): Remove misleading constraints.
6171 * config/aarch64/arm_neon.h
6172 (__aarch64_vget_lane_any): Define.
6173 (__aarch64_vget<q>_lane_<fpsu><8,16,32,64>): Likewise.
6174 (vget<q>_lane_<fpsu><8,16,32,64>): Use __aarch64_vget_lane macros.
6175 (vdup<bhsd>_lane_<su><8,16,32,64>): Likewise.
6176 * config/aarch64/iterators.md (VDQQH): New.
6180 2013-08-09 Eric Botcazou <ebotcazou@adacore.com>
6182 * configure.ac: Add GAS check for LEON instructions on SPARC.
6183 * configure: Regenerate.
6184 * config.in: Likewise.
6185 * config.gcc (with_cpu): Remove sparc-leon*-* and deal with LEON in the
6187 * config/sparc/sparc.opt (LEON, LEON3): New masks.
6188 * config/sparc/sparc.h (ASM_CPU32_DEFAULT_SPEC): Set to AS_LEON_FLAG
6190 (ASM_CPU_SPEC): Pass AS_LEON_FLAG if -mcpu=leon or -mcpu=leon3.
6191 (AS_LEON_FLAG): New macro.
6192 * config/sparc/sparc.c (sparc_option_override): Set MASK_LEON for leon
6193 and MASK_LEON3 for leon3 and unset them if HAVE_AS_LEON is not defined.
6194 Deal with LEON and LEON3 for the memory model.
6195 * config/sparc/sync.md (atomic_compare_and_swap<mode>): Enable if LEON3
6196 (atomic_compare_and_swap<mode>_1): Likewise.
6197 (*atomic_compare_and_swap<mode>_1): Likewise.
6199 2013-08-09 Zhenqiang Chen <zhenqiang.chen@linaro.org>
6201 * config/arm/neon.md (vcond): Fix floating-point vector
6202 comparisons against 0.
6204 2013-08-08 Vladimir Makarov <vmakarov@redhat.com>
6206 * lra-constraints.c (emit_spill_move): Remove assert.
6207 (process_alt_operands): Add more debugging
6208 output. Increase reject for spilling into memory. Decrease
6209 reject for reloading scratch.
6210 (split_reg): Use HARD_REGNO_CALLER_SAVE_MODE.
6212 2013-08-08 Steve Ellcey <sellcey@mips.com>
6214 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add nan2008.
6215 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Make mips16 and
6216 micromips incompatible. Add nan2008.
6217 (MULTILIB_DIRNAMES): Add nan2008.
6218 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
6219 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Make mips16
6220 and micromips incompatible. Add nan2008.
6221 (MULTILIB_DIRNAMES): Add nan2008.
6222 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
6224 2013-08-08 Richard Sandiford <rdsandiford@googlemail.com>
6226 PR rtl-optimization/58079
6227 * combine.c (combine_simplify_rtx): Avoid using SUBST if
6228 simplify_comparison has widened a comparison with an integer.
6230 2013-08-08 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6232 * config/arm/neon.md (movmisalign<mode>): Disable when we
6233 don't allow unaligned accesses.
6234 (*movmisalign<mode>_neon_store): Likewise.
6235 (*movmisalign<mode>_neon_load): Likewise.
6236 (*movmisalign<mode>_neon_store): Likewise.
6237 (*movmisalign<mode>_neon_load): Likewise.
6239 2013-08-08 Jan Hubicka <jh@suse.cz>
6241 * cgraphbuild.c (build_cgraph_edges): Do not walk into debugs.
6242 (make_pass_rebuild_cgraph_edges): Also clear references.
6243 * cgraph.c (verify_cgraph_node): Add basic ipa-ref verifier.
6244 * ipa-inline-transform.c (inline_transform): Remove all references
6246 * cgraphunit.c (expand_function): Remove all references after
6248 * ipa-ref.c (ipa_ref_has_aliases_p): Fix formatting.
6249 (ipa_find_reference): Rewrite to iterator.
6250 (remove_stmt_references): Likewise.
6251 (ipa_clear_stmts_in_references): New function.
6252 * ipa-ref.h (ipa_clear_stmts_in_references): Declare.
6253 * cgraphclones.c (cgraph_materialize_all_clones): Remove or
6255 * ipa-split.c (split_function): Remove references in split function.
6257 2013-08-08 Richard Earnshaw <rearnsha@arm.com>
6260 * config/arm/arm/neon.md (neon_vld1_dupdi): New expand pattern.
6261 (neon_vld1_dup<mode> VD iterator): Iterate over VD not VDX.
6263 2013-08-08 Richard Earnshaw <rearnsha@arm.com>
6266 * config/arm/arm.c (aapcs_vfp_allocate): Decompose the argument if the
6267 suggested mode for the assignment isn't compatible with the
6270 2013-08-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
6273 * config/arm/arm.h (MALLOC_ABI_ALIGNMENT): Define.
6275 2013-08-07 Xinliang David Li <davidxl@google.com>
6277 * config/i386/i386.opt: New option -mtune-ctrl=.
6278 * config/i386/x86-tune.def: New file.
6279 * config/i386/i386.h: include x86-tune.def.
6280 * config/i386/i386.c (ix86_option_override_internal):
6281 Parsing -mtune-ctrl= option and set tune features.
6283 2013-08-07 Oleg Endo <olegendo@gcc.gnu.org>
6286 * config/rs6000/rs6000.c (gen_2arg_fn_t): Remove typedef.
6287 (rs6000_emit_swdiv, rs6000_emit_swrsqrt): Don't cast result of GEN_FCN
6290 2013-08-07 Eric Botcazou <ebotcazou@adacore.com>
6292 * rtl.h (update_alignments): Declare.
6293 * final.c (grow_label_align): New function extracted from...
6294 (shorten_branches): ...here. Call it.
6295 (update_alignments): New function.
6296 * reorg.c (sibling_labels): New variable.
6297 (get_label_before): Add SIBLING parameter. If it is non-zero, push
6298 the new label along with it onto the sibling_labels vector.
6299 (fill_simple_delay_slots): Adjust call to get_label_before.
6300 (fill_slots_from_thread): Likewise.
6301 (relax_delay_slots): Likewise.
6302 (make_return_insns): Likewise.
6303 (dbr_schedule): Invoke update_alignment on the sibling_labels vector.
6305 2013-08-07 Eric Botcazou <ebotcazou@adacore.com>
6307 * diagnostic.c (diagnostic_classify_diagnostic): Accept zero index and
6308 document its semantics.
6309 (diagnostic_report_diagnostic): Adjust accordingly.
6311 2013-08-07 David Malcolm <dmalcolm@redhat.com>
6313 * config/sparc/sparc.c (insert_pass_work_around_errata): Move into...
6314 (sparc_option_override): ...and port to new C++ pass API.
6315 * config/sparc/t-sparc (sparc.o): Add dep on CONTEXT_H
6317 2013-08-07 Peter Bergner <bergner@vnet.ibm.com>
6319 * config/rs6000/rs6000.c (htm_expand_builtin) <case 0>: Remove.
6321 2013-08-06 Caroline Tice <cmtice@google.com>
6323 * gcc.c (VTABLE_VERIFICATION_SPEC): New definition.
6324 (LINK_COMMAND_SPEC): Add VTABLE_VERIFICATION_SPEC.
6325 * tree-pass.h: Add pass_vtable_verify.
6326 * varasm.c (assemble_variable): Add code to properly set the comdat
6327 section and name for the .vtable_map_vars section.
6328 (assemble_vtyv_preinit_initializer): New function.
6329 (default_sectin_type_flags): Make sure .vtable_map_vars section has
6331 * output.h: Add function decl for assemble_vtv_preinit_initializer.
6332 * vtable-verify.c: New file.
6333 * vtable-verify.h: New file.
6334 * flag-types.h (enum vtv_priority): Defintions for flag_vtable_verify
6335 initialiation levels.
6336 * timevar.def (TV_VTABLE_VERIFICATION): New definition.
6337 * passes.def: Insert pass_vtable_verify.
6338 * aclocal.m4: Reorder includes.
6339 * doc/invoke.texi: Document the -fvtable-verify=, -fvtv-debug, and
6340 -fvtv-counts options.
6341 * config/gnu-user.h (GNU_USER_TARGET_STARTFILE_SPEC): Add vtv_start*.o,
6342 as appropriate, if -fvtable-verify=... is used.
6343 (GNU_USER_TARGET_ENDFILE_SPEC): Add vtv_end*.o as appropriate, if
6344 -fvtable-verify=... is used.
6345 * Makefile.in (OBJS): Add vtable-verify.o to list.
6346 (vtable-verify.o): Add new build rule.
6347 (GTFILES): Add vtable-verify.c to list.
6348 * common.opt (fvtable-verify=): New flag.
6349 (vtv_priority): Values for fvtable-verify= flag.
6350 (fvtv-counts): New flag.
6351 (fvtv-debug): New flag.
6352 * tree.h (save_vtable_map_decl): New extern function decl.
6354 2013-08-07 David Malcolm <dmalcolm@redhat.com>
6356 * config/rl78/rl78.c (rl78_devirt_pass): Convert from a struct to...
6357 (pass_rl78_devirt): ...new subclass of rtl_opt_pass along with...
6358 (pass_data_rl78_devirt): ...new pass_data instance and...
6359 (make_pass_rl78_devirt): ...new function.
6360 (rl78_asm_file_start): Port pass registration to new C++ API.
6362 2013-08-07 David Malcolm <dmalcolm@redhat.com>
6364 * coretypes.h (rtl_opt_pass): Add.
6365 (gcc::context): Add.
6366 * config/epiphany/epiphany.c (pass_mode_switch_use): New.
6367 (epiphany_init): Port to new C++ pass API.
6368 (epiphany_optimize_mode_switching): Likewise.
6369 * pass_manager.h (pass_manager::get_pass_split_all_insns): New.
6370 (pass_manager::get_pass_mode_switching): New.
6371 (pass_manager::get_pass_peephole2): New.
6372 * mode-switching.c (pass_mode_switching): Add clone method.
6373 * recog.c (pass_peephole2): Add clone method.
6374 (pass_split_all_insns): Add clone method.
6376 2013-08-06 David Malcolm <dmalcolm@redhat.com>
6378 * config/mips/mips.c (insert_pass_mips_machine_reorg2): Move into...
6379 (mips_option_override): ...here, porting to new C++ API for passes.
6381 2013-08-06 Jan Hubicka <jh@suse.cz>
6383 * cgraph.c (cgraph_get_body): New function based on lto.c
6385 * cgraph.h (cgraph_get_body): Declare.
6386 * cgraphclones.c (cgraph_create_virtual_clone): Commonize WPA and
6388 * cgraphunit.c (expand_function): Get body prior expanding.
6389 * ipa.c (function_and_variable_visibility): Use gimple_has_body_p test.
6390 * lto-cgraph.c (lto_output_node): Do not stream bodies we don't
6392 * passes.c (do_per_function_toporder): Get body.
6393 * tree-inline.c (expand_call_inline): Get body prior inlining it.
6394 * tree-ssa-structalias.c (ipa_pta_execute): Get body; skip clones.
6396 2013-08-06 Martin Jambor <mjambor@suse.cz>
6399 * cgraphunit.c (cgraph_finalize_function): Assert that nested function
6400 is not re-finalized. Rename second parameter to no_collect.
6402 2013-08-06 Martin Jambor <mjambor@suse.cz>
6405 * gimple-ssa-strength-reduction.c (replace_ref): Make sure built
6406 MEM_REF has proper alignment information.
6408 2013-08-05 Oleg Endo <olegendo@gcc.gnu.org>
6411 * recog.h (rtx (*insn_gen_fn) (rtx, ...)): Replace typedef with new
6413 * expr.c (move_by_pieces_1, store_by_pieces_2): Replace argument
6414 rtx (*) (rtx, ...) with insn_gen_fn.
6415 * genoutput.c (output_insn_data): Cast gen_? function pointers to
6416 insn_gen_fn::stored_funcptr. Add initializer braces.
6418 2013-08-05 David Malcolm <dmalcolm@redhat.com>
6420 Rewrite how instances of passes are cloned to remove assumptions
6421 about their sizes (thus allowing pass subclasses to have
6422 additional data fields, albeit non-GC-managed ones at this point).
6424 * passes.c (make_pass_instance): Now that passes have clone
6425 methods, rewrite this function to eliminate XNEW and memcpy
6426 calls that used hardcoded sizes. Since this function no longer
6427 creates pass instances, rename it to...
6428 (add_pass_instance): ...this. Document the old way that passes were
6429 numbered and flagged, and rework this function to continue using it.
6430 (next_pass_1): Add an initial_pass argument for use by
6432 (position_pass): When adding multiple instances of a pass, use
6433 the pass's clone method, rather than relying on the XNEW/memcpy
6434 within the former make_pass_instance (now add_pass_instance).
6435 (pass_manager::pass_manager): When invoking next_pass_1, also supply
6436 the initial instance of the current pass within the pass manager.
6438 2013-08-05 David Malcolm <dmalcolm@redhat.com>
6440 This is the automated part of the conversion of passes from C
6441 structs to C++ classes.
6443 Patch autogenerated by refactor_passes.py from
6444 https://github.com/davidmalcolm/gcc-refactoring-scripts
6445 revision 03fe39476a4c4ea450b49e087cfa817b5f92021e
6447 * asan.c (pass_asan): Convert from a global struct to a subclass of
6448 gimple_opt_pass along with...
6449 (pass_data_asan): ...new pass_data instance and...
6450 (make_pass_asan): ...new function.
6451 (pass_asan_O0): Convert from a global struct to a subclass of
6452 gimple_opt_pass along with...
6453 (pass_data_asan_O0): ...new pass_data instance and...
6454 (make_pass_asan_O0): ...new function.
6455 * auto-inc-dec.c (pass_inc_dec): Convert from a global struct to a
6456 subclass of rtl_opt_pass along with...
6457 (pass_data_inc_dec): ...new pass_data instance and...
6458 (make_pass_inc_dec): ...new function.
6459 * bb-reorder.c (pass_reorder_blocks): Convert from a global struct to
6460 a subclass of rtl_opt_pass along with...
6461 (pass_data_reorder_blocks): ...new pass_data instance and...
6462 (make_pass_reorder_blocks): ...new function.
6463 (pass_duplicate_computed_gotos): Convert from a global struct to a
6464 subclass of rtl_opt_pass along with...
6465 (pass_data_duplicate_computed_gotos): ...new pass_data instance and...
6466 (make_pass_duplicate_computed_gotos): ...new function.
6467 (pass_partition_blocks): Convert from a global struct to a subclass of
6468 rtl_opt_pass along with...
6469 (pass_data_partition_blocks): ...new pass_data instance and...
6470 (make_pass_partition_blocks): ...new function.
6471 * bt-load.c (pass_branch_target_load_optimize1): Convert from a global
6472 struct to a subclass of rtl_opt_pass along with...
6473 (pass_data_branch_target_load_optimize1): ...new pass_data instance
6475 (make_pass_branch_target_load_optimize1): ...new function.
6476 (pass_branch_target_load_optimize2): Convert from a global struct to a
6477 subclass of rtl_opt_pass along with...
6478 (pass_data_branch_target_load_optimize2): ...new pass_data instance
6480 (make_pass_branch_target_load_optimize2): ...new function.
6481 * cfgcleanup.c (pass_jump): Convert from a global struct to a subclass
6482 of rtl_opt_pass along with...
6483 (pass_data_jump): ...new pass_data instance and...
6484 (make_pass_jump): ...new function.
6485 (pass_jump2): Convert from a global struct to a subclass of
6486 rtl_opt_pass along with...
6487 (pass_data_jump2): ...new pass_data instance and...
6488 (make_pass_jump2): ...new function.
6489 * cfgexpand.c (pass_expand): Convert from a global struct to a
6490 subclass of rtl_opt_pass along with...
6491 (pass_data_expand): ...new pass_data instance and...
6492 (make_pass_expand): ...new function.
6493 * cfgrtl.c (pass_free_cfg): Convert from a global struct to a subclass
6494 of rtl_opt_pass along with...
6495 (pass_data_free_cfg): ...new pass_data instance and...
6496 (make_pass_free_cfg): ...new function.
6497 (pass_into_cfg_layout_mode): Convert from a global struct to a
6498 subclass of rtl_opt_pass along with...
6499 (pass_data_into_cfg_layout_mode): ...new pass_data instance and...
6500 (make_pass_into_cfg_layout_mode): ...new function.
6501 (pass_outof_cfg_layout_mode): Convert from a global struct to a
6502 subclass of rtl_opt_pass along with...
6503 (pass_data_outof_cfg_layout_mode): ...new pass_data instance and...
6504 (make_pass_outof_cfg_layout_mode): ...new function.
6505 * cgraphbuild.c (pass_build_cgraph_edges): Convert from a global
6506 struct to a subclass of gimple_opt_pass along with...
6507 (pass_data_build_cgraph_edges): ...new pass_data instance and...
6508 (make_pass_build_cgraph_edges): ...new function.
6509 (pass_rebuild_cgraph_edges): Convert from a global struct to a
6510 subclass of gimple_opt_pass along with...
6511 (pass_data_rebuild_cgraph_edges): ...new pass_data instance and...
6512 (make_pass_rebuild_cgraph_edges): ...new function.
6513 (pass_remove_cgraph_callee_edges): Convert from a global struct to a
6514 subclass of gimple_opt_pass along with...
6515 (pass_data_remove_cgraph_callee_edges): ...new pass_data instance
6517 (make_pass_remove_cgraph_callee_edges): ...new function.
6518 * combine-stack-adj.c (pass_stack_adjustments): Convert from a global
6519 struct to a subclass of rtl_opt_pass along with...
6520 (pass_data_stack_adjustments): ...new pass_data instance and...
6521 (make_pass_stack_adjustments): ...new function.
6522 * combine.c (pass_combine): Convert from a global struct to a subclass
6523 of rtl_opt_pass along with...
6524 (pass_data_combine): ...new pass_data instance and...
6525 (make_pass_combine): ...new function.
6526 * compare-elim.c (pass_compare_elim_after_reload): Convert from a
6527 global struct to a subclass of rtl_opt_pass along with...
6528 (pass_data_compare_elim_after_reload): ...new pass_data instance
6530 (make_pass_compare_elim_after_reload): ...new function.
6531 * cprop.c (pass_rtl_cprop): Convert from a global struct to a subclass
6532 of rtl_opt_pass along with...
6533 (pass_data_rtl_cprop): ...new pass_data instance and...
6534 (make_pass_rtl_cprop): ...new function.
6535 * cse.c (pass_cse): Convert from a global struct to a subclass of
6536 rtl_opt_pass along with...
6537 (pass_data_cse): ...new pass_data instance and...
6538 (make_pass_cse): ...new function.
6539 (pass_cse2): Convert from a global struct to a subclass of
6540 rtl_opt_pass along with...
6541 (pass_data_cse2): ...new pass_data instance and...
6542 (make_pass_cse2): ...new function.
6543 (pass_cse_after_global_opts): Convert from a global struct to a
6544 subclass of rtl_opt_pass along with...
6545 (pass_data_cse_after_global_opts): ...new pass_data instance and...
6546 (make_pass_cse_after_global_opts): ...new function.
6547 * dce.c (pass_ud_rtl_dce): Convert from a global struct to a subclass
6548 of rtl_opt_pass along with...
6549 (pass_data_ud_rtl_dce): ...new pass_data instance and...
6550 (make_pass_ud_rtl_dce): ...new function.
6551 (pass_fast_rtl_dce): Convert from a global struct to a subclass of
6552 rtl_opt_pass along with...
6553 (pass_data_fast_rtl_dce): ...new pass_data instance and...
6554 (make_pass_fast_rtl_dce): ...new function.
6555 * df-core.c (pass_df_initialize_opt): Convert from a global struct to
6556 a subclass of rtl_opt_pass along with...
6557 (pass_data_df_initialize_opt): ...new pass_data instance and...
6558 (make_pass_df_initialize_opt): ...new function.
6559 (pass_df_initialize_no_opt): Convert from a global struct to a
6560 subclass of rtl_opt_pass along with...
6561 (pass_data_df_initialize_no_opt): ...new pass_data instance and...
6562 (make_pass_df_initialize_no_opt): ...new function.
6563 (pass_df_finish): Convert from a global struct to a subclass of
6564 rtl_opt_pass along with...
6565 (pass_data_df_finish): ...new pass_data instance and...
6566 (make_pass_df_finish): ...new function.
6567 * dse.c (pass_rtl_dse1): Convert from a global struct to a subclass of
6568 rtl_opt_pass along with...
6569 (pass_data_rtl_dse1): ...new pass_data instance and...
6570 (make_pass_rtl_dse1): ...new function.
6571 (pass_rtl_dse2): Convert from a global struct to a subclass of
6572 rtl_opt_pass along with...
6573 (pass_data_rtl_dse2): ...new pass_data instance and...
6574 (make_pass_rtl_dse2): ...new function.
6575 * dwarf2cfi.c (pass_dwarf2_frame): Convert from a global struct to a
6576 subclass of rtl_opt_pass along with...
6577 (pass_data_dwarf2_frame): ...new pass_data instance and...
6578 (make_pass_dwarf2_frame): ...new function.
6579 * except.c (pass_set_nothrow_function_flags): Convert from a global
6580 struct to a subclass of rtl_opt_pass along with...
6581 (pass_data_set_nothrow_function_flags): ...new pass_data instance
6583 (make_pass_set_nothrow_function_flags): ...new function.
6584 (pass_convert_to_eh_region_ranges): Convert from a global struct to a
6585 subclass of rtl_opt_pass along with...
6586 (pass_data_convert_to_eh_region_ranges): ...new pass_data instance
6588 (make_pass_convert_to_eh_region_ranges): ...new function.
6589 * final.c (pass_compute_alignments): Convert from a global struct to a
6590 subclass of rtl_opt_pass along with...
6591 (pass_data_compute_alignments): ...new pass_data instance and...
6592 (make_pass_compute_alignments): ...new function.
6593 (pass_final): Convert from a global struct to a subclass of
6594 rtl_opt_pass along with...
6595 (pass_data_final): ...new pass_data instance and...
6596 (make_pass_final): ...new function.
6597 (pass_shorten_branches): Convert from a global struct to a subclass of
6598 rtl_opt_pass along with...
6599 (pass_data_shorten_branches): ...new pass_data instance and...
6600 (make_pass_shorten_branches): ...new function.
6601 (pass_clean_state): Convert from a global struct to a subclass of
6602 rtl_opt_pass along with...
6603 (pass_data_clean_state): ...new pass_data instance and...
6604 (make_pass_clean_state): ...new function.
6605 * function.c (pass_instantiate_virtual_regs): Convert from a global
6606 struct to a subclass of rtl_opt_pass along with...
6607 (pass_data_instantiate_virtual_regs): ...new pass_data instance and...
6608 (make_pass_instantiate_virtual_regs): ...new function.
6609 (pass_leaf_regs): Convert from a global struct to a subclass of
6610 rtl_opt_pass along with...
6611 (pass_data_leaf_regs): ...new pass_data instance and...
6612 (make_pass_leaf_regs): ...new function.
6613 (pass_thread_prologue_and_epilogue): Convert from a global struct to a
6614 subclass of rtl_opt_pass along with...
6615 (pass_data_thread_prologue_and_epilogue): ...new pass_data instance
6617 (make_pass_thread_prologue_and_epilogue): ...new function.
6618 (pass_match_asm_constraints): Convert from a global struct to a
6619 subclass of rtl_opt_pass along with...
6620 (pass_data_match_asm_constraints): ...new pass_data instance and...
6621 (make_pass_match_asm_constraints): ...new function.
6622 * fwprop.c (pass_rtl_fwprop): Convert from a global struct to a
6623 subclass of rtl_opt_pass along with...
6624 (pass_data_rtl_fwprop): ...new pass_data instance and...
6625 (make_pass_rtl_fwprop): ...new function.
6626 (pass_rtl_fwprop_addr): Convert from a global struct to a subclass of
6627 rtl_opt_pass along with...
6628 (pass_data_rtl_fwprop_addr): ...new pass_data instance and...
6629 (make_pass_rtl_fwprop_addr): ...new function.
6630 * gcse.c (pass_rtl_pre): Convert from a global struct to a subclass of
6631 rtl_opt_pass along with...
6632 (pass_data_rtl_pre): ...new pass_data instance and...
6633 (make_pass_rtl_pre): ...new function.
6634 (pass_rtl_hoist): Convert from a global struct to a subclass of
6635 rtl_opt_pass along with...
6636 (pass_data_rtl_hoist): ...new pass_data instance and...
6637 (make_pass_rtl_hoist): ...new function.
6638 * gimple-low.c (pass_lower_cf): Convert from a global struct to a
6639 subclass of gimple_opt_pass along with...
6640 (pass_data_lower_cf): ...new pass_data instance and...
6641 (make_pass_lower_cf): ...new function.
6642 * gimple-ssa-strength-reduction.c (pass_strength_reduction): Convert
6643 from a global struct to a subclass of gimple_opt_pass along with...
6644 (pass_data_strength_reduction): ...new pass_data instance and...
6645 (make_pass_strength_reduction): ...new function.
6646 * ifcvt.c (pass_rtl_ifcvt): Convert from a global struct to a subclass
6647 of rtl_opt_pass along with...
6648 (pass_data_rtl_ifcvt): ...new pass_data instance and...
6649 (make_pass_rtl_ifcvt): ...new function.
6650 (pass_if_after_combine): Convert from a global struct to a subclass of
6651 rtl_opt_pass along with...
6652 (pass_data_if_after_combine): ...new pass_data instance and...
6653 (make_pass_if_after_combine): ...new function.
6654 (pass_if_after_reload): Convert from a global struct to a subclass of
6655 rtl_opt_pass along with...
6656 (pass_data_if_after_reload): ...new pass_data instance and...
6657 (make_pass_if_after_reload): ...new function.
6658 * init-regs.c (pass_initialize_regs): Convert from a global struct to
6659 a subclass of rtl_opt_pass along with...
6660 (pass_data_initialize_regs): ...new pass_data instance and...
6661 (make_pass_initialize_regs): ...new function.
6662 * ipa-cp.c (pass_ipa_cp): Convert from a global struct to a subclass
6663 of ipa_opt_pass_d along with...
6664 (pass_data_ipa_cp): ...new pass_data instance and...
6665 (make_pass_ipa_cp): ...new function.
6666 * ipa-inline-analysis.c (pass_inline_parameters): Convert from a
6667 global struct to a subclass of gimple_opt_pass along with...
6668 (pass_data_inline_parameters): ...new pass_data instance and...
6669 (make_pass_inline_parameters): ...new function.
6670 * ipa-inline.c (pass_early_inline): Convert from a global struct to a
6671 subclass of gimple_opt_pass along with...
6672 (pass_data_early_inline): ...new pass_data instance and...
6673 (make_pass_early_inline): ...new function.
6674 (pass_ipa_inline): Convert from a global struct to a subclass of
6675 ipa_opt_pass_d along with...
6676 (pass_data_ipa_inline): ...new pass_data instance and...
6677 (make_pass_ipa_inline): ...new function.
6678 * ipa-pure-const.c (pass_local_pure_const): Convert from a global
6679 struct to a subclass of gimple_opt_pass along with...
6680 (pass_data_local_pure_const): ...new pass_data instance and...
6681 (make_pass_local_pure_const): ...new function.
6682 (pass_ipa_pure_const): Convert from a global struct to a subclass of
6683 ipa_opt_pass_d along with...
6684 (pass_data_ipa_pure_const): ...new pass_data instance and...
6685 (make_pass_ipa_pure_const): ...new function.
6686 * ipa-reference.c (pass_ipa_reference): Convert from a global struct
6687 to a subclass of ipa_opt_pass_d along with...
6688 (pass_data_ipa_reference): ...new pass_data instance and...
6689 (make_pass_ipa_reference): ...new function.
6690 * ipa-split.c (pass_split_functions): Convert from a global struct to
6691 a subclass of gimple_opt_pass along with...
6692 (pass_data_split_functions): ...new pass_data instance and...
6693 (make_pass_split_functions): ...new function.
6694 (pass_feedback_split_functions): Convert from a global struct to a
6695 subclass of gimple_opt_pass along with...
6696 (pass_data_feedback_split_functions): ...new pass_data instance and...
6697 (make_pass_feedback_split_functions): ...new function.
6698 * ipa.c (pass_ipa_function_and_variable_visibility): Convert from a
6699 global struct to a subclass of simple_ipa_opt_pass along with...
6700 (pass_data_ipa_function_and_variable_visibility): ...new pass_data
6702 (make_pass_ipa_function_and_variable_visibility): ...new function.
6703 (pass_ipa_free_inline_summary): Convert from a global struct to a
6704 subclass of simple_ipa_opt_pass along with...
6705 (pass_data_ipa_free_inline_summary): ...new pass_data instance and...
6706 (make_pass_ipa_free_inline_summary): ...new function.
6707 (pass_ipa_whole_program_visibility): Convert from a global struct to a
6708 subclass of ipa_opt_pass_d along with...
6709 (pass_data_ipa_whole_program_visibility): ...new pass_data instance
6711 (make_pass_ipa_whole_program_visibility): ...new function.
6712 (pass_ipa_profile): Convert from a global struct to a subclass of
6713 ipa_opt_pass_d along with...
6714 (pass_data_ipa_profile): ...new pass_data instance and...
6715 (make_pass_ipa_profile): ...new function.
6716 (pass_ipa_cdtor_merge): Convert from a global struct to a subclass of
6717 ipa_opt_pass_d along with...
6718 (pass_data_ipa_cdtor_merge): ...new pass_data instance and...
6719 (make_pass_ipa_cdtor_merge): ...new function.
6720 * ira.c (pass_ira): Convert from a global struct to a subclass of
6721 rtl_opt_pass along with...
6722 (pass_data_ira): ...new pass_data instance and...
6723 (make_pass_ira): ...new function.
6724 (pass_reload): Convert from a global struct to a subclass of
6725 rtl_opt_pass along with...
6726 (pass_data_reload): ...new pass_data instance and...
6727 (make_pass_reload): ...new function.
6728 * jump.c (pass_cleanup_barriers): Convert from a global struct to a
6729 subclass of rtl_opt_pass along with...
6730 (pass_data_cleanup_barriers): ...new pass_data instance and...
6731 (make_pass_cleanup_barriers): ...new function.
6732 * loop-init.c (pass_loop2): Convert from a global struct to a subclass
6733 of rtl_opt_pass along with...
6734 (pass_data_loop2): ...new pass_data instance and...
6735 (make_pass_loop2): ...new function.
6736 (pass_rtl_loop_init): Convert from a global struct to a subclass of
6737 rtl_opt_pass along with...
6738 (pass_data_rtl_loop_init): ...new pass_data instance and...
6739 (make_pass_rtl_loop_init): ...new function.
6740 (pass_rtl_loop_done): Convert from a global struct to a subclass of
6741 rtl_opt_pass along with...
6742 (pass_data_rtl_loop_done): ...new pass_data instance and...
6743 (make_pass_rtl_loop_done): ...new function.
6744 (pass_rtl_move_loop_invariants): Convert from a global struct to a
6745 subclass of rtl_opt_pass along with...
6746 (pass_data_rtl_move_loop_invariants): ...new pass_data instance and...
6747 (make_pass_rtl_move_loop_invariants): ...new function.
6748 (pass_rtl_unswitch): Convert from a global struct to a subclass of
6749 rtl_opt_pass along with...
6750 (pass_data_rtl_unswitch): ...new pass_data instance and...
6751 (make_pass_rtl_unswitch): ...new function.
6752 (pass_rtl_unroll_and_peel_loops): Convert from a global struct to a
6753 subclass of rtl_opt_pass along with...
6754 (pass_data_rtl_unroll_and_peel_loops): ...new pass_data instance
6756 (make_pass_rtl_unroll_and_peel_loops): ...new function.
6757 (pass_rtl_doloop): Convert from a global struct to a subclass of
6758 rtl_opt_pass along with...
6759 (pass_data_rtl_doloop): ...new pass_data instance and...
6760 (make_pass_rtl_doloop): ...new function.
6761 * lower-subreg.c (pass_lower_subreg): Convert from a global struct to
6762 a subclass of rtl_opt_pass along with...
6763 (pass_data_lower_subreg): ...new pass_data instance and...
6764 (make_pass_lower_subreg): ...new function.
6765 (pass_lower_subreg2): Convert from a global struct to a subclass of
6766 rtl_opt_pass along with...
6767 (pass_data_lower_subreg2): ...new pass_data instance and...
6768 (make_pass_lower_subreg2): ...new function.
6769 * lto-streamer-out.c (pass_ipa_lto_gimple_out): Convert from a global
6770 struct to a subclass of ipa_opt_pass_d along with...
6771 (pass_data_ipa_lto_gimple_out): ...new pass_data instance and...
6772 (make_pass_ipa_lto_gimple_out): ...new function.
6773 (pass_ipa_lto_finish_out): Convert from a global struct to a subclass
6774 of ipa_opt_pass_d along with...
6775 (pass_data_ipa_lto_finish_out): ...new pass_data instance and...
6776 (make_pass_ipa_lto_finish_out): ...new function.
6777 * mode-switching.c (pass_mode_switching): Convert from a global struct
6778 to a subclass of rtl_opt_pass along with...
6779 (pass_data_mode_switching): ...new pass_data instance and...
6780 (make_pass_mode_switching): ...new function.
6781 * modulo-sched.c (pass_sms): Convert from a global struct to a
6782 subclass of rtl_opt_pass along with...
6783 (pass_data_sms): ...new pass_data instance and...
6784 (make_pass_sms): ...new function.
6785 * omp-low.c (pass_expand_omp): Convert from a global struct to a
6786 subclass of gimple_opt_pass along with...
6787 (pass_data_expand_omp): ...new pass_data instance and...
6788 (make_pass_expand_omp): ...new function.
6789 (pass_lower_omp): Convert from a global struct to a subclass of
6790 gimple_opt_pass along with...
6791 (pass_data_lower_omp): ...new pass_data instance and...
6792 (make_pass_lower_omp): ...new function.
6793 (pass_diagnose_omp_blocks): Convert from a global struct to a subclass
6794 of gimple_opt_pass along with...
6795 (pass_data_diagnose_omp_blocks): ...new pass_data instance and...
6796 (make_pass_diagnose_omp_blocks): ...new function.
6797 * passes.c (pass_early_local_passes): Convert from a global struct to
6798 a subclass of simple_ipa_opt_pass along with...
6799 (pass_data_early_local_passes): ...new pass_data instance and...
6800 (make_pass_early_local_passes): ...new function.
6801 (pass_all_early_optimizations): Convert from a global struct to a
6802 subclass of gimple_opt_pass along with...
6803 (pass_data_all_early_optimizations): ...new pass_data instance and...
6804 (make_pass_all_early_optimizations): ...new function.
6805 (pass_all_optimizations): Convert from a global struct to a subclass
6806 of gimple_opt_pass along with...
6807 (pass_data_all_optimizations): ...new pass_data instance and...
6808 (make_pass_all_optimizations): ...new function.
6809 (pass_all_optimizations_g): Convert from a global struct to a subclass
6810 of gimple_opt_pass along with...
6811 (pass_data_all_optimizations_g): ...new pass_data instance and...
6812 (make_pass_all_optimizations_g): ...new function.
6813 (pass_rest_of_compilation): Convert from a global struct to a subclass
6814 of rtl_opt_pass along with...
6815 (pass_data_rest_of_compilation): ...new pass_data instance and...
6816 (make_pass_rest_of_compilation): ...new function.
6817 (pass_postreload): Convert from a global struct to a subclass of
6818 rtl_opt_pass along with...
6819 (pass_data_postreload): ...new pass_data instance and...
6820 (make_pass_postreload): ...new function.
6821 * postreload-gcse.c (pass_gcse2): Convert from a global struct to a
6822 subclass of rtl_opt_pass along with...
6823 (pass_data_gcse2): ...new pass_data instance and...
6824 (make_pass_gcse2): ...new function.
6825 * postreload.c (pass_postreload_cse): Convert from a global struct to
6826 a subclass of rtl_opt_pass along with...
6827 (pass_data_postreload_cse): ...new pass_data instance and...
6828 (make_pass_postreload_cse): ...new function.
6829 * predict.c (pass_profile): Convert from a global struct to a subclass
6830 of gimple_opt_pass along with...
6831 (pass_data_profile): ...new pass_data instance and...
6832 (make_pass_profile): ...new function.
6833 (pass_strip_predict_hints): Convert from a global struct to a subclass
6834 of gimple_opt_pass along with...
6835 (pass_data_strip_predict_hints): ...new pass_data instance and...
6836 (make_pass_strip_predict_hints): ...new function.
6837 * recog.c (pass_peephole2): Convert from a global struct to a subclass
6838 of rtl_opt_pass along with...
6839 (pass_data_peephole2): ...new pass_data instance and...
6840 (make_pass_peephole2): ...new function.
6841 (pass_split_all_insns): Convert from a global struct to a subclass of
6842 rtl_opt_pass along with...
6843 (pass_data_split_all_insns): ...new pass_data instance and...
6844 (make_pass_split_all_insns): ...new function.
6845 (pass_split_after_reload): Convert from a global struct to a subclass
6846 of rtl_opt_pass along with...
6847 (pass_data_split_after_reload): ...new pass_data instance and...
6848 (make_pass_split_after_reload): ...new function.
6849 (pass_split_before_regstack): Convert from a global struct to a
6850 subclass of rtl_opt_pass along with...
6851 (pass_data_split_before_regstack): ...new pass_data instance and...
6852 (make_pass_split_before_regstack): ...new function.
6853 (pass_split_before_sched2): Convert from a global struct to a subclass
6854 of rtl_opt_pass along with...
6855 (pass_data_split_before_sched2): ...new pass_data instance and...
6856 (make_pass_split_before_sched2): ...new function.
6857 (pass_split_for_shorten_branches): Convert from a global struct to a
6858 subclass of rtl_opt_pass along with...
6859 (pass_data_split_for_shorten_branches): ...new pass_data instance
6861 (make_pass_split_for_shorten_branches): ...new function.
6862 * ree.c (pass_ree): Convert from a global struct to a subclass of
6863 rtl_opt_pass along with...
6864 (pass_data_ree): ...new pass_data instance and...
6865 (make_pass_ree): ...new function.
6866 * reg-stack.c (pass_stack_regs): Convert from a global struct to a
6867 subclass of rtl_opt_pass along with...
6868 (pass_data_stack_regs): ...new pass_data instance and...
6869 (make_pass_stack_regs): ...new function.
6870 (pass_stack_regs_run): Convert from a global struct to a subclass of
6871 rtl_opt_pass along with...
6872 (pass_data_stack_regs_run): ...new pass_data instance and...
6873 (make_pass_stack_regs_run): ...new function.
6874 * regcprop.c (pass_cprop_hardreg): Convert from a global struct to a
6875 subclass of rtl_opt_pass along with...
6876 (pass_data_cprop_hardreg): ...new pass_data instance and...
6877 (make_pass_cprop_hardreg): ...new function.
6878 * reginfo.c (pass_reginfo_init): Convert from a global struct to a
6879 subclass of rtl_opt_pass along with...
6880 (pass_data_reginfo_init): ...new pass_data instance and...
6881 (make_pass_reginfo_init): ...new function.
6882 * regmove.c (pass_regmove): Convert from a global struct to a subclass
6883 of rtl_opt_pass along with...
6884 (pass_data_regmove): ...new pass_data instance and...
6885 (make_pass_regmove): ...new function.
6886 * regrename.c (pass_regrename): Convert from a global struct to a
6887 subclass of rtl_opt_pass along with...
6888 (pass_data_regrename): ...new pass_data instance and...
6889 (make_pass_regrename): ...new function.
6890 * reorg.c (pass_delay_slots): Convert from a global struct to a
6891 subclass of rtl_opt_pass along with...
6892 (pass_data_delay_slots): ...new pass_data instance and...
6893 (make_pass_delay_slots): ...new function.
6894 (pass_machine_reorg): Convert from a global struct to a subclass of
6895 rtl_opt_pass along with...
6896 (pass_data_machine_reorg): ...new pass_data instance and...
6897 (make_pass_machine_reorg): ...new function.
6898 * sched-rgn.c (pass_sched): Convert from a global struct to a subclass
6899 of rtl_opt_pass along with...
6900 (pass_data_sched): ...new pass_data instance and...
6901 (make_pass_sched): ...new function.
6902 (pass_sched2): Convert from a global struct to a subclass of
6903 rtl_opt_pass along with...
6904 (pass_data_sched2): ...new pass_data instance and...
6905 (make_pass_sched2): ...new function.
6906 * stack-ptr-mod.c (pass_stack_ptr_mod): Convert from a global struct
6907 to a subclass of rtl_opt_pass along with...
6908 (pass_data_stack_ptr_mod): ...new pass_data instance and...
6909 (make_pass_stack_ptr_mod): ...new function.
6910 * store-motion.c (pass_rtl_store_motion): Convert from a global struct
6911 to a subclass of rtl_opt_pass along with...
6912 (pass_data_rtl_store_motion): ...new pass_data instance and...
6913 (make_pass_rtl_store_motion): ...new function.
6914 * tracer.c (pass_tracer): Convert from a global struct to a subclass
6915 of gimple_opt_pass along with...
6916 (pass_data_tracer): ...new pass_data instance and...
6917 (make_pass_tracer): ...new function.
6918 * trans-mem.c (pass_diagnose_tm_blocks): Convert from a global struct
6919 to a subclass of gimple_opt_pass along with...
6920 (pass_data_diagnose_tm_blocks): ...new pass_data instance and...
6921 (make_pass_diagnose_tm_blocks): ...new function.
6922 (pass_lower_tm): Convert from a global struct to a subclass of
6923 gimple_opt_pass along with...
6924 (pass_data_lower_tm): ...new pass_data instance and...
6925 (make_pass_lower_tm): ...new function.
6926 (pass_tm_init): Convert from a global struct to a subclass of
6927 gimple_opt_pass along with...
6928 (pass_data_tm_init): ...new pass_data instance and...
6929 (make_pass_tm_init): ...new function.
6930 (pass_tm_mark): Convert from a global struct to a subclass of
6931 gimple_opt_pass along with...
6932 (pass_data_tm_mark): ...new pass_data instance and...
6933 (make_pass_tm_mark): ...new function.
6934 (pass_tm_edges): Convert from a global struct to a subclass of
6935 gimple_opt_pass along with...
6936 (pass_data_tm_edges): ...new pass_data instance and...
6937 (make_pass_tm_edges): ...new function.
6938 (pass_tm_memopt): Convert from a global struct to a subclass of
6939 gimple_opt_pass along with...
6940 (pass_data_tm_memopt): ...new pass_data instance and...
6941 (make_pass_tm_memopt): ...new function.
6942 (pass_ipa_tm): Convert from a global struct to a subclass of
6943 simple_ipa_opt_pass along with...
6944 (pass_data_ipa_tm): ...new pass_data instance and...
6945 (make_pass_ipa_tm): ...new function.
6946 * tree-call-cdce.c (pass_call_cdce): Convert from a global struct to a
6947 subclass of gimple_opt_pass along with...
6948 (pass_data_call_cdce): ...new pass_data instance and...
6949 (make_pass_call_cdce): ...new function.
6950 * tree-cfg.c (pass_build_cfg): Convert from a global struct to a
6951 subclass of gimple_opt_pass along with...
6952 (pass_data_build_cfg): ...new pass_data instance and...
6953 (make_pass_build_cfg): ...new function.
6954 (pass_split_crit_edges): Convert from a global struct to a subclass of
6955 gimple_opt_pass along with...
6956 (pass_data_split_crit_edges): ...new pass_data instance and...
6957 (make_pass_split_crit_edges): ...new function.
6958 (pass_warn_function_return): Convert from a global struct to a
6959 subclass of gimple_opt_pass along with...
6960 (pass_data_warn_function_return): ...new pass_data instance and...
6961 (make_pass_warn_function_return): ...new function.
6962 (pass_warn_function_noreturn): Convert from a global struct to a
6963 subclass of gimple_opt_pass along with...
6964 (pass_data_warn_function_noreturn): ...new pass_data instance and...
6965 (make_pass_warn_function_noreturn): ...new function.
6966 (pass_warn_unused_result): Convert from a global struct to a subclass
6967 of gimple_opt_pass along with...
6968 (pass_data_warn_unused_result): ...new pass_data instance and...
6969 (make_pass_warn_unused_result): ...new function.
6970 * tree-cfgcleanup.c (pass_merge_phi): Convert from a global struct to
6971 a subclass of gimple_opt_pass along with...
6972 (pass_data_merge_phi): ...new pass_data instance and...
6973 (make_pass_merge_phi): ...new function.
6974 * tree-complex.c (pass_lower_complex): Convert from a global struct to
6975 a subclass of gimple_opt_pass along with...
6976 (pass_data_lower_complex): ...new pass_data instance and...
6977 (make_pass_lower_complex): ...new function.
6978 (pass_lower_complex_O0): Convert from a global struct to a subclass of
6979 gimple_opt_pass along with...
6980 (pass_data_lower_complex_O0): ...new pass_data instance and...
6981 (make_pass_lower_complex_O0): ...new function.
6982 * tree-eh.c (pass_lower_eh): Convert from a global struct to a
6983 subclass of gimple_opt_pass along with...
6984 (pass_data_lower_eh): ...new pass_data instance and...
6985 (make_pass_lower_eh): ...new function.
6986 (pass_refactor_eh): Convert from a global struct to a subclass of
6987 gimple_opt_pass along with...
6988 (pass_data_refactor_eh): ...new pass_data instance and...
6989 (make_pass_refactor_eh): ...new function.
6990 (pass_lower_resx): Convert from a global struct to a subclass of
6991 gimple_opt_pass along with...
6992 (pass_data_lower_resx): ...new pass_data instance and...
6993 (make_pass_lower_resx): ...new function.
6994 (pass_lower_eh_dispatch): Convert from a global struct to a subclass
6995 of gimple_opt_pass along with...
6996 (pass_data_lower_eh_dispatch): ...new pass_data instance and...
6997 (make_pass_lower_eh_dispatch): ...new function.
6998 (pass_cleanup_eh): Convert from a global struct to a subclass of
6999 gimple_opt_pass along with...
7000 (pass_data_cleanup_eh): ...new pass_data instance and...
7001 (make_pass_cleanup_eh): ...new function.
7002 * tree-emutls.c (pass_ipa_lower_emutls): Convert from a global struct
7003 to a subclass of simple_ipa_opt_pass along with...
7004 (pass_data_ipa_lower_emutls): ...new pass_data instance and...
7005 (make_pass_ipa_lower_emutls): ...new function.
7006 * tree-if-conv.c (pass_if_conversion): Convert from a global struct to
7007 a subclass of gimple_opt_pass along with...
7008 (pass_data_if_conversion): ...new pass_data instance and...
7009 (make_pass_if_conversion): ...new function.
7010 * tree-into-ssa.c (pass_build_ssa): Convert from a global struct to a
7011 subclass of gimple_opt_pass along with...
7012 (pass_data_build_ssa): ...new pass_data instance and...
7013 (make_pass_build_ssa): ...new function.
7014 * tree-loop-distribution.c (pass_loop_distribution): Convert from a
7015 global struct to a subclass of gimple_opt_pass along with...
7016 (pass_data_loop_distribution): ...new pass_data instance and...
7017 (make_pass_loop_distribution): ...new function.
7018 * tree-mudflap.c (pass_mudflap_1): Convert from a global struct to a
7019 subclass of gimple_opt_pass along with...
7020 (pass_data_mudflap_1): ...new pass_data instance and...
7021 (make_pass_mudflap_1): ...new function.
7022 (pass_mudflap_2): Convert from a global struct to a subclass of
7023 gimple_opt_pass along with...
7024 (pass_data_mudflap_2): ...new pass_data instance and...
7025 (make_pass_mudflap_2): ...new function.
7026 * tree-nomudflap.c (pass_mudflap_1): Convert from a global struct to a
7027 subclass of gimple_opt_pass along with...
7028 (pass_data_mudflap_1): ...new pass_data instance and...
7029 (make_pass_mudflap_1): ...new function.
7030 (pass_mudflap_2): Convert from a global struct to a subclass of
7031 gimple_opt_pass along with...
7032 (pass_data_mudflap_2): ...new pass_data instance and...
7033 (make_pass_mudflap_2): ...new function.
7034 * tree-nrv.c (pass_nrv): Convert from a global struct to a subclass of
7035 gimple_opt_pass along with...
7036 (pass_data_nrv): ...new pass_data instance and...
7037 (make_pass_nrv): ...new function.
7038 (pass_return_slot): Convert from a global struct to a subclass of
7039 gimple_opt_pass along with...
7040 (pass_data_return_slot): ...new pass_data instance and...
7041 (make_pass_return_slot): ...new function.
7042 * tree-object-size.c (pass_object_sizes): Convert from a global struct
7043 to a subclass of gimple_opt_pass along with...
7044 (pass_data_object_sizes): ...new pass_data instance and...
7045 (make_pass_object_sizes): ...new function.
7046 * tree-optimize.c (pass_cleanup_cfg_post_optimizing): Convert from a
7047 global struct to a subclass of gimple_opt_pass along with...
7048 (pass_data_cleanup_cfg_post_optimizing): ...new pass_data instance
7050 (make_pass_cleanup_cfg_post_optimizing): ...new function.
7051 (pass_fixup_cfg): Convert from a global struct to a subclass of
7052 gimple_opt_pass along with...
7053 (pass_data_fixup_cfg): ...new pass_data instance and...
7054 (make_pass_fixup_cfg): ...new function.
7055 * tree-pass.h (pass_mudflap_1): Replace declaration with that of...
7056 (make_pass_mudflap_1): ...new function.
7057 (pass_mudflap_2): Replace declaration with that of...
7058 (make_pass_mudflap_2): ...new function.
7059 (pass_asan): Replace declaration with that of...
7060 (make_pass_asan): ...new function.
7061 (pass_asan_O0): Replace declaration with that of...
7062 (make_pass_asan_O0): ...new function.
7063 (pass_tsan): Replace declaration with that of...
7064 (make_pass_tsan): ...new function.
7065 (pass_tsan_O0): Replace declaration with that of...
7066 (make_pass_tsan_O0): ...new function.
7067 (pass_lower_cf): Replace declaration with that of...
7068 (make_pass_lower_cf): ...new function.
7069 (pass_refactor_eh): Replace declaration with that of...
7070 (make_pass_refactor_eh): ...new function.
7071 (pass_lower_eh): Replace declaration with that of...
7072 (make_pass_lower_eh): ...new function.
7073 (pass_lower_eh_dispatch): Replace declaration with that of...
7074 (make_pass_lower_eh_dispatch): ...new function.
7075 (pass_lower_resx): Replace declaration with that of...
7076 (make_pass_lower_resx): ...new function.
7077 (pass_build_cfg): Replace declaration with that of...
7078 (make_pass_build_cfg): ...new function.
7079 (pass_early_tree_profile): Replace declaration with that of...
7080 (make_pass_early_tree_profile): ...new function.
7081 (pass_cleanup_eh): Replace declaration with that of...
7082 (make_pass_cleanup_eh): ...new function.
7083 (pass_sra): Replace declaration with that of...
7084 (make_pass_sra): ...new function.
7085 (pass_sra_early): Replace declaration with that of...
7086 (make_pass_sra_early): ...new function.
7087 (pass_early_ipa_sra): Replace declaration with that of...
7088 (make_pass_early_ipa_sra): ...new function.
7089 (pass_tail_recursion): Replace declaration with that of...
7090 (make_pass_tail_recursion): ...new function.
7091 (pass_tail_calls): Replace declaration with that of...
7092 (make_pass_tail_calls): ...new function.
7093 (pass_tree_loop): Replace declaration with that of...
7094 (make_pass_tree_loop): ...new function.
7095 (pass_tree_loop_init): Replace declaration with that of...
7096 (make_pass_tree_loop_init): ...new function.
7097 (pass_lim): Replace declaration with that of...
7098 (make_pass_lim): ...new function.
7099 (pass_tree_unswitch): Replace declaration with that of...
7100 (make_pass_tree_unswitch): ...new function.
7101 (pass_predcom): Replace declaration with that of...
7102 (make_pass_predcom): ...new function.
7103 (pass_iv_canon): Replace declaration with that of...
7104 (make_pass_iv_canon): ...new function.
7105 (pass_scev_cprop): Replace declaration with that of...
7106 (make_pass_scev_cprop): ...new function.
7107 (pass_empty_loop): Replace declaration with that of...
7108 (make_pass_empty_loop): ...new function.
7109 (pass_record_bounds): Replace declaration with that of...
7110 (make_pass_record_bounds): ...new function.
7111 (pass_graphite): Replace declaration with that of...
7112 (make_pass_graphite): ...new function.
7113 (pass_graphite_transforms): Replace declaration with that of...
7114 (make_pass_graphite_transforms): ...new function.
7115 (pass_if_conversion): Replace declaration with that of...
7116 (make_pass_if_conversion): ...new function.
7117 (pass_loop_distribution): Replace declaration with that of...
7118 (make_pass_loop_distribution): ...new function.
7119 (pass_vectorize): Replace declaration with that of...
7120 (make_pass_vectorize): ...new function.
7121 (pass_slp_vectorize): Replace declaration with that of...
7122 (make_pass_slp_vectorize): ...new function.
7123 (pass_complete_unroll): Replace declaration with that of...
7124 (make_pass_complete_unroll): ...new function.
7125 (pass_complete_unrolli): Replace declaration with that of...
7126 (make_pass_complete_unrolli): ...new function.
7127 (pass_parallelize_loops): Replace declaration with that of...
7128 (make_pass_parallelize_loops): ...new function.
7129 (pass_loop_prefetch): Replace declaration with that of...
7130 (make_pass_loop_prefetch): ...new function.
7131 (pass_iv_optimize): Replace declaration with that of...
7132 (make_pass_iv_optimize): ...new function.
7133 (pass_tree_loop_done): Replace declaration with that of...
7134 (make_pass_tree_loop_done): ...new function.
7135 (pass_ch): Replace declaration with that of...
7136 (make_pass_ch): ...new function.
7137 (pass_ccp): Replace declaration with that of...
7138 (make_pass_ccp): ...new function.
7139 (pass_phi_only_cprop): Replace declaration with that of...
7140 (make_pass_phi_only_cprop): ...new function.
7141 (pass_build_ssa): Replace declaration with that of...
7142 (make_pass_build_ssa): ...new function.
7143 (pass_build_alias): Replace declaration with that of...
7144 (make_pass_build_alias): ...new function.
7145 (pass_build_ealias): Replace declaration with that of...
7146 (make_pass_build_ealias): ...new function.
7147 (pass_dominator): Replace declaration with that of...
7148 (make_pass_dominator): ...new function.
7149 (pass_dce): Replace declaration with that of...
7150 (make_pass_dce): ...new function.
7151 (pass_dce_loop): Replace declaration with that of...
7152 (make_pass_dce_loop): ...new function.
7153 (pass_cd_dce): Replace declaration with that of...
7154 (make_pass_cd_dce): ...new function.
7155 (pass_call_cdce): Replace declaration with that of...
7156 (make_pass_call_cdce): ...new function.
7157 (pass_merge_phi): Replace declaration with that of...
7158 (make_pass_merge_phi): ...new function.
7159 (pass_split_crit_edges): Replace declaration with that of...
7160 (make_pass_split_crit_edges): ...new function.
7161 (pass_pre): Replace declaration with that of...
7162 (make_pass_pre): ...new function.
7163 (pass_profile): Replace declaration with that of...
7164 (make_pass_profile): ...new function.
7165 (pass_strip_predict_hints): Replace declaration with that of...
7166 (make_pass_strip_predict_hints): ...new function.
7167 (pass_lower_complex_O0): Replace declaration with that of...
7168 (make_pass_lower_complex_O0): ...new function.
7169 (pass_lower_complex): Replace declaration with that of...
7170 (make_pass_lower_complex): ...new function.
7171 (pass_lower_vector): Replace declaration with that of...
7172 (make_pass_lower_vector): ...new function.
7173 (pass_lower_vector_ssa): Replace declaration with that of...
7174 (make_pass_lower_vector_ssa): ...new function.
7175 (pass_lower_omp): Replace declaration with that of...
7176 (make_pass_lower_omp): ...new function.
7177 (pass_diagnose_omp_blocks): Replace declaration with that of...
7178 (make_pass_diagnose_omp_blocks): ...new function.
7179 (pass_expand_omp): Replace declaration with that of...
7180 (make_pass_expand_omp): ...new function.
7181 (pass_expand_omp_ssa): Replace declaration with that of...
7182 (make_pass_expand_omp_ssa): ...new function.
7183 (pass_object_sizes): Replace declaration with that of...
7184 (make_pass_object_sizes): ...new function.
7185 (pass_strlen): Replace declaration with that of...
7186 (make_pass_strlen): ...new function.
7187 (pass_fold_builtins): Replace declaration with that of...
7188 (make_pass_fold_builtins): ...new function.
7189 (pass_stdarg): Replace declaration with that of...
7190 (make_pass_stdarg): ...new function.
7191 (pass_early_warn_uninitialized): Replace declaration with that of...
7192 (make_pass_early_warn_uninitialized): ...new function.
7193 (pass_late_warn_uninitialized): Replace declaration with that of...
7194 (make_pass_late_warn_uninitialized): ...new function.
7195 (pass_cse_reciprocals): Replace declaration with that of...
7196 (make_pass_cse_reciprocals): ...new function.
7197 (pass_cse_sincos): Replace declaration with that of...
7198 (make_pass_cse_sincos): ...new function.
7199 (pass_optimize_bswap): Replace declaration with that of...
7200 (make_pass_optimize_bswap): ...new function.
7201 (pass_optimize_widening_mul): Replace declaration with that of...
7202 (make_pass_optimize_widening_mul): ...new function.
7203 (pass_warn_function_return): Replace declaration with that of...
7204 (make_pass_warn_function_return): ...new function.
7205 (pass_warn_function_noreturn): Replace declaration with that of...
7206 (make_pass_warn_function_noreturn): ...new function.
7207 (pass_cselim): Replace declaration with that of...
7208 (make_pass_cselim): ...new function.
7209 (pass_phiopt): Replace declaration with that of...
7210 (make_pass_phiopt): ...new function.
7211 (pass_forwprop): Replace declaration with that of...
7212 (make_pass_forwprop): ...new function.
7213 (pass_phiprop): Replace declaration with that of...
7214 (make_pass_phiprop): ...new function.
7215 (pass_tree_ifcombine): Replace declaration with that of...
7216 (make_pass_tree_ifcombine): ...new function.
7217 (pass_dse): Replace declaration with that of...
7218 (make_pass_dse): ...new function.
7219 (pass_nrv): Replace declaration with that of...
7220 (make_pass_nrv): ...new function.
7221 (pass_rename_ssa_copies): Replace declaration with that of...
7222 (make_pass_rename_ssa_copies): ...new function.
7223 (pass_sink_code): Replace declaration with that of...
7224 (make_pass_sink_code): ...new function.
7225 (pass_fre): Replace declaration with that of...
7226 (make_pass_fre): ...new function.
7227 (pass_check_data_deps): Replace declaration with that of...
7228 (make_pass_check_data_deps): ...new function.
7229 (pass_copy_prop): Replace declaration with that of...
7230 (make_pass_copy_prop): ...new function.
7231 (pass_vrp): Replace declaration with that of...
7232 (make_pass_vrp): ...new function.
7233 (pass_uncprop): Replace declaration with that of...
7234 (make_pass_uncprop): ...new function.
7235 (pass_return_slot): Replace declaration with that of...
7236 (make_pass_return_slot): ...new function.
7237 (pass_reassoc): Replace declaration with that of...
7238 (make_pass_reassoc): ...new function.
7239 (pass_rebuild_cgraph_edges): Replace declaration with that of...
7240 (make_pass_rebuild_cgraph_edges): ...new function.
7241 (pass_remove_cgraph_callee_edges): Replace declaration with that of...
7242 (make_pass_remove_cgraph_callee_edges): ...new function.
7243 (pass_build_cgraph_edges): Replace declaration with that of...
7244 (make_pass_build_cgraph_edges): ...new function.
7245 (pass_local_pure_const): Replace declaration with that of...
7246 (make_pass_local_pure_const): ...new function.
7247 (pass_tracer): Replace declaration with that of...
7248 (make_pass_tracer): ...new function.
7249 (pass_warn_unused_result): Replace declaration with that of...
7250 (make_pass_warn_unused_result): ...new function.
7251 (pass_diagnose_tm_blocks): Replace declaration with that of...
7252 (make_pass_diagnose_tm_blocks): ...new function.
7253 (pass_lower_tm): Replace declaration with that of...
7254 (make_pass_lower_tm): ...new function.
7255 (pass_tm_init): Replace declaration with that of...
7256 (make_pass_tm_init): ...new function.
7257 (pass_tm_mark): Replace declaration with that of...
7258 (make_pass_tm_mark): ...new function.
7259 (pass_tm_memopt): Replace declaration with that of...
7260 (make_pass_tm_memopt): ...new function.
7261 (pass_tm_edges): Replace declaration with that of...
7262 (make_pass_tm_edges): ...new function.
7263 (pass_split_functions): Replace declaration with that of...
7264 (make_pass_split_functions): ...new function.
7265 (pass_feedback_split_functions): Replace declaration with that of...
7266 (make_pass_feedback_split_functions): ...new function.
7267 (pass_strength_reduction): Replace declaration with that of...
7268 (make_pass_strength_reduction): ...new function.
7269 (pass_ipa_lower_emutls): Replace declaration with that of...
7270 (make_pass_ipa_lower_emutls): ...new function.
7271 (pass_ipa_function_and_variable_visibility): Replace declaration with
7273 (make_pass_ipa_function_and_variable_visibility): ...new function.
7274 (pass_ipa_tree_profile): Replace declaration with that of...
7275 (make_pass_ipa_tree_profile): ...new function.
7276 (pass_early_local_passes): Replace declaration with that of...
7277 (make_pass_early_local_passes): ...new function.
7278 (pass_ipa_whole_program_visibility): Replace declaration with that
7280 (make_pass_ipa_whole_program_visibility): ...new function.
7281 (pass_ipa_lto_gimple_out): Replace declaration with that of...
7282 (make_pass_ipa_lto_gimple_out): ...new function.
7283 (pass_ipa_increase_alignment): Replace declaration with that of...
7284 (make_pass_ipa_increase_alignment): ...new function.
7285 (pass_ipa_inline): Replace declaration with that of...
7286 (make_pass_ipa_inline): ...new function.
7287 (pass_ipa_free_lang_data): Replace declaration with that of...
7288 (make_pass_ipa_free_lang_data): ...new function.
7289 (pass_ipa_free_inline_summary): Replace declaration with that of...
7290 (make_pass_ipa_free_inline_summary): ...new function.
7291 (pass_ipa_cp): Replace declaration with that of...
7292 (make_pass_ipa_cp): ...new function.
7293 (pass_ipa_reference): Replace declaration with that of...
7294 (make_pass_ipa_reference): ...new function.
7295 (pass_ipa_pure_const): Replace declaration with that of...
7296 (make_pass_ipa_pure_const): ...new function.
7297 (pass_ipa_pta): Replace declaration with that of...
7298 (make_pass_ipa_pta): ...new function.
7299 (pass_ipa_lto_finish_out): Replace declaration with that of...
7300 (make_pass_ipa_lto_finish_out): ...new function.
7301 (pass_ipa_tm): Replace declaration with that of...
7302 (make_pass_ipa_tm): ...new function.
7303 (pass_ipa_profile): Replace declaration with that of...
7304 (make_pass_ipa_profile): ...new function.
7305 (pass_ipa_cdtor_merge): Replace declaration with that of...
7306 (make_pass_ipa_cdtor_merge): ...new function.
7307 (pass_cleanup_cfg_post_optimizing): Replace declaration with that
7309 (make_pass_cleanup_cfg_post_optimizing): ...new function.
7310 (pass_init_datastructures): Replace declaration with that of...
7311 (make_pass_init_datastructures): ...new function.
7312 (pass_fixup_cfg): Replace declaration with that of...
7313 (make_pass_fixup_cfg): ...new function.
7314 (pass_expand): Replace declaration with that of...
7315 (make_pass_expand): ...new function.
7316 (pass_instantiate_virtual_regs): Replace declaration with that of...
7317 (make_pass_instantiate_virtual_regs): ...new function.
7318 (pass_rtl_fwprop): Replace declaration with that of...
7319 (make_pass_rtl_fwprop): ...new function.
7320 (pass_rtl_fwprop_addr): Replace declaration with that of...
7321 (make_pass_rtl_fwprop_addr): ...new function.
7322 (pass_jump): Replace declaration with that of...
7323 (make_pass_jump): ...new function.
7324 (pass_jump2): Replace declaration with that of...
7325 (make_pass_jump2): ...new function.
7326 (pass_lower_subreg): Replace declaration with that of...
7327 (make_pass_lower_subreg): ...new function.
7328 (pass_cse): Replace declaration with that of...
7329 (make_pass_cse): ...new function.
7330 (pass_fast_rtl_dce): Replace declaration with that of...
7331 (make_pass_fast_rtl_dce): ...new function.
7332 (pass_ud_rtl_dce): Replace declaration with that of...
7333 (make_pass_ud_rtl_dce): ...new function.
7334 (pass_rtl_dce): Replace declaration with that of...
7335 (make_pass_rtl_dce): ...new function.
7336 (pass_rtl_dse1): Replace declaration with that of...
7337 (make_pass_rtl_dse1): ...new function.
7338 (pass_rtl_dse2): Replace declaration with that of...
7339 (make_pass_rtl_dse2): ...new function.
7340 (pass_rtl_dse3): Replace declaration with that of...
7341 (make_pass_rtl_dse3): ...new function.
7342 (pass_rtl_cprop): Replace declaration with that of...
7343 (make_pass_rtl_cprop): ...new function.
7344 (pass_rtl_pre): Replace declaration with that of...
7345 (make_pass_rtl_pre): ...new function.
7346 (pass_rtl_hoist): Replace declaration with that of...
7347 (make_pass_rtl_hoist): ...new function.
7348 (pass_rtl_store_motion): Replace declaration with that of...
7349 (make_pass_rtl_store_motion): ...new function.
7350 (pass_cse_after_global_opts): Replace declaration with that of...
7351 (make_pass_cse_after_global_opts): ...new function.
7352 (pass_rtl_ifcvt): Replace declaration with that of...
7353 (make_pass_rtl_ifcvt): ...new function.
7354 (pass_into_cfg_layout_mode): Replace declaration with that of...
7355 (make_pass_into_cfg_layout_mode): ...new function.
7356 (pass_outof_cfg_layout_mode): Replace declaration with that of...
7357 (make_pass_outof_cfg_layout_mode): ...new function.
7358 (pass_loop2): Replace declaration with that of...
7359 (make_pass_loop2): ...new function.
7360 (pass_rtl_loop_init): Replace declaration with that of...
7361 (make_pass_rtl_loop_init): ...new function.
7362 (pass_rtl_move_loop_invariants): Replace declaration with that of...
7363 (make_pass_rtl_move_loop_invariants): ...new function.
7364 (pass_rtl_unswitch): Replace declaration with that of...
7365 (make_pass_rtl_unswitch): ...new function.
7366 (pass_rtl_unroll_and_peel_loops): Replace declaration with that of...
7367 (make_pass_rtl_unroll_and_peel_loops): ...new function.
7368 (pass_rtl_doloop): Replace declaration with that of...
7369 (make_pass_rtl_doloop): ...new function.
7370 (pass_rtl_loop_done): Replace declaration with that of...
7371 (make_pass_rtl_loop_done): ...new function.
7372 (pass_web): Replace declaration with that of...
7373 (make_pass_web): ...new function.
7374 (pass_cse2): Replace declaration with that of...
7375 (make_pass_cse2): ...new function.
7376 (pass_df_initialize_opt): Replace declaration with that of...
7377 (make_pass_df_initialize_opt): ...new function.
7378 (pass_df_initialize_no_opt): Replace declaration with that of...
7379 (make_pass_df_initialize_no_opt): ...new function.
7380 (pass_reginfo_init): Replace declaration with that of...
7381 (make_pass_reginfo_init): ...new function.
7382 (pass_inc_dec): Replace declaration with that of...
7383 (make_pass_inc_dec): ...new function.
7384 (pass_stack_ptr_mod): Replace declaration with that of...
7385 (make_pass_stack_ptr_mod): ...new function.
7386 (pass_initialize_regs): Replace declaration with that of...
7387 (make_pass_initialize_regs): ...new function.
7388 (pass_combine): Replace declaration with that of...
7389 (make_pass_combine): ...new function.
7390 (pass_if_after_combine): Replace declaration with that of...
7391 (make_pass_if_after_combine): ...new function.
7392 (pass_ree): Replace declaration with that of...
7393 (make_pass_ree): ...new function.
7394 (pass_partition_blocks): Replace declaration with that of...
7395 (make_pass_partition_blocks): ...new function.
7396 (pass_match_asm_constraints): Replace declaration with that of...
7397 (make_pass_match_asm_constraints): ...new function.
7398 (pass_regmove): Replace declaration with that of...
7399 (make_pass_regmove): ...new function.
7400 (pass_split_all_insns): Replace declaration with that of...
7401 (make_pass_split_all_insns): ...new function.
7402 (pass_fast_rtl_byte_dce): Replace declaration with that of...
7403 (make_pass_fast_rtl_byte_dce): ...new function.
7404 (pass_lower_subreg2): Replace declaration with that of...
7405 (make_pass_lower_subreg2): ...new function.
7406 (pass_mode_switching): Replace declaration with that of...
7407 (make_pass_mode_switching): ...new function.
7408 (pass_sms): Replace declaration with that of...
7409 (make_pass_sms): ...new function.
7410 (pass_sched): Replace declaration with that of...
7411 (make_pass_sched): ...new function.
7412 (pass_ira): Replace declaration with that of...
7413 (make_pass_ira): ...new function.
7414 (pass_reload): Replace declaration with that of...
7415 (make_pass_reload): ...new function.
7416 (pass_clean_state): Replace declaration with that of...
7417 (make_pass_clean_state): ...new function.
7418 (pass_branch_prob): Replace declaration with that of...
7419 (make_pass_branch_prob): ...new function.
7420 (pass_value_profile_transformations): Replace declaration with that
7422 (make_pass_value_profile_transformations): ...new function.
7423 (pass_postreload_cse): Replace declaration with that of...
7424 (make_pass_postreload_cse): ...new function.
7425 (pass_gcse2): Replace declaration with that of...
7426 (make_pass_gcse2): ...new function.
7427 (pass_split_after_reload): Replace declaration with that of...
7428 (make_pass_split_after_reload): ...new function.
7429 (pass_branch_target_load_optimize1): Replace declaration with that
7431 (make_pass_branch_target_load_optimize1): ...new function.
7432 (pass_thread_prologue_and_epilogue): Replace declaration with that
7434 (make_pass_thread_prologue_and_epilogue): ...new function.
7435 (pass_stack_adjustments): Replace declaration with that of...
7436 (make_pass_stack_adjustments): ...new function.
7437 (pass_peephole2): Replace declaration with that of...
7438 (make_pass_peephole2): ...new function.
7439 (pass_if_after_reload): Replace declaration with that of...
7440 (make_pass_if_after_reload): ...new function.
7441 (pass_regrename): Replace declaration with that of...
7442 (make_pass_regrename): ...new function.
7443 (pass_cprop_hardreg): Replace declaration with that of...
7444 (make_pass_cprop_hardreg): ...new function.
7445 (pass_reorder_blocks): Replace declaration with that of...
7446 (make_pass_reorder_blocks): ...new function.
7447 (pass_branch_target_load_optimize2): Replace declaration with that
7449 (make_pass_branch_target_load_optimize2): ...new function.
7450 (pass_leaf_regs): Replace declaration with that of...
7451 (make_pass_leaf_regs): ...new function.
7452 (pass_split_before_sched2): Replace declaration with that of...
7453 (make_pass_split_before_sched2): ...new function.
7454 (pass_compare_elim_after_reload): Replace declaration with that of...
7455 (make_pass_compare_elim_after_reload): ...new function.
7456 (pass_sched2): Replace declaration with that of...
7457 (make_pass_sched2): ...new function.
7458 (pass_stack_regs): Replace declaration with that of...
7459 (make_pass_stack_regs): ...new function.
7460 (pass_stack_regs_run): Replace declaration with that of...
7461 (make_pass_stack_regs_run): ...new function.
7462 (pass_df_finish): Replace declaration with that of...
7463 (make_pass_df_finish): ...new function.
7464 (pass_compute_alignments): Replace declaration with that of...
7465 (make_pass_compute_alignments): ...new function.
7466 (pass_duplicate_computed_gotos): Replace declaration with that of...
7467 (make_pass_duplicate_computed_gotos): ...new function.
7468 (pass_variable_tracking): Replace declaration with that of...
7469 (make_pass_variable_tracking): ...new function.
7470 (pass_free_cfg): Replace declaration with that of...
7471 (make_pass_free_cfg): ...new function.
7472 (pass_machine_reorg): Replace declaration with that of...
7473 (make_pass_machine_reorg): ...new function.
7474 (pass_cleanup_barriers): Replace declaration with that of...
7475 (make_pass_cleanup_barriers): ...new function.
7476 (pass_delay_slots): Replace declaration with that of...
7477 (make_pass_delay_slots): ...new function.
7478 (pass_split_for_shorten_branches): Replace declaration with that of...
7479 (make_pass_split_for_shorten_branches): ...new function.
7480 (pass_split_before_regstack): Replace declaration with that of...
7481 (make_pass_split_before_regstack): ...new function.
7482 (pass_convert_to_eh_region_ranges): Replace declaration with that
7484 (make_pass_convert_to_eh_region_ranges): ...new function.
7485 (pass_shorten_branches): Replace declaration with that of...
7486 (make_pass_shorten_branches): ...new function.
7487 (pass_set_nothrow_function_flags): Replace declaration with that of...
7488 (make_pass_set_nothrow_function_flags): ...new function.
7489 (pass_dwarf2_frame): Replace declaration with that of...
7490 (make_pass_dwarf2_frame): ...new function.
7491 (pass_final): Replace declaration with that of...
7492 (make_pass_final): ...new function.
7493 (pass_rtl_seqabstr): Replace declaration with that of...
7494 (make_pass_rtl_seqabstr): ...new function.
7495 (pass_release_ssa_names): Replace declaration with that of...
7496 (make_pass_release_ssa_names): ...new function.
7497 (pass_early_inline): Replace declaration with that of...
7498 (make_pass_early_inline): ...new function.
7499 (pass_inline_parameters): Replace declaration with that of...
7500 (make_pass_inline_parameters): ...new function.
7501 (pass_update_address_taken): Replace declaration with that of...
7502 (make_pass_update_address_taken): ...new function.
7503 (pass_convert_switch): Replace declaration with that of...
7504 (make_pass_convert_switch): ...new function.
7505 * tree-profile.c (pass_ipa_tree_profile): Convert from a global struct
7506 to a subclass of simple_ipa_opt_pass along with...
7507 (pass_data_ipa_tree_profile): ...new pass_data instance and...
7508 (make_pass_ipa_tree_profile): ...new function.
7509 * tree-sra.c (pass_sra_early): Convert from a global struct to a
7510 subclass of gimple_opt_pass along with...
7511 (pass_data_sra_early): ...new pass_data instance and...
7512 (make_pass_sra_early): ...new function.
7513 (pass_sra): Convert from a global struct to a subclass of
7514 gimple_opt_pass along with...
7515 (pass_data_sra): ...new pass_data instance and...
7516 (make_pass_sra): ...new function.
7517 (pass_early_ipa_sra): Convert from a global struct to a subclass of
7518 gimple_opt_pass along with...
7519 (pass_data_early_ipa_sra): ...new pass_data instance and...
7520 (make_pass_early_ipa_sra): ...new function.
7521 * tree-ssa-ccp.c (pass_ccp): Convert from a global struct to a
7522 subclass of gimple_opt_pass along with...
7523 (pass_data_ccp): ...new pass_data instance and...
7524 (make_pass_ccp): ...new function.
7525 (pass_fold_builtins): Convert from a global struct to a subclass of
7526 gimple_opt_pass along with...
7527 (pass_data_fold_builtins): ...new pass_data instance and...
7528 (make_pass_fold_builtins): ...new function.
7529 * tree-ssa-copy.c (pass_copy_prop): Convert from a global struct to a
7530 subclass of gimple_opt_pass along with...
7531 (pass_data_copy_prop): ...new pass_data instance and...
7532 (make_pass_copy_prop): ...new function.
7533 * tree-ssa-copyrename.c (pass_rename_ssa_copies): Convert from a
7534 global struct to a subclass of gimple_opt_pass along with...
7535 (pass_data_rename_ssa_copies): ...new pass_data instance and...
7536 (make_pass_rename_ssa_copies): ...new function.
7537 * tree-ssa-dce.c (pass_dce): Convert from a global struct to a
7538 subclass of gimple_opt_pass along with...
7539 (pass_data_dce): ...new pass_data instance and...
7540 (make_pass_dce): ...new function.
7541 (pass_dce_loop): Convert from a global struct to a subclass of
7542 gimple_opt_pass along with...
7543 (pass_data_dce_loop): ...new pass_data instance and...
7544 (make_pass_dce_loop): ...new function.
7545 (pass_cd_dce): Convert from a global struct to a subclass of
7546 gimple_opt_pass along with...
7547 (pass_data_cd_dce): ...new pass_data instance and...
7548 (make_pass_cd_dce): ...new function.
7549 * tree-ssa-dom.c (pass_dominator): Convert from a global struct to a
7550 subclass of gimple_opt_pass along with...
7551 (pass_data_dominator): ...new pass_data instance and...
7552 (make_pass_dominator): ...new function.
7553 (pass_phi_only_cprop): Convert from a global struct to a subclass of
7554 gimple_opt_pass along with...
7555 (pass_data_phi_only_cprop): ...new pass_data instance and...
7556 (make_pass_phi_only_cprop): ...new function.
7557 * tree-ssa-dse.c (pass_dse): Convert from a global struct to a
7558 subclass of gimple_opt_pass along with...
7559 (pass_data_dse): ...new pass_data instance and...
7560 (make_pass_dse): ...new function.
7561 * tree-ssa-forwprop.c (pass_forwprop): Convert from a global struct to
7562 a subclass of gimple_opt_pass along with...
7563 (pass_data_forwprop): ...new pass_data instance and...
7564 (make_pass_forwprop): ...new function.
7565 * tree-ssa-ifcombine.c (pass_tree_ifcombine): Convert from a global
7566 struct to a subclass of gimple_opt_pass along with...
7567 (pass_data_tree_ifcombine): ...new pass_data instance and...
7568 (make_pass_tree_ifcombine): ...new function.
7569 * tree-ssa-loop-ch.c (pass_ch): Convert from a global struct to a
7570 subclass of gimple_opt_pass along with...
7571 (pass_data_ch): ...new pass_data instance and...
7572 (make_pass_ch): ...new function.
7573 * tree-ssa-loop.c (pass_tree_loop): Convert from a global struct to a
7574 subclass of gimple_opt_pass along with...
7575 (pass_data_tree_loop): ...new pass_data instance and...
7576 (make_pass_tree_loop): ...new function.
7577 (pass_tree_loop_init): Convert from a global struct to a subclass of
7578 gimple_opt_pass along with...
7579 (pass_data_tree_loop_init): ...new pass_data instance and...
7580 (make_pass_tree_loop_init): ...new function.
7581 (pass_lim): Convert from a global struct to a subclass of
7582 gimple_opt_pass along with...
7583 (pass_data_lim): ...new pass_data instance and...
7584 (make_pass_lim): ...new function.
7585 (pass_tree_unswitch): Convert from a global struct to a subclass of
7586 gimple_opt_pass along with...
7587 (pass_data_tree_unswitch): ...new pass_data instance and...
7588 (make_pass_tree_unswitch): ...new function.
7589 (pass_predcom): Convert from a global struct to a subclass of
7590 gimple_opt_pass along with...
7591 (pass_data_predcom): ...new pass_data instance and...
7592 (make_pass_predcom): ...new function.
7593 (pass_vectorize): Convert from a global struct to a subclass of
7594 gimple_opt_pass along with...
7595 (pass_data_vectorize): ...new pass_data instance and...
7596 (make_pass_vectorize): ...new function.
7597 (pass_graphite): Convert from a global struct to a subclass of
7598 gimple_opt_pass along with...
7599 (pass_data_graphite): ...new pass_data instance and...
7600 (make_pass_graphite): ...new function.
7601 (pass_graphite_transforms): Convert from a global struct to a subclass
7602 of gimple_opt_pass along with...
7603 (pass_data_graphite_transforms): ...new pass_data instance and...
7604 (make_pass_graphite_transforms): ...new function.
7605 (pass_check_data_deps): Convert from a global struct to a subclass of
7606 gimple_opt_pass along with...
7607 (pass_data_check_data_deps): ...new pass_data instance and...
7608 (make_pass_check_data_deps): ...new function.
7609 (pass_iv_canon): Convert from a global struct to a subclass of
7610 gimple_opt_pass along with...
7611 (pass_data_iv_canon): ...new pass_data instance and...
7612 (make_pass_iv_canon): ...new function.
7613 (pass_scev_cprop): Convert from a global struct to a subclass of
7614 gimple_opt_pass along with...
7615 (pass_data_scev_cprop): ...new pass_data instance and...
7616 (make_pass_scev_cprop): ...new function.
7617 (pass_record_bounds): Convert from a global struct to a subclass of
7618 gimple_opt_pass along with...
7619 (pass_data_record_bounds): ...new pass_data instance and...
7620 (make_pass_record_bounds): ...new function.
7621 (pass_complete_unroll): Convert from a global struct to a subclass of
7622 gimple_opt_pass along with...
7623 (pass_data_complete_unroll): ...new pass_data instance and...
7624 (make_pass_complete_unroll): ...new function.
7625 (pass_complete_unrolli): Convert from a global struct to a subclass of
7626 gimple_opt_pass along with...
7627 (pass_data_complete_unrolli): ...new pass_data instance and...
7628 (make_pass_complete_unrolli): ...new function.
7629 (pass_parallelize_loops): Convert from a global struct to a subclass
7630 of gimple_opt_pass along with...
7631 (pass_data_parallelize_loops): ...new pass_data instance and...
7632 (make_pass_parallelize_loops): ...new function.
7633 (pass_loop_prefetch): Convert from a global struct to a subclass of
7634 gimple_opt_pass along with...
7635 (pass_data_loop_prefetch): ...new pass_data instance and...
7636 (make_pass_loop_prefetch): ...new function.
7637 (pass_iv_optimize): Convert from a global struct to a subclass of
7638 gimple_opt_pass along with...
7639 (pass_data_iv_optimize): ...new pass_data instance and...
7640 (make_pass_iv_optimize): ...new function.
7641 (pass_tree_loop_done): Convert from a global struct to a subclass of
7642 gimple_opt_pass along with...
7643 (pass_data_tree_loop_done): ...new pass_data instance and...
7644 (make_pass_tree_loop_done): ...new function.
7645 * tree-ssa-math-opts.c (pass_cse_reciprocals): Convert from a global
7646 struct to a subclass of gimple_opt_pass along with...
7647 (pass_data_cse_reciprocals): ...new pass_data instance and...
7648 (make_pass_cse_reciprocals): ...new function.
7649 (pass_cse_sincos): Convert from a global struct to a subclass of
7650 gimple_opt_pass along with...
7651 (pass_data_cse_sincos): ...new pass_data instance and...
7652 (make_pass_cse_sincos): ...new function.
7653 (pass_optimize_bswap): Convert from a global struct to a subclass of
7654 gimple_opt_pass along with...
7655 (pass_data_optimize_bswap): ...new pass_data instance and...
7656 (make_pass_optimize_bswap): ...new function.
7657 (pass_optimize_widening_mul): Convert from a global struct to a
7658 subclass of gimple_opt_pass along with...
7659 (pass_data_optimize_widening_mul): ...new pass_data instance and...
7660 (make_pass_optimize_widening_mul): ...new function.
7661 * tree-ssa-phiopt.c (pass_phiopt): Convert from a global struct to a
7662 subclass of gimple_opt_pass along with...
7663 (pass_data_phiopt): ...new pass_data instance and...
7664 (make_pass_phiopt): ...new function.
7665 (pass_cselim): Convert from a global struct to a subclass of
7666 gimple_opt_pass along with...
7667 (pass_data_cselim): ...new pass_data instance and...
7668 (make_pass_cselim): ...new function.
7669 * tree-ssa-phiprop.c (pass_phiprop): Convert from a global struct to a
7670 subclass of gimple_opt_pass along with...
7671 (pass_data_phiprop): ...new pass_data instance and...
7672 (make_pass_phiprop): ...new function.
7673 * tree-ssa-pre.c (pass_pre): Convert from a global struct to a
7674 subclass of gimple_opt_pass along with...
7675 (pass_data_pre): ...new pass_data instance and...
7676 (make_pass_pre): ...new function.
7677 (pass_fre): Convert from a global struct to a subclass of
7678 gimple_opt_pass along with...
7679 (pass_data_fre): ...new pass_data instance and...
7680 (make_pass_fre): ...new function.
7681 * tree-ssa-reassoc.c (pass_reassoc): Convert from a global struct to a
7682 subclass of gimple_opt_pass along with...
7683 (pass_data_reassoc): ...new pass_data instance and...
7684 (make_pass_reassoc): ...new function.
7685 * tree-ssa-sink.c (pass_sink_code): Convert from a global struct to a
7686 subclass of gimple_opt_pass along with...
7687 (pass_data_sink_code): ...new pass_data instance and...
7688 (make_pass_sink_code): ...new function.
7689 * tree-ssa-strlen.c (pass_strlen): Convert from a global struct to a
7690 subclass of gimple_opt_pass along with...
7691 (pass_data_strlen): ...new pass_data instance and...
7692 (make_pass_strlen): ...new function.
7693 * tree-ssa-structalias.c (pass_build_alias): Convert from a global
7694 struct to a subclass of gimple_opt_pass along with...
7695 (pass_data_build_alias): ...new pass_data instance and...
7696 (make_pass_build_alias): ...new function.
7697 (pass_build_ealias): Convert from a global struct to a subclass of
7698 gimple_opt_pass along with...
7699 (pass_data_build_ealias): ...new pass_data instance and...
7700 (make_pass_build_ealias): ...new function.
7701 (pass_ipa_pta): Convert from a global struct to a subclass of
7702 simple_ipa_opt_pass along with...
7703 (pass_data_ipa_pta): ...new pass_data instance and...
7704 (make_pass_ipa_pta): ...new function.
7705 * tree-ssa-uncprop.c (pass_uncprop): Convert from a global struct to a
7706 subclass of gimple_opt_pass along with...
7707 (pass_data_uncprop): ...new pass_data instance and...
7708 (make_pass_uncprop): ...new function.
7709 * tree-ssa-uninit.c (pass_late_warn_uninitialized): Convert from a
7710 global struct to a subclass of gimple_opt_pass along with...
7711 (pass_data_late_warn_uninitialized): ...new pass_data instance and...
7712 (make_pass_late_warn_uninitialized): ...new function.
7713 * tree-ssa.c (pass_init_datastructures): Convert from a global struct
7714 to a subclass of gimple_opt_pass along with...
7715 (pass_data_init_datastructures): ...new pass_data instance and...
7716 (make_pass_init_datastructures): ...new function.
7717 (pass_early_warn_uninitialized): Convert from a global struct to a
7718 subclass of gimple_opt_pass along with...
7719 (pass_data_early_warn_uninitialized): ...new pass_data instance and...
7720 (make_pass_early_warn_uninitialized): ...new function.
7721 (pass_update_address_taken): Convert from a global struct to a
7722 subclass of gimple_opt_pass along with...
7723 (pass_data_update_address_taken): ...new pass_data instance and...
7724 (make_pass_update_address_taken): ...new function.
7725 * tree-ssanames.c (pass_release_ssa_names): Convert from a global
7726 struct to a subclass of gimple_opt_pass along with...
7727 (pass_data_release_ssa_names): ...new pass_data instance and...
7728 (make_pass_release_ssa_names): ...new function.
7729 * tree-stdarg.c (pass_stdarg): Convert from a global struct to a
7730 subclass of gimple_opt_pass along with...
7731 (pass_data_stdarg): ...new pass_data instance and...
7732 (make_pass_stdarg): ...new function.
7733 * tree-switch-conversion.c (pass_convert_switch): Convert from a
7734 global struct to a subclass of gimple_opt_pass along with...
7735 (pass_data_convert_switch): ...new pass_data instance and...
7736 (make_pass_convert_switch): ...new function.
7737 * tree-tailcall.c (pass_tail_recursion): Convert from a global struct
7738 to a subclass of gimple_opt_pass along with...
7739 (pass_data_tail_recursion): ...new pass_data instance and...
7740 (make_pass_tail_recursion): ...new function.
7741 (pass_tail_calls): Convert from a global struct to a subclass of
7742 gimple_opt_pass along with...
7743 (pass_data_tail_calls): ...new pass_data instance and...
7744 (make_pass_tail_calls): ...new function.
7745 * tree-vect-generic.c (pass_lower_vector): Convert from a global
7746 struct to a subclass of gimple_opt_pass along with...
7747 (pass_data_lower_vector): ...new pass_data instance and...
7748 (make_pass_lower_vector): ...new function.
7749 (pass_lower_vector_ssa): Convert from a global struct to a subclass of
7750 gimple_opt_pass along with...
7751 (pass_data_lower_vector_ssa): ...new pass_data instance and...
7752 (make_pass_lower_vector_ssa): ...new function.
7753 * tree-vectorizer.c (pass_slp_vectorize): Convert from a global struct
7754 to a subclass of gimple_opt_pass along with...
7755 (pass_data_slp_vectorize): ...new pass_data instance and...
7756 (make_pass_slp_vectorize): ...new function.
7757 (pass_ipa_increase_alignment): Convert from a global struct to a
7758 subclass of simple_ipa_opt_pass along with...
7759 (pass_data_ipa_increase_alignment): ...new pass_data instance and...
7760 (make_pass_ipa_increase_alignment): ...new function.
7761 * tree-vrp.c (pass_vrp): Convert from a global struct to a subclass of
7762 gimple_opt_pass along with...
7763 (pass_data_vrp): ...new pass_data instance and...
7764 (make_pass_vrp): ...new function.
7765 * tree.c (pass_ipa_free_lang_data): Convert from a global struct to a
7766 subclass of simple_ipa_opt_pass along with...
7767 (pass_data_ipa_free_lang_data): ...new pass_data instance and...
7768 (make_pass_ipa_free_lang_data): ...new function.
7769 * tsan.c (pass_tsan): Convert from a global struct to a subclass of
7770 gimple_opt_pass along with...
7771 (pass_data_tsan): ...new pass_data instance and...
7772 (make_pass_tsan): ...new function.
7773 (pass_tsan_O0): Convert from a global struct to a subclass of
7774 gimple_opt_pass along with...
7775 (pass_data_tsan_O0): ...new pass_data instance and...
7776 (make_pass_tsan_O0): ...new function.
7777 * var-tracking.c (pass_variable_tracking): Convert from a global
7778 struct to a subclass of rtl_opt_pass along with...
7779 (pass_data_variable_tracking): ...new pass_data instance and...
7780 (make_pass_variable_tracking): ...new function.
7781 * web.c (pass_web): Convert from a global struct to a subclass of
7782 rtl_opt_pass along with...
7783 (pass_data_web): ...new pass_data instance and...
7784 (make_pass_web): ...new function.
7785 * config/epiphany/epiphany.h (pass_mode_switch_use): Replace
7786 declaration with that of...
7787 (make_pass_mode_switch_use): ...new function.
7788 (pass_resolve_sw_modes): Replace declaration with that of...
7789 (make_pass_resolve_sw_modes): ...new function.
7790 * config/epiphany/mode-switch-use.c (pass_mode_switch_use): Convert
7791 from a global struct to a subclass of rtl_opt_pass along with...
7792 (pass_data_mode_switch_use): ...new pass_data instance and...
7793 (make_pass_mode_switch_use): ...new function.
7794 * config/epiphany/resolve-sw-modes.c (pass_resolve_sw_modes): Convert
7795 from a global struct to a subclass of rtl_opt_pass along with...
7796 (pass_data_resolve_sw_modes): ...new pass_data instance and...
7797 (make_pass_resolve_sw_modes): ...new function.
7798 * config/i386/i386.c (pass_insert_vzeroupper): Convert from a global
7799 struct to a subclass of rtl_opt_pass along with...
7800 (pass_data_insert_vzeroupper): ...new pass_data instance and...
7801 (make_pass_insert_vzeroupper): ...new function.
7802 * config/sparc/sparc.c (pass_work_around_errata): Convert from a
7803 global struct to a subclass of rtl_opt_pass along with...
7804 (pass_data_work_around_errata): ...new pass_data instance and...
7805 (make_pass_work_around_errata): ...new function.
7806 * config/mips/mips.c (pass_mips_machine_reorg2): Convert from a global
7807 struct to a subclass of rtl_opt_pass along with...
7808 (pass_data_mips_machine_reorg2): ...new pass_data instance and...
7809 (make_pass_mips_machine_reorg2): ...new function.
7811 2013-08-05 David Malcolm <dmalcolm@redhat.com>
7813 * passes.c (pass_manager::operator new): New.
7815 2013-08-05 David Malcolm <dmalcolm@redhat.com>
7817 Handwritten part of conversion of passes to C++ classes.
7819 * Makefile.in (PASS_MANAGER_H): Add dep on pass-instances.def.
7820 (toplev.o): Add dep on PASS_MANAGER_H.
7821 * cgraphunit.c (cgraph_process_new_functions): Rework invocation
7822 of early local pases to reflect this moving from a global to a
7823 member of gcc::pass_manager.
7824 (cgraph_add_new_function): Likewise.
7825 * lto-cgraph.c (lto_output_node): Update for conversion of
7826 struct ipa_opt_pass_d to a C++ subclass of opt_pass.
7827 * passes.c (opt_pass::clone): New.
7828 (opt_pass::gate): New.
7829 (opt_pass::execute): New.
7830 (opt_pass::opt_pass): New.
7831 (pass_manager::execute_early_local_passes): New.
7832 (pass_manager::execute_pass_mode_switching): new.
7833 (finish_optimization_passes): Convert to...
7834 (pass_manager::finish_optimization_passes): ...this.
7835 (finish_optimization_passes): Update for conversion of passes to
7837 (register_dump_files_1): Use has_gate since we cannot portably
7838 check a vtable entry against NULL.
7839 (dump_one_pass): Likewise.
7840 (ipa_write_summaries_2): Likewise.
7841 (ipa_write_optimization_summaries_1): Likewise.
7842 (ipa_read_summaries_1): Likewise.
7843 (ipa_read_optimization_summaries_1): Likewise.
7844 (execute_ipa_stmt_fixups): Likewise.
7845 (pass_manager::pass_manager): Rewrite pass-creation, invoking
7846 pass-creation functions rather than wiring up globals, and
7847 storing the results in fields of pass_manager generated using
7849 (pass_manager::dump_profile_report): Update for conversion of
7850 passes to C++ classes.
7851 (pass_manager::execute_ipa_summary_passes): Likewise.
7852 (execute_one_ipa_transform_pass): Likewise.
7853 (execute_one_pass): Use has_gate and has_execute since we cannot
7854 portably check a vtable entry against NULL.
7855 * pass_manager.h (pass_manager::finish_optimization_passes): New.
7856 (pass_manager): Use pass-instances.def to add fields for the
7857 various pass instances.
7858 * toplev.c (finalize): Update for move of
7859 finish_optimization_passes to a method of gcc::pass_manager.
7860 * toplev.h (finish_optimization_passes): Move to method of class
7862 * tree-pass.h (struct pass_data): New.
7863 (opt_pass): Convert to C++ class, make it a subclass of pass_data.
7864 (opt_pass::gate): Convert to virtual function.
7865 (opt_pass::~opt_pass): New.
7866 (opt_pass::clone): New.
7867 (opt_pass::execute): Convert to virtual function.
7868 (opt_pass::opt_pass): New.
7869 (opt_pass::ctxt_): new.
7870 (gimple_opt_pass): Convert to subclass of opt_pass.
7871 (gimple_opt_pass::gimple_opt_pass): New.
7872 (rtl_opt_pass): Convert to subclass of opt_pass.
7873 (rtl_opt_pass::rtl_opt_pass): New.
7874 (ipa_opt_pass_d): Convert to subclass of opt_pass.
7875 (ipa_opt_pass_d::ipa_opt_pass_d): New.
7876 (simple_ipa_opt_pass): Convert to subclass of opt_pass.
7877 (simple_ipa_opt_pass::simple_ipa_opt_pass): New.
7878 * config/i386/i386.c (rest_of_handle_insert_vzeroupper): Rework
7879 invocation of pass_mode_switching to reflect this moving from a
7880 global to a member of gcc::pass_manager.
7881 (ix86_option_override): Rework how pass_insert_vzeroupper is
7882 added to the pass_manager to reflect autogenerated changes.
7883 * config/i386/t-i386 (i386.o) Add deps on CONTEXT_H and PASS_MANAGER_H.
7885 2013-08-05 Richard Earnshaw <rearnsha@arm.com>
7887 PR rtl-optimization/57708
7888 * recog.c (peep2_find_free_register): Validate all regs in a
7891 2013-08-05 Jan Hubicka <jh@suse.cz>
7894 * cgraph.c (verify_cgraph_node): Accept local flags from other
7896 * ipa.c (symtab_remove_unreachable_nodes): Do not clear local flag.
7897 (function_and_variable_visibility): Likewise.
7898 * trans-mem.c (ipa_tm_create_version): TM versions are not local.
7900 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
7902 * graph.c (init_graph_slim_pretty_print): Remove.
7903 (print_graph_cfg): Do not call it. Use local pretty printer.
7904 (start_graph_dump): Likewise.
7906 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
7908 * gimple-pretty-print.c (buffer): Remove.
7909 (initialized): Likewise.
7910 (maybe_init_pretty_print): Likewise.
7911 (print_gimple_stmt): Do not call it. Use non-static local
7912 pretty_printer variable.
7913 (print_gimple_expr): Likewise.
7914 (print_gimple_seq): Likewise.
7915 (gimple_dump_bb): Likewise.
7917 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
7919 * asan.c (asan_pp): Remove.
7920 (asan_pp_initialized): Likewise.
7921 (asan_pp_initialize): Likewise.
7922 (asan_pp_string): Take a pretty_printer parameter. Adjust callers.
7923 (asan_emit_stack_protection): Tidy. Use local pretty printer.
7924 (asan_add_global): Likewise.
7926 2013-08-04 Gabriel Dos Reis <gdr@integrable-solutions.net>
7928 * pretty-print.h (pp_base): Remove. Adjust dependent macros.
7929 * diagnostic.h (diagnostic_flush_buffer): Adjust.
7930 * pretty-print.c (pp_formatted_text_data): Likewise.
7931 (pp_indent): Rename from pp_base_indent.
7932 (pp_format): Rename from pp_base_format.
7933 (pp_output_formatted_text): Rename from pp_base_output_formatted_text.
7934 (pp_format_verbatim): Rename from pp_base_format_verbatim.
7935 (pp_flush): Rename from pp_base_flush.
7936 (pp_set_line_maximum_length): Rename from
7937 pp_base_set_line_maximum_length.
7938 (pp_clear_output_area): Rename from pp_base_clear_output_area.
7939 (pp_set_prefix): Rename from pp_base_set_prefix.
7940 (pp_destroy_prefix): Rename from pp_base_destroy_prefix.
7941 (pp_emit_prefix): Rename from pp_base_emit_prefix.
7942 (pp_append_text): Rename from pp_base_append_text.
7943 (pp_formatted_text): Rename from pp_base_formatted_text.
7944 (pp_last_position_in_text): Rename from pp_base_last_position_in_text.
7945 (pp_remaining_character_count_for_line): Rename from
7946 pp_base_remaining_character_count_for_line.
7947 (pp_newline): Rename from pp_base_newline.
7948 (pp_character): Rename from pp_base_character.
7949 (pp_string): Rename from pp_base_string.
7950 (pp_maybe_space): Rename from pp_base_maybe_space.
7951 * asan.c (asan_pp_string): Adjust.
7952 (asan_emit_stack_protection): Likewise.
7953 (asan_add_global): Likewise.
7954 * sched-vis.c (str_pattern_slim): Adjust pretty printer function call.
7955 * tree-mudflap.c (mf_varname_tree): Likewise.
7956 * tree-pretty-print.c (pp_tree_identifier): Rename from
7957 pp_base_tree_identifier.
7958 * tree-pretty-print.h (pp_tree_identifier): Remove macro definition.
7959 Declare as function.
7961 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
7963 * pretty-print.h (pp_bar_bar): New.
7964 (pp_ampersand_ampersand): Likewise.
7965 (pp_less_equal): Likewise.
7966 (pp_greater_equal): Likewise.
7967 * gimple-pretty-print.c (dump_ternary_rhs): Use specialized pretty
7968 printer functions instead of pp_string or operators and punctuators.
7969 (dump_gimple_call): Likewise.
7970 (dump_gimple_omp_for): Likewise.
7971 (dump_gimple_transaction): Likewise.
7972 (dump_gimple_phi): Likewise.
7973 (pp_gimple_stmt_1): Likewise.
7974 * sched-vis.c (print_insn): Likewise.
7975 * tree-mudflap.c (mf_varname_tree): Likewise.
7976 * tree-pretty-print.c (dump_block_node): Likewise.
7977 (dump_generic_node): Likewise.
7979 2013-08-02 Jan Hubicka <jh@suse.cz>
7981 * lto-cgraph.c (compute_ltrans_boundary): Add abstract origins into
7983 * lto-streamer-out.c (tree_is_indexable): Results decls and
7984 parm decls are not indexable.
7985 (DFS_write_tree_body): Do not follow args and results.
7986 (hash_tree): Likewise.
7987 (output_functions): Rearrange so struct function is needed
7988 only when real body is output; be able to also ouptut abstract
7989 functions; output DECL_ARGUMENTS and DECL_RESULT.
7990 (lto_output): When not in WPA, ale store abstract functions.
7991 (write_symbol): Do not care about RESULT_DECL.
7992 (output_symbol_p): Handle correctly sbtract decls.
7993 * lto-streamer-in.c (input_function): Rearrange so struct
7994 function can be NULL at entry; allow streaming of
7995 functions w/o body; store DECL_ARGUMENTS and DECL_RESULT.
7996 * ipa.c (symtab_remove_unreachable_nodes): Silence confused
7997 sanity check during LTO.
7998 * tree-streamer-out.c (write_ts_decl_non_common_tree_pointers): Skip
7999 RESULT_DECl and DECL_ARGUMENTS.
8000 * tree-streamer-in.c (lto_input_ts_decl_non_common_tree_pointers):
8003 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
8005 * pretty-print.h (pp_underscore): New.
8007 * gimple-pretty-print.c (dump_unary_rhs): Use specialized pretty
8008 printer functions instead of pp_character.
8009 (dump_binary_rhs): Likewise.
8010 (dump_ternary_rhs): Likewise.
8011 (dump_gimple_call_args): Likewise.
8012 (pp_points_to_solution): Likewise.
8013 (dump_gimple_call): Likewise.
8014 (dump_gimple_switch): Likewise.
8015 (dump_gimple_cond): Likewise.
8016 (dump_gimple_bind): Likewise.
8017 (dump_gimple_try): Likewise.
8018 (dump_gimple_omp_for): Likewise.
8019 (dump_gimple_omp_continue): Likewise.
8020 (dump_gimple_omp_single): Likewise.
8021 (dump_gimple_omp_sections): Likewise.
8022 (dump_gimple_omp_block): Likewise.
8023 (dump_gimple_omp_critical): Likewise.
8024 (dump_gimple_transaction): Likewise.
8025 (dump_gimple_asm): Likewise.
8026 (dump_gimple_phi): Likewise.
8027 (dump_gimple_omp_parallel): Likewise.
8028 (dump_gimple_omp_task): Likewise.
8029 (dump_gimple_omp_atomic_load): Likewise.
8030 (dump_gimple_omp_atomic_store): Likewise.
8031 (dump_gimple_mem_ops): Likewise.
8032 (pp_gimple_stmt_1): Likewise.
8033 (pp_cfg_jump): Likewise.
8034 (dump_implicit_edges): Likewise.
8035 (gimple_dump_bb_for_graph): Likewise.
8036 * graph.c (draw_cfg_node): Likewise.
8037 * langhooks.c (lhd_print_error_function): Likewise.
8038 * sched-vis.c (print_exp): Likewise.
8039 (print_value): Likewise.
8040 (print_pattern): Likewise.
8041 (print_insn): Likewise.
8042 (rtl_dump_bb_for_graph): Likewise.
8043 * tree-pretty-print.c (dump_function_declaration): Likewise.
8044 (dump_array_domain): Likewise.
8045 (dump_omp_clause): Likewise.
8046 (dump_location): Likewise.
8047 (dump_generic_node): Likewise.
8048 (print_struct_decl): Likewise.
8049 * diagnostic.c (diagnostic_show_locus): Use pp_space.
8051 2013-08-03 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
8053 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Update
8054 candidate table when replacing a candidate statement.
8055 (replace_rhs_if_not_dup): Likewise.
8056 (replace_one_candidate): Likewise.
8058 2013-08-02 Jan Hubicka <jh@suse.cz>
8059 Martin Liska <marxin.liska@gmail.com>
8061 * cgraphunit.c (add_new_function): Fix logic when adding from
8063 (assemble_thunk): Rename to ...
8064 (expand_thunk); .. this one; export; get it working with
8065 general functions; make produced gimple valid.
8066 * cgraph.h (expand_thunk): Declare.
8068 2013-08-02 Jan Hubicka <jh@suse.cz>
8070 * ipa-cp.c (gather_context_independent_values): Use
8071 ipa_get_param_move_cost.
8072 (get_replacement_map): Remove PARAM; move parameter folding
8074 (create_specialized_node): Update.
8075 * ipa-prop.c (ipa_populate_param_decls): Do not look for origins;
8076 assert that we have gimple body; update move_cost.
8077 (count_formal_params): Assert that we have gimple body.
8078 (ipa_dump_param): New function.
8079 (ipa_alloc_node_params): Break out from ...
8080 (ipa_initialize_node_params): ... here.
8081 (ipa_get_vector_of_formal_parms): ICE when used in WPA.
8082 (ipa_write_node_info): Stream move costs.
8083 (ipa_read_node_info): Read move costs.
8084 (ipa_update_after_lto_read): Do not recompute node params.
8085 * ipa-prop.h (ipa_param_descriptor): Add move_cost.
8086 (ipa_get_param): Check we are not in WPA.
8087 (ipa_get_param_move_cost): New.
8088 * tree-inline.c (tree_function_versioning): Fold replacement as needed.
8089 * ipa-inline-analysis.c (inline_node_duplication_hook): Expect only
8090 parm numbers to be present.
8092 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
8094 PR rtl-optimization/58048
8095 * lra-constraints.c (process_alt_operands): Don't check asm
8096 operand on register.
8098 2013-08-02 Eric Botcazou <ebotcazou@adacore.com>
8100 * config/sparc/sparc.c (sparc_emit_membar_for_model) <SMM_TSO>: Add
8101 the implied StoreLoad barrier for atomic operations if before.
8103 2013-08-02 Jan Hubicka <jh@suse.cz>
8104 Martin Liska <marxin.liska@gmail.com>
8106 * cgraph.c (cgraph_function_body_availability): Do not check
8108 * cgraph.h (symtab_for_node_and_aliases, symtab_nonoverwritable_alias,
8109 symtab_node_availability): Declare.
8110 * ipa.c (can_replace_by_local_alias): New.
8111 (function_and_variable_visibility): Use it.
8112 * symtab.c (symtab_for_node_and_aliases,
8113 symtab_nonoverwritable_alias_1, symtab_nonoverwritable_alias): New.
8115 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
8117 PR rtl-optimization/57963
8118 * lra-constraints.c (reverse_equiv_p, contains_reloaded_insn_p): New.
8119 (lra_constraints): Use them.
8121 2013-08-02 Sofiane Naci <sofiane.naci@arm.com>
8123 * config/arm/types.md (define_attr "type"): Add "load_acq"
8125 * config/arm/cortex-a53.md (cortex_a53_load1): Update for attribute
8127 (cortex_a53_store1): Likewise.
8129 2013-08-01 Jan Hubicka <jh@suse.cz>
8131 * ipa.c (symtab_remove_unreachable_nodes): Nodes in other
8132 partitions are not needed.
8134 2013-08-01 Uros Bizjak <ubizjak@gmail.com>
8136 * config/i386/i386.h (MAYBE_NON_Q_CLASS_P): New.
8137 * config/i386/i386.c (ix86_secondary_reload): Use INTEGER_CLASS_P and
8138 MAYBE_NON_Q_CLASS_P where appropriate.
8140 2013-08-01 Jan Hubicka <jh@suse.cz>
8142 * cgraph.h (release_function_body): Declare.
8143 * tree.c (free_lang_data_in_decl): Free, parameters and return values
8144 of unused delcarations.
8146 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8148 * config/arm/arm.md (minmax_arithsi_non_canon): Emit canonical
8149 RTL form when subtracting a constant.
8151 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8153 * config/arm/arm.md (peepholes for eq (reg1) (reg2/imm)):
8154 Generate canonical plus rtx with negated immediate instead of minus
8156 * config/arm/arm.c (thumb2_reorg): Handle ADCS <Rd>, <Rn> case.
8158 2013-08-01 Jan Hubicka <jh@suse.cz>
8160 * cgraph.c (cgraph_release_function_body): Use used_as_abstract_origin.
8161 (cgraph_release_function_body): Likewise.
8162 (cgraph_can_remove_if_no_direct_calls_p): Likewise.
8163 * cgraph.h (cgrpah_node): Rename abstract_and_needed
8164 to used_as_abstract_origin.
8165 * tree-inline-transfrom.c (can_remove_node_now_p_1): Do not remove
8166 symbols used as abstract origins.
8167 * cgraphunit.c (analyze_functions): Update.
8168 * ipa.c (symtab_remove_unreachable_nodes): Recompute
8169 used_as_abstract_origin.
8170 * tree-inline.c (tree_function_versioning): Update
8171 used_as_abstract_origin; be ready for DECL_RESULT and
8172 DECL_ARGUMENTS to be NULL.
8174 * lto-symtab.c (lto_symtab_merge_symbols): Merge duplicated nodes
8175 for abstract functions.
8176 * cgraph.h (symtab_real_symbol_p): Abstract declarations are not
8179 2013-08-01 Jan Hubicka <jh@suse.cz>
8181 * profile.c (compute_value_histograms): Fix thinko.
8183 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
8185 * config.gcc (aarch64*-*-*): Add aarch-common.o to extra_objs. Add
8186 aarch-common-protos.h to extra_headers.
8187 (aarch64*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
8188 * config/aarch64/aarch64.md: Include "../arm/cortex-a53.md".
8189 * config/aarch64/t-aarch64 (aarch-common.o): Define.
8191 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
8193 * config/aarch64/aarch64.md (define_attr "type"): Delete.
8194 Include "../arm/types.md". Define "type" attribute for all patterns.
8195 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update for
8198 2013-07-31 Michael Meissner <meissner@linux.vnet.ibm.com>
8200 * config/rs6000/predicates.md (fusion_gpr_addis): New predicates
8201 to support power8 load fusion.
8202 (fusion_gpr_mem_load): Likewise.
8204 * config/rs6000/rs6000-modes.def (PTImode): Update a comment.
8206 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): New
8207 declarations for power8 load fusion.
8208 (emit_fusion_gpr_load): Likewise.
8210 * config/rs6000/rs6000.c (rs6000_option_override_internal): If
8211 tuning for power8, turn on fusion mode by default. Turn on sign
8212 extending fusion mode if normal fusion mode is on, and we are at
8214 (fusion_gpr_load_p): New function, return true if we can fuse an
8215 addis instruction with a dependent load to a GPR.
8216 (emit_fusion_gpr_load): Emit the instructions for power8 load
8219 * config/rs6000/vsx.md (VSX_M2): New iterator for fusion peepholes.
8220 (VSX load fusion peepholes): New peepholes to fuse together an
8221 addi instruction with a VSX load instruction.
8223 * config/rs6000/rs6000.md (GPR load fusion peepholes): New
8224 peepholes to fuse an addis instruction with a load to a GPR base
8225 register. If we are supporting sign extending fusions, convert
8226 sign extending loads to zero extending loads and add an explicit
8229 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
8231 * config.gcc (arm*-*-*): Add aarch-common.o to extra_objs. Add
8232 aarch-common-protos.h to extra_headers.
8233 (arm*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
8234 * config/arm/arm.c (arm_early_load_addr_dep): Move from here to ...
8235 (arm_early_store_addr_dep): Likewise.
8236 (arm_no_early_alu_shift_dep): Likewise.
8237 (arm_no_early_alu_shift_value_dep): Likewise.
8238 (arm_no_early_mul_dep): Likewise.
8239 (arm_no_early_store_addr_dep): Likewise.
8240 (arm_mac_accumulator_is_mul_result): Likewise.
8241 (arm_mac_accumulator_is_result): Likewise.
8242 * config/arm/aarch-common.c: ... here. New file.
8243 * config/arm/arm-protos.h (arm_early_load_addr_dep): Move from
8245 (arm_early_store_addr_dep): Likewise.
8246 (arm_no_early_alu_shift_dep): Likewise.
8247 (arm_no_early_alu_shift_value_dep): Likewise.
8248 (arm_no_early_mul_dep): Likewise.
8249 (arm_no_early_store_addr_dep): Likewise.
8250 (arm_mac_accumulator_is_mul_result): Likewise.
8251 (arm_mac_accumulator_is_result): Likewise.
8252 * config/arm/aarch-common-protos.h: ... here. New file.
8253 * config/arm/t-arm (aarch-common.o): Define.
8255 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
8257 * config/arm/arm.md: Include new file "types.md".
8258 (define_attr "type"): Move from here to ...
8259 (define_attr "mul32"): Likewise.
8260 (define_attr "mul64"): Likewise.
8261 * config/arm/types.md: ... here. New file.
8263 2013-07-31 Sebastian Huber <sebastian.huber@embedded-brains.de>
8265 * config.gcc (*-*-rtems*): Use __cxa_atexit by default.
8266 * config/rs6000/rtems.h (TARGET_LIBGCC_SDATA_SECTION): Define.
8268 2013-07-31 Jan-Benedict Glaw <jbglaw@lug-owl.de>
8270 * gen-pass-instances.awk: Fix offset of substr().
8272 2013-07-31 David Malcolm <dmalcolm@redhat.com>
8274 * Makefile.in (pass-instances.def): New.
8275 (passes.o): Replace dependency on passes.def with one on
8278 * gen-pass-instances.awk: New.
8280 * passes.c (pass_manager::pass_manager): Use pass-instances.def
8281 rather than passes.def, updating local definition of NEXT_PASS
8282 macro to add an extra NUM parameter (currently unused).
8284 2013-07-30 David Malcolm <dmalcolm@redhat.com>
8286 * Makefile.in (PASS_MANAGER_H): New.
8287 (lto-cgraph.o): Depend on CONTEXT_H and PASS_MANAGER_H.
8288 (passes.o): Likewise.
8289 (statistics.o): Likewise.
8290 (cgraphunit.o): Likewise.
8291 (context.o): Depend on PASS_MANAGER_H.
8293 * pass_manager.h: New.
8295 * cgraphunit.c (cgraph_add_new_function): Update for moves
8296 of globals to fields of pass_manager.
8297 (analyze_function): Likewise.
8298 (expand_function): Likewise.
8299 (ipa_passes): Likewise.
8300 (compile): Likewise.
8302 * context.c (context::context): New.
8303 * context.h (context::context): New.
8304 (context::get_passes): New.
8305 (context::passes_): New.
8307 * lto-cgraph.c (input_node): Update for moves of globals to
8308 fields of pass_manager.
8310 * passes.c (all_passes): Remove, in favor of a field of the
8311 same name within the new class pass_manager.
8312 (all_small_ipa_passes): Likewise.
8313 (all_lowering_passes): Likewise.
8314 (all_regular_ipa_passes): Likewise.
8315 (all_late_ipa_passes): Likewise.
8316 (all_lto_gen_passes): Likewise.
8317 (passes_by_id): Likewise.
8318 (passes_by_id_size): Likewise.
8319 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
8320 the new class pass_manager.
8321 (set_pass_for_id): Convert to...
8322 (pass_manager::set_pass_for_id): ...method.
8323 (get_pass_for_id): Convert to...
8324 (pass_manager::get_pass_for_id): ...method.
8325 (register_one_dump_file): Move body of implementation into...
8326 (pass_manager::register_one_dump_file): ...here.
8327 (register_dump_files_1): Convert to...
8328 (pass_manager::register_dump_files_1): ...method.
8329 (register_dump_files): Convert to...
8330 (pass_manager::register_dump_files): ...method.
8331 (create_pass_tab): Update for moves of globals to fields of
8333 (dump_passes): Move body of implementation into...
8334 (pass_manager::dump_passes): ...here.
8335 (register_pass): Move body of implementation into...
8336 (pass_manager::register_pass): ...here.
8337 (init_optimization_passes): Convert into...
8338 (pass_manager::pass_manager): ...constructor for new
8339 pass_manager class, and initialize the pass_lists array.
8340 (check_profile_consistency): Update for moves of globals to
8341 fields of pass_manager.
8342 (dump_profile_report): Move body of implementation into...
8343 (pass_manager::dump_profile_report): ...here.
8344 (ipa_write_summaries_1): Update for moves of pass lists from
8345 being globals to fields of pass_manager.
8346 (ipa_write_optimization_summaries): Likewise.
8347 (ipa_read_summaries): Likewise.
8348 (ipa_read_optimization_summaries): Likewise.
8349 (execute_all_ipa_stmt_fixups): Likewise.
8351 * statistics.c (statistics_fini): Update for moves of globals to
8352 fields of pass_manager.
8354 * toplev.c (general_init): Replace call to
8355 init_optimization_passes with construction of the pass_manager
8358 * tree-pass.h (all_passes): Remove, in favor of a field of the
8359 same name within the new class pass_manager.
8360 (all_small_ipa_passes): Likewise.
8361 (all_lowering_passes): Likewise.
8362 (all_regular_ipa_passes): Likewise.
8363 (all_lto_gen_passes): Likewise.
8364 (all_late_ipa_passes): Likewise.
8365 (passes_by_id): Likewise.
8366 (passes_by_id_size): Likewise.
8367 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
8368 the new class pass_manager.
8369 (get_pass_for_id): Remove.
8371 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
8373 * config.gcc (arm): Require 64-bit host-wide-int for all ARM target
8376 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
8378 * arm.md (mulhi3): New expand pattern.
8380 2013-07-30 Jan Hubicka <jh@suse.cz>
8381 Martin Liska <marxin.liska@gmail.com>
8383 * profile.c (compute_value_histograms): Do not ICE when
8384 there is mismatch only on some counters.
8386 2013-07-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
8388 PR rtl-optimization/57637
8389 * function.c (move_insn_for_shrink_wrap): Also check the
8390 GEN set of the LIVE problem for the liveness analysis
8391 if it exists, otherwise give up.
8393 2013-07-29 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
8395 PR tree-optimization/57993
8396 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Record
8397 replaced statement in the candidate table.
8398 (phi_add_costs): Return infinite cost when the hidden basis does
8399 not dominate all phis on which the candidate is dependent.
8400 (replace_one_candidate): Record replaced statement in the
8403 2013-07-29 Joern Rennecke <joern.rennecke@embecosm.com>
8405 * config/epiphany/epiphany.md (*isub_i+2): New peephole.
8406 (ashlv2si3): New expander.
8407 (*ashlv2si3_i): New define_insn_and_split.
8408 * predicates.md (float_operation): Allow patterns with three
8411 PR rtl-optimization/58021
8412 * mode-switching.c (create_pre_exit): Always split off preceding
8413 insns if we are not at the basic block head.
8415 2013-07-29 Maciej W. Rozycki <macro@codesourcery.com>
8417 * config/mips/linux.h (GLIBC_DYNAMIC_LINKER): Handle `-mnan=2008'.
8418 (UCLIBC_DYNAMIC_LINKER): New macro.
8419 * config/mips/linux64.h (GLIBC_DYNAMIC_LINKER32): Handle
8421 (GLIBC_DYNAMIC_LINKER64, GLIBC_DYNAMIC_LINKERN32): Likewise.
8422 (UCLIBC_DYNAMIC_LINKER32): Undefine macro first. Handle
8424 (UCLIBC_DYNAMIC_LINKER64): Redefine macro.
8425 (UCLIBC_DYNAMIC_LINKERN32): Likewise.
8426 * config/mips/mips-modes.def: Remove RESET_FLOAT_FORMAT calls
8427 for SF and DF modes. Use ieee_quad_format for TF mode.
8428 * config/mips/mips-opts.h (mips_ieee_754_setting): New enum.
8429 * config/mips/mips.c (mips_file_start): Output a `.nan' directive.
8430 (mips_option_override): Handle `-mnan=legacy'.
8431 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Handle
8432 `-mabs=2008' and `-mnan=2008'.
8433 (OPTION_DEFAULT_SPECS): Add "nan" default.
8434 (ASM_SPEC): Handle `-mnan='.
8435 [!HAVE_AS_NAN] (HAVE_AS_NAN): New macro.
8436 * config/mips/mips.md (abs<mode>2): Handle `-mabs=2008', update
8437 comment accordingly.
8438 (neg<mode>2): Likewise.
8439 * config/mips/mips.opt (mabs, mnan): New options.
8440 * doc/install.texi (Configuration): Document `--with-nan=' option.
8441 * doc/invoke.texi (Option Summary): List MIPS `-mabs=' and
8443 (MIPS Options): Document them.
8444 * config.gcc <mips*-*-*>: Handle `--with-nan='.
8445 * configure.ac <mips*-*-*>: Check for GAS `-mnan=2008' support.
8446 * configure: Regenerate.
8447 * config.in: Regenerate.
8449 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
8451 * config/i386/i386.md (float post-reload splitters): Do not check
8452 for subregs of SSE registers.
8454 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
8455 H.J. Lu <hongjiu.lu@intel.com>
8459 * config/i386/i386.md (post-reload splitter
8460 to avoid partial SSE reg dependency stalls): New pattern.
8462 2013-07-29 Dominik Vogt <vogt@linux.vnet.ibm.com>
8464 * config/s390/s390.md ("movcc"): Swap load and store instructions.
8466 2013-07-27 Joern Rennecke <joern.rennecke@embecosm.com>
8468 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
8469 Also reserve space for saving UNKNOWN_REGNUM for leaf functions.
8471 2013-07-26 Cary Coutant <ccoutant@google.com>
8473 * dwarf2out.c (die_checksum_ordered): Don't include template
8474 instantiations in signature.
8475 (is_template_parameter): New function.
8476 (is_template_instantiation): New function.
8477 (generate_skeleton_bottom_up): Don't include template instantiations
8479 (generate_skeleton): Likewise.
8480 (break_out_comdat_types): Move recursive call to break out nested
8482 (prune_unused_types_mark_generic_parms_dies): Call
8483 is_template_parameter.
8485 2013-07-26 Ian Bolton <ian.bolton@arm.com>
8487 * config/aarch64/aarch64.md (neg<mode>2): Offer alternative that
8488 uses vector registers.
8489 * config/aarch64/iterators.md: Add attributes rtn and vas.
8491 2013-07-26 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8492 Richard Earnshaw <richard.earnshaw@arm.com>
8494 * combine.c (simplify_comparison): Re-canonicalize operands
8496 * config/arm/arm.md (movcond_addsi): New splitter.
8498 2013-07-25 Sterling Augustine <saugustine@google.com>
8500 * dwarf2out.c (size_of_pubnames): Move code to...
8501 (include_pubname_in_output): ...here. New.
8502 (want_pubnames): Rearrange.
8503 (output_pubnames): Call include_pubname_in_output. Move assertion.
8505 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
8507 * doc/extend.texi: Fix return types for __builtin_ia32_cmp*s builtins.
8509 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
8512 * doc/extend.texi: Remove obsolete builtins. Fix
8513 typo for __builtin_ia32_loadss and __builtin_ia32_cmpnltss.
8515 2013-07-25 Jan Hubicka <jh@suse.cz>
8517 * cgraph.c (release_function_body): Break out from ...
8518 (cgraph_release_function_body): ... this one; also release DECL_RESULT
8520 * ipa-cp.c (get_replacement_map): Add parm_num argument; do not set
8521 old_tree in the map.
8522 (create_specialized_node): Update.
8523 * lto-cgraph.c (output_node_opt_summary): Do not translate old_tree
8525 * cgraphclones.c (cgraph_create_virtual_clone): Do not copy
8526 DECL_ARGUMENTS, DECL_INITIAL and DECL_RESULT.
8527 * ipa-prop.c (ipa_populate_param_decls): Look for origin of clones.
8528 * tree-inline.c (initialize_cfun): Initialize DECL_ARGUMENTS and
8531 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8533 * config/arm/arm.md (arm_addsi3, addsi3_carryin_<optab>,
8534 addsi3_carryin_alt2_<optab>): Correct output template.
8536 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8538 * config/arm/arm-fixed.md (ssmulsa3, usmulusa3):
8539 Adjust for arm_restrict_it.
8540 Remove trailing whitespace.
8542 2013-07-25  Mark Kettenis  <kettenis@openbsd.org>
8544 * config/pa/pa.c (pa_trampoline_init): Emit __enable_execute_stack
8545 libcall if HAVE_ENABLE_EXECUTE_STACK is defined.
8547 * config.gcc (hppa-*-openbsd*): Don't set tmake_file.
8549 2013-07-25 Vladimir Makarov <vmakarov@redhat.com>
8551 PR rtl-optimization/57960
8552 * lra-constraints.c (process_alt_operands): Use the right mode
8553 when checking strict_low.
8555 2013-07-25 Jan Hubicka <jh@suse.cz>
8557 * lto-symtab.c (lto_cgraph_replace_node): Release function body.
8558 * cgraph.c (cgraph_remove_node): Do not release function body
8559 when in cgraph streaming.
8560 * ipa.c (process_references, symtab_remove_unreachable_nodes): Objects
8561 in other partitions are not considered reachable; fix handling of
8564 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
8566 * config/arm/arm.md (*sibcall_insn): Remove unnecessary space.
8568 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
8573 * config/arm/arm.md ("*sibcall_insn): Replace use of
8574 Ss with US. Adjust output for v5 and v4t.
8575 (*sibcall_value_insn): Likewise and loosen predicate on operand0.
8577 * config/arm/constraints.md ("Ss"): Rename to US.
8579 2013-07-25 Terry Guo <terry.guo@arm.com>
8581 * config/arm/arm.c (thumb1_size_rtx_costs): Assign proper cost for
8582 shift_add/shift_sub0/shift_sub1 RTXs.
8584 2013-07-24 Bill Schmidt <wschmidt@linux.ibm.com>
8585 Anton Blanchard <anton@au1.ibm.com>
8587 * config/rs6000/altivec.md (altivec_vpkpx): Handle little endian.
8588 (altivec_vpks<VI_char>ss): Likewise.
8589 (altivec_vpks<VI_char>us): Likewise.
8590 (altivec_vpku<VI_char>us): Likewise.
8591 (altivec_vpku<VI_char>um): Likewise.
8593 2013-07-24 David Malcolm <dmalcolm@redhat.com>
8595 Introduce context class.
8597 * Makefile.in (CONTEXT_H): New.
8598 (OBJS): Add context.o.
8599 (toplev.o): Add CONTEXT_H to dependencies.
8602 * toplev.c (general_init): Create the singleton gcc::context instance.
8608 2013-07-24 Joern Rennecke <joern.rennecke@embecosm.com>
8610 PR rtl-optimization/57968
8611 * mode-switching.c (create_pre_exit): Allow instructions that
8612 don't set a return register to need a non-exit mode.
8614 2013-07-24 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
8615 Anton Blanchard <anton@au1.ibm.com>
8617 * config/rs6000/vector.md (vec_realign_load_<mode>): Reorder input
8618 operands to vperm for little endian.
8619 * config/rs6000/rs6000.c (rs6000_expand_builtin): Use lvsr instead
8620 of lvsl to create the control mask for a vperm for little endian.
8622 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8623 Anton Blanchard <anton@au1.ibm.com>
8625 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Reverse
8626 two operands for little-endian.
8628 2013-07-23 Steve Ellcey <sellcey@mips.com>
8630 * config/mips/mips.c (mips_case_values_threshold): New.
8631 (TARGET_CASE_VALUES_THRESHOLD): Define.
8633 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8634 Anton Blanchard <anton@au1.ibm.com>
8636 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Correct
8637 selection of field for vector splat in little endian mode.
8639 2013-07-23 Michael Meissner <meissner@linux.vnet.ibm.com>
8641 * config/rs6000/vector.md (xor<mode>3): Move 128-bit boolean
8642 expanders to rs6000.md.
8643 (ior<mode>3): Likewise.
8644 (and<mode>3): Likewise.
8645 (one_cmpl<mode>2): Likewise.
8646 (nor<mode>3): Likewise.
8647 (andc<mode>3): Likewise.
8648 (eqv<mode>3): Likewise.
8649 (nand<mode>3): Likewise.
8650 (orc<mode>3): Likewise.
8652 * config/rs6000/rs6000-protos.h (rs6000_split_logical): New
8655 * config/rs6000/rs6000.c (rs6000_split_logical_inner): Add support
8656 to split multi-word logical operations.
8657 (rs6000_split_logical_di): Likewise.
8658 (rs6000_split_logical): Likewise.
8660 * config/rs6000/vsx.md (VSX_L2): Delete, no longer used.
8661 (vsx_and<mode>3_32bit): Move 128-bit logical insns to rs6000.md,
8662 and allow TImode operations in 32-bit.
8663 (vsx_and<mode>3_64bit): Likewise.
8664 (vsx_ior<mode>3_32bit): Likewise.
8665 (vsx_ior<mode>3_64bit): Likewise.
8666 (vsx_xor<mode>3_32bit): Likewise.
8667 (vsx_xor<mode>3_64bit): Likewise.
8668 (vsx_one_cmpl<mode>2_32bit): Likewise.
8669 (vsx_one_cmpl<mode>2_64bit): Likewise.
8670 (vsx_nor<mode>3_32bit): Likewise.
8671 (vsx_nor<mode>3_64bit): Likewise.
8672 (vsx_andc<mode>3_32bit): Likewise.
8673 (vsx_andc<mode>3_64bit): Likewise.
8674 (vsx_eqv<mode>3_32bit): Likewise.
8675 (vsx_eqv<mode>3_64bit): Likewise.
8676 (vsx_nand<mode>3_32bit): Likewise.
8677 (vsx_nand<mode>3_64bit): Likewise.
8678 (vsx_orc<mode>3_32bit): Likewise.
8679 (vsx_orc<mode>3_64bit): Likewise.
8681 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Always allow vector
8682 logical types in GPRs.
8684 * config/rs6000/altivec.md (altivec_and<mode>3): Move 128-bit
8685 logical insns to rs6000.md, and allow TImode operations in
8687 (altivec_ior<mode>3): Likewise.
8688 (altivec_xor<mode>3): Likewise.
8689 (altivec_one_cmpl<mode>2): Likewise.
8690 (altivec_nor<mode>3): Likewise.
8691 (altivec_andc<mode>3): Likewise.
8693 * config/rs6000/rs6000.md (BOOL_128): New mode iterators and mode
8694 attributes for moving the 128-bit logical operations into
8696 (BOOL_REGS_OUTPUT): Likewise.
8697 (BOOL_REGS_OP1): Likewise.
8698 (BOOL_REGS_OP2): Likewise.
8699 (BOOL_REGS_UNARY): Likewise.
8700 (BOOL_REGS_AND_CR0): Likewise.
8701 (one_cmpl<mode>2): Add support for DI logical operations on
8702 32-bit, splitting the operations to 32-bit.
8706 (and<mode>3, 128-bit types): Rewrite 2013-06-06 logical operator
8707 changes to combine the 32/64-bit code, allow logical operations on
8708 TI mode in 32-bit, and to use similar match_operator patterns like
8709 scalar mode uses. Combine the Altivec and VSX code for logical
8710 operations, and move it here.
8711 (ior<mode>3, 128-bit types): Likewise.
8712 (xor<mode>3, 128-bit types): Likewise.
8713 (one_cmpl<mode>3, 128-bit types): Likewise.
8714 (nor<mode>3, 128-bit types): Likewise.
8715 (andc<mode>3, 128-bit types): Likewise.
8716 (eqv<mode>3, 128-bit types): Likewise.
8717 (nand<mode>3, 128-bit types): Likewise.
8718 (orc<mode>3, 128-bit types): Likewise.
8719 (and<mode>3_internal): Likewise.
8720 (bool<mode>3_internal): Likewise.
8721 (boolc<mode>3_internal1): Likewise.
8722 (boolc<mode>3_internal2): Likewise.
8723 (boolcc<mode>3_internal1): Likewise.
8724 (boolcc<mode>3_internal2): Likewise.
8725 (eqv<mode>3_internal1): Likewise.
8726 (eqv<mode>3_internal2): Likewise.
8727 (one_cmpl1<mode>3_internal): Likewise.
8729 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
8731 * config/microblaze/microblaze.c (microblaze_expand_prologue):
8732 Rename flag_stack_usage to flag_stack_usage_info.
8734 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
8736 * config/microblaze/sync.md: New file.
8737 * config/microblaze/microblaze.md: Include sync.md
8738 * config/microblaze/microblaze.c: Add print_operand 'y'.
8739 * config/microblaze/constraints.md: Add memory_contraint
8740 'Q' which is a single register.
8742 2013-07-23 Eric Botcazou <ebotcazou@adacore.com>
8744 * doc/invoke.texi (SPARC Options): Document new leon3 processor value.
8746 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
8748 * reload.c (find_reloads): Exit loop once we find this operand
8749 cannot be reloaded somehow for this alternative.
8751 * reload.c (find_reloads): Exit loop once we find a hard register.
8753 * rtlanal.c (computed_jump_p): Exit loop once we find label
8756 * i386.c (ix86_pad_returns): Exit loop after setting replace.
8758 * cfgloopmanip.c (remove_path): Exit loop after setting
8761 * gensupport.c (subst_dup): Avoid loop if code is not
8762 MATCH_DUP nor MATCH_OP_DUP.
8764 2013-07-23 Nicklas Bo Jensen <nbjensen@gmail.com>
8766 * doc/md.texi (Machine-Specific Peephole Optimizers): Fix a typo.
8768 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
8770 * config/aarch64/aarch64.c (aarch64_hard_regno_mode_ok): Also return
8771 true for SP_REGNUM if mode == ptr_mode.
8772 * config/aarch64/aarch64.h (ADDITIONAL_REGISTER_NAMES): Add "wsp"
8773 with value R0_REGNUM + 31.
8775 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
8777 * config/aarch64/aarch64.c (aarch64_pad_arg_upward): In big-endian,
8778 pad pointer-typed argument downward.
8780 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
8782 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Define _ILP32
8783 and __ILP32__ when the ILP32 model is in use.
8785 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
8787 * config/aarch64/aarch64.c (POINTER_BYTES): New define.
8788 (aarch64_load_symref_appropriately): In the case of
8789 SYMBOL_SMALL_ABSOLUTE, use the mode of 'dest' instead of Pmode
8790 to generate new rtx; likewise to the case of SYMBOL_SMALL_GOT.
8791 (aarch64_expand_mov_immediate): In the case of SYMBOL_FORCE_TO_MEM,
8792 change to pass 'ptr_mode' to force_const_mem and zero-extend 'mem'
8793 if 'mode' doesn't equal to 'ptr_mode'.
8794 (aarch64_output_mi_thunk): Add an assertion on the alignment of
8795 'vcall_offset'; change to call aarch64_emit_move differently depending
8796 on whether 'Pmode' equals to 'ptr_mode' or not; use 'POINTER_BYTES'
8797 to calculate the upper bound of 'vcall_offset'.
8798 (aarch64_cannot_force_const_mem): Change to also return true if
8800 (aarch64_legitimize_reload_address): In the case of large
8801 displacements, add new local variable 'xmode' and an assertion
8802 based on it; change to use 'xmode' to generate the new rtx and
8804 (aarch64_asm_trampoline_template): Change to generate the template
8805 differently depending on TARGET_ILP32 or not; change to use
8806 'POINTER_BYTES' in the argument passed to assemble_aligned_integer.
8807 (aarch64_trampoline_size): Removed.
8808 (aarch64_trampoline_init): Add new local constant 'tramp_code_sz'
8809 and replace immediate literals with it. Change to use 'ptr_mode'
8810 instead of 'DImode' and call convert_memory_address if the mode
8811 of 'fnaddr' doesn't equal to 'ptr_mode'.
8812 (aarch64_elf_asm_constructor): Change to use assemble_aligned_integer
8814 (aarch64_elf_asm_destructor): Likewise.
8815 * config/aarch64/aarch64.h (TRAMPOLINE_SIZE): Change to be dependent
8816 on TARGET_ILP32 instead of aarch64_trampoline_size.
8817 * config/aarch64/aarch64.md (movsi_aarch64): Add new alternatives
8818 of 'mov' between WSP and W registers as well as 'adr' and 'adrp'.
8819 (loadwb_pair<GPI:mode>_<PTR:mode>): Rename to ...
8820 (loadwb_pair<GPI:mode>_<P:mode>): ... this. Replace PTR with P.
8821 (storewb_pair<GPI:mode>_<PTR:mode>): Likewise; rename to ...
8822 (storewb_pair<GPI:mode>_<P:mode>): ... this.
8823 (add_losym): Change to 'define_expand' and call gen_add_losym_<mode>
8824 depending on the value of 'mode'.
8825 (add_losym_<mode>): New.
8826 (ldr_got_small_<mode>): New, based on ldr_got_small.
8827 (ldr_got_small): Remove.
8828 (ldr_got_small_sidi): New.
8829 * config/aarch64/iterators.md (P): New.
8830 (PTR): Change to 'ptr_mode' in the condition.
8832 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
8834 * config.gcc (aarch64*-*-*): Support --with-abi.
8835 (aarch64*-*-elf): Support --with-multilib-list.
8836 (aarch64*-*-linux*): Likewise.
8837 (supported_defaults): Add abi to aarch64*-*-*.
8838 * configure.ac: Mention AArch64 for --with-multilib-list.
8839 * configure: Re-generated.
8840 * config/aarch64/biarchilp32.h: New file.
8841 * config/aarch64/biarchlp64.h: New file.
8842 * config/aarch64/aarch64-elf.h (ENDIAN_SPEC): New define.
8844 (MULTILIB_DEFAULTS): Ditto.
8845 (DRIVER_SELF_SPECS): Ditto.
8846 (ASM_SPEC): Update to also substitute -mabi.
8847 * config/aarch64/aarch64-elf-raw.h (LINK_SPEC): Add linker script
8848 file whose name depends on -mabi= and -mbig-endian.
8849 * config/aarch64/aarch64.h (LONG_TYPE_SIZE): Change to depend on
8851 (POINTER_SIZE): New define.
8852 (POINTERS_EXTEND_UNSIGNED): Ditto.
8853 (enum aarch64_abi_type): New enumeration tag.
8854 (AARCH64_ABI_LP64, AARCH64_ABI_ILP32): New enumerators.
8855 (AARCH64_ABI_DEFAULT): Define to AARCH64_ABI_LP64 if undefined.
8856 (TARGET_ILP32): New define.
8857 * config/aarch64/aarch64.opt (mabi): New.
8859 (ilp32, lp64): New values for -mabi.
8860 * config/aarch64/t-aarch64 (comma): New define.
8861 (MULTILIB_OPTIONS): Ditto.
8862 (MULTILIB_DIRNAMES): Ditto.
8863 * config/aarch64/t-aarch64-linux (MULTIARCH_DIRNAME): New define.
8864 * doc/invoke.texi: Document -mabi for AArch64.
8866 2013-07-23 Georg-Johann Lay <avr@gjlay.de>
8868 * config/avr/avr.md: Explain asm print modifier 'r' for REG.
8870 2013-07-22 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
8871 Anton Blanchard <anton@au1.ibm.com>
8873 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Fix
8874 endianness when selecting field to splat.
8876 2013-07-22 Eric Christopher <echristo@gmail.com>
8878 * dwarf2out.c (die_odr_checksum): New function to use
8879 CHECKSUM_ macros and ULEB128 for DIE tag.
8880 (generate_type_signature): Use.
8882 2013-07-22 Eric Botcazou <ebotcazou@adacore.com>
8884 * config.gcc (sparc*-*-*): Accept leon3 processor.
8885 (sparc-leon*-*): Merge with sparc*-*-* and add leon3 support.
8886 * doc/invoke.texi (SPARC Options): Adjust -mfix-ut699 entry.
8887 * config/sparc/sparc-opts.h (enum processor_type): Add PROCESSOR_LEON3.
8888 * config/sparc/sparc.opt (enum processor_type): Add leon3.
8889 (mfix-ut699): Adjust comment.
8890 * config/sparc/sparc.h (TARGET_CPU_leon3): New define.
8891 (CPP_CPU32_DEFAULT_SPEC): Add leon3 support.
8892 (CPP_CPU_SPEC): Likewise.
8893 (ASM_CPU_SPEC): Likewise.
8894 * config/sparc/sparc.c (leon3_cost): New constant.
8895 (sparc_option_override): Add leon3 support.
8896 (mem_ref): New function.
8897 (sparc_gate_work_around_errata): Return true if -mfix-ut699 is enabled.
8898 (sparc_do_work_around_errata): Look into the instruction in the delay
8899 slot and adjust accordingly. Add fix for the data cache nullify issues
8900 of the UT699. Change insertion position for the NOP.
8901 * config/sparc/leon.md (leon_fpalu, leon_fpmds, write_buf): Delete.
8902 (leon3_load): New reservation.
8903 (leon_store): Bump latency to 2.
8904 (grfpu): New automaton.
8905 (grfpu_alu): New unit.
8906 (grfpu_ds): Likewise.
8907 (leon_fp_alu): Adjust.
8908 (leon_fp_mult): Delete.
8909 (leon_fp_div): Split into leon_fp_divs and leon_fp_divd.
8910 (leon_fp_sqrt): Split into leon_fp_sqrts and leon_fp_sqrtd.
8911 * config/sparc/sparc.md (cpu): Add leon3.
8912 * config/sparc/sync.md (atomic_exchangesi): Disable if -mfix-ut699.
8914 (atomic_test_and_set): Likewise.
8917 2013-07-22 Jürgen Urban <JuergenUrban@gmx.de>
8919 * config.gcc (mips*-*-*): Add --with-fpu support. Make single the
8920 default for R5900 targets.
8921 * config/mips/mips.h (OPTION_DEFAULT_SPECS): Handle --with-fpu.
8922 (ISA_HAS_LDC1_SDC1): Set to false for TARGET_MIPS5900.
8923 * config/mips/mips.c (mips_option_override): Report an error for
8924 -march=r5900 -mhard-float -mdouble-float. Use spu_single_format
8925 for -march=r5900 -mhard-float.
8927 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
8929 * df-problems.c (can_move_insns_across): Exit loop once we
8930 find a non-fixed, non-global register.
8932 * ipa-pure-const.c (propagate_nothrow): Exit loop after
8935 * omega.c (omega_eliminate_red): Break after setting red_found.
8936 (omega_problem_has_red_equations): Similarly after setting found.
8937 (omega_query_variable): Similarly after setting coupled.
8939 2013-07-22 Marek Polacek <polacek@redhat.com>
8941 * gimplify.c: Don't include gimple.h twice.
8943 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8945 * config/arm/constraints.md (Pd): Allow TARGET_THUMB
8946 instead of TARGET_THUMB1.
8947 (Pz): New constraint.
8948 * config/arm/arm.md (arm_addsi3): Add alternatives for 16-bit
8950 (compare_negsi_si): Likewise.
8951 (compare_addsi2_op0): Likewise.
8952 (compare_addsi2_op1): Likewise.
8953 (addsi3_carryin_<optab>): Likewise.
8954 (addsi3_carryin_alt2_<optab>): Likewise.
8955 (addsi3_carryin_shift_<optab>): Disable cond_exec variant
8956 for arm_restrict_it.
8957 (subsi3_carryin): Likewise.
8958 (arm_subsi3_insn): Add alternatives for 16-bit encoding.
8959 (minmax_arithsi): Disable for arm_restrict_it.
8960 (minmax_arithsi_non_canon): Adjust for arm_restrict_it.
8961 (satsi_<SAT:code>): Disable cond_exec variant for arm_restrict_it.
8962 (satsi_<SAT:code>_shift): Likewise.
8963 (arm_shiftsi3): Add alternative for 16-bit encoding.
8964 (arm32_movhf): Disable for arm_restrict_it.
8965 (arm_cmpdi_unsigned): Add alternatives for 16-bit encoding.
8966 (arm_movtas_ze): Disable cond_exec variant for arm_restrict_it.
8968 2013-07-22 Sofiane Naci <sofiane.naci@arm.com>
8970 * config/arm/arm.md (attribute "insn"): Delete.
8971 (attribute "type"): Add "mov_imm", "mov_reg", "mov_shift",
8972 "mov_shift_reg", "mvn_imm", "mvn_reg", "mvn_shift" and "mvn_shift_reg".
8973 (not_shiftsi): Update for attribute change.
8974 (not_shiftsi_compare0): Likewise.
8975 (not_shiftsi_compare0_scratch): Likewise.
8976 (arm_one_cmplsi2): Likewise.
8977 (thumb1_one_cmplsi2): Likewise.
8978 (notsi_compare0): Likewise.
8979 (notsi_compare0_scratch): Likewise.
8980 (thumb1_movdi_insn): Likewise.
8981 (arm_movsi_insn): Likewise.
8982 (movhi_insn_arch4): Likewise.
8983 (movhi_bytes): Likewise.
8984 (arm_movqi_insn): Likewise.
8985 (thumb1_movqi_insn): Likewise.
8986 (arm32_movhf): Likewise.
8987 (thumb1_movhf): Likewise.
8988 (arm_movsf_soft_insn): Likewise.
8989 (thumb1_movsf_insn): Likewise.
8990 (thumb_movdf_insn): Likewise.
8991 (movsicc_insn): Likewise.
8992 (movsfcc_soft_insn): Likewise.
8993 (and_scc): Likewise.
8994 (cond_move): Likewise.
8995 (if_move_not): Likewise.
8996 (if_not_move): Likewise.
8997 (if_shift_move): Likewise.
8998 (if_move_shift): Likewise.
8999 (if_shift_shift): Likewise.
9000 (if_not_arith): Likewise.
9001 (if_arith_not): Likewise.
9002 (cond_move_not): Likewise.
9003 * config/arm/neon.md (neon_mov<mode>): Update for attribute change.
9004 (neon_mov<mode>): Likewise.
9005 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute change.
9006 (thumb2_movsi_vfp): Likewise.
9007 (movsf_vfp): Likewise.
9008 (thumb2_movsf_vfp): Likewise.
9009 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
9011 (cortexa7_older_only): Likewise.
9012 (cortexa7_younger): Likewise.
9013 * config/arm/arm1020e.md (1020alu_op): Update for attribute change.
9014 (1020alu_shift_op): Likewise.
9015 (1020alu_shift_reg_op): Likewise.
9016 * config/arm/arm1026ejs.md (alu_op): Update for attribute change.
9017 (alu_shift_op): Likewise.
9018 (alu_shift_reg_op): Likewise.
9019 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute change.
9020 (11_alu_shift_op): Likewise.
9021 (11_alu_shift_reg_op): Likewise.
9022 * config/arm/arm926ejs.md (9_alu_op): Update for attribute change.
9023 (9_alu_shift_reg_op): Likewise.
9024 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
9026 (cortex_a15_alu_shift): Likewise.
9027 (cortex_a15_alu_shift_reg): Likewise.
9028 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute change.
9029 (cortex_a5_alu_shift): Likewise.
9030 * config/arm/cortex-a53.md (cortex_a53_alu): Update for attribute
9032 (cortex_a53_alu_shift): Likewise.
9033 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
9035 (cortex_a7_alu_reg): Likewise.
9036 (cortex_a7_alu_shift): Likewise.
9037 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
9038 (cortex_a8_alu_shift): Likewise.
9039 (cortex_a8_alu_shift_reg): Likewise.
9040 (cortex_a8_mov): Likewise.
9041 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
9042 (cortex_a9_dp_shift): Likewise.
9043 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute change.
9044 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute change.
9045 (cortex_r4_mov): Likewise.
9046 (cortex_r4_alu_shift): Likewise.
9047 (cortex_r4_alu_shift_reg): Likewise.
9048 * config/arm/fa526.md (526_alu_op): Update for attribute change.
9049 (526_alu_shift_op): Likewise.
9050 * config/arm/fa606te.md (606te_alu_op): Update for attribute change.
9051 * config/arm/fa626te.md (626te_alu_op): Update for attribute change.
9052 (626te_alu_shift_op): Likewise.
9053 * config/arm/fa726te.md (726te_shift_op): Update for attribute change.
9054 (726te_alu_op): Likewise.
9055 (726te_alu_shift_op): Likewise.
9056 (726te_alu_shift_reg_op): Likewise.
9057 * config/arm/fmp626.md (mp626_alu_op): Update for attribute change.
9058 (mp626_alu_shift_op): Likewise.
9059 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute change.
9060 (pj4_alu_e1_conds): Likewise.
9061 (pj4_alu): Likewise.
9062 (pj4_alu_conds): Likewise.
9063 (pj4_shift): Likewise.
9064 (pj4_shift_conds): Likewise.
9065 (pj4_alu_shift): Likewise.
9066 (pj4_alu_shift_conds): Likewise.
9068 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9070 * config/arm/predicates.md (shiftable_operator_strict_it):
9072 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si):
9073 Disable cond_exec version for arm_restrict_it.
9074 (thumb2_smaxsi3): Convert to generate cond_exec.
9075 (thumb2_sminsi3): Likewise.
9076 (thumb32_umaxsi3): Likewise.
9077 (thumb2_uminsi3): Likewise.
9078 (thumb2_abssi2): Adjust constraints for arm_restrict_it.
9079 (thumb2_neg_abssi2): Likewise.
9080 (thumb2_mov_scc): Add alternative for 16-bit encoding.
9081 (thumb2_movsicc_insn): Adjust alternatives.
9082 (thumb2_mov_negscc): Disable for arm_restrict_it.
9083 (thumb2_mov_negscc_strict_it): New pattern.
9084 (thumb2_mov_notscc_strict_it): New pattern.
9085 (thumb2_mov_notscc): Disable for arm_restrict_it.
9086 (thumb2_ior_scc): Likewise.
9087 (thumb2_ior_scc_strict_it): New pattern.
9088 (thumb2_cond_move): Adjust for arm_restrict_it.
9089 (thumb2_cond_arith): Disable for arm_restrict_it.
9090 (thumb2_cond_arith_strict_it): New pattern.
9091 (thumb2_cond_sub): Adjust for arm_restrict_it.
9092 (thumb2_movcond): Likewise.
9093 (thumb2_extendqisi_v6): Disable cond_exec variant for arm_restrict_it.
9094 (thumb2_zero_extendhisi2_v6): Likewise.
9095 (thumb2_zero_extendqisi2_v6): Likewise.
9096 (orsi_notsi_si): Likewise.
9097 (orsi_not_shiftsi_si): Likewise.
9099 2013-07-22 Georg-Johann Lay <avr@gjlay.de>
9101 * config/avr/avr.c (avr_out_xload): No SBIS around LPM so that
9102 instruction sequence is 1 byte shorter.
9104 2013-07-22 Uros Bizjak <ubizjak@gmail.com>
9106 * config/i386/i386.md (nonlocal_goto_receiver): Delete insn if
9107 it is not needed after split.
9109 2013-07-20 Iain Sandoe <iain@codesourcery.com>
9112 * config/i386/i386.c (output_set_got) [TARGET_MACHO]: Adjust to emit a
9113 second label for nonlocal goto receivers. Don't output pic base labels
9114 unless we're producing PIC; mark that action unreachable().
9115 (ix86_save_reg): If the function contains a nonlocal label, save the
9117 * config/darwin-protos.h (machopic_should_output_picbase_label): New.
9118 * config/darwin.c (emitted_pic_label_num): New GTY.
9119 (update_pic_label_number_if_needed): New.
9120 (machopic_output_function_base_name): Adjust for nonlocal receiver
9122 (machopic_should_output_picbase_label): New.
9123 * config/i386/i386.md (enum unspecv): UNSPECV_NLGR: New.
9124 (nonlocal_goto_receiver): New insn and split.
9126 2013-07-20 James Greenhalgh <james.greenhalgh@arm.com>
9128 * config/aarch64/aarch64-builtins.c
9129 (aarch64_fold_builtin): Fold abs in all modes.
9130 * config/aarch64/aarch64-simd-builtins.def
9131 (abs): Enable for all modes.
9132 * config/aarch64/arm_neon.h
9133 (vabs<q>_s<8,16,32,64): Rewrite using builtins.
9134 (vabs_f64): Add missing intrinsic.
9136 2013-07-19 Ian Bolton <ian.bolton@arm.com>
9138 * config/aarch64/arm_neon.h (vabs_s64): New function
9140 2013-07-19 Georg-Johann Lay <avr@gjlay.de>
9143 * config/avr/avr-fixed.md (round<mode>3_const): Turn expander to insn.
9144 * config/avr/avr.md (adjust_len): Add `round'.
9145 * config/avr/avr-protos.h (avr_out_round): New prototype.
9146 (avr_out_plus): Add `out_label' argument.
9147 * config/avr/avr.c (avr_out_plus_1): Add `out_label' argument.
9148 (avr_out_plus): Pass down `out_label' to avr_out_plus_1.
9149 Handle the case where `insn' is just a pattern.
9150 (avr_out_bitop): Handle the case where `insn' is just a pattern.
9151 (avr_out_round): New function.
9152 (avr_adjust_insn_length): Handle ADJUST_LEN_ROUND.
9154 2013-07-18 David Holsgrove <david.holsgrove@xilinx.com>
9156 * config/microblaze/microblaze.c (microblaze_expand_prologue):
9157 Add check for flag_stack_usage to handle -fstack-usage support
9159 2013-07-18 Pat Haugen <pthaugen@us.ibm.com>
9161 * config/rs6000/rs6000.c (rs6000_option_override_internal): Adjust flag
9162 interaction for new Power8 flags and VSX.
9164 2013-07-18 Sriraman Tallam <tmsriram@google.com>
9167 * tree-inline.c (expand_call_inline): Emit errors during
9168 early_inlining only if optimization is not turned on.
9170 2013-07-18 David Malcolm <dmalcolm@redhat.com>
9174 * passes.c (init_optimization_passes): Move the construction of
9175 the pass hierarchy into a new passes.def file.
9177 * Makefile.in (passes.o): Add dependency on passes.def.
9179 2013-07-18 David Malcolm <dmalcolm@redhat.com>
9181 * passes.c (init_optimization_passes): Introduce macros for
9182 constructing the tree of passes (INSERT_PASSES_AFTER,
9183 PUSH_INSERT_PASSES_WITHIN, POP_INSERT_PASSES,
9184 TERMINATE_PASS_LIST).
9186 2013-07-18 Vladimir Makarov <vmakarov@redhat.com>
9187 Wei Mi <wmi@google.com>
9189 PR rtl-optimization/57878
9190 * lra-assigns.c (assign_by_spills): Move non_reload_pseudos to the
9192 (reload_pseudo_compare_func): Check nregs first for reload
9195 2013-07-18 David Malcolm <dmalcolm@redhat.com>
9197 * tree-pass.h (pass_ipa_lto_wpa_fixup): Remove redundant decl.
9199 2013-07-18 Po-Chun Chang <pchang9@cs.wisc.edu>
9201 * read-rtl.c (validate_const_int): Once an invalid character is
9202 seen, quit the loop.
9204 * gengtype.c (write_roots): Similarly once we find the "deletable"
9205 or "if_marked" option.
9207 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
9209 * config/arm/arm.md (attribute "insn"): Delete values "mrs", "msr",
9210 "xtab" and "sat". Move value "clz" from here to ...
9211 (attriubte "type"): ... here.
9212 (satsi_<SAT:code>): Delete "insn" attribute.
9213 (satsi_<SAT:code>_shift): Likewise.
9214 (arm_zero_extendqisi2addsi): Likewise.
9215 (arm_extendqisi2addsi): Likewise.
9216 (clzsi2): Update for attribute changes.
9217 (rbitsi2): Likewise.
9218 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Delete "insn"
9220 (arm_usatsihi): Likewise.
9221 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
9223 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
9225 * config/arm/arm.md (attribute "type"): Rename "simple_alu_imm" to
9226 "arlo_imm". Rename "alu_reg" to "arlo_reg". Rename "simple_alu_shift"
9227 to "extend". Split "alu_shift" into "shift" and "arlo_shift". Split
9228 "alu_shift_reg" into "shift_reg" and "arlo_shift_reg". List types
9229 in alphabetical order.
9230 (attribute "core_cycles"): Update for attribute changes.
9231 (arm_addsi3): Likewise.
9232 (addsi3_compare0): Likewise.
9233 (addsi3_compare0_scratch): Likewise.
9234 (addsi3_compare_op1): Likewise.
9235 (addsi3_compare_op2): Likewise.
9236 (compare_addsi2_op0): Likewise.
9237 (compare_addsi2_op1): Likewise.
9238 (addsi3_carryin_shift_<optab>): Likewise.
9239 (subsi3_carryin_shift): Likewise.
9240 (rsbsi3_carryin_shift): Likewise.
9241 (arm_subsi3_insn): Likewise.
9242 (subsi3_compare0): Likewise.
9243 (subsi3_compare): Likewise.
9244 (arm_andsi3_insn): Likewise.
9245 (thumb1_andsi3_insn): Likewise.
9246 (andsi3_compare0): Likewise.
9247 (andsi3_compare0_scratch): Likewise.
9248 (zeroextractsi_compare0_scratch
9249 (andsi_not_shiftsi_si): Likewise.
9250 (iorsi3_insn): Likewise.
9251 (iorsi3_compare0): Likewise.
9252 (iorsi3_compare0_scratch): Likewise.
9253 (arm_xorsi3): Likewise.
9254 (thumb1_xorsi3_insn): Likewise.
9255 (xorsi3_compare0): Likewise.
9256 (xorsi3_compare0_scratch): Likewise.
9257 (satsi_<SAT:code>_shift): Likewise.
9259 (arm_shiftsi3): Likewise.
9260 (shiftsi3_compare0): Likewise.
9261 (not_shiftsi): Likewise.
9262 (not_shiftsi_compare0): Likewise.
9263 (not_shiftsi_compare0_scratch): Likewise.
9264 (arm_one_cmplsi2): Likewise.
9265 (thumb_one_complsi2): Likewise.
9266 (notsi_compare0): Likewise.
9267 (notsi_compare0_scratch): Likewise.
9268 (thumb1_zero_extendhisi2): Likewise.
9269 (arm_zero_extendhisi2): Likewise.
9270 (arm_zero_extendhisi2_v6): Likewise.
9271 (arm_zero_extendhisi2addsi): Likewise.
9272 (thumb1_zero_extendqisi2): Likewise.
9273 (thumb1_zero_extendqisi2_v6): Likewise.
9274 (arm_zero_extendqisi2): Likewise.
9275 (arm_zero_extendqisi2_v6): Likewise.
9276 (arm_zero_extendqisi2addsi): Likewise.
9277 (thumb1_extendhisi2): Likewise.
9278 (arm_extendhisi2): Likewise.
9279 (arm_extendhisi2_v6): Likewise.
9280 (arm_extendqisi): Likewise.
9281 (arm_extendqisi_v6): Likewise.
9282 (arm_extendqisi2addsi): Likewise.
9283 (thumb1_extendqisi2): Likewise.
9284 (thumb1_movdi_insn): Likewise.
9285 (arm_movsi_insn): Likewise.
9286 (movsi_compare0): Likewise.
9287 (movhi_insn_arch4): Likewise.
9288 (movhi_bytes): Likewise.
9289 (arm_movqi_insn): Likewise.
9290 (thumb1_movqi_insn): Likewise.
9291 (arm32_movhf): Likewise.
9292 (thumb1_movhf): Likewise.
9293 (arm_movsf_soft_insn): Likewise.
9294 (thumb1_movsf_insn): Likewise.
9295 (movdf_soft_insn): Likewise.
9296 (thumb_movdf_insn): Likewise.
9297 (arm_cmpsi_insn): Likewise.
9298 (cmpsi_shiftsi): Likewise.
9299 (cmpsi_shiftsi_swp): Likewise.
9300 (arm_cmpsi_negshiftsi_si): Likewise.
9301 (movsicc_insn): Likewise.
9302 (movsfcc_soft_insn): Likewise.
9303 (arith_shiftsi): Likewise.
9304 (arith_shiftsi_compare0
9305 (arith_shiftsi_compare0_scratch
9306 (sub_shiftsi): Likewise.
9307 (sub_shiftsi_compare0
9308 (sub_shiftsi_compare0_scratch
9309 (and_scc): Likewise.
9310 (cond_move): Likewise.
9311 (if_plus_move): Likewise.
9312 (if_move_plus): Likewise.
9313 (if_move_not): Likewise.
9314 (if_not_move): Likewise.
9315 (if_shift_move): Likewise.
9316 (if_move_shift): Likewise.
9317 (if_shift_shift): Likewise.
9318 (if_not_arith): Likewise.
9319 (if_arith_not): Likewise.
9320 (cond_move_not): Likewise.
9321 (thumb1_ashlsi3): Set type attribute.
9322 (thumb1_ashrsi3): Likewise.
9323 (thumb1_lshrsi3): Likewise.
9324 (thumb1_rotrsi3): Likewise.
9325 (shiftsi3_compare0_scratch): Likewise.
9326 * config/arm/neon.md (neon_mov<mode>): Update for attribute changes.
9327 (neon_mov<mode>): Likewise.
9328 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si): Update for
9330 (thumb2_movsi_insn): Likewise.
9331 (thumb2_cmpsi_neg_shiftsi): Likewise.
9332 (thumb2_extendqisi_v6): Likewise.
9333 (thumb2_zero_extendhisi2_v6): Likewise.
9334 (thumb2_zero_extendqisi2_v6): Likewise.
9335 (thumb2_shiftsi3_short): Likewise.
9336 (thumb2_addsi3_compare0_scratch): Likewise.
9337 (orsi_not_shiftsi_si): Likewise.
9338 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
9339 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Update for attribute
9341 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
9342 (1020alu_shift_op): Likewise.
9343 (1020alu_shift_reg_op): Likewise.
9344 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
9345 (alu_shift_op): Likewise.
9346 (alu_shift_reg_op): Likewise.
9347 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
9348 (11_alu_shift_op): Likewise.
9349 (11_alu_shift_reg_op): Likewise.
9350 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
9351 (9_alu_shift_reg_op): Likewise.
9352 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
9354 (cortex_a15_alu_shift): Likewise.
9355 (cortex_a15_alu_shift_reg): Likewise.
9356 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute
9358 (cortex_a5_alu_shift): Likewise.
9359 * config/arm/cortex-a53.md (cortex_a53_alu) : Update for attribute
9361 (cortex_a53_alu_shift): Likewise.
9362 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
9364 (cortex_a7_alu_reg): Likewise.
9365 (cortex_a7_alu_shift): Likewise.
9366 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute
9368 (cortex_a8_alu_shift): Likewise.
9369 (cortex_a8_alu_shift_reg): Likewise.
9370 (cortex_a8_mov): Likewise.
9371 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
9372 (cortex_a9_dp_shift): Likewise.
9373 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute
9375 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute
9377 (cortex_r4_mov): Likewise.
9378 (cortex_r4_alu_shift): Likewise.
9379 (cortex_r4_alu_shift_reg): Likewise.
9380 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
9381 (526_alu_shift_op): Likewise.
9382 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
9383 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
9384 (626te_alu_shift_op): Likewise.
9385 * config/arm/fa726te.md (726te_shift_op): Update for attribute changes.
9386 (726te_alu_op): Likewise.
9387 (726te_alu_shift_op): Likewise.
9388 (726te_alu_shift_reg_op): Likewise.
9389 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
9390 (mp626_alu_shift_op): Likewise.
9391 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute changes.
9392 (pj4_alu_e1_conds): Likewise.
9393 (pj4_alu): Likewise.
9394 (pj4_alu_conds): Likewise.
9395 (pj4_shift): Likewise.
9396 (pj4_shift_conds): Likewise.
9397 (pj4_alu_shift): Likewise.
9398 (pj4_alu_shift_conds): Likewise.
9399 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
9401 (cortexa7_older_only): Likewise.
9402 (cortexa7_younger): Likewise.
9404 2013-07-18 David Malcolm <dmalcolm@redhat.com>
9406 * ipa-pure-const.c (generate_summary): Rename to...
9407 (pure_const_generate_summary): ... this.
9409 2013-07-17 Iain Sandoe <iain@codesourcery.com>
9411 * config/rs6000/darwin.h (REGISTER_NAMES): Add HTM registers.
9413 2013-07-17 Yvan Roux <yvan.roux@linaro.org>
9416 * config/arm/arm.c (gen_movmem_ldrd_strd): Fix unaligned load/store
9419 2013-07-17 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
9421 * config/s390/s390.c: (s390_expand_builtin): Allow -mhtm to be
9422 enabled without -march=zEC12.
9423 * config/s390/s390.h (TARGET_HTM): Do not require EC12 machine
9426 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
9428 * config/mips/mips.h (ISA_HAS_FP4): Correct formatting.
9429 (ISA_HAS_FP_MADD4_MSUB4): Also enable for ISA_MIPS32R2.
9430 (ISA_HAS_NMADD4_NMSUB4): Remove the MODE argument; rewrite in
9431 terms of ISA_HAS_FP4, and also enable for ISA_MIPS32R2.
9432 (ISA_HAS_NMADD3_NMSUB3): Remove the MODE argument.
9433 * config/mips/mips.c (mips_rtx_costs) <PLUS>: Check for
9434 ISA_HAS_FP_MADD4_MSUB4 || ISA_HAS_FP_MADD3_MSUB3 rather than
9436 <MINUS, NEG>: Update according to changes to ISA_HAS_NMADD4_NMSUB4
9437 and ISA_HAS_NMADD3_NMSUB3.
9438 * config/mips/mips.md (nmadd4<mode>, nmadd3<mode>): Likewise.
9439 (nmadd4<mode>_fastmath, nmadd3<mode>_fastmath): Likewise.
9440 (nmsub4<mode>, nmsub3<mode>): Likewise.
9441 (nmsub4<mode>_fastmath, nmsub3<mode>_fastmath): Likewise.
9443 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
9445 * config/mips/mips.h (ISA_HAS_NMADD4_NMSUB4): Remove
9446 TARGET_MIPS5400 checking.
9448 2013-07-16 Jakub Jelinek <jakub@redhat.com>
9449 Peter Bergner <bergner@vnet.ibm.com>
9451 * config/rs6000/rs6000.h (FIRST_PSEUDO_REGISTERS): Mention HTM
9452 registers in the comment.
9453 (DWARF_FRAME_REGISTERS): Subtract also the 3 HTM registers.
9454 (DWARF_REG_TO_UNWIND_COLUMN): Use DWARF_FRAME_REGISTERS
9455 rather than FIRST_PSEUDO_REGISTERS.
9457 2013-07-16 Peter Bergner <bergner@vnet.ibm.com>
9459 * config/rs6000/rs6000.c (rs6000_option_override_internal): Do not
9460 enable extra ISA flags with TARGET_HTM.
9462 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
9464 * config/mips/mips.h (ISA_HAS_MULS, ISA_HAS_MSAC, ISA_HAS_MACC):
9467 2013-07-15 Cong Hou <congh@google.com>
9469 * tree-vect-data-refs.c (dr_group_sort_cmp): Do not use hash function
9470 in compare function for sorting.
9472 2013-07-15 Peter Bergner <bergner@vnet.ibm.com>
9474 * config.gcc (powerpc*-*-*): Install htmintrin.h and htmxlintrin.h.
9475 * config/rs6000/t-rs6000 (MD_INCLUDES): Add htm.md.
9476 * config/rs6000/rs6000.opt: Add -mhtm option.
9477 * config/rs6000/rs6000-cpus.def (POWERPC_MASKS): Add OPTION_MASK_HTM.
9478 (ISA_2_7_MASKS_SERVER): Add OPTION_MASK_HTM.
9479 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
9480 __HTM__ if the HTM instructions are available.
9481 * config/rs6000/predicates.md (u3bit_cint_operand, u10bit_cint_operand,
9482 htm_spr_reg_operand): New define_predicates.
9483 * config/rs6000/rs6000.md (define_attr "type"): Add htm.
9484 (TFHAR_REGNO, TFIAR_REGNO, TEXASR_REGNO): New define_constants.
9486 * config/rs6000/rs6000-builtin.def (BU_HTM_0, BU_HTM_1, BU_HTM_2,
9487 BU_HTM_3, BU_HTM_SPR0, BU_HTM_SPR1): Add support macros for defining
9488 HTM builtin functions.
9489 * config/rs6000/rs6000.c (RS6000_BUILTIN_H): New macro.
9490 (rs6000_reg_names, alt_reg_names): Add HTM SPR register names.
9491 (rs6000_init_hard_regno_mode_ok): Add support for HTM instructions.
9492 (rs6000_builtin_mask_calculate): Likewise.
9493 (rs6000_option_override_internal): Likewise.
9494 (bdesc_htm): Add new HTM builtin support.
9495 (htm_spr_num): New function.
9496 (htm_spr_regno): Likewise.
9497 (rs6000_htm_spr_icode): Likewise.
9498 (htm_expand_builtin): Likewise.
9499 (htm_init_builtins): Likewise.
9500 (rs6000_expand_builtin): Add support for HTM builtin functions.
9501 (rs6000_init_builtins): Likewise.
9502 (rs6000_invalid_builtin, rs6000_opt_mask): Add support for -mhtm
9504 * config/rs6000/rs6000.h (ASM_CPU_SPEC): Add support for -mhtm.
9505 (TARGET_HTM, MASK_HTM): Define macros.
9506 (FIRST_PSEUDO_REGISTER): Adjust for new HTM SPR registers.
9507 (FIXED_REGISTERS): Likewise.
9508 (CALL_USED_REGISTERS): Likewise.
9509 (CALL_REALLY_USED_REGISTERS): Likewise.
9510 (REG_ALLOC_ORDER): Likewise.
9511 (enum reg_class): Likewise.
9512 (REG_CLASS_NAMES): Likewise.
9513 (REG_CLASS_CONTENTS): Likewise.
9514 (REGISTER_NAMES): Likewise.
9515 (ADDITIONAL_REGISTER_NAMES): Likewise.
9516 (RS6000_BTC_SPR, RS6000_BTC_VOID, RS6000_BTC_32BIT, RS6000_BTC_64BIT,
9517 RS6000_BTC_MISC_MASK, RS6000_BTM_HTM): New macros.
9518 (RS6000_BTM_COMMON): Add RS6000_BTM_HTM.
9519 * config/rs6000/htm.md: New file.
9520 * config/rs6000/htmintrin.h: New file.
9521 * config/rs6000/htmxlintrin.h: New file.
9523 2013-07-15 Marcus Shawcroft <marcus.shawcroft@arm.com>
9525 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
9526 Define SYMBOL_TINY_GOT, update comment.
9527 * config/aarch64/aarch64.c
9528 (aarch64_load_symref_appropriately): Handle SYMBOL_TINY_GOT.
9529 (aarch64_expand_mov_immediate): Likewise.
9530 (aarch64_print_operand): Likewise.
9531 (aarch64_classify_symbol): Likewise.
9532 * config/aarch64/aarch64.md (UNSPEC_GOTTINYPIC): Define.
9533 (ldr_got_tiny): Define.
9535 2013-07-13 Tobias Grosser <tobias@grosser.es>
9537 PR tree-optimization/54094
9538 * graphite-clast-to-gimple.c (translate_clast_for_loop): Derive the
9539 scheduling dimension for the parallelism check from the polyhedral
9540 information in the AST.
9541 * graphite-dependences.c (carries_deps): Do not assume the schedule is
9544 2013-07-13 Jason Merrill <jason@redhat.com>
9546 * print-tree.c (debug_vec_tree): Use debug_raw.
9547 (debug_raw (vec<tree, va_gc> &)): New.
9548 (debug_raw (vec<tree, va_gc> *)): New.
9549 * tree.h: Declare them.
9551 2013-07-13 Bin Cheng <bin.cheng@arm.com>
9553 * ifcvt.c (ifcvt_after_combine): New static variable.
9554 (cheap_bb_rtx_cost_p): Set scale to REG_BR_PROB_BASE when optimizing
9556 (if_convert): New parameter after_combine. Set ifcvt_after_combine.
9557 (rest_of_handle_if_conversion, rest_of_handle_if_after_combine,
9558 rest_of_handle_if_after_reload): Pass new argument for if_convert.
9560 2013-07-12 Maciej W. Rozycki <macro@codesourcery.com>
9562 * config/mips/mips.c (mips_expand_call): Remove empty statement.
9564 2013-07-12 Michael Matz <matz@suse.de>
9567 * convert.c (convert_to_real): Reject non-float inner types.
9569 2013-07-12 Tejas Belagod <tejas.belagod@arm.com>
9571 * config/aarch64/aarch64-protos.h
9572 (aarch64_simd_immediate_valid_for_move): Remove.
9573 * config/aarch64/aarch64.c (simd_immediate_info): New member.
9574 (aarch64_simd_valid_immediate): Recognize idioms for shifting ones
9576 (aarch64_output_simd_mov_immediate): Print the correct shift specifier.
9578 2013-07-11 Steve Ellcey <sellcey@mips.com>
9580 * config/mips/mips.c (mips_conditional_register_usage): Do not
9581 use t[0-7] registers in MIPS16 mode when optimizing for size.
9583 2013-07-11 Sriraman Tallam <tmsriram@google.com>
9585 * config/i386/i386.c (dispatch_function_versions): Fix array
9586 indexing of function_version_info to match actual_versions.
9588 2013-07-11 Teresa Johnson <tejohnson@google.com>
9590 * vec.h (struct va_gc): Move release out-of-line.
9591 (va_gc::release): Call ggc_free on released vec.
9593 2013-07-11 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
9595 * config/rs6000/rs6000.md (""*tls_gd_low<TLSmode:tls_abi_suffix>"):
9596 Require GOT register as additional operand in UNSPEC.
9597 ("*tls_ld_low<TLSmode:tls_abi_suffix>"): Likewise.
9598 ("*tls_got_dtprel_low<TLSmode:tls_abi_suffix>"): Likewise.
9599 ("*tls_got_tprel_low<TLSmode:tls_abi_suffix>"): Likewise.
9600 ("*tls_gd<TLSmode:tls_abi_suffix>"): Update splitter.
9601 ("*tls_ld<TLSmode:tls_abi_suffix>"): Likewise.
9602 ("tls_got_dtprel_<TLSmode:tls_abi_suffix>"): Likewise.
9603 ("tls_got_tprel_<TLSmode:tls_abi_suffix>"): Likewise.
9605 2013-07-11 Georg-Johann Lay <avr@gjlay.de>
9608 * config/avr/avr.c (avr_set_current_function): Sanity-check signal
9609 name seen by assembler/linker rather if available.
9611 2013-07-11 Andreas Schwab <schwab@suse.de>
9613 * config/aarch64/aarch64-linux.h (CPP_SPEC): Define.
9615 2013-07-10 Vladimir Makarov <vmakarov@redhat.com>
9617 * lra-constraints.c (curr_insn_transform): Switch off optional reloads.
9619 2013-07-10 Joseph Myers <joseph@codesourcery.com>
9621 * doc/tm.texi.in: Move hook documentation to ....
9622 * target.def: ... here.
9624 * doc/tm.texi.in (TARGET_CANONICALIZE_COMPARISON): Remove stray
9626 * doc/tm.texi: Regenerate.
9628 2013-07-10 Paolo Carlini <paolo.carlini@oracle.com>
9631 * doc/invoke.texi: Document -Wconditionally-supported.
9633 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
9636 * config/avr/avr.c (avr_prologue_setup_frame): Trunk -size to mode
9639 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
9642 * config/avr/avr-mcus.def (atmega16hva, atmega16hva2, atmega16hvb)
9643 (atmega16m1, atmega16u4, atmega32a, atmega32c1, atmega32hvb)
9644 (atmega32m1, atmega32u4, atmega32u6, atmega64c1, atmega64m1):
9645 Remove duplicate devices.
9646 * config/avr/gen-avr-mmcu-texi.c (print_mcus): Fail on duplicate MCUs.
9647 * config/avr/t-multilib: Regenerate.
9648 * config/avr/avr-tables.opt: Regenerate.
9649 * doc/avr-mmcu.texi: Regenerate.
9651 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
9654 * config/avr/avr.opt (Waddr-space-convert): Fix typo.
9656 2013-07-10 Graham Stott <graham.stott@btinternet.com>
9658 * config/mips/mips.c (mips_rtx_costs): Very slightly increase
9659 the cost of MULT when optimizing for size.
9661 2013-07-10 Jan-Benedict Glaw <jbglaw@lug-owl.de>
9663 * config/cr16/cr16-protos.h: Don't include target.h.
9665 2013-07-09 Joseph Myers <joseph@codesourcery.com>
9667 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Only
9668 adjust register size for TDmode and TFmode for VSX registers.
9670 2013-07-08 Kai Tietz <ktietz@redhat.com>
9673 * config/i386/i386.c (TARGET_FUNCTION_ATTRIBUTE_INLINABLE_P): Define as
9674 hook_bool_const_tree_true.
9676 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
9678 * config/s390/s390.c: Replace F*_REGNUM with FPR*_REGNUM.
9679 * config/s390/s390.h: Remove F*_REGNUM macro definitions.
9680 * config/s390/s390.md: Define FPR*_REGNUM constants.
9681 Fix FPR2_REGNUM constant (18 -> 17).
9682 ("*trunc<BFP:mode><DFP_ALL:mode>2")
9683 ("*trunc<DFP_ALL:mode><BFP:mode>2")
9684 ("trunc<BFP:mode><DFP_ALL:mode>2")
9685 ("trunc<DFP_ALL:mode><BFP:mode>2")
9686 ("*extend<BFP:mode><DFP_ALL:mode>2")
9687 ("*extend<DFP_ALL:mode><BFP:mode>2")
9688 ("extend<BFP:mode><DFP_ALL:mode>2")
9689 ("extend<DFP_ALL:mode><BFP:mode>2"): Replace FPR2_REGNUM with
9692 2013-07-08 Graham Stott <graham.stott@btinternet.com>
9694 * Makefile.in: (c-family-warn): Define to $(STRICT_WARN)
9696 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
9698 * config/s390/s390.c: Rename cfun_set_fpr_bit to cfun_set_fpr_save
9699 and cfun_fpr_bit_p to cfun_fpr_save_p.
9700 (s390_frame_area, s390_register_info, s390_frame_info)
9701 (s390_emit_prologue, s390_emit_epilogue)
9702 (s390_conditional_register_usage): Use the *_REGNUM macros for FPR
9704 * config/s390/s390.h: Define *_REGNUM macros for floating point
9707 2013-07-08 Eric Botcazou <ebotcazou@adacore.com>
9709 * Makefile.in (tree-ssa-reassoc.o): Add dependency on $(PARAMS_H).
9711 2013-07-08 Po-Chun Chang <pchang9@cs.wisc.edu>
9713 PR rtl-optimization/57786
9714 * combine.c (distribute_notes) <case REG_DEAD>: Change all_used to bool
9715 and break out of the loop when it is set to false.
9717 2013-07-08 Jakub Jelinek <jakub@redhat.com>
9720 * simplify-rtx.c (simplify_unary_operation_1) <case ZERO_EXTEND>:
9721 Simplify (zero_extend:SI (subreg:QI (and:SI (reg:SI)
9722 (const_int 63)) 0)).
9723 * combine.c (make_extraction): Create ZERO_EXTEND or SIGN_EXTEND
9724 using simplify_gen_unary instead of gen_rtx_*_EXTEND.
9725 * config/i386/i386.md (*jcc_bt<mode>_1): New define_insn_and_split.
9727 PR rtl-optimization/57829
9728 * simplify-rtx.c (simplify_binary_operation_1) <case IOR>: Ensure that
9729 mask bits outside of mode are just sign-extension from mode to HWI.
9731 2013-07-08 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
9733 * config/i386/i386-opts.h (enum stringop_alg): Add vector_loop.
9734 * config/i386/i386.c (expand_set_or_movmem_via_loop): Use
9735 adjust_address instead of change_address to keep info about alignment.
9736 (emit_strmov): Remove.
9737 (emit_memmov): New function.
9738 (expand_movmem_epilogue): Refactor to properly handle bigger sizes.
9739 (expand_movmem_epilogue): Likewise and return updated rtx for
9741 (expand_constant_movmem_prologue): Likewise and return updated rtx for
9742 destination and source.
9743 (decide_alignment): Refactor, handle vector_loop.
9744 (ix86_expand_movmem): Likewise.
9745 (ix86_expand_setmem): Likewise.
9746 * config/i386/i386.opt (Enum): Add vector_loop to option stringop_alg.
9748 2013-07-07 Uros Bizjak <ubizjak@gmail.com>
9750 * config/i386/driver-i386.c (host_detect_local_cpu): Do not check
9751 signature_TM2_ebx, it interferes with signature_INTEL_ebx.
9753 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
9755 * config/i386/sse.md (sse_movlhps): Change alternative 3
9756 of operand 2 to "m".
9758 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
9761 * config/i386/sse.md (iptr): New mode attribute.
9762 (sse2_movq128): Add pointer size overrides for Intel asm dialect.
9763 (<sse>_vm<plusminus_insn><mode>3): Ditto.
9764 (<sse>_vmmul<mode>3): Ditto.
9765 (<sse>_vmdiv<mode>3): Ditto.
9766 (sse_vmrcpv4sf2): Ditto.
9767 (<sse>_vmsqrt<mode>2): Ditto.
9768 (sse_vmrsqrtv4sf2): Ditto.
9769 (<sse>_vm<code><mode>3): Ditto.
9770 (avx_vmcmp<mode>3): Ditto.
9771 (<sse>_vmmaskcmp<mode>3): Ditto.
9772 (<sse>_comi): Ditto.
9773 (<sse>_ucomi): Ditto.
9774 (*xop_vmfrcz_<mode>): Ditto.
9775 (*fmai_fmadd_<mode>): Ditto.
9776 (*fmai_fmsub_<mode>): Ditto.
9777 (*fmai_fnmadd_<mode>): Ditto.
9778 (*fmai_fnmsub_<mode>): Ditto.
9779 (*fma4i_vmfmadd_<mode>): Ditto.
9780 (*fma4i_vmfmsub_<mode>): Ditto.
9781 (*fma4i_vmfnmadd_<mode>): Ditto.
9782 (*fma4i_vmfnmsub_<mode>): Ditto.
9783 (*xop_vmfrcz_<mode>): Ditto.
9784 (sse_cvtps2pi): Ditto.
9785 (sse_cvttps2pi): Ditto.
9786 (sse_cvtss2si): Ditto.
9787 (sse_cvtss2si_2): Ditto.
9788 (sse_cvtss2siq_2): Ditto.
9789 (sse_cvttss2si): Ditto.
9790 (sse_cvttss2siq): Ditto.
9791 (sse_cvtsd2si): Ditto.
9792 (sse_cvtsd2si_2): Ditto.
9793 (sse_cvtsd2siq_2): Ditto.
9794 (sse_cvttsd2si): Ditto.
9795 (sse_cvttsd2siq): Ditto.
9796 (sse_cvtsd2ss): Ditto.
9797 (sse_cvtss2sd): Ditto.
9798 (avx2_pbroadcast<mode>): Ditto.
9799 (avx2_pbroadcast<mode>_1): Ditto.
9800 (*avx_vperm_broadcast_v4sf): Ditto.
9802 (sse_movhlps): Ditto for movlp[sd]/movhp[sd] alternatives.
9803 (sse_movlhps): Ditto.
9804 (sse_storehps): Ditto.
9805 (sse_loadhps): Ditto.
9806 (sse_storelps): Ditto.
9807 (sse_loadlps): Ditto.
9808 (*vec_concatv4sf): Ditto.
9809 (*vec_interleave_highv2df): Ditto.
9810 (*vec_interleave_lowv2df): Ditto.
9811 (*vec_extractv2df_1_sse): Ditto.
9812 (*vec_extractv2df_0_sse): Ditto.
9813 (sse2_storelpd): Ditto.
9814 (sse2_loadlpd): Ditto.
9815 (sse2_movsd): Ditto.
9816 (*vec_concatv4si): Ditto.
9817 (vec_concatv2di): Ditto.
9819 * config/i386/mmx.md (mmx_punpcklbw): Add pointer size overrides
9820 for Intel asm dialect.
9821 (mmx_punpcklwd): Ditto.
9822 (mmx_punpckldq): Ditto.
9824 * config/i386/i386.c (ix86_print_operand) ['H']: Output 'qword ptr'
9825 for intel assembler dialect.
9827 2013-07-06 Jakub Jelinek <jakub@redhat.com>
9830 * fold-const.c (tree_call_nonnegative_warnv_p): Return true
9831 for BUILT_IN_C{LZ,LRSB}*.
9832 * tree.h (CASE_INT_FN): Add FN##IMAX case.
9833 * tree-vrp.c (extract_range_basic): Handle
9834 BUILT_IN_{FFS,PARITY,POPCOUNT,C{LZ,TZ,LRSB}}*. For
9835 BUILT_IN_CONSTANT_P if argument isn't (D) of PARM_DECL,
9836 fall thru to code calling set_value*.
9837 * builtins.c (expand_builtin): Remove *IMAX cases.
9838 (fold_builtin_bitop): For BUILT_IN_CLRSB* return NULL_TREE
9839 if width is bigger than 2*HWI.
9841 2013-07-05 Vladimir Makarov <vmakarov@redhat.com>
9843 PR rtl-optimization/55342
9844 * lra-int.h (lra_subreg_reload_pseudos): New.
9845 * lra.c: Add undoing optional reloads to the block diagram.
9846 (lra_subreg_reload_pseudos): New.
9847 (lra_optional_reload_pseudos): Change comments.
9848 (lra): Init and clear lra_subreg_reload_pseudos. Clear
9849 lra_optional_reload_pseudos after undo transformations.
9850 * lra-assigns.c (pseudo_prefix_title): New.
9851 (lra_setup_reg_renumber): Use it.
9852 (spill_for): Ditto. Check subreg reload pseudos too.
9853 (assign_by_spills): Consider subreg reload pseudos too.
9854 * lra-constraints.c (simplify_operand_subreg): Use
9855 lra_subreg_reload_pseudos instead of lra_optional_reload_pseudos.
9856 (curr_insn_transform): Recognize and do optional reloads.
9857 (undo_optional_reloads): New.
9858 (lra_undo_inheritance): Call undo_optional_reloads.
9860 2013-07-05 Thomas Quinot <quinot@adacore.com>
9862 * tree-complex.c (expand_complex_operations_1): Fix typo.
9864 2013-07-04 Tejas Belagod <tejas.belagod@arm.com>
9866 * config/aarch64/aarch64-protos.h (cpu_vector_cost): New.
9867 (tune_params): New member 'const vec_costs'.
9868 * config/aarch64/aarch64.c (generic_vector_cost): New.
9869 (generic_tunings): New member 'generic_vector_cost'.
9870 (aarch64_builtin_vectorization_cost): New.
9871 (aarch64_add_stmt_cost): New.
9872 (TARGET_VECTORIZE_ADD_STMT_COST): New.
9873 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST): New.
9875 2013-07-03 Jakub Jelinek <jakub@redhat.com>
9878 * config/i386/predicates.md (vsib_address_operand): Disallow
9879 SYMBOL_REF or LABEL_REF in parts.disp if TARGET_64BIT && flag_pic.
9881 2013-07-03 Hans-Peter Nilsson <hp@bitrange.com>
9884 * stmt.c (expand_nl_goto_receiver): Remove almost-copy of
9885 expand_builtin_setjmp_receiver.
9886 (expand_label): Adjust, call expand_builtin_setjmp_receiver
9887 with NULL for the label parameter.
9888 * builtins.c (expand_builtin_setjmp_receiver): Don't clobber
9889 the frame-pointer. Adjust comments.
9890 [HAVE_builtin_setjmp_receiver]: Emit builtin_setjmp_receiver
9891 only if LABEL is non-NULL.
9893 2013-07-03 Yufeng Zhang <yufeng.zhang@arm.com>
9895 * config/aarch64/aarch64.h (enum arm_abi_type): Remove.
9896 (ARM_ABI_AAPCS64): Ditto.
9898 (ARM_DEFAULT_ABI): Ditto.
9900 2013-07-03 James Greenhalgh <james.greenhalgh@arm.com>
9902 * config/aarch64/aarch64-builtins.c
9903 (aarch64_simd_expand_builtin): Handle AARCH64_SIMD_STORE1.
9904 * config/aarch64/aarch64-simd-builtins.def (ld1): New.
9906 * config/aarch64/aarch64-simd.md
9907 (aarch64_ld1<VALL:mode>): New.
9908 (aarch64_st1<VALL:mode>): Likewise.
9909 * config/aarch64/arm_neon.h
9910 (vld1<q>_<fpsu><8, 16, 32, 64>): Convert to RTL builtins.
9912 2013-07-02 Sriraman Tallam <tmsriram@google.com>
9914 * config/i386/i386.c (gate_insert_vzeroupper): Check if
9916 (ix86_option_override_internal):Turn on all -mavx target flags by
9917 default as they are dependent on AVX anyway.
9919 2013-07-02 Cary Coutant <ccoutant@google.com>
9921 * dwarf2out.c (loc_checksum): Call hash_loc_operands for a
9923 (loc_checksum_ordered): Likewise.
9924 (hash_loc_operands): Remove inline keyword.
9926 2013-07-02 Jakub Jelinek <jakub@redhat.com>
9928 PR tree-optimization/57741
9929 * tree-vect-loop.c (vect_is_simple_iv_evolution): Disallow
9930 non-INTEGRAL_TYPE_P non-SCALAR_FLOAT_TYPE_P SSA_NAME step_exprs,
9931 or SCALAR_FLOAT_TYPE_P SSA_NAMEs if !flag_associative_math.
9932 Allow REAL_CST step_exprs if flag_associative_math.
9933 (get_initial_def_for_induction): Handle SCALAR_FLOAT_TYPE_P step_expr.
9935 2013-07-02 Ian Bolton <ian.bolton@arm.com>
9937 * config/aarch64/aarch64-simd.md (absdi2): Support abs for DI mode.
9939 2013-07-02 Ian Bolton <ian.bolton@arm.com>
9941 * config/aarch64/aarch64.md (*extr_insv_reg<mode>): New pattern.
9943 2013-07-02 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9945 * config/arm/arm.md (arm_andsi3_insn): Add alternatives for 16-bit
9947 (iorsi3_insn): Likewise.
9948 (arm_xorsi3): Likewise.
9950 2013-07-01 Sofiane Naci <sofiane.naci@arm.com>
9952 * arm.md (attribute "wtype"): Delete. Move attribute values from here
9954 (attribute "type"): ... here, and prefix with "wmmx_".
9955 (attribute "core_cycles"): Update for attribute changes.
9956 * iwmmxt.md (tbcstv8qi): Update for attribute changes.
9957 (tbcstv4hi): Likewise.
9958 (tbcstv2si): Likewise.
9959 (iwmmxt_iordi3): Likewise.
9960 (iwmmxt_xordi3): Likewise.
9961 (iwmmxt_anddi3): Likewise.
9962 (iwmmxt_nanddi3): Likewise.
9963 (iwmmxt_arm_movdi): Likewise.
9964 (iwmmxt_movsi_insn): Likewise.
9965 (mov<mode>_internal): Likewise.
9966 (and<mode>3_iwmmxt): Likewise.
9967 (ior<mode>3_iwmmxt): Likewise.
9968 (xor<mode>3_iwmmxt): Likewise.
9969 (add<mode>3_iwmmxt): Likewise.
9970 (ssaddv8qi3): Likewise.
9971 (ssaddv4hi3): Likewise.
9972 (ssaddv2si3): Likewise.
9973 (usaddv8qi3): Likewise.
9974 (usaddv4hi3): Likewise.
9975 (usaddv2si3): Likewise.
9976 (sub<mode>3_iwmmxt): Likewise.
9977 (sssubv8qi3): Likewise.
9978 (sssubv4hi3): Likewise.
9979 (sssubv2si3): Likewise.
9980 (ussubv8qi3): Likewise.
9981 (ussubv4hi3): Likewise.
9982 (ussubv2si3): Likewise.
9983 (mulv4hi3_iwmmxt): Likewise.
9984 (smulv4hi3_highpart): Likewise.
9985 (umulv4hi3_highpart): Likewise.
9986 (iwmmxt_wmacs): Likewise.
9987 (iwmmxt_wmacsz): Likewise.
9988 (iwmmxt_wmacu): Likewise.
9989 (iwmmxt_wmacuz): Likewise.
9990 (iwmmxt_clrdi): Likewise.
9991 (iwmmxt_clrv8qi): Likewise.
9992 (iwmmxt_clr4hi): Likewise.
9993 (iwmmxt_clr2si): Likewise.
9994 (iwmmxt_uavgrndv8qi3): Likewise.
9995 (iwmmxt_uavgrndv4hi3): Likewise.
9996 (iwmmxt_uavgv8qi3): Likewise.
9997 (iwmmxt_uavgv4hi3): Likewise.
9998 (iwmmxt_tinsrb): Likewise.
9999 (iwmmxt_tinsrh): Likewise.
10000 (iwmmxt_tinsrw): Likewise.
10001 (iwmmxt_textrmub): Likewise.
10002 (iwmmxt_textrmsb): Likewise.
10003 (iwmmxt_textrmuh): Likewise.
10004 (iwmmxt_textrmsh): Likewise.
10005 (iwmmxt_textrmw): Likewise.
10006 (iwmxxt_wshufh): Likewise.
10007 (eqv8qi3): Likewise.
10008 (eqv4hi3): Likewise.
10009 (eqv2si3): Likewise.
10010 (gtuv8qi3): Likewise.
10011 (gtuv4hi3): Likewise.
10012 (gtuv2si3): Likewise.
10013 (gtv8qi3): Likewise.
10014 (gtv4hi3): Likewise.
10015 (gtv2si3): Likewise.
10016 (smax<mode>3_iwmmxt): Likewise.
10017 (umax<mode>3_iwmmxt): Likewise.
10018 (smin<mode>3_iwmmxt): Likewise.
10019 (umin<mode>3_iwmmxt): Likewise.
10020 (iwmmxt_wpackhss): Likewise.
10021 (iwmmxt_wpackwss): Likewise.
10022 (iwmmxt_wpackdss): Likewise.
10023 (iwmmxt_wpackhus): Likewise.
10024 (iwmmxt_wpackwus): Likewise.
10025 (iwmmxt_wpackdus): Likewise.
10026 (iwmmxt_wunpckihb): Likewise.
10027 (iwmmxt_wunpckihh): Likewise.
10028 (iwmmxt_wunpckihw): Likewise.
10029 (iwmmxt_wunpckilb): Likewise.
10030 (iwmmxt_wunpckilh): Likewise.
10031 (iwmmxt_wunpckilw): Likewise.
10032 (iwmmxt_wunpckehub): Likewise.
10033 (iwmmxt_wunpckehuh): Likewise.
10034 (iwmmxt_wunpckehuw): Likewise.
10035 (iwmmxt_wunpckehsb): Likewise.
10036 (iwmmxt_wunpckehsh): Likewise.
10037 (iwmmxt_wunpckehsw): Likewise.
10038 (iwmmxt_wunpckelub): Likewise.
10039 (iwmmxt_wunpckeluh): Likewise.
10040 (iwmmxt_wunpckeluw): Likewise.
10041 (iwmmxt_wunpckelsb): Likewise.
10042 (iwmmxt_wunpckelsh): Likewise.
10043 (iwmmxt_wunpckelsw): Likewise.
10044 (ror<mode>3): Likewise.
10045 (ashr<mode>3_iwmmxt): Likewise.
10046 (lshr<mode>3_iwmmxt): Likewise.
10047 (ashl<mode>3_iwmmxt): Likewise.
10048 (ror<mode>3_di): Likewise.
10049 (ashr<mode>3_di): Likewise.
10050 (lshr<mode>3_di): Likewise.
10051 (ashl<mode>3_di): Likewise.
10052 (iwmmxt_wmadds): Likewise.
10053 (iwmmxt_wmaddu): Likewise.
10054 (iwmmxt_tmia): Likewise.
10055 (iwmmxt_tmiaph): Likewise.
10056 (iwmmxt_tmiabb): Likewise.
10057 (iwmmxt_tmiatb): Likewise.
10058 (iwmmxt_tmiabt): Likewise.
10059 (iwmmxt_tmiatt): Likewise.
10060 (iwmmxt_tmovmskb): Likewise.
10061 (iwmmxt_tmovmskh): Likewise.
10062 (iwmmxt_tmovmskw): Likewise.
10063 (iwmmxt_waccb): Likewise.
10064 (iwmmxt_wacch): Likewise.
10065 (iwmmxt_waccw): Likewise.
10066 (iwmmxt_waligni): Likewise.
10067 (iwmmxt_walignr): Likewise.
10068 (iwmmxt_walignr0): Likewise.
10069 (iwmmxt_walignr1): Likewise.
10070 (iwmmxt_walignr2): Likewise.
10071 (iwmmxt_walignr3): Likewise.
10072 (iwmmxt_wsadb): Likewise.
10073 (iwmmxt_wsadh): Likewise.
10074 (iwmmxt_wsadbz): Likewise.
10075 (iwmmxt_wsadhz): Likewise.
10076 * iwmmxt2.md (iwmmxt_wabs<mode>3): Update for attribute changes.
10077 (iwmmxt_wabsdiffb): Likewise.
10078 (iwmmxt_wabsdiffh): Likewise.
10079 (iwmmxt_wabsdiffw): Likewise.
10080 (iwmmxt_waddsubhx): Likewise
10081 (iwmmxt_wsubaddhx): Likewise.
10082 (addc<mode>3): Likewise.
10083 (iwmmxt_avg4): Likewise.
10084 (iwmmxt_avg4r): Likewise.
10085 (iwmmxt_wmaddsx): Likewise.
10086 (iwmmxt_wmaddux): Likewise.
10087 (iwmmxt_wmaddsn): Likewise.
10088 (iwmmxt_wmaddun): Likewise.
10089 (iwmmxt_wmulwsm): Likewise.
10090 (iwmmxt_wmulwum): Likewise.
10091 (iwmmxt_wmulsmr): Likewise.
10092 (iwmmxt_wmulumr): Likewise.
10093 (iwmmxt_wmulwsmr): Likewise.
10094 (iwmmxt_wmulwumr): Likewise.
10095 (iwmmxt_wmulwl): Likewise.
10096 (iwmmxt_wqmulm): Likewise.
10097 (iwmmxt_wqmulwm): Likewise.
10098 (iwmmxt_wqmulmr): Likewise.
10099 (iwmmxt_wqmulwmr): Likewise.
10100 (iwmmxt_waddbhusm): Likewise.
10101 (iwmmxt_waddbhusl): Likewise.
10102 (iwmmxt_wqmiabb): Likewise.
10103 (iwmmxt_wqmiabt): Likewise.
10104 (iwmmxt_wqmiatb): Likewise.
10105 (iwmmxt_wqmiatt): Likewise.
10106 (iwmmxt_wqmiabbn): Likewise.
10107 (iwmmxt_wqmiabtn): Likewise.
10108 (iwmmxt_wqmiatbn): Likewise.
10109 (iwmmxt_wqmiattn): Likewise.
10110 (iwmmxt_wmiabb): Likewise.
10111 (iwmmxt_wmiabt): Likewise.
10112 (iwmmxt_wmiatb): Likewise.
10113 (iwmmxt_wmiatt): Likewise.
10114 (iwmmxt_wmiabbn): Likewise.
10115 (iwmmxt_wmiabtn): Likewise.
10116 (iwmmxt_wmiatbn): Likewise.
10117 (iwmmxt_wmiattn): Likewise.
10118 (iwmmxt_wmiawbb): Likewise.
10119 (iwmmxt_wmiawbt): Likewise.
10120 (iwmmxt_wmiawtb): Likewise.
10121 (iwmmxt_wmiawtt): Likewise.
10122 (iwmmxt_wmiawbbn): Likewise.
10123 (iwmmxt_wmiawbtn): Likewise.
10124 (iwmmxt_wmiawtbn): Likewise.
10125 (iwmmxt_wmiawttn): Likewise.
10126 (iwmmxt_wmerge): Likewise.
10127 (iwmmxt_tandc<mode>3): Likewise.
10128 (iwmmxt_torc<mode>3): Likewise.
10129 (iwmmxt_torvsc<mode>3): Likewise.
10130 (iwmmxt_textrc<mode>3): Likewise.
10131 * marvell-f-iwmmxt.md (wmmxt_shift): Update for attribute changes.
10132 (wmmxt_pack): Likewise.
10133 (wmmxt_mult_c1): Likewise.
10134 (wmmxt_mult_c2): Likewise.
10135 (wmmxt_alu_c1): Likewise.
10136 (wmmxt_alu_c2): Likewise.
10137 (wmmxt_alu_c3): Likewise.
10138 (wmmxt_transfer_c1): Likewise.
10139 (wmmxt_transfer_c2): Likewise.
10140 (wmmxt_transfer_c3): Likewise.
10141 (marvell_f_iwmmxt_wstr): Likewise.
10142 (marvell_f_iwmmxt_wldr): Likewise.
10144 2013-06-29 Yufeng Zhang <yufeng.zhang@arm.com>
10146 * config/aarch64/aarch64.c: Remove junk from the beginning of the file.
10148 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
10151 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
10152 * lra-constraints.c (need_for_split_p): Check call used hard regs
10153 living through calls.
10155 * lra-constraints.c (inherit_in_ebb): Reset live_hard_regs for
10156 call used regs for call insn.
10158 2013-06-28 Jakub Jelinek <jakub@redhat.com>
10161 * config/i386/i386.c (ix86_expand_builtin): If target == NULL and
10162 mode is VOIDmode, don't create a VOIDmode pseudo to copy result into.
10164 2013-06-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
10166 * builtins.def: Fixed the function type of CILKPLUS_BUILTIN.
10168 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
10170 * lra-constraints.c (need_for_split_p): Check call used hard regs
10171 living through calls.
10173 2013-06-28 Michael Meissner <meissner@linux.vnet.ibm.com>
10176 * config/rs6000/rs6000.h (MODES_TIEABLE_P): Do not allow PTImode
10177 to tie with any other modes. Eliminate Altivec vector mode tests,
10178 since these are a subset of ALTIVEC or VSX vector modes. Simplify
10179 code, to return 0 if testing MODE2 for a condition, if we've
10180 already tested MODE1 for the same condition.
10182 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
10184 * config/aarch64/aarch64.c (aarch64_cannot_force_const_mem): Adjust
10187 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
10189 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
10190 Update comment w.r.t SYMBOL_TINY_ABSOLUTE.
10192 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
10194 * config/aarch64/aarch64-protos.h (aarch64_classify_symbol_expression):
10196 (aarch64_symbolic_constant_p): Remove.
10197 * config/aarch64/aarch64.c (aarch64_classify_symbol_expression): Remove
10198 static. Fix line length and white space.
10199 (aarch64_symbolic_constant_p): Remove.
10200 * config/aarch64/predicates.md (aarch64_valid_symref):
10201 Use aarch64_classify_symbol_expression.
10203 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10205 * config/arm/constraints.md (Ts): New constraint.
10206 * config/arm/arm.md (arm_movqi_insn): Add alternatives for
10208 (compare_scc): Use "Ts" constraint for operand 0.
10209 (ior_scc_scc): Likewise.
10210 (and_scc_scc): Likewise.
10211 (and_scc_scc_nodom): Likewise.
10212 (ior_scc_scc_cmp): Likewise for operand 7.
10213 (and_scc_scc_cmp): Likewise.
10214 * config/arm/thumb2.md (thumb2_movsi_insn):
10215 Add alternatives for 16-bit encodings.
10216 (thumb2_movhi_insn): Likewise.
10217 (thumb2_movsicc_insn): Likewise.
10218 (thumb2_and_scc): Take 'and' outside cond_exec. Use "Ts" constraint.
10219 (thumb2_negscc): Use "Ts" constraint.
10220 Move mvn instruction outside cond_exec block.
10221 * config/arm/vfp.md (thumb2_movsi_vfp): Add alternatives
10222 for 16-bit encodings.
10224 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10226 * config/arm/arm.md (arm_mulsi3_v6): Add alternative for 16-bit
10228 (mulsi3addsi_v6): Disable predicable variant for arm_restrict_it.
10229 (mulsi3subsi): Likewise.
10230 (mulsidi3adddi): Likewise.
10231 (mulsidi3_v6): Likewise.
10232 (umulsidi3_v6): Likewise.
10233 (umulsidi3adddi_v6): Likewise.
10234 (smulsi3_highpart_v6): Likewise.
10235 (umulsi3_highpart_v6): Likewise.
10236 (mulhisi3tb): Likewise.
10237 (mulhisi3bt): Likewise.
10238 (mulhisi3tt): Likewise.
10239 (maddhisi4): Likewise.
10240 (maddhisi4tb): Likewise.
10241 (maddhisi4tt): Likewise.
10242 (maddhidi4): Likewise.
10243 (maddhidi4tb): Likewise.
10244 (maddhidi4tt): Likewise.
10245 (zeroextractsi_compare0_scratch): Likewise.
10246 (insv_zero): Likewise.
10247 (insv_t2): Likewise.
10248 (anddi_notzesidi_di): Likewise.
10249 (anddi_notsesidi_di): Likewise.
10250 (andsi_notsi_si): Likewise.
10251 (iordi_zesidi_di): Likewise.
10252 (xordi_zesidi_di): Likewise.
10253 (andsi_iorsi3_notsi): Likewise.
10254 (smax_0): Likewise.
10255 (smax_m1): Likewise.
10256 (smin_0): Likewise.
10257 (not_shiftsi): Likewise.
10258 (unaligned_loadsi): Likewise.
10259 (unaligned_loadhis): Likewise.
10260 (unaligned_loadhiu): Likewise.
10261 (unaligned_storesi): Likewise.
10262 (unaligned_storehi): Likewise.
10263 (extv_reg): Likewise.
10264 (extzv_t2): Likewise.
10265 (divsi3): Likewise.
10266 (udivsi3): Likewise.
10267 (arm_zero_extendhisi2addsi): Likewise.
10268 (arm_zero_extendqisi2addsi): Likewise.
10269 (compareqi_eq0): Likewise.
10270 (arm_extendhisi2_v6): Likewise.
10271 (arm_extendqisi2addsi): Likewise.
10272 (arm_movt): Likewise.
10273 (thumb2_ldrd): Likewise.
10274 (thumb2_ldrd_base): Likewise.
10275 (thumb2_ldrd_base_neg): Likewise.
10276 (thumb2_strd): Likewise.
10277 (thumb2_strd_base): Likewise.
10278 (thumb2_strd_base_neg): Likewise.
10279 (arm_negsi2): Add alternative for 16-bit encoding.
10280 (arm_one_cmplsi2): Likewise.
10282 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10284 * config/arm/predicates.md (arm_cond_move_operator): New predicate.
10285 * config/arm/arm.md (movsfcc): Use arm_cond_move_operator predicate.
10286 (movdfcc): Likewise.
10287 * config/arm/vfp.md (*thumb2_movsf_vfp):
10288 Disable predication for arm_restrict_it.
10289 (*thumb2_movsfcc_vfp): Disable for arm_restrict_it.
10290 (*thumb2_movdfcc_vfp): Likewise.
10291 (*abssf2_vfp, *absdf2_vfp, *negsf2_vfp, *negdf2_vfp,*addsf3_vfp,
10292 *adddf3_vfp, *subsf3_vfp, *subdf3_vfpc, *divsf3_vfp,*divdf3_vfp,
10293 *mulsf3_vfp, *muldf3_vfp, *mulsf3negsf_vfp, *muldf3negdf_vfp,
10294 *mulsf3addsf_vfp, *muldf3adddf_vfp, *mulsf3subsf_vfp,
10295 *muldf3subdf_vfp, *mulsf3negsfaddsf_vfp, *fmuldf3negdfadddf_vfp,
10296 *mulsf3negsfsubsf_vfp, *muldf3negdfsubdf_vfp, *fma<SDF:mode>4,
10297 *fmsub<SDF:mode>4, *fnmsub<SDF:mode>4, *fnmadd<SDF:mode>4,
10298 *extendsfdf2_vfp, *truncdfsf2_vfp, *extendhfsf2, *truncsfhf2,
10299 *truncsisf2_vfp, *truncsidf2_vfp, fixuns_truncsfsi2, fixuns_truncdfsi2,
10300 *floatsisf2_vfp, *floatsidf2_vfp, floatunssisf2, floatunssidf2,
10301 *sqrtsf2_vfp, *sqrtdf2_vfp, *cmpsf_vfp, *cmpsf_trap_vfp, *cmpdf_vfp,
10302 *cmpdf_trap_vfp, <vrint_pattern><SDF:mode>2):
10303 Disable predication for arm_restrict_it.
10305 2013-06-28 Kirill Yukhin <kirill.yukhin@intel.com>
10307 * config/i386/bmiintrin.h (_bextr_u32): New.
10308 (_bextr_u64): Ditto.
10310 2013-06-27 Richard Sandiford <rdsandiford@googlemail.com>
10312 * config.gcc (mips*-mti-elf*, mips*-sde-elf*, mips64r5900-*-elf*)
10313 (mips64r5900el-*-elf*): Include mips/n32-elf.h.
10314 * config/mips/sde.h (LOCAL_LABEL_PREFIX, NO_DOLLAR_IN_LABEL)
10315 (LONG_DOUBLE_TYPE_SIZE, LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Move to...
10316 * config/mips/n32-elf.h: ...this new file.
10318 2013-06-27 Marc Glisse <marc.glisse@inria.fr>
10321 * config/i386/i386.c (enum ix86_builtins, bdesc_args): Remove
10322 IX86_BUILTIN_CMPNGTSS and IX86_BUILTIN_CMPNGESS.
10324 2013-06-27 Catherine Moore <clm@codesourcery.com>
10326 * config/mips/mips-tables.opt: Regenerate.
10327 * config/mips/mips-cpus.def: Add m14ke and m14kec.
10328 * config/mips/mips.h (BASE_DRIVER_SELF_SPECS): m14ke* implies -mdspr2.
10329 * doc/invoke.texi: Add -m14kc.
10331 2013-06-27 Jakub Jelinek <jakub@redhat.com>
10334 * config/i386/i386.md (bmi_bextr_<mode>): Swap predicates and
10335 constraints of operand 1 and 2.
10338 * config/i386/i386.md (bmi2_bzhi_<mode>3): Swap AND arguments
10339 to match RTL canonicalization. Swap predicates and
10340 constraints of operand 1 and 2.
10342 2013-06-27 Vladimir Makarov <vmakarov@redhat.com>
10344 * lra-constraints.c (inherit_in_ebb): Process static hard regs too.
10345 Process OP_INOUT regs for splitting too.
10347 2013-06-27 Jakub Jelinek <jakub@redhat.com>
10349 * tree-vect-stmts.c (vectorizable_store): Move ptr_incr var
10350 decl before the loop, initialize to NULL.
10351 (vectorizable_load): Initialize ptr_incr to NULL.
10353 2013-06-27 Martin Jambor <mjambor@suse.cz>
10356 * ipa-ref.h (ipa_maybe_record_reference): Declare.
10357 * ipa-ref.c (ipa_maybe_record_reference): New function.
10358 * cgraphclones.c (cgraph_create_virtual_clone): Use it.
10359 * ipa-cp.c (create_specialized_node): Record potential references from
10361 * Makefile.in (ipa-ref.o): Add IPA_REF_H to dependencies.
10363 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
10365 * config/aarch64/aarch64.c (aarch64_force_temporary): Add an extra
10366 parameter 'mode' of type 'enum machine_mode mode'; change to pass
10367 'mode' to force_reg.
10368 (aarch64_add_offset): Update calls to aarch64_force_temporary.
10369 (aarch64_expand_mov_immediate): Likewise.
10371 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
10373 * config/aarch64/aarch64.c (aarch64_add_offset): Change to pass
10374 'mode' to aarch64_plus_immediate and gen_rtx_PLUS.
10376 2013-06-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
10378 * config/s390/s390.c: Rename UNSPEC_CCU_TO_INT to
10379 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
10380 (struct machine_function): Add tbegin_p.
10381 (s390_canonicalize_comparison): Fold CC mode compares to
10382 conditional jump if possible.
10383 (s390_emit_jump): Return the emitted jump.
10384 (s390_branch_condition_mask, s390_branch_condition_mnemonic):
10385 Handle CCRAWmode compares.
10386 (s390_option_override): Default to -mhtm if available.
10387 (s390_reg_clobbered_rtx): Handle floating point regs as well.
10388 (s390_regs_ever_clobbered): Use s390_regs_ever_clobbered also for
10389 FPRs instead of df_regs_ever_live_p.
10390 (s390_optimize_nonescaping_tx): New function.
10391 (s390_init_frame_layout): Extend clobbered_regs array to cover
10393 (s390_emit_prologue): Call s390_optimize_nonescaping_tx.
10394 (s390_expand_tbegin): New function.
10395 (enum s390_builtin): New enum definition.
10396 (code_for_builtin): New array definition.
10397 (s390_init_builtins): New function.
10398 (s390_expand_builtin): New function.
10399 (TARGET_INIT_BUILTINS): Define.
10400 (TARGET_EXPAND_BUILTIN): Define.
10401 * common/config/s390/s390-common.c (processor_flags_table): Add PF_TX.
10402 * config/s390/predicates.md (s390_comparison): Handle CCRAWmode.
10403 (s390_alc_comparison): Likewise.
10404 * config/s390/s390-modes.def: Add CCRAWmode.
10405 * config/s390/s390.h (processor_flags): Add PF_TX.
10406 (TARGET_CPU_HTM): Define macro.
10407 (TARGET_HTM): Define macro.
10408 (TARGET_CPU_CPP_BUILTINS): Define __HTM__ for htm.
10409 * config/s390/s390.md: Rename UNSPEC_CCU_TO_INT to
10410 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
10411 (UNSPECV_TBEGIN, UNSPECV_TBEGINC, UNSPECV_TEND, UNSPECV_TABORT)
10412 (UNSPECV_ETND, UNSPECV_NTSTG, UNSPECV_PPA): New unspecv enum values.
10413 (TBEGIN_MASK, TBEGINC_MASK): New constants.
10414 ("*cc_to_int"): Move up.
10415 ("*mov<mode>cc", "*cjump_64", "*cjump_31"): Accept integer
10416 constants other than 0.
10417 ("*ccraw_to_int"): New insn and splitter definition.
10418 ("tbegin", "tbegin_nofloat", "tbegin_retry")
10419 ("tbegin_retry_nofloat", "tbeginc", "tend", "tabort")
10420 ("tx_assist"): New expander.
10421 ("tbegin_1", "tbegin_nofloat_1", "*tbeginc_1", "*tend_1")
10422 ("*tabort_1", "etnd", "ntstg", "*ppa"): New insn definition.
10423 * config/s390/s390.opt: Add -mhtm option.
10424 * config/s390/s390-protos.h (s390_emit_jump): Add return type.
10425 * config/s390/htmxlintrin.h: New file.
10426 * config/s390/htmintrin.h: New file.
10427 * config/s390/s390intrin.h: New file.
10428 * doc/extend.texi: Document htm builtins.
10429 * config.gcc: Add the new header files to extra_headers.
10431 2013-06-26 Thomas Schwinge <thomas@codesourcery.com>
10433 * config/i386/gnu.h [TARGET_LIBC_PROVIDES_SSP]
10434 (TARGET_CAN_SPLIT_STACK, TARGET_THREAD_SPLIT_STACK_OFFSET): Undefine.
10436 2013-06-26 Michael Meissner <meissner@linux.vnet.ibm.com>
10437 Pat Haugen <pthaugen@us.ibm.com>
10438 Peter Bergner <bergner@vnet.ibm.com>
10440 * config/rs6000/power8.md: New.
10441 * config/rs6000/rs6000-cpus.def (RS6000_CPU table): Adjust processor
10442 setting for power8 entry.
10443 * config/rs6000/t-rs6000 (MD_INCLUDES): Add power8.md.
10444 * config/rs6000/rs6000.c (is_microcoded_insn, is_cracked_insn): Adjust
10445 test for Power4/Power5 only.
10446 (insn_must_be_first_in_group, insn_must_be_last_in_group): Add Power8
10448 (force_new_group): Adjust comment.
10449 * config/rs6000/rs6000.md: Include power8.md.
10451 2013-06-26 Greta Yorsh <Greta.Yorsh@arm.com>
10453 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Define macro.
10454 * config/arm/arm-protos.h (arm_max_conditional_execute): New
10456 (tune_params): Update comment.
10457 * config/arm/arm.c (arm_cortex_a15_tune): Set max_cond_insns to 2.
10458 (arm_max_conditional_execute): New function.
10459 (thumb2_final_prescan_insn): Use max_insn_skipped and
10460 MAX_INSN_PER_IT_BLOCK to compute maximum instructions in a block.
10462 2013-06-25 Jakub Jelinek <jakub@redhat.com>
10464 PR tree-optimization/57705
10465 * tree-vect-loop.c (vect_is_simple_iv_evolution): Allow
10466 SSA_NAME step, provided that it is not defined inside the loop.
10467 (vect_analyze_scalar_cycles_1): Disallow SSA_NAME step in nested loop.
10468 (get_initial_def_for_induction): Handle SSA_NAME IV step.
10470 2013-06-25 Martin Jambor <mjambor@suse.cz>
10472 PR middle-end/57670
10473 * cgraph.h (cgraph_indirect_call_info): New flag member_ptr.
10474 * ipa-prop.c (ipa_print_node_jump_functions): Mark member pointer
10476 (ipa_note_param_call): Initialize member_ptr flag.
10477 (ipa_analyze_indirect_call_uses): Set member_ptr flag.
10478 (ipa_make_edge_direct_to_target): Bail out if member_ptr is set.
10479 (ipa_write_indirect_edge_info): Stream member_ptr flag.
10480 (ipa_read_indirect_edge_info): Likewise.
10482 2013-06-25 Richard Biener <rguenther@suse.de>
10484 PR middle-end/56977
10485 * passes.c (init_optimization_passes): Move pass_fold_builtins
10486 and pass_dce earlier with -Og.
10488 2013-06-25 Eric Botcazou <ebotcazou@adacore.com>
10490 * expr.c (expand_expr_real_1) <ARRAY_REF>: Fix formatting glitches.
10491 <BIT_FIELD_REF>: Remove trailing TAB.
10492 * varasm.c (output_constructor_bitfield): Fix formatting glitch and
10495 2013-06-24 Martin Jambor <mjambor@suse.cz>
10497 PR tree-optimization/57358
10498 * ipa-prop.c (ipa_func_spec_opts_forbid_analysis_p): New function.
10499 (ipa_compute_jump_functions_for_edge): Bail out if it returns true.
10500 (ipa_analyze_params_uses): Generate pessimistic info when true.
10502 2013-06-24 Martin Jambor <mjambor@suse.cz>
10504 PR tree-optimization/57539
10505 * cgraphclones.c (cgraph_clone_node): Add parameter new_inlined_to, set
10506 global.inlined_to of the new node to it. All callers changed.
10507 * ipa-inline-transform.c (clone_inlined_nodes): New variable
10508 inlining_into, pass it to cgraph_clone_node.
10509 * ipa-prop.c (ipa_propagate_indirect_call_infos): Do not call
10510 ipa_free_edge_args_substructures.
10511 (ipa_edge_duplication_hook): Only add edges from inlined nodes to
10512 rdesc linked list. Do not assert rdesc edges have inlined caller.
10513 Assert we have found an rdesc in the rdesc list.
10515 2013-06-24 Richard Biener <rguenther@suse.de>
10517 * pointer-set.h (struct pointer_set_t): Move here from pointer-set.c.
10518 (pointer_set_lookup): Declare.
10519 (class pointer_map): New template class implementing a
10520 generic pointer to T map.
10521 (pointer_map<T>::pointer_map, pointer_map<T>::~pointer_map,
10522 pointer_map<T>::contains, pointer_map<T>::insert,
10523 pointer_map<T>::traverse): New functions.
10524 * pointer-set.c (struct pointer_set_t): Moved to pointer-set.h.
10525 (pointer_set_lookup): New function.
10526 (pointer_set_contains): Use pointer_set_lookup.
10527 (pointer_set_insert): Likewise.
10528 (insert_aux): Remove.
10529 (struct pointer_map_t): Embed a pointer_set_t.
10530 (pointer_map_create): Adjust.
10531 (pointer_map_destroy): Likewise.
10532 (pointer_map_contains): Likewise.
10533 (pointer_map_insert): Likewise.
10534 (pointer_map_traverse): Likewise.
10535 * tree-streamer.h (struct streamer_tree_cache_d): Use a
10536 pointer_map<unsigned> instead of a pointer_map_t.
10537 * tree-streamer.c (streamer_tree_cache_insert_1): Adjust.
10538 (streamer_tree_cache_lookup): Likewise.
10539 (streamer_tree_cache_create): Likewise.
10540 (streamer_tree_cache_delete): Likewise.
10541 * lto-streamer.h (struct lto_tree_ref_encoder): Use a
10542 pointer_map<unsigned> instead of a pointer_map_t.
10543 (lto_init_tree_ref_encoder): Adjust.
10544 (lto_destroy_tree_ref_encoder): Likewise.
10545 * lto-section-out.c (lto_output_decl_index): Likewise.
10546 (lto_record_function_out_decl_state): Likewise.
10547 * dominance.c (iterate_fix_dominators): Use pointer_map<int>.
10549 2013-06-24 Richard Biener <rguenther@suse.de>
10551 PR tree-optimization/57488
10552 * tree-ssa-pre.c (insert): Clear NEW sets before each iteration.
10554 2013-06-24 Alan Modra <amodra@gmail.com>
10556 * config/rs6000/rs6000.c (vspltis_constant): Correct for little-endian.
10557 (gen_easy_altivec_constant): Likewise.
10558 * config/rs6000/predicates.md (easy_vector_constant_add_self,
10559 easy_vector_constant_msb): Likewise.
10561 2013-06-23 Jakub Jelinek <jakub@redhat.com>
10564 * common/config/i386/i386-common.c (ix86_handle_option): For OPT_mlzcnt
10565 add missing return true.
10567 2013-06-23 Oleg Endo <olegendo@gcc.gnu.org>
10570 * config/sh/predicates.md (general_extend_operand): Invoke
10571 general_movsrc_operand for memory operands.
10572 (general_movsrc_operand): Allow reg+reg addressing, do not use
10573 general_operand for memory operands.
10575 2013-06-23 Sriraman Tallam <tmsriram@google.com>
10577 * config/i386/i386.c (ix86_pragma_target_parse): Restore target
10578 when current target options does not apply.
10579 * config/i386/i386-protos.h (ix86_reset_previous_fndecl): New function.
10580 * config/i386/i386.c (ix86_reset_previous_fndecl): Ditto.
10581 * config/i386/bmiintrin.h: Pass appropriate target
10582 attributes to header.
10583 * config/i386/mmintrin.h: Ditto.
10584 * config/i386/nmmintrin.h: Ditto.
10585 * config/i386/avx2intrin.h: Ditto.
10586 * config/i386/fxsrintrin.h: Ditto.
10587 * config/i386/tbmintrin.h: Ditto.
10588 * config/i386/xsaveintrin.h: Ditto.
10589 * config/i386/f16cintrin.h: Ditto.
10590 * config/i386/xtestintrin.h: Ditto.
10591 * config/i386/xsaveoptintrin.h: Ditto.
10592 * config/i386/bmi2intrin.h: Ditto.
10593 * config/i386/lzcntintrin.h: Ditto.
10594 * config/i386/smmintrin.h: Ditto.
10595 * config/i386/wmmintrin.h: Ditto.
10596 * config/i386/x86intrin.h: Remove all header include guards.
10597 * config/i386/prfchwintrin.h: Ditto.
10598 * config/i386/pmmintrin.h: Ditto.
10599 * config/i386/tmmintrin.h: Ditto.
10600 * config/i386/xmmintrin.h: Ditto.
10601 * config/i386/popcntintrin.h: Ditto.
10602 * config/i386/rdseedintrin.h: Ditto.
10603 * config/i386/ammintrin.h: Ditto.
10604 * config/i386/emmintrin.h: Ditto.
10605 * config/i386/immintrin.h: Remove all header include guards.
10606 * config/i386/fma4intrin.h: Ditto.
10607 * config/i386/lwpintrin.h: Ditto.
10608 * config/i386/xopintrin.h: Ditto.
10609 * config/i386/ia32intrin.h: Ditto.
10610 * config/i386/avxintrin.h: Ditto.
10611 * config/i386/rtmintrin.h: Ditto.
10612 * config/i386/fmaintrin.h: Ditto.
10613 * config/i386/mm3dnow.h: Ditto.
10615 2013-06-22 Sriraman Tallam <tmsriram@google.com>
10617 * common/config/i386/i386-common.c: Handle LZCNT.
10619 2013-06-22 Andi Kleen <ak@linux.intel.com>
10621 * doc/extend.texi: Use __atomic_store_n instead of
10622 __atomic_store in HLE example.
10624 2013-06-22 Oleg Endo <olegendo@gcc.gnu.org>
10626 * config/sh/sh.c: Remove <cstdlib> workaround.
10628 2013-06-21 Andi Kleen <ak@linux.intel.com>
10630 * doc/extend.texi: Dont use __atomic_clear in HLE example. Fix typo.
10632 2013-06-21 Andi Kleen <ak@linux.intel.com>
10634 * doc/extend.texi: Document that __atomic_clear and
10635 __atomic_test_and_set should only be used with bool.
10637 2013-06-20 Jan Hubicka <jh@suse.cz>
10639 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Use
10640 types_same_for_odr.
10641 * tree.c (decls_same_for_odr): New function.
10642 (same_for_edr): New function.
10643 (types_same_for_odr): New function.
10644 (get_binfo_at_offset): Use it.
10645 * tree.h (types_same_for_odr): Declare.
10647 2013-06-20 Oleg Endo <olegendo@gcc.gnu.org>
10648 Jason Merrill <jason@redhat.com>
10650 * system.h: Include <cstdlib> as well as <stdlib.h>.
10652 2013-06-20 Uros Bizjak <ubizjak@gmail.com>
10655 * config/i386/i386.c (construct_container): Report error if
10656 long double is used with disabled x87 float returns.
10658 2013-06-20 Jan Hubicka <jh@suse.cz>
10660 * lto-cgraph.c (input_symtab): Do not set cgraph state.
10662 2013-06-20 Joern Rennecke <joern.rennecke@embecosm.com>
10664 PR rtl-optimization/57425
10665 PR rtl-optimization/57569
10666 * alias.c (write_dependence_p): Remove parameters mem_mode and
10667 canon_mem_addr. Add parameters x_mode, x_addr and x_canonicalized.
10668 Changed all callers.
10669 (canon_anti_dependence): Get comments and semantics in sync.
10670 Add parameter mem_canonicalized. Changed all callers.
10671 * rtl.h (canon_anti_dependence): Update prototype.
10673 2013-06-20 Richard Biener <rguenther@suse.de>
10675 * data-streamer-in.c (streamer_read_uhwi): Optimize single
10676 byte case, inline streamer_read_uchar and defer section
10679 2013-06-20 Richard Biener <rguenther@suse.de>
10681 PR tree-optimization/57584
10682 * tree-ssa-loop-niter.c (expand_simple_operations): Avoid including
10683 SSA names into the expanded expression that take part in
10684 abnormal coalescing.
10686 2013-06-19 Sharad Singhai <singhai@google.com>
10688 * gcov.c (print_usage): Handle new option.
10689 (process_args): Ditto.
10690 (get_gcov_intermediate_filename): New function.
10691 (output_intermediate_file): New function.
10692 (output_gcov_file): New function
10693 (generate_results): Handle new option.
10694 (release_function): Relase demangled name.
10695 (read_graph_file): Handle demangled name.
10696 (output_lines): Ditto.
10697 * doc/gcov.texi: Document gcov intermediate format.
10699 2013-06-19 Vladimir Makarov <vmakarov@redhat.com>
10702 * lra.c (emit_add3_insn, emit_add2_insn): New functions.
10703 (lra_emit_add): Use the functions. Add comment about Y as an
10706 2013-06-19 David Edelsohn <dje.gcc@gmail.com>
10709 * collect2.c (collect_atexit): New.
10710 (collect_exit): Delete.
10711 (main): Register collect_atexit with atexit.
10712 (collect_wait): Change collect_exit to exit.
10714 * collect2.h (collect_exit): Delete.
10715 * tlink.c (do_tlink): Rename exit to ret. Change collect_exit to exit.
10717 2013-06-19 Wei Mi <wmi@google.com>
10719 PR rtl-optimization/57518
10720 * ira.c (set_paradoxical_subreg): Set pdx_subregs[regno]
10721 if regno is used in paradoxical subreg.
10722 (update_equiv_regs): Check pdx_subregs[regno] before
10723 set a reg to be equivalent with a mem.
10725 2013-06-19 Matthias Klose <doko@ubuntu.com>
10728 * file-find.h (find_a_file): Add a mode parameter.
10729 * file-find.c (find_a_file): Likewise.
10730 * gcc-ar.c (main): Call find_a_file with R_OK for the plugin,
10731 with X_OK for the executables.
10732 * collect2.c (main): Call find_a_file with X_OK.
10734 2013-06-19 Steve Ellcey <sellcey@mips.com>
10737 * config/mips/mips.md (casesi_internal_mips16_<mode>):
10738 Use NEXT_INSN instead of next_real_insn.
10740 2013-06-19 Jan Hubicka <jh@suse.cz>
10742 * cgraph.h (const_value_known_p): Replace by ...
10743 (ctor_for_folding): .. this one.
10744 * cgraphunit.c (process_function_and_variable_attributes): Use it.
10745 * lto-cgraph.c (compute_ltrans_boundary): Use ctor_for_folding.
10746 * expr.c (expand_expr_real_1): Likewise.
10747 (string_constant): Likewise.
10748 * tree-ssa-loop-ivcanon.c (constant_after_peeling): Likewise.
10749 * ipa.c (process_references): Likewise.
10750 (symtab_remove_unreachable_nodes): Likewise.
10751 * ipa-inline-analysis.c (param_change_prob): Likewise.
10752 * gimple-fold.c (canonicalize_constructor_val): Likewise.
10753 (get_base_constructor): Likwise.
10754 * varpool.c (varpool_remove_node): Likewise.
10755 (varpool_remove_initializer): LIkewise.
10756 (dump_varpool_node): LIkwise.
10757 (const_value_known_p): Rewrite to ...
10758 (ctor_for_folding): ... this one.
10760 2013-06-19 Jakub Jelinek <jakub@redhat.com>
10763 * gcc-ar.c (main): If not CROSS_DIRECTORY_STRUCTURE, look for
10764 PERSONALITY in $PATH derived prefixes.
10766 2013-06-19 Jeff Law <law@redhat.com>
10768 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): Fix typo
10771 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): New function.
10772 (simplify_bitwise_binary): Use it to simpify certain binary ops on
10775 2013-06-19 Sofiane Naci <sofiane.naci@arm.com>
10777 * config/arm/vfp.md: Move VFP instruction classification documentation
10779 * config/arm/arm.md: ... here. Update instruction classification
10782 2013-06-19 Richard Earnshaw <rearnsha@arm.com>
10784 arm.md (split for eq(reg, 0)): Add variants for ARMv5 and Thumb2.
10785 (peepholes for eq(reg, not-0)): Ensure condition register is dead after
10786 pattern. Use more efficient sequences on ARMv5 and Thumb2.
10788 2013-06-19 Steven Bosscher <steven@gcc.gnu.org>
10791 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
10792 with NEXT_INSN. Use tablejump_p to check for jump table data
10795 2013-06-19 Paolo Carlini <paolo.carlini@oracle.com>
10798 * doc/cpp.texi [Standard Predefined Macros, __cplusplus]: Document
10799 that now in C++ the value is correct per the C++ standards.
10801 2013-06-19 Richard Biener <rguenther@suse.de>
10803 * expr.c (expand_expr_real_1): Use SCOPE_FILE_SCOPE_P to check
10804 for global context.
10806 2013-06-19 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
10809 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
10812 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
10813 with next_active_insn.
10815 2013-06-18 Sriraman Tallam <tmsriram@google.com>
10817 * ipa-inline.c (inline_always_inline_functions): Pretend always_inline
10818 functions are inlined during failures to flag an error.
10819 * tree-inline.c (expand_call_inline): Allow the error to be flagged
10820 in early inline pass.
10822 2013-06-18 H.J. Lu <hongjiu.lu@intel.com>
10824 * config/i386/i386.c (initial_ix86_tune_features): Fix a typo
10827 2013-06-18 Julian Brown <julian@codesourcery.com>
10829 * config/arm/arm.c (neon_vector_mem_operand): Add strict argument.
10830 Permit virtual register pre-reload if !strict.
10831 (coproc_secondary_reload_class): Adjust for neon_vector_mem_operand
10833 * config/arm/arm-protos.h (neon_vector_mem_operand): Adjust
10835 * config/arm/neon.md (movmisalign<mode>): Use
10836 neon_perm_struct_or_reg_operand instead of
10837 neon_struct_or_register_operand.
10838 (*movmisalign<mode>_neon_load, *movmisalign<mode>_neon_store): Use
10839 neon_permissive_struct_operand instead of neon_struct_operand.
10840 * config/arm/constraints.md (Un, Um, Us): Adjust calls to
10841 neon_vector_mem_operand.
10842 * config/arm/predicates.md (neon_struct_operand): Adjust call to
10843 neon_vector_mem_operand.
10844 (neon_permissive_struct_operand): New.
10845 (neon_struct_or_register_operand): Rename to...
10846 (neon_perm_struct_or_reg_operand): This. Adjust call to
10847 neon_vector_mem_operand.
10849 2013-06-18 Richard Biener <rguenther@suse.de>
10851 * Makefile.in (LTO_STREAMER_H): Add pointer-set.h dependency.
10852 * lto-streamer.h: Include pointer-set.h.
10853 (struct lto_decl_slot): Remove.
10854 (struct lto_tree_ref_encoder): Make tree_hash_table a pointer-map.
10855 Remove next_index entry.
10856 (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
10857 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
10858 (lto_init_tree_ref_encoder): Adjust.
10859 (lto_destroy_tree_ref_encoder): Likewise.
10860 * lto-section-out.c (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
10861 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
10862 (lto_output_decl_index): Adjust.
10863 (lto_new_out_decl_state): Likewise.
10864 (lto_record_function_out_decl_state): Likewise.
10865 * lto-streamer-out.c (copy_function): Likewise.
10867 2013-06-18 Richard Biener <rguenther@suse.de>
10869 * Makefile.in (cgraphunit.o): Add $(CFGLOOP_H) dependency.
10870 * cgraphunit.c: Include cfgloop.h.
10871 (init_lowered_empty_function): Initialize the loop tree.
10872 (assemble_thunk): Insert new BBs into loops.
10874 2013-06-18 Richard Biener <rguenther@suse.de>
10876 * tree-streamer.h (streamer_tree_cache_create): Adjust prototype.
10877 * tree-streamer.c (streamer_tree_cache_create): Make maintaining
10878 the map from cache entry to cache index optional.
10879 (streamer_tree_cache_replace_tree): Adjust accordingly.
10880 (streamer_tree_cache_append): Likewise.
10881 (streamer_tree_cache_delete): Likewise.
10882 * lto-streamer-in.c (lto_data_in_create): Do not maintain the
10883 streamer cache map from cache entry to cache index.
10884 * lto-streamer-out.c (create_output_block): Adjust.
10886 2013-06-18 Sofiane Naci <sofiane.naci@arm.com>
10888 * config/arm/arm.md (attribute "insn"): Move multiplication and
10889 division attributes to...
10890 (attribute "type"): ... here. Remove mult.
10891 (attribute "mul32"): New attribute.
10892 (attribute "mul64"): Add umaal.
10893 (*arm_mulsi3): Update attributes.
10894 (*arm_mulsi3_v6): Likewise.
10895 (*thumb_mulsi3): Likewise.
10896 (*thumb_mulsi3_v6): Likewise.
10897 (*mulsi3_compare0): Likewise.
10898 (*mulsi3_compare0_v6): Likewise.
10899 (*mulsi_compare0_scratch): Likewise.
10900 (*mulsi_compare0_scratch_v6): Likewise.
10901 (*mulsi3addsi): Likewise.
10902 (*mulsi3addsi_v6): Likewise.
10903 (*mulsi3addsi_compare0): Likewise.
10904 (*mulsi3addsi_compare0_v6): Likewise.
10905 (*mulsi3addsi_compare0_scratch): Likewise.
10906 (*mulsi3addsi_compare0_scratch_v6): Likewise.
10907 (*mulsi3subsi): Likewise.
10908 (*mulsidi3adddi): Likewise.
10909 (*mulsi3addsi_v6): Likewise.
10910 (*mulsidi3adddi_v6): Likewise.
10911 (*mulsidi3_nov6): Likewise.
10912 (*mulsidi3_v6): Likewise.
10913 (*umulsidi3_nov6): Likewise.
10914 (*umulsidi3_v6): Likewise.
10915 (*umulsidi3adddi): Likewise.
10916 (*umulsidi3adddi_v6): Likewise.
10917 (*smulsi3_highpart_nov6): Likewise.
10918 (*smulsi3_highpart_v6): Likewise.
10919 (*umulsi3_highpart_nov6): Likewise.
10920 (*umulsi3_highpart_v6): Likewise.
10921 (mulhisi3): Likewise.
10922 (*mulhisi3tb): Likewise.
10923 (*mulhisi3bt): Likewise.
10924 (*mulhisi3tt): Likewise.
10925 (maddhisi4): Likewise.
10926 (*maddhisi4tb): Likewise.
10927 (*maddhisi4tt): Likewise.
10928 (maddhidi4): Likewise.
10929 (*maddhidi4tb): Likewise.
10930 (*maddhidi4tt): Likewise.
10931 (divsi3): Likewise.
10932 (udivsi3): Likewise.
10933 * config/arm/thumb2.md (thumb2_mulsi_short): Update attributes.
10934 (thumb2_mulsi_short_compare0): Likewise.
10935 (thumb2_mulsi_short_compare0_scratch): Likewise.
10936 * config/arm/arm1020e.md (1020mult1): Update attribute change.
10937 (1020mult2): Likewise.
10938 (1020mult3): Likewise.
10939 (1020mult4): Likewise.
10940 (1020mult5): Likewise.
10941 (1020mult6): Likewise.
10942 * config/arm/cortex-a15.md (cortex_a15_mult32): Update attribute
10944 (cortex_a15_mult64): Likewise.
10945 (cortex_a15_sdiv): Likewise.
10946 (cortex_a15_udiv): Likewise.
10947 * config/arm/arm1026ejs.md (mult1): Update attribute change.
10953 * config/arm/marvell-pj4.md (pj4_ir_mul): Update attribute change.
10954 (pj4_ir_div): Likewise.
10955 * config/arm/arm1136jfs.md (11_mult1): Update attribute change.
10956 (11_mult2): Likewise.
10957 (11_mult3): Likewise.
10958 (11_mult4): Likewise.
10959 (11_mult5): Likewise.
10960 (11_mult6): Likewise.
10961 (11_mult7): Likewise.
10962 * config/arm/cortex-a8.md (cortex_a8_mul): Update attribute change.
10963 (cortex_a8_mla): Likewise.
10964 (cortex_a8_mull): Likewise.
10965 (cortex_a8_smulwy): Likewise.
10966 (cortex_a8_smlald): Likewise.
10967 * config/arm/cortex-m4.md (cortex_m4_alu): Update attribute change.
10968 * config/arm/cortex-r4.md (cortex_r4_mul_4): Update attribute change.
10969 (cortex_r4_mul_3): Likewise.
10970 (cortex_r4_mla_4): Likewise.
10971 (cortex_r4_mla_3): Likewise.
10972 (cortex_r4_smlald): Likewise.
10973 (cortex_r4_mull): Likewise.
10974 (cortex_r4_sdiv): Likewise.
10975 (cortex_r4_udiv): Likewise.
10976 * config/arm/cortex-a7.md (cortex_a7_mul): Update attribute change.
10977 (cortex_a7_idiv): Likewise.
10978 * config/arm/arm926ejs.md (9_mult1): Update attribute change.
10979 (9_mult2): Likewise.
10980 (9_mult3): Likewise.
10981 (9_mult4): Likewise.
10982 (9_mult5): Likewise.
10983 (9_mult6): Likewise.
10984 * config/arm/cortex-a53.md (cortex_a53_mul): Update attribute change.
10985 (cortex_a53_sdiv): Likewise.
10986 (cortex_a53_udiv): Likewise.
10987 * config/arm/fa726te.md (726te_mult_op): Update attribute change.
10988 * config/arm/fmp626.md (mp626_mult1): Update attribute change.
10989 (mp626_mult2): Likewise.
10990 (mp626_mult3): Likewise.
10991 (mp626_mult4): Likewise.
10992 * config/arm/fa526.md (526_mult1): Update attribute change.
10993 (526_mult2): Likewise.
10994 * config/arm/arm-generic.md (mult): Update attribute change.
10995 (mult_ldsched_strongarm): Likewise.
10996 (mult_ldsched): Likewise.
10997 (multi_cycle): Likewise.
10998 * config/arm/cortex-a5.md (cortex_a5_mul): Update attribute change.
10999 * config/arm/fa606te.md (606te_mult1): Update attribute change.
11000 (606te_mult2): Likewise.
11001 (606te_mult3): Likewise.
11002 (606te_mult4): Likewise.
11003 * config/arm/cortex-a9.md (cortex_a9_mult16): Update attribute change.
11004 (cortex_a9_mac16): Likewise.
11005 (cortex_a9_multiply): Likewise.
11006 (cortex_a9_mac): Likewise.
11007 (cortex_a9_multiply_long): Likewise.
11008 * config/arm/fa626te.md (626te_mult1): Update attribute change.
11009 (626te_mult2): Likewise.
11010 (626te_mult3): Likewise.
11011 (626te_mult4): Likewise.
11013 2013-06-18 Richard Biener <rguenther@suse.de>
11016 * lto-symtab.c (lto_symtab_merge_decls): Process nodes properly.
11018 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
11021 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
11022 with next_active_insn.
11024 2013-06-18 Alan Modra <amodra@gmail.com>
11026 * config/rs6000/rs6000.h (enum data_align): New.
11027 (LOCAL_ALIGNMENT, DATA_ALIGNMENT): Use rs6000_data_alignment.
11028 (DATA_ABI_ALIGNMENT): Define.
11029 (CONSTANT_ALIGNMENT): Correct comment.
11030 * config/rs6000/rs6000-protos.h (rs6000_data_alignment): Declare.
11031 * config/rs6000/rs6000.c (rs6000_data_alignment): New function.
11033 2013-06-17 David Malcolm <dmalcolm@redhat.com>
11035 * ggc-page.c (ggc_pch_write_object) <d>: Remove erroneous
11036 ATTRIBUTE_UNUSED marking.
11038 2013-06-17 Sofiane Naci <sofiane.naci@arm.com>
11040 * config/aarch64/aarch64-simd.md (aarch64_dup_lane<mode>): Add r<-w
11041 alternative and update.
11042 (aarch64_dup_lanedi): Delete.
11043 * config/aarch64/arm_neon.h (vdup<bhsd>_lane_*): Update.
11044 * config/aarch64/aarch64-simd-builtins.def: Update.
11046 2013-06-17 Richard Biener <rguenther@suse.de>
11048 * lto-streamer.h (enum LTO_tags): Add LTO_tree_scc.
11049 (lto_input_scc): Declare.
11050 (lto_input_tree_1): Likewise.
11051 (struct lto_stats_d): Add num_tree_bodies_output and
11052 num_pickle_refs_output.
11053 * lto-streamer-in.c (lto_read_body): Use streamer_tree_cache_get_tree.
11054 (lto_read_tree_1): Split out from ...
11055 (lto_read_tree): ... this.
11056 (lto_input_scc): New function.
11057 (lto_input_tree_1): Split out from ...
11058 (lto_input_tree): ... this. Handle LTO_tree_scc.
11059 (lto_data_in_create): Create the streamer cache without hashes.
11060 * lto-streamer-out.c (create_output_block): Create the streamer
11061 cache with hashes when not doing WPA.
11062 (lto_write_tree_1): Split out from ...
11063 (lto_write_tree): ... this.
11064 (get_symbol_initial_value): New function.
11065 (lto_output_tree_1): Split out from ...
11066 (lto_output_tree): ... this. Write trees as series of SCCs
11067 using a DFS walk via DFS_write_tree.
11068 (struct sccs, struct scc_entry): New types.
11069 (next_dfs_num, sccstack, sccstate, sccstate_obstack): New globals.
11070 (DFS_write_tree_body): New function.
11071 (DFS_write_tree): Likewise.
11072 (hash_tree): Likewise.
11073 (scc_entry_compare): Likewise.
11074 (hash_scc): Likewise.
11075 (tree_is_indexable): DEBUG_EXPR_DECLs are local entities.
11076 * tree-streamer-in.c (lto_input_ts_list_tree_pointers): Stream
11077 TREE_CHAIN as regular reference.
11078 (streamer_read_integer_cst): Remove.
11079 (streamer_get_pickled_tree): Adjust.
11080 * tree-streamer-out.c (streamer_write_chain): Disable streaming
11081 of DECL_EXTERNALs in BLOCK_VARS for now.
11082 (write_ts_list_tree_pointers): Stream TREE_CHAIN as regular
11084 * tree-streamer.c (streamer_tree_cache_add_to_node_array):
11085 Add hash value argument and record that if hashes are recorded
11087 (streamer_tree_cache_insert_1): Adjust.
11088 (streamer_tree_cache_insert): Likewise.
11089 (streamer_tree_cache_insert_at): Rename to ...
11090 (streamer_tree_cache_replace_tree): ... this and adjust.
11091 (streamer_tree_cache_append): Adjust.
11092 (record_common_node): Likewise.
11093 (streamer_tree_cache_create): Add argument whether to
11094 record hash values together with trees.
11095 (streamer_tree_cache_delete): Adjust.
11096 * tree-streamer.h (struct streamer_tree_cache_d): Add
11098 (streamer_read_integer_cst): Remove.
11099 (streamer_tree_cache_insert): Adjust.
11100 (streamer_tree_cache_append): Likewise.
11101 (streamer_tree_cache_insert_at): Rename to ...
11102 (streamer_tree_cache_replace_tree): ... this and adjust.
11103 (streamer_tree_cache_create): Add argument whether to record hashes.
11104 (streamer_tree_cache_get): Rename to ...
11105 (streamer_tree_cache_get_tree): ... this.
11106 (streamer_tree_cache_get_hash): New function.
11107 * tree.c (cache_integer_cst): New function.
11108 * tree.h (cache_integer_cst): Declare.
11109 (ANON_AGGRNAME_FORMAT, ANON_AGGRNAME_P): Move here from cp/cp-tree.h.
11110 * lto-symtab.c (lto_varpool_replace_node): Only release
11111 DECL_INITIAL of non-prevailing decls.
11112 * varpool.c (varpool_remove_initializer): Do not release
11113 DECL_INITIAL when we are still in CGRAPH_LTO_STREAMING.
11115 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
11117 * config/mips/mips.h (ISA_HAS_MUL3): Include TARGET_MIPS5900.
11118 (ISA_HAS_MULT, ISA_HAS_DMULT, ISA_HAS_DIV, ISA_HAS_DDIV): New macros.
11119 * config/mips/mips.md (mul<mode>3, mul<mode>3_internal)
11120 (mul<mode>3_r4000): Require ISA_HAS_<D>MULT.
11121 (mul<mode>3_mul3): Handle TARGET_MIPS5900.
11122 (mulsidi3_64bit_dmul): Remove redundant TARGET_64BIT test.
11123 (<su>muldi3_highpart, <su>muldi3_highpart_internal, <u>mulditi3)
11124 (<u>mulditi3_internal, <u>mulditi3_r4000): Require ISA_HAS_DMULT
11125 instead of TARGET_64BIT.
11126 (divmod<mode>4, udivmod<mode>4, <u>divmod<GPR:mode>4_hilo_<HILO:mode>):
11127 Require ISA_HAS_<D>DIV.
11129 2013-06-16 Richard Sandiford <rdsandiford@googlemail.com>
11131 * config.gcc (mips*-mti-linux*, mips64*-*-linux*, mipsisa64*-*-linux*)
11132 (mips*-*-linux*): Move default with_llsc setting to where other
11134 (mips*-*-vxworks*): Move with_arch default from with_cpu block to
11136 (mips64r5900-*-*, mips64r5900el-*-*, mipsr5900-*-*, mipsr5900el-*-*):
11137 Likewise. Remove default with_tune setting. Move default float
11138 setting to its own block. Handle with_llsc in the same block as above.
11140 2013-06-16 Joern Rennecke <joern.rennecke@embecosm.com>
11142 PR rtl-optimization/57425
11143 PR rtl-optimization/57569
11144 * alias.c (write_dependence_p): Add new parameters mem_mode,
11145 canon_mem_addr and mem_canonicalized. Change type of writep to bool.
11146 Changed all callers.
11147 (canon_anti_dependence): New function.
11148 * cse.c (check_dependence): Use canon_anti_dependence.
11149 * cselib.c (cselib_invalidate_mem): Likewise.
11150 * rtl.h (canon_anti_dependence): Declare.
11152 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
11154 * config/mips/mips.h (ISA_HAS_LL_SC): Exclude TARGET_MIPS5900.
11155 * config/mips/mips.c (mips_start_ll_sc_sync_block): Output
11156 ".set mips3" for 64-bit targets.
11158 2013-06-15 Dehao Chen <dehao@google.com>
11160 * tree-flow.h (gimple_check_call_matching_types): Add new argument.
11161 * gimple-low.c (gimple_check_call_matching_types): Likewise.
11162 (gimple_check_call_args): Likewise.
11163 * value-prof.c (check_ic_target): Likewise.
11164 * ipa-inline.c (early_inliner): Likewise.
11165 * ipa-prop.c (update_indirect_edges_after_inlining): Likewise.
11166 * cgraph.c (cgraph_create_edge_1): Likewise.
11167 (cgraph_make_edge_direct): Likewise.
11169 2013-06-14 Michael Meissner <meissner@linux.vnet.ibm.com>
11172 * config/rs6000/rs6000.md (mov<mode>_ppc64): Call
11173 rs6000_output_move_128bit to handle emitting quad memory
11174 operations. Set attribute length to 8 bytes.
11176 2013-06-14 Vidya Praveen <vidyapraveen@arm.com>
11178 * config/aarch64/aarch64-simd.md (aarch64_<su>mlal_lo<mode>):
11180 (aarch64_<su>mlal_hi<mode>, aarch64_<su>mlsl_lo<mode>): Likewise.
11181 (aarch64_<su>mlsl_hi<mode>, aarch64_<su>mlal<mode>): Likewise.
11182 (aarch64_<su>mlsl<mode>): Likewise.
11184 2013-06-14 Mike Stump <mikestump@comcast.net>
11186 * Makefile.in (TARGET_H): Add insn-codes.h.
11188 2013-06-14 Alan Modra <amodra@gmail.com>
11190 PR middle-end/57134
11191 PR middle-end/57586
11192 * expr.c (expand_expr_real_1 <normal_inner_ref>): Pass
11193 EXPAND_MEMORY and EXPAND_WRITE to recursive call. Don't use
11194 bitfield expansion when EXPAND_MEMORY.
11195 (expand_expr_real_1 <VIEW_CONVERT_EXPR>): Pass modifier likewise.
11197 2013-06-13 Michael Meissner <meissner@linux.vnet.ibm.com>
11199 * config/rs6000/rs6000.c (rs6000_option_override_internal): Move
11200 test for clearing quad memory on 32-bit later.
11202 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
11204 * fold-const.c (negate_expr_p): Handle VECTOR_CST.
11205 (fold_negate_expr): Likewise.
11206 (fold_real_zero_addition_p): Handle vectors.
11207 (fold_binary_loc) <PLUS_EXPR, MINUS_EXPR>: Likewise.
11209 2013-06-14 Alan Modra <amodra@gmail.com>
11211 * varasm.c (force_const_mem): Revert 2013-06-07 change.
11213 2013-06-13 Jan Hubicka <jh@suse.cz>
11215 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
11216 Local comdats are not externally visible.
11217 * symtab.c (dump_symtab_base): Dump externally visible.
11218 (verify_symtab_base): Verify back links in the symtab hash.
11220 2013-06-13 Bin Cheng <bin.cheng@arm.com>
11222 * fold-const.c (operand_equal_p): Consider NOP_EXPR and
11223 CONVERT_EXPR as equal nodes.
11225 2013-06-13 Bin Cheng <bin.cheng@arm.com>
11227 * rtlanal.c (noop_move_p): Check the code to be executed for COND_EXEC.
11229 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
11231 * tree-ssa-forwprop.c (simplify_bitwise_binary, associate_plusminus):
11232 Generalize to complex and vector.
11233 * tree.c (build_all_ones_cst): New function.
11234 * tree.h (build_all_ones_cst): Declare it.
11236 2013-06-13 Alan Modra <amodra@gmail.com>
11238 * config/rs6000/rs6000.h (LONG_DOUBLE_LARGE_FIRST): Define.
11239 * config/rs6000/rs6000.md (signbittf2): New insn.
11240 (extenddftf2_internal): Use LONG_DOUBLE_LARGE_FIRST.
11241 (abstf2_internal, cmptf_internal2): Likewise.
11242 * config/rs6000/spe.md (spe_abstf2_cmp, spe_abstf2_tst): Likewise.
11244 2013-06-12 Michael Meissner <meissner@linux.vnet.ibm.com>
11245 Pat Haugen <pthaugen@us.ibm.com>
11246 Peter Bergner <bergner@vnet.ibm.com>
11248 * config/rs6000/rs6000.c (emit_load_locked): Add support for
11249 power8 byte, half-word, and quad-word atomic instructions.
11250 (emit_store_conditional): Likewise.
11251 (rs6000_expand_atomic_compare_and_swap): Likewise.
11252 (rs6000_expand_atomic_op): Likewise.
11254 * config/rs6000/sync.md (larx): Add new modes for power8.
11256 (AINT): New mode iterator to include TImode as well as normal
11257 integer modes on power8.
11258 (fetchop_pred): Use int_reg_operand instead of gpc_reg_operand so
11259 that VSX registers are not considered. Use AINT mode iterator
11260 instead of INT1 to allow inclusion of quad word atomic operations
11262 (load_locked<mode>): Likewise.
11263 (store_conditional<mode>): Likewise.
11264 (atomic_compare_and_swap<mode>): Likewise.
11265 (atomic_exchange<mode>): Likewise.
11266 (atomic_nand<mode>): Likewise.
11267 (atomic_fetch_<fetchop_name><mode>): Likewise.
11268 (atomic_nand_fetch<mode>): Likewise.
11269 (mem_thread_fence): Use gen_loadsync_<mode> instead of enumerating
11271 (ATOMIC): On power8, add QImode, HImode modes.
11272 (load_locked<QHI:mode>_si): Varients of load_locked for QI/HI
11273 modes that promote to SImode.
11274 (load_lockedti): Convert TImode arguments to PTImode, so that we
11275 get a guaranteed even/odd register pair.
11276 (load_lockedpti): Likewise.
11277 (store_conditionalti): Likewise.
11278 (store_conditionalpti): Likewise.
11280 * config/rs6000/rs6000.md (QHI): New mode iterator for power8
11281 atomic load/store instructions.
11284 2013-06-12 Richard Sandiford <rdsandiford@googlemail.com>
11286 * config/mips/mips.md (extended_mips16): Include GOT and constant-pool
11288 (insn_count): New attribute, with most cases extracted from...
11289 (length): ...here. Redefine most cases in terms of insn_count.
11290 (single_insn): Delete.
11291 (can_delay): Use insn_count to check for single instructions.
11292 (*mul<mode>3_r4300, mul<mode>3_r4000, *mul_acc_si, *mul_acc_si_r3900)
11293 (*msac_using_macc, *mul_sub_si, <u>mulsidi3_32bit_r4000)
11294 (<u>mulsidi3_64bit_r4000, <su>muldi3_highpart_internal)
11295 (<su>mulsi3_highpart_split, <su>muldi3_highpart_internal)
11296 (<u>mulditi3_r4000, *div<mode>3, *recip<mode>3, divmod<mode>4)
11297 (udivmod<mode>4, sqrt<mode>2, *rsqrt<mode>a, *rsqrt<mode>b)
11298 (fix_truncdfsi2_macro, fix_truncsfsi2_macro, *lea_high64)
11299 (*lea64, cprestore_<mode>, clear_hazard_<mode>, <unnamed insn>)
11300 (casesi_internal_mips16_<mode>, *tls_get_tp_<mode>_split)
11301 (tls_get_tp_mips16, *tls_get_tp_mips16_call_<mode>): Use "insn_count"
11302 rather than "length".
11303 (tls_get_tp_<mode>): Likewise. Remove redundant "no_delay" attribute.
11304 * config/mips/mips-ps-3d.md (mips_c_cond_4s, mips_cabs_cond_4s):
11305 Use "insn_count" rather than "length".
11306 * config/mips/mips-dsp.md
11307 (mips_l<SHORT:size><u>x_ext<GPR:mode>_<P:mode>)
11308 (mips_l<GPR:size>x_<P:mode>, *mips_lw<u>x_<P:mode>_ext): Remove
11311 2013-06-12 Marc Glisse <marc.glisse@inria.fr>
11313 PR tree-optimization/57361
11314 * tree-ssa-dse.c (dse_possible_dead_store_p): Handle self-assignment.
11316 2013-06-12 Sofiane Naci <sofiane.naci@arm.com>
11318 * config/aarch64/aarch64-simd.md (aarch64_combine<mode>): Convert
11320 (aarch64_simd_combine<mode>): New instruction expansion.
11321 * config/aarch64/aarch64-protos.h (aarch64_split_simd_combine): New
11322 function prototype.
11323 * config/aarch64/aarch64.c (aarch64_split_combine): New function.
11324 * config/aarch64/iterators.md (Vdbl): Add entry for DF.
11326 2013-06-12 Jan Hubicka <jh@suse.cz>
11328 * cgraph.c (verify_edge_corresponds_to_fndecl): Be lax about
11329 decl has when in streaming stage.
11330 * lto-symtab.c (lto_symtab_merge_symbols): Likewise.
11331 * cgraph.h (cgraph_state): Add CGRAPH_LTO_STREAMING.
11333 2013-06-12 Roland Stigge <stigge@antcom.de>
11336 * config/rs6000/t-linux (MULTIARCH_DIRNAME): Fix SPE version detection.
11338 2013-06-12 Jakub Jelinek <jakub@redhat.com>
11340 PR tree-optimization/57537
11341 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): If
11342 vect_handle_widen_op_by_const, convert oprnd1 to half_type1.
11344 2013-06-12 Richard Biener <rguenther@suse.de>
11346 * data-streamer.h (streamer_write_char_stream): CSE
11347 obs->current_pointer.
11348 * data-streamer-out.c (streamer_write_uhwi_stream): Inline
11349 streamer_write_char_stream manually and optimize the resulting loop.
11350 (streamer_write_hwi_stream): Likewise.
11352 2013-06-12 Jan Hubicka <jh@suse.cz>
11354 * lto-symtab.c (lto_symtab_merge_symbols): Populate symtab hashtable.
11355 * cgraph.h (varpool_create_empty_node): Declare.
11356 * lto-cgraph.c (input_node, input_varpool_node): Forcingly create
11358 * symtab.c (symtab_unregister_node): Be lax about missin entries
11360 (symtab_get_node): Update comment.
11361 * varpool.c (varpool_create_empty_node): Break out from ...
11362 (varpool_node_for_decl): ... here.
11363 * lto-streamer.h (lto_file_decl_data): Add RESOLUTION_MAP.
11365 2013-06-12 Eric Botcazou <ebotcazou@adacore.com>
11367 * expr.c (expand_expr_real_1) <TARGET_MEM_REF>: Use straight-line flow.
11368 <MEM_REF>: Use 'type' instead of TREE_TYPE (exp) and tidy up the first
11369 part. Use straight-line flow at the end.
11370 <COMPONENT_REF>: Remove superfluous else.
11371 <VIEW_CONVERT_EXPR>: Use 'type' instead of TREE_TYPE (exp).
11373 2013-06-12 Jakub Jelinek <jakub@redhat.com>
11376 * varasm.c (decl_binds_to_current_def_p): Call binds_local_p
11377 target hook even for !TREE_PUBLIC decls. If no resolution info
11378 is available, return false for common and external decls.
11380 2013-06-12 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
11382 * config/rl78/constraints.md (U): New constraint.
11383 * config/rl78/rl78.md (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): Add
11386 2013-06-11 Michael Meissner <meissner@linux.vnet.ibm.com>
11389 * config/rs6000/driver-rs6000.c (elf_platform): Make buffer static
11390 to allow returning address to AT_PLATFORM name.
11392 2013-06-11 Jan Hubicka <jh@suse.cz>
11394 * cgraph.c (cgraph_create_function_alias): Set weakref flag.
11395 * cgraph.h (symtab_node_base): Add weakref flag.
11396 * cgraphunit.c (cgraph_reset_node): Clear weakref flag.
11397 (handle_alias_pairs): Set weakref flag, do not set DECL_EXTERNAL.
11398 (output_weakrefs): Use weakref flag.
11399 * fold-const.c (simple_operand_p): Handle WEAK.
11400 * gimple-fold.c (can_refer_decl_in_current_unit_p): Drop weakref.
11401 * ipa.c (varpool_externally_visible_p): Drop weakref.
11402 (function_and_variable_visibility): Update comment; fix weakref
11403 sanity checks; do not clear DECL_WEAK on them.
11404 * lto-cgraph.c (lto_output_node): update.
11405 (lto_output_varpool_node): Update.
11406 (input_overwrite_node): Update.
11407 (input_node): Update.
11408 (input_varpool_node): Update.
11409 * lto-symtab.c (lto_symtab_symbol_p): Do not special case weakrefs.
11410 (lto_symtab_merge_symbols): Add sanity check.
11411 (lto_symtab_prevailing_decl): Do not special case weakrefs.
11412 * passes.c (rest_of_decl_compilation): Set static flag, too.
11413 * symtab.c (dump_symtab_base): Dump weakref.
11414 (verify_symtab_base): Sanity check weakrefs.
11415 (symtab_make_decl_local): Remove duplicated code.
11416 (symtab_alias_ultimate_target): Simplify.
11417 * varpool.c (varpool_create_variable_alias): Set weakref flag.
11419 2013-06-11 Tom de Vries <tom@codesourcery.com>
11421 * genautomata.c (gen_regexp_sequence): Handle els_num == -1. Handle
11422 sequence_vect == NULL.
11424 2013-06-11 DJ Delorie <dj@redhat.com>
11426 * config/rl78/rl78.c (TARGET_UNWIND_WORD_MODE): Define.
11427 (rl78_unwind_word_mode): New.
11429 2013-06-11 David Malcolm <dmalcolm@redhat.com>
11431 * final.c (debug_prefix_maps): Make static.
11433 2013-06-11 David Malcolm <dmalcolm@redhat.com>
11435 * function.c (initial_trampoline): Remove stray copy.
11437 2013-06-11 Sofiane Naci <sofiane.naci@arm.com>
11439 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update.
11441 2013-06-11 Martin Jambor <mjambor@suse.cz>
11443 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that param_index is
11444 within bounds at the beginning of the function.
11446 2013-06-11 Alan Modra <amodra@gmail.com>
11448 * varasm.c (get_section): Don't die on !DECL_P decl. Tidy error
11450 (get_named_section): Don't NULL !DECL_P decl.
11452 2013-06-11 Igor Zamyatin <igor.zamyatin@intel.com>
11454 * doc/invoke.texi (core-avx2): Document.
11456 (atom): Updated with MOVBE.
11458 2013-06-11 Richard Biener <rguenther@suse.de>
11460 * collect2.c (main): Do not redirect ld stdout/stderr when debugging.
11462 2013-06-11 Anton Blanchard <anton@samba.org>
11464 * config/rs6000/rs6000.c (rs6000_adjust_atomic_subword): Calculate
11465 correct shift value in little-endian mode.
11467 2013-06-11 Jakub Jelinek <jakub@redhat.com>
11470 * varasm.c (get_variable_align): Move #endif to the right place.
11472 2013-06-10 Cary Coutant <ccoutant@google.com>
11474 * dwarf2out.c (hash_external_ref): Use die_symbol or signature
11475 for hash so that hash table traversal order is deterministic.
11477 2013-06-10 Michael Meissner <meissner@linux.vnet.ibm.com>
11478 Pat Haugen <pthaugen@us.ibm.com>
11479 Peter Bergner <bergner@vnet.ibm.com>
11481 * config/rs6000/vector.md (GPR move splitter): Do not split moves
11482 of vectors in GPRS if they are direct moves or quad word load or
11485 * config/rs6000/rs6000-protos.h (rs6000_output_move_128bit): Add
11487 (direct_move_p): Likewise.
11488 (quad_load_store_p): Likewise.
11490 * config/rs6000/rs6000.c (enum rs6000_reg_type): Simplify register
11491 classes into bins based on the physical register type.
11492 (reg_class_to_reg_type): Likewise.
11493 (IS_STD_REG_TYPE): Likewise.
11494 (IS_FP_VECT_REG_TYPE): Likewise.
11495 (reload_fpr_gpr): Arrays to determine what insn to use if we can
11496 use direct move instructions.
11497 (reload_gpr_vsx): Likewise.
11498 (reload_vsx_gpr): Likewise.
11499 (rs6000_init_hard_regno_mode_ok): Precalculate the register type
11500 information that is a simplification of register classes. Also
11501 precalculate direct move reload helpers.
11502 (direct_move_p): New function to return true if the operation can
11503 be done as a direct move instruciton.
11504 (quad_load_store_p): New function to return true if the operation
11505 is a quad memory operation.
11506 (rs6000_legitimize_address): If quad memory, only allow register
11507 indirect for TImode addresses.
11508 (rs6000_legitimate_address_p): Likewise.
11509 (enum reload_reg_type): Delete, replace with rs6000_reg_type.
11510 (rs6000_reload_register_type): Likewise.
11511 (register_to_reg_type): Return register type.
11512 (rs6000_secondary_reload_simple_move): New helper function for
11513 secondary reload and secondary memory needed to identify anything
11514 that is a simple move, and does not need reloading.
11515 (rs6000_secondary_reload_direct_move): New helper function for
11516 secondary reload to identify cases that can be done with several
11517 instructions via the direct move instructions.
11518 (rs6000_secondary_reload_move): New helper function for secondary
11519 reload to identify moves between register types that can be done.
11520 (rs6000_secondary_reload): Add support for quad memory operations
11521 and for direct move.
11522 (rs6000_secondary_memory_needed): Likewise.
11523 (rs6000_debug_secondary_memory_needed): Change argument names.
11524 (rs6000_output_move_128bit): New function to return the move to
11525 use for 128-bit moves, including knowing about the various
11526 limitations of quad memory operations.
11528 * config/rs6000/vsx.md (vsx_mov<mode>): Add support for quad
11529 memory operations. call rs6000_output_move_128bit for the actual
11530 instruciton(s) to generate.
11531 (vsx_movti_64bit): Likewise.
11533 * config/rs6000/rs6000.md (UNSPEC_P8V_FMRGOW): New unspec values.
11534 (UNSPEC_P8V_MTVSRWZ): Likewise.
11535 (UNSPEC_P8V_RELOAD_FROM_GPR): Likewise.
11536 (UNSPEC_P8V_MTVSRD): Likewise.
11537 (UNSPEC_P8V_XXPERMDI): Likewise.
11538 (UNSPEC_P8V_RELOAD_FROM_VSX): Likewise.
11539 (UNSPEC_FUSION_GPR): Likewise.
11540 (FMOVE128_GPR): New iterator for direct move.
11541 (f32_lv): New mode attribute for load/store of SFmode/SDmode values.
11542 (f32_sv): Likewise.
11543 (f32_dm): Likewise.
11544 (zero_extend<mode>di2_internal1): Add support for power8 32-bit
11545 loads and direct move instructions.
11546 (zero_extendsidi2_lfiwzx): Likewise.
11547 (extendsidi2_lfiwax): Likewise.
11548 (extendsidi2_nocell): Likewise.
11549 (floatsi<mode>2_lfiwax): Likewise.
11550 (lfiwax): Likewise.
11551 (floatunssi<mode>2_lfiwzx): Likewise.
11552 (lfiwzx): Likewise.
11553 (fix_trunc<mode>_stfiwx): Likewise.
11554 (fixuns_trunc<mode>_stfiwx): Likewise.
11555 (mov<mode>_hardfloat, 32-bit floating point): Likewise.
11556 (mov<move>_hardfloat64, 64-bit floating point): Likewise.
11557 (parity<mode>2_cmpb): Set length/type attr.
11558 (unnamed shift right patterns, mov<mode>_internal2): Change type attr
11559 for 'mr.' to fast_compare.
11560 (bpermd_<mode>): Change type attr to popcnt.
11561 (p8_fmrgow_<mode>): New insns for power8 direct move support.
11562 (p8_mtvsrwz_1): Likewise.
11563 (p8_mtvsrwz_2): Likewise.
11564 (reload_fpr_from_gpr<mode>): Likewise.
11565 (p8_mtvsrd_1): Likewise.
11566 (p8_mtvsrd_2): Likewise.
11567 (p8_xxpermdi_<mode>): Likewise.
11568 (reload_vsx_from_gpr<mode>): Likewise.
11569 (reload_vsx_from_gprsf): Likewise.
11570 (p8_mfvsrd_3_<mode>): LIkewise.
11571 (reload_gpr_from_vsx<mode>): Likewise.
11572 (reload_gpr_from_vsxsf): Likewise.
11573 (p8_mfvsrd_4_disf): Likewise.
11574 (multi-word GPR splits): Do not split direct moves or quad memory
11577 2013-06-10 David Malcolm <dmalcolm@redhat.com>
11579 * tree-into-ssa.c (interesting_blocks): Make static.
11581 2013-06-10 Jakub Jelinek <jakub@redhat.com>
11584 * varasm.c (align_variable): Don't use DATA_ALIGNMENT or
11585 CONSTANT_ALIGNMENT if !decl_binds_to_current_def_p (decl).
11586 Use DATA_ABI_ALIGNMENT for that case instead if defined.
11587 (get_variable_align): New function.
11588 (get_variable_section, emit_bss, emit_common,
11589 assemble_variable_contents, place_block_symbol): Use
11590 get_variable_align instead of DECL_ALIGN.
11591 (assemble_noswitch_variable): Add align argument, use it
11592 instead of DECL_ALIGN.
11593 (assemble_variable): Adjust caller. Use get_variable_align
11594 instead of DECL_ALIGN.
11595 * config/i386/i386.h (DATA_ALIGNMENT): Adjust x86_data_alignment
11597 (DATA_ABI_ALIGNMENT): Define.
11598 * config/i386/i386-protos.h (x86_data_alignment): Adjust prototype.
11599 * config/i386/i386.c (x86_data_alignment): Add opt argument. If
11600 opt is false, only return the psABI mandated alignment increase.
11601 * config/c6x/c6x.h (DATA_ALIGNMENT): Renamed to...
11602 (DATA_ABI_ALIGNMENT): ... this.
11603 * config/mmix/mmix.h (DATA_ALIGNMENT): Renamed to...
11604 (DATA_ABI_ALIGNMENT): ... this.
11605 * config/mmix/mmix.c (mmix_data_alignment): Adjust function comment.
11606 * config/s390/s390.h (DATA_ALIGNMENT): Renamed to...
11607 (DATA_ABI_ALIGNMENT): ... this.
11608 * doc/tm.texi.in (DATA_ABI_ALIGNMENT): Document.
11609 * doc/tm.texi: Regenerated.
11611 2013-06-10 Uros Bizjak <ubizjak@gmail.com>
11613 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Also use
11614 cmp_code to construct REG_EQUAL note.
11616 2013-06-09 Jakub Jelinek <jakub@redhat.com>
11619 * config/i386/i386.md (TARGET_READ_MODIFY_WRITE peepholes): Ensure
11620 that operands[2] doesn't overlap with operands[0].
11622 2013-06-09 David Edelsohn <dje.gcc@gmail.com>
11623 Jan Hubicka <jh@suse.cz>
11625 * config/rs6000/rs6000.c (print_operand, 'z'): Remove historical
11626 hack to mark symbols as used.
11628 2013-06-08 Vladimir Makarov <vmakarov@redhat.com>
11630 PR rtl-optimization/57559
11631 * lra-constraints.c (process_alt_operands): Don't discourage
11632 memory with known offset for offsetable memory constraint.
11633 * lra.c (lra_emit_add): Exchange y and z for 2-op add insn.
11635 2013-06-08 Eric Botcazou <ebotcazou@adacore.com>
11637 * varasm.c (struct oc_local_state): Reorder fields.
11638 (output_constructor_bitfield): Replace OUTER parameter with BIT_OFFSET
11639 and adjust accordingly.
11640 (output_constructor): Reorder initialization code and adjust call to
11641 output_constructor_bitfield.
11643 2013-06-07 Jan Hubicka <jh@suse.cz>
11645 * symtab.c (symtab_resolve_alias): Do not remove alias attribute.
11647 2013-06-07 David Malcolm <dmalcolm@redhat.com>
11649 * tree-object-size.c (unknown): Make const.
11651 2013-06-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
11653 * config/s390/s390.md (cpu_facility): Add cpu_zarch.
11654 ("*movmem_short", "*clrmem_short", "*cmpmem_short): Use cpu_zarch
11655 for last alternative in the cpu_facility attribute.
11657 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11660 * config/arm/arm.md (*xordi3_insn): Change to insn_and_split.
11661 (xordi3): Change operand 2 constraint to arm_xordi_operand.
11662 * config/arm/arm.c (const_ok_for_dimode_op): Handle XOR.
11663 * config/arm/constraints.md (Dg): New constraint.
11664 * config/arm/neon.md (xordi3_neon): Remove.
11665 (neon_veor<mode>): Generate xordi3 instead of xordi3_neon.
11666 * config/arm/predicates.md (arm_xordi_operand): New predicate.
11668 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11670 * config/arm/arm.md (anddi3_insn): Remove duplicate alternatives.
11671 Clean up alternatives.
11673 2013-06-07 Alan Modra <amodra@gmail.com>
11675 * config/rs6000/rs6000.c (setup_incoming_varargs): Round up
11678 2013-06-07 Alan Modra <amodra@gmail.com>
11680 * varasm.c (force_const_mem): Assert mode is not VOID or BLK.
11682 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11684 * config/arm/constraints.md (Df): New constraint.
11685 * config/arm/arm.md (iordi3_insn): Use Df constraint instead of De.
11686 Correct length attribute for last two alternatives.
11688 2013-06-07 Alan Modra <amodra@gmail.com>
11690 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
11691 override user -mfp-in-toc.
11692 (offsettable_ok_by_alignment): Consider just the current access
11693 rather than the whole object, unless BLKmode. Handle
11694 CONSTANT_POOL_ADDRESS_P constants that lack a decl too.
11695 (use_toc_relative_ref): Allow CONSTANT_POOL_ADDRESS_P constants
11696 for -mcmodel=medium.
11697 * config/rs6000/linux64.h (SUBSUBTARGET_OVERRIDE_OPTIONS): Don't
11698 override user -mfp-in-toc or -msum-in-toc. Default to
11699 -mno-fp-in-toc for -mcmodel=medium.
11701 2013-06-06 DJ Delorie <dj@redhat.com>
11703 * config/rl78/rl78.c (rl78_valid_pointer_mode): New, implements
11704 TARGET_VALID_POINTER_MODE.
11706 2013-06-06 Michael Meissner <meissner@linux.vnet.ibm.com>
11707 Pat Haugen <pthaugen@us.ibm.com>
11708 Peter Bergner <bergner@vnet.ibm.com>
11710 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
11711 Document new power8 builtins.
11713 * config/rs6000/vector.md (and<mode>3): Add a clobber/scratch of a
11714 condition code register, to allow 128-bit logical operations to be
11715 done in the VSX or GPR registers.
11716 (nor<mode>3): Use the canonical form for nor.
11717 (eqv<mode>3): Add expanders for power8 xxleqv, xxlnand, xxlorc,
11718 vclz*, and vpopcnt* vector instructions.
11719 (nand<mode>3): Likewise.
11720 (orc<mode>3): Likewise.
11721 (clz<mode>2): LIkewise.
11722 (popcount<mode>2): Likewise.
11724 * config/rs6000/predicates.md (int_reg_operand): Rework tests so
11725 that only the GPRs are recognized.
11727 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
11728 support for new power8 builtins.
11730 * config/rs6000/rs6000-builtin.def (xscvspdpn): Add new power8
11732 (xscvdpspn): Likewise.
11738 (vpopcnt): Likewise.
11739 (vpopcntb): Likewise.
11740 (vpopcnth): Likewise.
11741 (vpopcntw): Likewise.
11742 (vpopcntd): Likewise.
11744 (vmrgew): Likewise.
11745 (vmrgow): Likewise.
11747 (eqv_v16qi3): Likewise.
11748 (eqv_v8hi3): Likewise.
11749 (eqv_v4si3): Likewise.
11750 (eqv_v2di3): Likewise.
11751 (eqv_v4sf3): Likewise.
11752 (eqv_v2df3): Likewise.
11754 (nand_v16qi3): Likewise.
11755 (nand_v8hi3): Likewise.
11756 (nand_v4si3): Likewise.
11757 (nand_v2di3): Likewise.
11758 (nand_v4sf3): Likewise.
11759 (nand_v2df3): Likewise.
11761 (orc_v16qi3): Likewise.
11762 (orc_v8hi3): Likewise.
11763 (orc_v4si3): Likewise.
11764 (orc_v2di3): Likewise.
11765 (orc_v4sf3): Likewise.
11766 (orc_v2df3): Likewise.
11768 * config/rs6000/rs6000.c (rs6000_option_override_internal): Only
11769 allow power8 quad mode in 64-bit.
11770 (rs6000_builtin_vectorized_function): Add support to vectorize
11771 ISA 2.07 count leading zeros, population count builtins.
11772 (rs6000_expand_vector_init): On ISA 2.07 use xscvdpspn to form
11773 V4SF vectors instead of xscvdpsp to avoid IEEE related traps.
11774 (builtin_function_type): Add vgbbd builtin function which takes an
11776 (altivec_expand_vec_perm_const): Add support for new power8 merge
11779 * config/rs6000/vsx.md (VSX_L2): New iterator for 128-bit types,
11780 that does not include TImdoe for use with 32-bit.
11781 (UNSPEC_VSX_CVSPDPN): Support for power8 xscvdpspn and xscvspdpn
11783 (UNSPEC_VSX_CVDPSPN): Likewise.
11784 (vsx_xscvdpspn): Likewise.
11785 (vsx_xscvspdpn): Likewise.
11786 (vsx_xscvdpspn_scalar): Likewise.
11787 (vsx_xscvspdpn_directmove): Likewise.
11788 (vsx_and<mode>3): Split logical operations into 32-bit and
11789 64-bit. Add support to do logical operations on TImode as well as
11790 VSX vector types. Allow logical operations to be done in either
11791 VSX registers or in general purpose registers in 64-bit mode. Add
11792 splitters if GPRs were used. For AND, add clobber of CCmode to
11793 allow use of ANDI on GPRs. Rewrite nor to use the canonical RTL
11795 (vsx_and<mode>3_32bit): Likewise.
11796 (vsx_and<mode>3_64bit): Likewise.
11797 (vsx_ior<mode>3): Likewise.
11798 (vsx_ior<mode>3_32bit): Likewise.
11799 (vsx_ior<mode>3_64bit): Likewise.
11800 (vsx_xor<mode>3): Likewise.
11801 (vsx_xor<mode>3_32bit): Likewise.
11802 (vsx_xor<mode>3_64bit): Likewise.
11803 (vsx_one_cmpl<mode>2): Likewise.
11804 (vsx_one_cmpl<mode>2_32bit): Likewise.
11805 (vsx_one_cmpl<mode>2_64bit): Likewise.
11806 (vsx_nor<mode>3): Likewise.
11807 (vsx_nor<mode>3_32bit): Likewise.
11808 (vsx_nor<mode>3_64bit): Likewise.
11809 (vsx_andc<mode>3): Likewise.
11810 (vsx_andc<mode>3_32bit): Likewise.
11811 (vsx_andc<mode>3_64bit): Likewise.
11812 (vsx_eqv<mode>3_32bit): Add support for power8 xxleqv, xxlnand,
11813 and xxlorc instructions.
11814 (vsx_eqv<mode>3_64bit): Likewise.
11815 (vsx_nand<mode>3_32bit): Likewise.
11816 (vsx_nand<mode>3_64bit): Likewise.
11817 (vsx_orc<mode>3_32bit): Likewise.
11818 (vsx_orc<mode>3_64bit): Likewise.
11820 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Update comment.
11822 * config/rs6000/altivec.md (UNSPEC_VGBBD): Add power8 vgbbd
11824 (p8_vmrgew): Add power8 vmrgew and vmrgow instructions.
11825 (p8_vmrgow): Likewise.
11826 (altivec_and<mode>3): Add clobber of CCmode to allow AND using
11827 GPRs to be split under VSX.
11828 (p8v_clz<mode>2): Add power8 count leading zero support.
11829 (p8v_popcount<mode>2): Add power8 population count support.
11830 (p8v_vgbbd): Add power8 gather bits by bytes by doubleword
11833 * config/rs6000/rs6000.md (eqv<mode>3): Add support for powerp eqv
11836 * config/rs6000/altivec.h (vec_eqv): Add defines to export power8
11838 (vec_nand): Likewise.
11839 (vec_vclz): Likewise.
11840 (vec_vclzb): Likewise.
11841 (vec_vclzd): Likewise.
11842 (vec_vclzh): Likewise.
11843 (vec_vclzw): Likewise.
11844 (vec_vgbbd): Likewise.
11845 (vec_vmrgew): Likewise.
11846 (vec_vmrgow): Likewise.
11847 (vec_vpopcnt): Likewise.
11848 (vec_vpopcntb): Likewise.
11849 (vec_vpopcntd): Likewise.
11850 (vec_vpopcnth): Likewise.
11851 (vec_vpopcntw): Likewise.
11853 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
11855 PR rtl-optimization/57468
11856 * config/i386/i386.c (inline_secondary_memory_needed): Ignore
11859 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
11861 PR rtl-optimization/57459
11862 * lra-constraints.c (update_ebb_live_info): Fix typo for operand
11863 type when setting live regs.
11865 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
11867 * config/s390/s390.opt (mlra): New option.
11868 * config/s390/s390.c (s390_decompose_address): Check displacement
11869 for all registers for LRA.
11870 (s390_secondary_reload): Don't used secondary reloads for LRA.
11871 (s390_lra_p): New function.
11872 (TARGET_LRA_P): Define.
11873 * config/s390/s390.md (*movmem_short, *clrmem_short): Change value
11874 of attribute cpu_facility to zarch for the last alternative.
11875 (*cmpmem_short): Ditto.
11877 2013-06-06 Eric Botcazou <ebotcazou@adacore.com>
11879 * config/arm/arm.c (arm_r3_live_at_start_p): New predicate.
11880 (arm_compute_static_chain_stack_bytes): Use it. Tidy up.
11881 (arm_expand_prologue): Likewise.
11883 2013-06-06 Teresa Johnson <tejohnson@google.com>
11886 * ifcvt.c (find_if_case_1): Replace BB_COPY_PARTITION with assert
11887 as this is now done by redirect_edge_and_branch_force.
11888 * function.c (thread_prologue_and_epilogue_insns): Insert new bb after
11889 barriers, and fix interaction with splitting.
11890 * emit-rtl.c (try_split): Copy REG_CROSSING_JUMP notes.
11891 * cfgcleanup.c (try_forward_edges): Fix early return value to properly
11892 reflect changes made in the routine.
11893 * bb-reorder.c (emit_barrier_after_bb): Move to cfgrtl.c.
11894 (fix_up_fall_thru_edges): Remove incorrect check for bb layout order
11895 since this is called in cfglayout mode, and replace partition fixup
11896 with assert as that is now done by force_nonfallthru_and_redirect.
11897 (add_reg_crossing_jump_notes): Handle the fact that some jumps may
11898 already be marked with region crossing note.
11899 (insert_section_boundary_note): Make non-static, gate on flag
11900 has_bb_partition, rewrite to also check for multiple partitions.
11901 (rest_of_handle_reorder_blocks): Remove call to
11902 insert_section_boundary_note, now done later during free_cfg.
11903 (duplicate_computed_gotos): Don't duplicate partition crossing edge.
11904 * bb-reorder.h (insert_section_boundary_note): Declare.
11905 * Makefile.in (cfgrtl.o): Depend on bb-reorder.h
11906 * cfgrtl.c (rest_of_pass_free_cfg): If partitions exist
11907 invoke insert_section_boundary_note.
11908 (try_redirect_by_replacing_jump): Remove unnecessary
11909 check for region crossing note.
11910 (fixup_partition_crossing): New function.
11911 (rtl_redirect_edge_and_branch): Fixup partition boundaries.
11912 (emit_barrier_after_bb): Move here from bb-reorder.c, handle insertion
11913 in non-cfglayout mode.
11914 (force_nonfallthru_and_redirect): Fixup partition boundaries,
11915 remove old code that tried to do this. Emit barrier correctly
11916 when we are in cfglayout mode.
11917 (last_bb_in_partition): New function.
11918 (rtl_split_edge): Correctly fixup partition boundaries.
11919 (commit_one_edge_insertion): Remove old code that tried to
11920 fixup region crossing edge since this is now handled in
11921 split_block, and set up insertion point correctly since
11922 block may now end in a jump.
11923 (verify_hot_cold_block_grouping): Guard against checking when not in
11924 linearized RTL mode.
11925 (rtl_verify_edges): Add checks for incorrect/missing REG_CROSSING_JUMP
11927 (rtl_verify_flow_info_1): Move verify_hot_cold_block_grouping to
11928 rtl_verify_flow_info, so not called in cfglayout mode.
11929 (rtl_verify_flow_info): Move verify_hot_cold_block_grouping here.
11930 (fixup_reorder_chain): Remove old code that attempted to fixup region
11931 crossing note as this is now handled in force_nonfallthru_and_redirect.
11932 (duplicate_insn_chain): Don't duplicate switch section notes.
11933 (rtl_can_remove_branch_p): Remove unnecessary check for region crossing
11935 * basic-block.h (emit_barrier_after_bb): Declare.
11937 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11939 * config/arm/arm-fixed.md (add<mode>3,usadd<mode>3,ssadd<mode>3,
11940 sub<mode>3, ussub<mode>3, sssub<mode>3, arm_ssatsihi_shift,
11941 arm_usatsihi): Adjust alternatives for arm_restrict_it.
11943 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11945 * config/arm/arm-ldmstm.ml: Set "predicable_short_it" to "no"
11947 * config/arm/ldmstm.md: Regenerate.
11949 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11951 * config/arm/sync.md (atomic_loaddi_1):
11952 Disable predication for arm_restrict_it.
11953 (arm_load_exclusive<mode>): Likewise.
11954 (arm_load_exclusivesi): Likewise.
11955 (arm_load_exclusivedi): Likewise.
11956 (arm_load_acquire_exclusive<mode>): Likewise.
11957 (arm_load_acquire_exclusivesi): Likewise.
11958 (arm_load_acquire_exclusivedi): Likewise.
11959 (arm_store_exclusive<mode>): Likewise.
11960 (arm_store_exclusive<mode>): Likewise.
11961 (arm_store_release_exclusivedi): Likewise.
11962 (arm_store_release_exclusive<mode>): Likewise.
11964 2013-06-06 Richard Biener <rguenther@suse.de>
11966 * lto-streamer.h (enum LTO_tags): Move LTO_tree_pickle_reference
11968 (lto_tag_is_tree_code_p): Adjust.
11969 (lto_tag_is_gimple_code_p): Likewise.
11970 (lto_gimple_code_to_tag): Likewise.
11971 (lto_tag_to_gimple_code): Likewise.
11972 (lto_tree_code_to_tag): Likewise.
11973 (lto_tag_to_tree_code): Likewise.
11974 * data-streamer.h (streamer_write_hwi_in_range): Use
11975 uhwi streaming to stream the normalized range.
11976 (streamer_read_hwi_in_range): Likewise.
11978 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11980 * config/arm/arm.md (enabled_for_depr_it): New attribute.
11981 (predicable_short_it): Likewise.
11982 (predicated): Likewise.
11983 (enabled): Handle above.
11984 (define_cond_exec): Set predicated attribute to yes.
11986 2013-06-05 Mike Stump <mikestump@comcast.net>
11988 * gdbinit.in (__FUNCTION__): Add.
11990 2013-06-05 Uros Bizjak <ubizjak@gmail.com>
11992 * config/alpha/alpha.c (alpha_emit_conditional_move): Swap all
11993 GE, GT, GEU and GTU compares, modulo DImode compares with zero.
11995 2013-06-05 Jan Hubicka <jh@suse.cz>
11997 * varasm.c (mark_decl_referenced): Revert the removal until targets
12000 2013-06-05 David Edelsohn <dje.gcc@gmail.com>
12002 * config/rs6000/rs6000.c (print_operand, 'z'): Use DECL_PRESERVE_P
12003 instead of mark_decl_referenced.
12005 2013-06-05 Jan Hubicka <jh@suse.cz>
12007 * cgraph.c (cgraph_remove_node): Clear forced_by_abi.
12008 (cgraph_node_cannot_be_local_p_1): Honnor symbol.forced_by_abi
12009 and symtab_used_from_object_file_p.
12010 (cgraph_make_node_local_1): Clear forced_by_abi.
12011 (cgraph_can_remove_if_no_direct_calls_and): Use forced_by_abi
12012 * cgraph.h (symtab_node_base): Add forced_by_abi.
12013 (decide_is_variable_needed): Remove.
12014 (varpool_can_remove_if_no_refs): Honnor symbol.forced_by_abi.
12015 * cgraphunit.c (cgraph_decide_is_function_needed): Rename to ..
12016 (decide_is_symbol_needed): ... this one; handle symbols in general;
12017 always analyze virtuals; honnor forced_by_abi.
12018 (cgraph_finalize_function): Update.
12019 (varpool_finalize_decl): Update.
12020 (symbol_defined_and_needed): Remove.
12021 (analyze_functions): Update.
12022 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
12023 output_refs, input_overwrite_node): Handle forced_by_abi.
12024 * ipa.c (cgraph_address_taken_from_non_vtable_p): Rename to ...
12025 (address_taken_from_non_vtable_p): ... this one.
12026 (comdat_can_be_unshared_p_1): New function.
12027 (cgraph_comdat_can_be_unshared_p): Rename to ...
12028 (comdat_can_be_unshared_p): ... this one; handle symbols in general.
12029 (varpool_externally_visible_p): Use comdat_can_be_unshared_p.
12030 (function_and_variable_visibility): Clear forced_by_abi as needed.
12031 * trans-mem.c (ipa_tm_mark_forced_by_abi_node): New functoin.
12032 (ipa_tm_create_version_alias, ipa_tm_create_version): Update.
12033 * symtab.c (dump_symtab_base): Dump forced_by_abi.
12034 * varpool.c (decide_is_variable_needed): Remove.
12036 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12038 * config/arm/arm.c (MAX_INSN_PER_IT_BLOCK): New macro.
12039 (arm_option_override): Override arm_restrict_it where appropriate.
12040 (thumb2_final_prescan_insn): Use MAX_INSN_PER_IT_BLOCK.
12041 * config/arm/arm.opt (mrestrict-it): New command-line option.
12042 * doc/invoke.texi: Document -mrestrict-it.
12044 2013-06-05 David Malcolm <dmalcolm@redhat.com>
12046 * tsan.c (tsan_atomic_table): Make const.
12048 2013-06-05 Richard Biener <rguenther@suse.de>
12050 * tree-streamer.c (streamer_tree_cache_insert_1): Update the
12051 index associated with the tree we are supposed to replace.
12052 * tree-streamer-out.c (pack_ts_base_value_fields): Output
12053 TREE_ASM_WRITTEN as zero for everything but SSA names.
12055 2013-06-05 David Malcolm <dmalcolm@redhat.com>
12057 * tree-ssa-structalias.c (call_stmt_vars): Make static.
12059 2013-06-04 Jan Hubicka <jh@suse.cz>
12061 * lto-cgraph.c (get_alias_symbol): Remove weakref sanity check.
12062 (input_node, input_varpool_node): Handle correctly external same
12064 * ipa.c (symtab_remove_unreachable_nodes): Do not remove external
12065 nodes at ltrans stage.
12067 2013-06-04 Jan Hubicka <jh@suse.cz>
12069 * ipa-inline.c (update_caller_keys): Fix availability test.
12070 (update_callee_keys): Likewise.
12071 * symtab.c (symtab_alias_ultimate_target): Make availaiblity logic
12072 to follow ELF standard.
12074 2013-06-04 Jürgen Urban <JuergenUrban@gmx.de>
12076 * config.gcc (mipsr5900-*-elf*, mipsr5900el-*-elf*, mips64r5900-*-elf*)
12077 (mips64r5900el-*-elf*): New configurations.
12078 * config/mips/mips-cpus.def (r5900): New processor.
12079 * config/mips/mips-tables.opt: Regenerate.
12080 * config/mips/mips.c (mips_rtx_cost_data): Add an R5900 entry.
12081 (mips_issue_rate): Handle PROCESSOR_R5900.
12082 (mips_reorg_process_insns): Force reorder mode for the R5900.
12083 * config/mips/mips.h (TARGET_MIPS5900): Define.
12084 (ISA_HAS_CONDMOVE, ISA_HAS_PREFETCH, ISA_HAS_HILO_INTERLOCKS): Include
12086 (ISA_HAS_LOAD_DELAY, ISA_HAS_XFER_DELAY, ISA_HAS_FCMP_DELAY): Exclude
12088 * config/mips/mips.md (processor): Add r5900.
12089 (MOVECC): Disallow CCmode conditions for TARGET_MIPS5900.
12091 2013-06-04 Ian Bolton <ian.bolton@arm.com>
12093 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Call
12094 into function to generate MOVI instruction.
12095 * config/aarch64/aarch64.c (aarch64_simd_container_mode): New function.
12096 (aarch64_preferred_simd_mode): Turn into wrapper.
12097 (aarch64_output_scalar_simd_mov_immediate): New function.
12098 * config/aarch64/aarch64-protos.h: Add prototype for above.
12100 2013-06-04 Ian Bolton <ian.bolton@arm.com>
12102 * config/aarch64/aarch64.c (simd_immediate_info): Remove
12103 element_char member.
12104 (sizetochar): Return signed char.
12105 (aarch64_simd_valid_immediate): Remove elchar and other
12106 unnecessary variables.
12107 (aarch64_output_simd_mov_immediate): Take rtx instead of &rtx.
12108 Calculate element_char as required.
12109 * config/aarch64/aarch64-protos.h: Update and move prototype
12110 for aarch64_output_simd_mov_immediate.
12111 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>):
12114 2013-06-04 Ian Bolton <ian.bolton@arm.com>
12116 * config/aarch64/aarch64.c (simd_immediate_info): Struct to hold
12117 information completed by aarch64_simd_valid_immediate.
12118 (aarch64_legitimate_constant_p): Update arguments.
12119 (aarch64_simd_valid_immediate): Work with struct rather than many
12121 (aarch64_simd_scalar_immediate_valid_for_move): Update arguments.
12122 (aarch64_simd_make_constant): Update arguments.
12123 (aarch64_output_simd_mov_immediate): Work with struct rather than
12124 many pointers. Output immediate directly rather than as operand.
12125 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
12127 * config/aarch64/constraints.md (Dn): Update arguments.
12129 2013-06-04 Ian Bolton <ian.bolton@arm.com>
12131 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): No
12133 (aarch64_simd_immediate_valid_for_move): Remove.
12134 (aarch64_simd_scalar_immediate_valid_for_move): Update call.
12135 (aarch64_simd_make_constant): Update call.
12136 (aarch64_output_simd_mov_immediate): Update call.
12137 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
12139 * config/aarch64/constraints.md (Dn): Update call.
12141 2013-06-04 Ian Bolton <ian.bolton@arm.com>
12143 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Change
12144 return type to bool for prototype.
12145 (aarch64_legitimate_constant_p): Check for true instead of not -1.
12146 (aarch64_simd_valid_immediate): Fix up each return to return a bool.
12147 (aarch64_simd_immediate_valid_for_move): Update retval for bool.
12149 2013-06-04 Catherine Moore <clm@codesourcery.com>
12151 * config/mips/mips.opt (meva): New.
12152 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Define __mips_eva.
12153 (ASM_SPEC): Handle -meva.
12154 * doc/invoke.texi (meva): Document.
12156 2013-06-04 Alan Modra <amodra@gmail.com>
12158 * config/rs6000/rs6000.c (output_toc): Correct little-endian float
12161 2013-06-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12163 * rtl.def: Add extra fourth optional field to define_cond_exec.
12164 * gensupport.c (process_one_cond_exec): Process attributes from
12166 * doc/md.texi: Document fourth field in define_cond_exec.
12168 2013-06-04 Eric Botcazou <ebotcazou@adacore.com>
12170 * expmed.c (extract_bit_field_1): In the larger-than-a-word case, factor
12171 out the processing order as in store_bit_field_1.
12173 2013-06-04 Jan Hubicka <jh@suse.cz>
12175 PR middle-end/57500
12176 * cgraphunit.c (cgraph_process_same_body_aliases): Create
12177 non-VAR_DECL node if it does not exist yet.
12179 2013-06-03 Richard Sandiford <rdsandiford@googlemail.com>
12181 * config.gcc (mipsisa64sr71k-*-elf*, mipsisa64sb1-*-elf*)
12182 (mipsisa64sb1el-*-elf*, mips64-*-elf*, mips64el-*-elf*)
12183 (mips64orion-*-elf*, mips64orionel-*-elf*): Remove
12184 target_cpu_default setting.
12186 2013-06-03 Teresa Johnson <tejohnson@google.com>
12188 * dumpfile.c (opt_info_switch_p): Change -fopt-info
12189 default to -fopt-info=optimized instead of all.
12190 * doc/invoke.texi: Ditto.
12191 * tree-vectorizer.c (vectorize_loops): Emit loop vectorization
12192 success under MSG_OPTIMIZED_LOCATIONS, and use dump_printf_loc.
12193 (execute_vect_slp): Emit BB vectorization success under
12194 MSG_OPTIMIZED_LOCATIONS.
12195 * tree-vect-slp.c (vect_slp_transform_bb): Change
12196 MSG_OPTIMIZED_LOCATIONS to MSG_NOTE.
12197 * tree-vect-loop.c (vect_transform_loop): Ditto.
12199 2013-06-03 Jason Merrill <jason@redhat.com>
12202 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
12203 Use TARGET_EXPR for C++.
12205 2013-06-03 Jakub Jelinek <jakub@redhat.com>
12207 PR rtl-optimization/57268
12208 * sched-deps.c (sched_analyze_2): Don't flush_pending_lists
12209 if DEBUG_INSN_P (insn).
12212 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
12214 PR rtl-optimization/57268
12215 * sched-deps.c (sched_analyze_2): Flush dependence lists if
12216 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
12218 2013-06-03 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
12220 * config/i386/i386.c (ix86_lea_outperforms): Fix formatting.
12221 (ix86_avoid_lea_for_addr): Likewise.
12222 (exact_dependency_1): Likewise.
12223 (ix86_adjust_cost): Likewise.
12224 (swap_top_of_ready_list): Fix formatting and !reload_completed check
12226 (do_reorder_for_imul): Fix typo, formatting and
12227 !reload_completed check removed.
12228 (ix86_sched_reorder): Fix typo and formatting.
12229 (fold_builtin_cpu): Move M_INTEL_SLM at the end of processor types
12232 2013-06-03 Sofiane Naci <sofiane.naci@arm.com>
12234 * config/aarch64/aarch64.md (*movdi_aarch64): Define "simd" attribute.
12236 2013-06-03 Eric Botcazou <ebotcazou@adacore.com>
12238 * varasm.c (output_constant) <CONSTRUCTOR>: Minor formatting tweak.
12239 <STRING_CST>: Likewise.
12240 <VECTOR_CST>: Likewise.
12242 2013-06-01 Janus Weil <janus@gcc.gnu.org>
12243 Mikael Morin <mikael@gcc.gnu.org>
12245 * configure.ac: Add AC_HEADER_TIOCGWINSZ macro.
12246 * config.in: Regenerated.
12247 * configure: Regenerated.
12249 2013-06-01 Jan Hubicka <jh@suse.cz>
12251 PR middle-end/57366
12252 * cgraphunit.c (compile): When weakref is not supported,
12253 set up transparent aliases before final output pass.
12254 * varasm.c (assemble_alias): Do not try to do it here.
12256 2013-06-01 Jan Hubicka <jh@suse.cz>
12258 PR middle-end/57467
12259 * passes.c (for_per_function): Skip unanalyzed functions.
12261 2013-06-01 Jan Hubicka <jh@suse.cz>
12263 * lto-symtab.c (lto_symtab_merge_cgraph_nodes_1): Rename to ...
12264 (lto_symtab_merge_symbols_1): ... this one.
12265 (lto_symtab_merge_cgraph_nodes): Rename to ...
12266 (lto_symtab_merge_symbols): ... this one; simplify.
12267 * cgraph.c (same_body_aliases_done): Rename to ...
12268 (cpp_implicit_aliases_done): ... this one.
12269 (cgraph_create_function_alias): Update.
12270 (cgraph_same_body_alias): Update.
12271 (dump_cgraph_node): Remove alias dumping; simplify thunk dumping.
12272 (verify_edge_corresponds_to_fndecl): Simplify.
12273 * cgraph.h (symtab_node_base): Add cpp_implicit_alias, alias_target.
12274 (cgraph_node): Remove same_body_alias.
12275 (varpool_node): Remove alias_of and extra_name_alias.
12276 (same_body_aliases_done): Rename to ..
12277 (cpp_implicit_aliases_done): ... this one.
12278 (symtab_alias_ultimate_target): Add default parameter.
12279 (symtab_resolve_alias): New function.
12280 (fixup_same_cpp_alias_visibility): Declare.
12281 (cgraph_function_node): Add default parameter.
12282 (cgraph_node_asm_name): Likewise.
12283 (cgraph_function_or_thunk_node): Add default parameter; do
12284 not ICE when it is NULL.
12285 (varpool_variable_node): Likewise.
12286 * tree-emutls.c (create_emultls_var): Update.
12287 (ipa_lower_emutls): Update.
12288 * cgraphunit.c (cgraph_decide_is_function_needed): Update.
12289 (cgraph_reset_node): Reset alias info.
12290 (cgraph_finalize_function): Update.
12291 (fixup_same_cpp_alias_visibility): Move to symtab.c.
12292 (analyze_function): Simplify.
12293 (cgraph_process_same_body_aliases): Simplify.
12294 (analyze_functions): Fixup same body aliases.
12295 (handle_alias_pairs): Simplify.
12296 (assemble_thunk): Update.
12297 (assemble_thunks_and_aliases): Update.
12298 (output_weakrefs): Rewrite.
12299 * lto-cgraph.c (lto_output_node): Rewrite alias handling.
12300 (lto_output_varpool_node): Likewise.
12301 (compute_ltrans_boundary): Remve assert.
12302 (get_alias_symbol): New functoin.
12303 (input_node): Rewrite alias handling.
12304 (input_varpool_node): Likewise.
12305 * ipa-pure-const.c (propagate_pure_const): Fix formating.
12306 * ipa.c (process_references): Handle weakrefs correctly.
12307 (symtab_remove_unreachable_nodes): Likewise.
12308 * trans-mem.c (get_cg_data): Update.
12309 (ipa_tm_create_version_alias): Update.
12310 (ipa_tm_execute): Update.
12311 * symtab.c (dump_symtab_base): Dump aliases.
12312 (verify_symtab_base): Verify aliases.
12313 (symtab_node_availability): New function.
12314 (symtab_alias_ultimate_target): Simplify.
12315 (fixup_same_cpp_alias_visibility): Move here from cgraphunit.c;
12316 handle all the fixup cases.
12317 (symtab_resolve_alias): New function.
12318 * passes.c (ipa_write_summaries): Handle weakrefs.
12319 * varpool.c (varpool_analyze_node): Simplify.
12320 (assemble_aliases): Update.
12321 (varpool_create_variable_alias): Simplify.
12322 (varpool_extra_name_alias): Simplify.
12323 * lto-streamer.h (lto_symtab_merge_cgraph_nodes): Rename to...
12324 (lto_symtab_merge_symbols): ... this one.
12326 2013-06-01 Dinar Temirbulatov <dinar@kugelworks.com>
12329 PR rtl-optimization/57268
12330 * sched-deps.c (sched_analyze_2): Flush dependence lists if
12331 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
12333 2013-06-01 Tobias Burnus <burnus@net-b.de>
12335 Partially reverted:
12336 2013-05-31 Tobias Burnus <burnus@net-b.de>
12338 PR middle-end/57073
12339 * tree-ssa-math-opts.c (execute_cse_sincos): Move check
12342 2013-05-31 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 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
12350 * config/rs6000/predicates.md (rs6000_cbranch_operator): Accept some
12351 unordered comparison operators when -fno-trapping-math is in effect
12353 * config/rs6000/rs6000.c (rs6000_generate_compare): Remove dead code
12354 and implement unordered comparison operators properly on the e500.
12356 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
12358 * simplify-rtx.c (simplify_byte_swapping_operation): Use proper macro
12359 for constant scalar integers.
12360 (simplify_relational_operation_1): Likewise.
12362 2013-05-31 Segher Boessenkool <segher@kernel.crashing.org>
12364 * config/rs6000/rs6000-opts.h (enum processor_type): Reorder.
12365 * config/rs6000/rs6000.md (cpu): Reorder. Split long line.
12368 2013-05-31 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
12369 Igor Zamyatin <igor.zamyatin@intel.com>
12371 Silvermont (SLM) architecture performance tuning.
12372 * config/i386/i386.h (enum ix86_tune_indices): Add
12373 X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS.
12374 (TARGET_SPLIT_MEM_OPND_FOR_FP_CONVERTS): New define.
12376 * config/i386/i386.c (initial_ix86_tune_features)
12377 <X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS>: Initialize.
12378 (ix86_lea_outperforms): Handle Silvermont tuning.
12379 (ix86_avoid_lea_for_add): Add new argument to ix86_lea_outperforms
12381 (ix86_use_lea_for_mov): Likewise.
12382 (ix86_avoid_lea_for_addr): Likewise.
12383 (ix86_lea_for_add_ok): Likewise.
12384 (exact_dependency_1): New function.
12385 (exact_store_load_dependency): Likewise.
12386 (ix86_adjust_cost): Handle Silvermont tuning.
12387 (do_reoder_for_imul): Likewise.
12388 (swap_top_of_ready_list): New function.
12389 (ix86_sched_reorder): Changed to handle Silvermont tuning.
12391 * config/i386/i386.md (peepholes that split memory operand in fp
12394 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
12396 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
12397 Remove un-necessary braces.
12399 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
12401 * config/aarch64/aarch64.c (aarch64_classify_symbol):
12402 Use SYMBOL_TINY_ABSOLUTE for AARCH64_CMODEL_TINY_PIC.
12404 2013-05-31 Tobias Burnus <burnus@net-b.de>
12406 PR middle-end/57073
12407 * tree-ssa-math-opts.c (execute_cse_sincos): Move check further up.
12409 2013-05-31 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12412 * config/arm/arm.c (const_ok_for_dimode_op): Handle IOR.
12413 * config/arm/arm.md (*iordi3_insn): Change to insn_and_split.
12414 * config/arm/neon.md (iordi3_neon): Remove.
12415 (neon_vorr<mode>): Generate iordi3 instead of iordi3_neon.
12416 * config/arm/predicates.md (imm_for_neon_logic_operand):
12417 Move to earlier in the file.
12418 (neon_logic_op2): Likewise.
12419 (arm_iordi_operand_neon): New predicate.
12421 2013-05-31 Richard Biener <rguenther@suse.de>
12423 PR tree-optimization/57478
12424 PR tree-optimization/57453
12425 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Uses in PHI nodes
12428 2013-05-31 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
12430 * config/rl78/rl78.md (mulqi3,mulhi3): New define_expands.
12431 (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): New define_insns.
12433 2013-05-30 Tobias Burnus <burnus@net-b.de>
12434 Thomas Koenig <tkoenig@gcc.gnu.org>
12436 PR middle-end/57073
12437 * tree-ssa-math-opts.c (execute_cse_sincos): Optimize
12438 powi (-1.0, k) to (k & 1) ? -1.0 : 1.0.
12440 2013-05-30 Steven Bosscher <steven@gcc.gnu.org>
12442 * rtlanal.c (tablejump_p): Expect table and label to be adjacent.
12444 2013-05-30 Vladimir Makarov <vmakarov@redhat.com>
12446 * target.def (register_usage_leveling_p): New hook.
12447 * targhooks.c (default_register_usage_leveling_p): New.
12448 * targhooks.h (default_register_usage_leveling_p): New prototype.
12449 * lra-assigns.c (register_usage_leveling_p): Use the hook.
12450 * doc/tm.texi.in (TARGET_REGISTER_USAGE_LEVELING_P): New hook.
12451 * doc/tm.texi: Update.
12452 * config/i386/i386.c (TARGET_REGISTER_USAGE_LEVELING_P): Define.
12454 2013-05-30 Ian Bolton <ian.bolton@arm.com>
12456 * config/aarch64/aarch64.md (insv<mode>): New define_expand.
12457 (*insv_reg<mode>): New define_insn.
12459 2013-05-30 Joern Rennecke <joern.rennecke@embecosm.com>
12461 PR rtl-optimization/57439
12462 * postreload.c (move2add_valid_value_p): Check that we have
12463 a zero subreg_regno_offset when accessing the register in
12464 the requested mode.
12466 2013-05-30 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
12467 Igor Zamyatin <igor.zamyatin@intel.com>
12469 Silvermont (SLM) architecture pipeline model, tuning and
12471 * config.gcc: Add slm config options and target.
12473 * config/i386/slm.md: New.
12475 * config/i386/driver-i386.c (host_detect_local_cpu): Check movbe.
12477 * config/i386/i386-c.c (ix86_target_macros_internal): New case
12479 (ix86_target_macros_internal): Likewise.
12481 * config/i386/i386.c (slm_cost): New cost.
12482 (m_SLM): New macro flag.
12483 (initial_ix86_tune_features): Set m_SLM.
12484 (x86_accumulate_outgoing_args): Likewise.
12485 (x86_arch_always_fancy_math_387): Likewise.
12486 (processor_target_table): Add slm cost.
12487 (cpu_names): Add slm cpu name.
12488 (x86_option_override_internal): Set SLM ISA.
12489 (ix86_issue_rate): New case PROCESSOR_SLM.
12490 (ia32_multipass_dfa_lookahead): Likewise.
12491 (fold_builtin_cpu): Add slm.
12493 * config/i386/i386.h (TARGET_SLM): New target macro.
12494 (target_cpu_default): Add TARGET_CPU_DEFAULT_slm.
12495 (processor_type): Add PROCESSOR_SLM.
12497 * config/i386/i386.md (cpu): Add new value "slm".
12498 (slm.md): Include slm.md.
12500 2013-05-30 Bernd Schmidt <bernds@codesourcery.com>
12501 Zhenqiang Chen <zhenqiang.chen@linaro.org>
12503 * config/arm/arm-protos.h: Add and update function protos.
12504 * config/arm/arm.c (use_simple_return_p): New added.
12505 (thumb2_expand_return): Check simple_return flag.
12506 * config/arm/arm.md: Add simple_return and conditional simple_return.
12507 * config/arm/iterators.md: Add iterator for return and simple_return.
12509 2013-05-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
12511 * config/arm/arm.c (arm_add_cfa_adjust_cfa_note): New added.
12512 (arm_emit_multi_reg_pop): Add REG_CFA_ADJUST_CFA notes.
12513 (arm_emit_vfp_multi_reg_pop): Likewise.
12514 (thumb2_emit_ldrd_pop): Likewise.
12515 (arm_expand_epilogue): Add misc REG_CFA notes.
12516 (arm_unwind_emit): Skip REG_CFA_ADJUST_CFA and REG_CFA_RESTORE.
12518 2013-05-29 Lawrence Crowl <crowl@google.com>
12520 * config/arm/t-arm: Update for below.
12522 * config/arm/arm.c (arm_libcall_uses_aapcs_base::libcall_htab):
12523 Change type to hash_table. Update dependent calls and types.
12525 * config/i386/t-cygming: Update for below.
12527 * config/i386/t-interix: Update for below.
12529 * config/i386/winnt.c (i386_pe_section_type_flags::htab):
12530 Change type to hash_table. Update dependent calls and types.
12531 (i386_find_on_wrapper_list::wrappers): Likewise.
12533 * config/ia64/t-ia64: Update for below.
12535 * config/ia64/ia64.c (bundle_state_table):
12536 Change type to hash_table. Update dependent calls and types.
12538 * config/mips/mips.c (mips_reorg_process_insns::htab):
12539 Change type to hash_table. Update dependent calls and types.
12541 * config/sol2.c (solaris_comdat_htab):
12542 Change type to hash_table. Update dependent calls and types.
12544 * config/t-sol2: Update for above.
12546 2013-05-29 Teresa Johnson <tejohnson@google.com>
12548 * passes.c (dump_passes): Use FOR_EACH_FUNCTION since
12549 functions are not yet marked as defined.
12551 2013-05-29 Michael Meissner <meissner@linux.vnet.ibm.com>
12552 Pat Haugen <pthaugen@us.ibm.com>
12553 Peter Bergner <bergner@vnet.ibm.com>
12555 * config/rs6000/vector.md (VEC_I): Add support for new power8 V2DI
12559 (vrotl<mode>3): Likewise.
12560 (vashl<mode>3): Likewise.
12561 (vlshr<mode>3): Likewise.
12562 (vashr<mode>3): Likewise.
12564 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
12565 support for power8 V2DI builtins.
12567 * config/rs6000/rs6000-builtin.def (abs_v2di): Add support for
12568 power8 V2DI builtins.
12569 (vupkhsw): Likewise.
12570 (vupklsw): Likewise.
12571 (vaddudm): Likewise.
12572 (vminsd): Likewise.
12573 (vmaxsd): Likewise.
12574 (vminud): Likewise.
12575 (vmaxud): Likewise.
12576 (vpkudum): Likewise.
12577 (vpksdss): Likewise.
12578 (vpkudus): Likewise.
12579 (vpksdus): Likewise.
12584 (vsubudm): Likewise.
12585 (vcmpequd): Likewise.
12586 (vcmpgtsd): Likewise.
12587 (vcmpgtud): Likewise.
12588 (vcmpequd_p): Likewise.
12589 (vcmpgtsd_p): Likewise.
12590 (vcmpgtud_p): Likewise.
12591 (vupkhsw): Likewise.
12592 (vupklsw): Likewise.
12593 (vaddudm): Likewise.
12594 (vmaxsd): Likewise.
12595 (vmaxud): Likewise.
12596 (vminsd): Likewise.
12597 (vminud): Likewise.
12598 (vpksdss): Likewise.
12599 (vpksdus): Likewise.
12600 (vpkudum): Likewise.
12601 (vpkudus): Likewise.
12606 (vsubudm): Likewise.
12608 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Add
12609 support for power8 V2DI instructions.
12611 * config/rs6000/altivec.md (UNSPEC_VPKUHUM): Add support for
12612 power8 V2DI instructions. Combine pack and unpack insns to use an
12613 iterator for each mode. Check whether a particular mode supports
12614 Altivec instructions instead of just checking TARGET_ALTIVEC.
12615 (UNSPEC_VPKUWUM): Likewise.
12616 (UNSPEC_VPKSHSS): Likewise.
12617 (UNSPEC_VPKSWSS): Likewise.
12618 (UNSPEC_VPKUHUS): Likewise.
12619 (UNSPEC_VPKSHUS): Likewise.
12620 (UNSPEC_VPKUWUS): Likewise.
12621 (UNSPEC_VPKSWUS): Likewise.
12622 (UNSPEC_VPACK_SIGN_SIGN_SAT): Likewise.
12623 (UNSPEC_VPACK_SIGN_UNS_SAT): Likewise.
12624 (UNSPEC_VPACK_UNS_UNS_SAT): Likewise.
12625 (UNSPEC_VPACK_UNS_UNS_MOD): Likewise.
12626 (UNSPEC_VUPKHSB): Likewise.
12627 (UNSPEC_VUNPACK_HI_SIGN): Likewise.
12628 (UNSPEC_VUNPACK_LO_SIGN): Likewise.
12629 (UNSPEC_VUPKHSH): Likewise.
12630 (UNSPEC_VUPKLSB): Likewise.
12631 (UNSPEC_VUPKLSH): Likewise.
12633 (VI_char): Likewise.
12634 (VI_scalar): Likewise.
12635 (VI_unit): Likewise.
12637 (VP_small): Likewise.
12638 (VP_small_lc): Likewise.
12639 (VU_char): Likewise.
12640 (add<mode>3): Likewise.
12641 (altivec_vaddcuw): Likewise.
12642 (altivec_vaddu<VI_char>s): Likewise.
12643 (altivec_vadds<VI_char>s): Likewise.
12644 (sub<mode>3): Likewise.
12645 (altivec_vsubcuw): Likewise.
12646 (altivec_vsubu<VI_char>s): Likewise.
12647 (altivec_vsubs<VI_char>s): Likewise.
12648 (altivec_vavgs<VI_char>): Likewise.
12649 (altivec_vcmpbfp): Likewise.
12650 (altivec_eq<mode>): Likewise.
12651 (altivec_gt<mode>): Likewise.
12652 (altivec_gtu<mode>): Likewise.
12653 (umax<mode>3): Likewise.
12654 (smax<mode>3): Likewise.
12655 (umin<mode>3): Likewise.
12656 (smin<mode>3): Likewise.
12657 (altivec_vpkuhum): Likewise.
12658 (altivec_vpkuwum): Likewise.
12659 (altivec_vpkshss): Likewise.
12660 (altivec_vpkswss): Likewise.
12661 (altivec_vpkuhus): Likewise.
12662 (altivec_vpkshus): Likewise.
12663 (altivec_vpkuwus): Likewise.
12664 (altivec_vpkswus): Likewise.
12665 (altivec_vpks<VI_char>ss): Likewise.
12666 (altivec_vpks<VI_char>us): Likewise.
12667 (altivec_vpku<VI_char>us): Likewise.
12668 (altivec_vpku<VI_char>um): Likewise.
12669 (altivec_vrl<VI_char>): Likewise.
12670 (altivec_vsl<VI_char>): Likewise.
12671 (altivec_vsr<VI_char>): Likewise.
12672 (altivec_vsra<VI_char>): Likewise.
12673 (altivec_vsldoi_<mode>): Likewise.
12674 (altivec_vupkhsb): Likewise.
12675 (altivec_vupkhs<VU_char>): Likewise.
12676 (altivec_vupkls<VU_char>): Likewise.
12677 (altivec_vupkhsh): Likewise.
12678 (altivec_vupklsb): Likewise.
12679 (altivec_vupklsh): Likewise.
12680 (altivec_vcmpequ<VI_char>_p): Likewise.
12681 (altivec_vcmpgts<VI_char>_p): Likewise.
12682 (altivec_vcmpgtu<VI_char>_p): Likewise.
12683 (abs<mode>2): Likewise.
12684 (vec_unpacks_hi_v16qi): Likewise.
12685 (vec_unpacks_hi_v8hi): Likewise.
12686 (vec_unpacks_lo_v16qi): Likewise.
12687 (vec_unpacks_hi_<VP_small_lc>): Likewise.
12688 (vec_unpacks_lo_v8hi): Likewise.
12689 (vec_unpacks_lo_<VP_small_lc>): Likewise.
12690 (vec_pack_trunc_v8h): Likewise.
12691 (vec_pack_trunc_v4si): Likewise.
12692 (vec_pack_trunc_<mode>): Likewise.
12694 * config/rs6000/altivec.h (vec_vaddudm): Add defines for power8
12696 (vec_vmaxsd): Likewise.
12697 (vec_vmaxud): Likewise.
12698 (vec_vminsd): Likewise.
12699 (vec_vminud): Likewise.
12700 (vec_vpksdss): Likewise.
12701 (vec_vpksdus): Likewise.
12702 (vec_vpkudum): Likewise.
12703 (vec_vpkudus): Likewise.
12704 (vec_vrld): Likewise.
12705 (vec_vsld): Likewise.
12706 (vec_vsrad): Likewise.
12707 (vec_vsrd): Likewise.
12708 (vec_vsubudm): Likewise.
12709 (vec_vupkhsw): Likewise.
12710 (vec_vupklsw): Likewise.
12712 2013-05-29 Jan Hubicka <jh@suse.cz>
12714 * cgraph.h (symtab_node_base): Add definition, alias and analyzed
12715 flags; reorder rest of fields in more consistent way.
12716 (varpool_node): Remove analyzed, finalized and alias.
12717 (cgraph_ndoe): Likewise.
12718 (symtab_alias_ultimate_target): New function.
12719 (cgraph_function_node): Move offline.
12720 (cgraph_reset_node): Declare.
12721 (cgraph_comdat_can_be_unshared_p): Remove.
12722 (varpool_remove_initializer): Declare.
12723 (varpool_first_defined_variable, varpool_next_defined_variable
12724 cgraph_first_defined_function, cgraph_next_defined_function): Update.
12725 (cgraph_function_with_gimple_body_p): Update.
12726 (varpool_all_refs_explicit_p): Update.
12727 (symtab_alias_target): New function.
12728 (cgraph_alias_aliased_node, varpool_alias_aliased_node): Rename to ...
12729 (cgraph_alias_target, varpool_alias_target): .. this one; simplify.
12730 (cgraph_function_or_thunk_node): Simplify using
12731 symtab_alias_ultimate_target.
12732 (varpool_variable_node): Likewise.
12733 * cgraph.c (cgraph_create_function_alias): Update.
12734 (cgraph_add_thunk): Update.
12735 (cgraph_remove_node): Update.
12736 (dump_cgraph_node): Do not dump removed flags.
12737 (cgraph_function_body_availability): Update.
12738 (cgraph_propagate_frequency): Update.
12739 (verify_cgraph_node): Check sanity of local flag.
12740 (cgraph_function_node): Move here from cgraph.h; revamp for
12741 cgraph_function_or_thunk_node.
12742 * lto-symtab.c (lto_varpool_replace_node): Update.
12743 (lto_symtab_resolve_can_prevail_p): Update.
12744 (lto_symtab_merge_cgraph_nodes): Update.
12745 * ipa-cp.c (determine_versionability, initialize_node_lattices,
12746 propagate_constants_accross_call, devirtualization_time_bonus,
12747 ipcp_propagate_stage): Update.
12748 * tree-emutls.c (create_emultls_var, ipa_lower_emutls): Update.
12749 * ipa-inline-transform.c (clone_inlined_nodes,
12750 preserve_function_body_p): Update.
12751 * ipa-reference.c (propagate): Update.
12752 (write_node_summary_p): Update.
12753 * toplev.c (wrapup_global_declaration_2): Update.
12754 * cgraphunit.c (cgraph_analyze_function): Rename to ...
12755 (analyze_function) ... this one.
12756 (cgraph_process_new_functions): Update.
12757 (cgraph_reset_node): Export.
12758 (cgraph_finalize_function): Update.
12759 (cgraph_add_new_function): Update.
12760 (process_function_and_variable_attributes): Update.
12761 (varpool_finalize_decl): Update.
12762 (symbol_finalized): Remove.
12763 (symbol_finalized_and_needed): Rename to ...
12764 (symbol_defined_and_needed): ... update.
12765 (cgraph_analyze_functions): Update.
12766 (handle_alias_pairs): Update.
12767 (mark_functions_to_output): Update.
12768 (assemble_thunk): Update.
12769 (output_in_order): Update.
12770 (output_weakrefs): Update.
12771 (finalize_compilation_unit): Update.
12772 * lto-cgraph.c (reachable_from_other_partition_p, lto_output_node,
12773 lto_output_varpool_node, compute_ltrans_boundary, input_overwrite_node,
12774 input_node, input_varpool_node): Update.
12775 * dbxout.c (dbxout_expand_expr): Update.
12776 * cgraphclones.c (cgraph_clone_node): Update.
12777 (cgraph_copy_node_for_versioning): Update.
12778 (cgraph_materialize_clone): Update.
12779 (cgraph_materialize_all_clones): Update.
12780 * ipa-pure-const.c (analyze_function, pure_const_write_summary,
12781 propagate_pure_const, propagate_nothrow): Update.
12782 * lto-streamer-out.c (lto_output, write_symbol): Update.
12783 * ipa-utils.c (ipa_reverse_postorder): Update.
12784 * ipa-inline.c (can_inline_edge_p): Update.
12785 (update_caller_keys, ipa_inline): Update.
12786 * dwarf2out.c (reference_to_unused,
12787 premark_types_used_by_global_vars_helper): Update.
12788 * tree-eh.c (tree_could_trap_p): Update.
12789 * ipa-split.c (consider_split, execute_split_functions): Update.
12790 * ipa.c (cgraph_non_local_node_p_1, cgraph_local_node_p,
12791 has_addr_references_p): Update; move ahead in file for better
12793 (process_references): Simplify.
12794 (symtab_remove_unreachable_nodes): Update; cleanup way function/var
12795 bodies are removed.
12796 (cgraph_comdat_can_be_unshared_p): Make static.
12797 (cgraph_externally_visible_p): Update.
12798 (varpool_externally_visible_p): Update.
12799 (function_and_variable_visibility): Update.
12800 * trans-mem.c (get_cg_data, ipa_tm_mayenterirr_function,
12801 ipa_tm_mark_force_output_node): Update.
12802 * ipa-inline-analysis.c (dump_inline_summary, initialize_inline_failed,
12803 estimate_edge_devirt_benefit, inline_generate_summary,
12804 inline_write_summary): Update.
12805 * gimple-fold.c (can_refer_decl_in_current_unit_p): Update.
12806 * ipa-prop.c (ipa_compute_jump_functions): Update.
12807 (ipa_print_node_params, ipa_prop_read_section,
12808 ipa_update_after_lto_read, read_replacements_section): Update.
12809 * varasm.c (mark_decl_referenced): Update.
12810 (assemble_alias, dump_tm_clone_pairs): Update.
12811 * tree-inline.c (copy_bb): Update.
12812 (estimate_num_insns, optimize_inline_calls, tree_function_versioning):
12814 * symtab.c (dump_symtab_base): Print new flags.
12815 (verify_symtab_base): Verify new flags.
12816 (symtab_alias_ultimate_target): New function.
12817 * tree-ssa-structalias.c (get_constraint_for_ssa_var,
12818 create_variable_info_for, associate_varinfo_to_alias, ipa_pta_execute):
12820 * passes.c (ipa_write_summaries, ipa_write_optimization_summaries):
12822 * i386.c (ix86_get_function_versions_dispatcher,
12823 ix86_generate_version_dispatcher_body): Update.
12824 (fold_builtin_cpu): Use varpool_add_new_variable.
12825 * varpool.c (varpool_remove_initializer): Break out from ...
12826 (varpool_remove_node): ... this one.
12827 (dump_varpool_node, varpool_node_for_asm,
12828 cgraph_variable_initializer_availability, varpool_analyze_node,
12829 varpool_assemble_decl, varpool_remove_unreferenced_decls,
12830 varpool_finalize_named_section_flags, varpool_create_variable_alias):
12833 2013-05-29 Jan Hubicka <jh@suse.cz>
12835 * passes.c (init_optimization_passes): Move OMP expansion into lowering.
12837 2013-05-29 Easwaran Raman <eraman@google.com>
12839 PR tree-optimization/57442
12840 * tree-ssa-reassoc.c (appears_later_in_bb): Return correct value
12841 when control exits the main loop.
12843 2013-05-29 Sandeep Kumar Singh <Sandeep.Singh2@kpitcummins.com>
12845 * rx/rx.h (TARGET_CPU_CPP_BUILTINS): Add macros for RX100, RX200,
12847 * rx/rx.opt: Add macro for rx100 with string rx100 and value RX100.
12848 * rx/rx-opts.h (rx_cpu_types): Add new cpu type rx100.
12849 * rx/t-rx: Add rx100 under multi library matches option for nofpu
12852 2013-05-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
12854 PR tree-optimization/57441
12855 * gimple-ssa-strength-reduction.c (analyze_candidates_and_replace):
12856 Don't limit size of incr_vec to number of candidates.
12858 2013-05-29 Steve Ellcey <sellcey@imgtec.com>
12860 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add micromips
12861 and mips16 directories.
12862 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Add micromips and mips16.
12863 (MULTILIB_DIRNAMES): Ditto.
12864 (MULTILIB_EXCEPTIONS): Add new exceptions.
12865 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Add micromips.
12866 (MULTILIB_DIRNAMES): Ditto.
12867 (MULTILIB_EXCEPTIONS): Add new exceptions.
12869 2012-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
12870 Marcus Shawcroft <marcus.shawcroft@arm.com>
12872 * config/aarch64/aarch64-protos.h (aarch64_symbol_type): Define
12873 SYMBOL_TINY_ABSOLUTE.
12874 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately): Handle
12875 SYMBOL_TINY_ABSOLUTE.
12876 (aarch64_expand_mov_immediate): Likewise.
12877 (aarch64_classify_symbol): Likewise.
12878 (aarch64_mov_operand_p): Remove ATTRIBUTE_UNUSED.
12879 Permit SYMBOL_TINY_ABSOLUTE.
12880 * config/aarch64/predicates.md (aarch64_mov_operand): Permit CONST.
12882 2013-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
12883 Marcus Shawcroft <marcus.shawcroft@arm.com>
12885 * config/aarch64/aarch64.c (aarch64_classify_symbol): Remove comment.
12886 Refactor if/switch. Replace gcc_assert with if.
12888 2013-05-29 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
12890 * config/i386/i386.c (initial_ix86_tune_features): Enable
12891 FP Reassociation for AMD bdver1 and bdver2.
12893 2013-05-29 Martin Jambor <mjambor@suse.cz>
12895 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REF, REALPART_EXPR
12896 and IMAGPART_EXPR do not occur within other handled_components.
12898 2013-05-29 Richard Biener <rguenther@suse.de>
12900 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Guard vinfo
12901 access on whether the use is in the BB we currently try to
12903 (vect_bb_vectorization_profitable_p): Pass the BB we currently
12904 vectorize to vect_bb_slp_scalar_cost.
12906 2013-05-29 Richard Biener <rguenther@suse.de>
12908 * tree-vect-slp.c (vect_bb_slp_scalar_cost): New function
12909 computing scalar cost offsetted by stmts that are kept live
12911 (vect_bb_vectorization_profitable_p): Use vect_bb_slp_scalar_cost
12912 for computation of scalar cost.
12914 2013-05-28 Steve Ellcey <sellcey@mips.com>
12916 * config/mips/mips-cpus.def (mips32r2): Change processor type.
12918 2013-05-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
12920 * doc/extend.texi (C Extensions): Added documentation about Cilk Plus
12921 array notation built-in reduction functions.
12922 * doc/passes.texi (Passes): Added documentation about changes done
12924 * doc/invoke.texi (C Dialect Options): Added documentation about
12925 the -fcilkplus flag.
12926 * Makefile.in (C_COMMON_OBJS): Added c-family/array-notation-common.o.
12927 (BUILTINS_DEF): Depend on cilkplus.def.
12928 * builtins.def: Include cilkplus.def. Define DEF_CILKPLUS_BUILTIN.
12929 * builtin-types.def: Define BT_FN_INT_PTR_PTR_PTR.
12930 * cilkplus.def: New file.
12932 2013-05-28 Joern Rennecke <joern.rennecke@embecosm.com>
12934 PR rtl-optimization/57439
12935 * postreload.c (move2add_use_add2_insn): Use gen_lowpart_common.
12937 2013-05-28 Easwaran Raman <eraman@google.com>
12939 PR tree-optimization/57337
12940 * tree-ssa-reassoc.c (appears_later_in_bb): New function.
12941 (find_insert_point): Correctly identify the insertion point
12942 when two statements with the same UID is compared.
12944 2013-05-28 Richard Biener <rguenther@suse.de>
12946 PR tree-optimization/56787
12947 * tree-vect-data-refs.c (vect_analyze_data_refs): Drop clobbers
12948 from the list of data references.
12949 * tree-vect-loop.c (vect_determine_vectorization_factor): Skip
12951 (vect_analyze_loop_operations): Likewise.
12952 (vect_transform_loop): Remove clobbers.
12954 2013-05-28 Martin Jambor <mjambor@suse.cz>
12956 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REFs, IMAGPART_EXPRs
12957 and REALPART_EXPRs have scalar type.
12959 2013-05-28 Richard Biener <rguenther@suse.de>
12961 PR tree-optimization/57411
12962 * tree-ssa-copy.c (may_propagate_copy): Cannot propagate
12964 * tree-ssa-dom.c (eliminate_const_or_copy): Special-case
12965 virtual operand propagation.
12967 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
12969 * config/sparc/sparc.c (sparc_expand_vec_perm_bmask): Use %g0 as
12970 destination register for bmasksi_vis.
12971 (vector_init_bshuffle): Likewise.
12972 * config/sparc/sparc.md (vec_perm_constv8qi): Likewise.
12974 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
12976 * doc/invoke.texi (SPARC Options): Document -mfix-ut699.
12977 * builtins.c (expand_builtin_mathfn) <BUILT_IN_SQRT>: Try to widen the
12978 mode if the instruction isn't available in the original mode.
12979 * config/sparc/sparc.opt (mfix-ut699): New option.
12980 * config/sparc/sparc.md (muldf3_extend): Disable if -mfix-ut699.
12981 (divdf3): Turn into expander.
12982 (divdf3_nofix): New insn.
12983 (divdf3_fix): Likewise.
12984 (divsf3): Disable if -mfix-ut699.
12985 (sqrtdf2): Turn into expander.
12986 (sqrtdf2_nofix): New insn.
12987 (sqrtdf2_fix): Likewise.
12988 (sqrtsf2): Disable if -mfix-ut699.
12990 2013-05-27 Richard Biener <rguenther@suse.de>
12992 PR middle-end/57412
12993 * omp-low.c (expand_omp_atomic_pipeline): Use the correct latch
12994 block for the new loop.
12996 2013-05-27 Richard Biener <rguenther@suse.de>
12998 PR tree-optimization/57343
12999 * tree-ssa-loop-niter.c (number_of_iterations_ne_max): Do not
13000 use multiple_of_p if not TYPE_OVERFLOW_UNDEFINED.
13001 (number_of_iterations_cond): Do not build the folded tree.
13003 2013-05-27 Richard Biener <rguenther@suse.de>
13006 PR middle-end/57381
13007 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
13008 OEP_CONSTANT_ADDRESS_OF retained.
13010 PR tree-optimization/57417
13011 * tree-ssa-sccvn.c (vn_reference_fold_indirect): Fix test
13012 for unchanged base.
13013 (set_ssa_val_to): Compare addresses using
13014 get_addr_base_and_unit_offset.
13016 2013-05-27 Joern Rennecke <joern.rennecke@embecosm.com>
13018 PR rtl-optimization/56833
13019 * postreload.c (move2add_record_mode): New function.
13020 (move2add_record_sym_value, move2add_valid_value_p): Likewise.
13021 (move2add_use_add2_insn): Use move2add_record_sym_value.
13022 (move2add_use_add3_insn): Likewise.
13023 (reload_cse_move2add): Use move2add_valid_value_p and
13024 move2add_record_mode. Invalidate call-clobbered and REG_INC
13025 affected regs by setting reg_mode to VOIDmode.
13026 (move2add_note_store): Don't pretend the inside of a SUBREG is
13027 the actual destination. Invalidate single/leading registers by
13028 setting reg_mode to VOIDmode.
13029 Use move2add_record_sym_value, move2add_valid_value_p and
13030 move2add_record_mode.
13032 2013-05-27 Richard Biener <rguenther@suse.de>
13034 PR tree-optimization/57396
13035 * tree-affine.c (double_int_constant_multiple_p): Properly
13036 return false for val == 0 and div != 0.
13038 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
13040 * config/mips/mips.h: Use #elif in preprocessor conditions.
13042 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
13045 * config/mips/constraints.md (kl): New constraint.
13046 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Delete.
13047 (divmod<mode>4_internal): Rename to divmod<mode>4. Use "kl" as the
13048 constraint for operand 0. Split after CSE for MIPS16. Emit a move
13049 from LO for MIPS16.
13050 (udivmod<mode>4_internal): Likewise udivmod<mode>4.
13052 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
13055 * config/mips/mips.c (mips_can_inline_p): New function.
13056 (TARGET_CAN_INLINE_P): Define.
13058 2013-05-25 Steven Bosscher <steven@gcc.gnu.org>
13060 * sched-int.h (ds_t, dw_t): Make unsigned int.
13061 Fix documentation that describes how all the ds_t bits are used.
13062 Reserve the last bit for delayed-branch scheduling.
13063 (BITS_PER_DEP_STATUS): Move to ds_t typedef.
13064 (BITS_PER_DEP_WEAK): Fix definition and documentation.
13065 (gen_dep_weak_1): Remove prototype.
13066 * sched-deps.c (get_dep_weak_1): Make static.
13067 * target.def (speculate_insn, needs_block_p, gen_spec_check,
13068 get_insn_spec_ds, get_insn_checked_ds): Adjust hook prototypes.
13069 * doc/tm.texi: Regenerate.
13070 * config/ia64/ia64.c (ia64_needs_block_p): Update prototype.
13072 2013-05-24 Steven Bosscher <steven@gcc.gnu.org>
13075 * haifa-sched.c (sched_extend_bb): Ignore DEBUG_INSNs.
13077 2013-05-24 Nathan Sidwell <nathan@codesourcery.com>
13078 Sandra Loosemore <sandra@codesourcery.com>
13080 * config.gcc (powerpc-*): Allow native for with-cpu.
13082 2013-05-24 Jeff Law <law@redhat.com>
13084 PR tree-optimization/57124
13085 * tree-vrp.c (simplify_cond_using_ranges): Only simplify a
13086 conversion feeding a condition if the range has an overflow
13087 if -fstrict-overflow. Add warnings for when we do make the
13090 2013-05-24 Dehao Chen <dehao@google.com>
13092 * tree-cfg.c (locus_discrim_map): Fix the typo.
13093 (locus_discrim_hasher): Likewise.
13094 (locus_discrim_hasher::hash): Likewise.
13095 (locus_discrim_hasher::equal): Likewise.
13097 2013-05-24 Martin Jambor <mjambor@suse.cz>
13099 PR tree-optimization/57294
13100 * cgraph.h (ipa_record_stmt_references): Declare.
13101 * cgraphbuild.c (ipa_record_stmt_references): New function.
13102 (build_cgraph_edges): Use ipa_record_stmt_references.
13103 (rebuild_cgraph_edges): Likewise.
13104 (cgraph_rebuild_references): Likewise.
13105 * ipa-prop.c (ipa_modify_call_arguments): Discard references
13106 associated with the old statement and build references from the
13107 newly built statements.
13108 * ipa-ref.c (ipa_remove_stmt_references): New function.
13109 * ipa-ref.h (ipa_remove_stmt_references): Declare.
13111 2013-05-24 Vladimir Makarov <vmakarov@redhat.com>
13113 * lra-constraints.c (emit_spill_move): Use smaller mode for
13115 (check_and_process_move): Consider mem-reg moves for secondary
13117 (curr_insn_transform): Don't lose insns emitted before for
13118 secondary memory moves.
13119 (inherit_in_ebb): Mark defined reg. Add usage only if it is not a
13120 reg set up in the current insn.
13122 2013-05-24 Dehao Chen <dehao@google.com>
13124 * tree-cfg.c (locus_descrim_hasher::hash): Change discriminator
13126 (locus_descrim_hasher::equal): Likewise.
13127 (build_gimple_cfg): New discriminator assignment algorithm.
13128 (make_edges): Likewise.
13129 (next_discriminator_for_locus): Likewise.
13130 (same_line_p): Likewise.
13131 (assign_discriminators): Likewise.
13132 (make_cond_expr_edges): Likewise.
13133 (make_gimple_switch_edges): Likewise.
13134 (make_goto_expr_edges): Likewise.
13135 (make_gimple_asm_edges): Likewise.
13137 2013-05-24 Ian Bolton <ian.bolton@arm.com>
13139 * config/aarch64/aarch64.c (aarch64_print_operand): Change the
13140 X format specifier to only display bottom 16 bits.
13141 * config/aarch64/aarch64.md (insv_imm<mode>): Allow any size of
13142 immediate to match for operand 2, since it will be masked.
13144 2013-05-24 Richard Biener <rguenther@suse.de>
13146 PR tree-optimization/57287
13147 * tree-ssa-uninit.c (compute_uninit_opnds_pos): Disregard
13148 all SSA names that occur in abnormal PHIs.
13150 2013-05-24 Alexander Ivchenko <alexander.ivchenko@intel.com>
13153 * tree-ssa-sccvn.c (fully_constant_vn_reference_p): Check
13154 that index is not negative.
13156 2013-05-24 Eric Botcazou <ebotcazou@adacore.com>
13158 PR rtl-optimization/55177
13159 * simplify-rtx.c (simplify_unary_operation_1) <NOT>: Deal with BSWAP.
13160 (simplify_byte_swapping_operation): New.
13161 (simplify_binary_operation_1): Call it for AND, IOR and XOR.
13162 (simplify_relational_operation_1): Deal with BSWAP.
13164 2013-05-23 Richard Henderson <rth@redhat.com>
13167 * config/i386/i386.c (ix86_seh_fixup_eh_fallthru): New.
13168 (ix86_reorg): Call it.
13170 2013-05-23 Uros Bizjak <ubizjak@gmail.com>
13173 * config/alpha/alpha.md (unspec): Add UNSPEC_XFLT_COMPARE.
13174 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Construct
13175 REG_EQUAL note as UNSPEC_XFLT_COMPARE unspec.
13177 2013-05-23 Christian Bruel <christian.bruel@st.com>
13180 * config/arm/arm.c (arm_dwarf_register_span): Do not use dbx number.
13182 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
13183 Marcus Shawcroft <marcus.shawcroft@arm.com>
13185 * config/aarch64/aarch64.md (*movdi_aarch64): Replace Usa with S.
13186 * config/aarch64/constraints.md (Usa): Remove.
13187 * doc/md.texi (AArch64 Usa): Remove.
13189 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
13190 Marcus Shawcroft <marcus.shawcroft@arm.com>
13192 * config/aarch64/aarch64-protos.h (aarch64_mov_operand_p): Define.
13193 * config/aarch64/aarch64.c (aarch64_mov_operand_p): Define.
13194 * config/aarch64/predicates.md (aarch64_const_address): Remove.
13195 (aarch64_mov_operand): Use aarch64_mov_operand_p.
13197 2013-05-23 Vidya Praveen <vidyapraveen@arm.com>
13199 * config/aarch64/aarch64-simd.md (clzv4si2): Support for CLZ
13200 instruction (AdvSIMD).
13201 * config/aarch64/aarch64-builtins.c
13202 (aarch64_builtin_vectorized_function): Handler for BUILT_IN_CLZ.
13203 * config/aarch64/aarch-simd-builtins.def: Entry for CLZ.
13205 2013-05-23 Martin Jambor <mjambor@suse.cz>
13207 PR middle-end/57347
13208 * tree.h (contains_bitfld_component_ref_p): Declare.
13209 * tree-sra.c (contains_bitfld_comp_ref_p): Move...
13210 * tree.c (contains_bitfld_component_ref_p): ...here. Adjust its
13212 * ipa-prop.c (determine_known_aggregate_parts): Check that LHS does
13213 not access a bit-field. Assert all final offsets are byte-aligned.
13215 2013-05-23 Richard Biener <rguenther@suse.de>
13217 PR tree-optimization/57380
13218 * tree-ssa-phiprop.c (propagate_with_phi): Do not require at
13219 least one invariant or re-used load.
13220 * passes.c (init_optimization_passes): Move pass_phiprop before
13223 2013-05-23 James Greenhalgh <james.greenhalgh@arm.com>
13225 * config/aarch64/aarch64-simd.md
13226 (aarch64_cm<optab>di): Add clobber of CC_REGNUM to unsplit pattern.
13228 2013-05-23 Richard Biener <rguenther@suse.de>
13230 PR middle-end/57381
13231 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
13232 OEP_CONSTANT_ADDRESS_OF retained.
13234 2013-05-23 Jakub Jelinek <jakub@redhat.com>
13236 PR middle-end/57344
13237 * expmed.c (store_split_bit_field): If op0 is a REG or SUBREG of a REG,
13238 don't lower unit. Handle unit not being always BITS_PER_WORD.
13240 2013-05-23 Richard Biener <rguenther@suse.de>
13242 PR rtl-optimization/57341
13243 * ira.c (validate_equiv_mem_from_store): Use anti_dependence
13244 instead of true_dependence.
13246 2013-05-22 David Malcolm <dmalcolm@redhat.com>
13248 * bb-reorder.c (branch_threshold): Make const.
13249 (exec_threshold): Ditto.
13251 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
13252 Pat Haugen <pthaugen@us.ibm.com>
13253 Peter Bergner <bergner@vnet.ibm.com>
13255 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions): Add
13256 documentation for the power8 crypto builtins.
13258 * config/rs6000/t-rs6000 (MD_INCLUDES): Add crypto.md.
13260 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_1): Add support
13261 macros for defining power8 builtin functions.
13262 (BU_P8V_AV_2): Likewise.
13263 (BU_P8V_AV_P): Likewise.
13264 (BU_P8V_VSX_1): Likewise.
13265 (BU_P8V_OVERLOAD_1): Likewise.
13266 (BU_P8V_OVERLOAD_2): Likewise.
13267 (BU_CRYPTO_1): Likewise.
13268 (BU_CRYPTO_2): Likewise.
13269 (BU_CRYPTO_3): Likewise.
13270 (BU_CRYPTO_OVERLOAD_1): Likewise.
13271 (BU_CRYPTO_OVERLOAD_2): Likewise.
13272 (XSCVSPDP): Fix typo, point to the correct instruction.
13273 (VCIPHER): Add power8 crypto builtins.
13274 (VCIPHERLAST): Likewise.
13275 (VNCIPHER): Likewise.
13276 (VNCIPHERLAST): Likewise.
13277 (VPMSUMB): Likewise.
13278 (VPMSUMH): Likewise.
13279 (VPMSUMW): Likewise.
13280 (VPERMXOR_V2DI): Likewise.
13281 (VPERMXOR_V4SI: Likewise.
13282 (VPERMXOR_V8HI: Likewise.
13283 (VPERMXOR_V16QI: Likewise.
13284 (VSHASIGMAW): Likewise.
13285 (VSHASIGMAD): Likewise.
13286 (VPMSUM): Likewise.
13287 (VPERMXOR): Likewise.
13288 (VSHASIGMA): Likewise.
13290 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
13291 __CRYPTO__ if the crypto instructions are available.
13292 (altivec_overloaded_builtins): Add support for overloaded power8
13295 * config/rs6000/rs6000.c (rs6000_expand_ternop_builtin): Add
13296 support for power8 crypto builtins.
13297 (builtin_function_type): Likewise.
13298 (altivec_init_builtins): Add support for builtins that take vector
13299 long long (V2DI) arguments.
13301 * config/rs6000/crypto.md: New file, define power8 crypto
13304 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
13305 Pat Haugen <pthaugen@us.ibm.com>
13306 Peter Bergner <bergner@vnet.ibm.com>
13308 * doc/invoke.texi (Option Summary): Add power8 options.
13309 (RS/6000 and PowerPC Options): Likewise.
13311 * doc/md.texi (PowerPC and IBM RS6000 constraints): Update to use
13312 constraints.md instead of rs6000.h. Reorder w* constraints. Add
13313 wm, wn, wr documentation.
13315 * config/rs6000/constraints.md (wm): New constraint for VSX
13316 registers if direct move instructions are enabled.
13317 (wn): New constraint for no registers.
13318 (wq): New constraint for quad word even GPR registers.
13319 (wr): New constraint if 64-bit instructions are enabled.
13320 (wv): New constraint if power8 vector instructions are enabled.
13321 (wQ): New constraint for quad word memory locations.
13323 * config/rs6000/predicates.md (const_0_to_15_operand): New
13324 constraint for 0..15 for crypto instructions.
13325 (gpc_reg_operand): If VSX allow registers in VSX registers as well
13326 as GPR and floating point registers.
13327 (int_reg_operand): New predicate to match only GPR registers.
13328 (base_reg_operand): New predicate to match base registers.
13329 (quad_int_reg_operand): New predicate to match even GPR registers
13330 for quad memory operations.
13331 (vsx_reg_or_cint_operand): New predicate to allow vector logical
13332 operations in both GPR and VSX registers.
13333 (quad_memory_operand): New predicate for quad memory operations.
13334 (reg_or_indexed_operand): New predicate for direct move support.
13336 * config/rs6000/rs6000-cpus.def (ISA_2_5_MASKS_EMBEDDED):
13337 Inherit from ISA_2_4_MASKS, not ISA_2_2_MASKS.
13338 (ISA_2_7_MASKS_SERVER): New mask for ISA 2.07 (i.e. power8).
13339 (POWERPC_MASKS): Add power8 options.
13340 (power8 cpu): Use ISA_2_7_MASKS_SERVER instead of specifying the
13343 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
13344 Define _ARCH_PWR8 and __POWER8_VECTOR__ for power8.
13346 * config/rs6000/rs6000.opt (-mvsx-timode): Add documentation.
13347 (-mpower8-fusion): New power8 options.
13348 (-mpower8-fusion-sign): Likewise.
13349 (-mpower8-vector): Likewise.
13350 (-mcrypto): Likewise.
13351 (-mdirect-move): Likewise.
13352 (-mquad-memory): Likewise.
13354 * config/rs6000/rs6000.c (power8_cost): Initial definition for power8.
13355 (rs6000_hard_regno_mode_ok): Make PTImode only match even GPR
13357 (rs6000_debug_reg_print): Print the base register class if -mdebug=reg.
13358 (rs6000_debug_vector_unit): Add p8_vector.
13359 (rs6000_debug_reg_global): If -mdebug=reg, print power8 constraint
13360 definitions. Also print fusion state.
13361 (rs6000_init_hard_regno_mode_ok): Set up power8 constraints.
13362 (rs6000_builtin_mask_calculate): Add power8 builtin support.
13363 (rs6000_option_override_internal): Add support for power8.
13364 (rs6000_common_init_builtins): Add debugging for skipped builtins
13365 if -mdebug=builtin.
13366 (rs6000_adjust_cost): Add power8 support.
13367 (rs6000_issue_rate): Likewise.
13368 (insn_must_be_first_in_group): Likewise.
13369 (insn_must_be_last_in_group): Likewise.
13370 (force_new_group): Likewise.
13371 (rs6000_register_move_cost): Likewise.
13372 (rs6000_opt_masks): Likewise.
13374 * config/rs6000/rs6000.h (ASM_CPU_POWER8_SPEC): If we don't have a
13375 power8 capable assembler, default to power7 options.
13376 (TARGET_DIRECT_MOVE): Likewise.
13377 (TARGET_CRYPTO): Likewise.
13378 (TARGET_P8_VECTOR): Likewise.
13379 (VECTOR_UNIT_P8_VECTOR_P): Define power8 vector support.
13380 (VECTOR_UNIT_VSX_OR_P8_VECTOR_P): Likewise.
13381 (VECTOR_MEM_P8_VECTOR_P): Likewise.
13382 (VECTOR_MEM_VSX_OR_P8_VECTOR_P): Likewise.
13383 (VECTOR_MEM_ALTIVEC_OR_VSX_P): Likewise.
13384 (TARGET_XSCVDPSPN): Likewise.
13385 (TARGET_XSCVSPDPN): Likewsie.
13386 (TARGET_SYNC_HI_QI): Likewise.
13387 (TARGET_SYNC_TI): Likewise.
13388 (MASK_CRYPTO): Likewise.
13389 (MASK_DIRECT_MOVE): Likewise.
13390 (MASK_P8_FUSION): Likewise.
13391 (MASK_P8_VECTOR): Likewise.
13392 (REG_ALLOC_ORDER): Move fr13 to be lower in priority so that the TFmode
13393 temporary used by some of the direct move instructions to get two FP
13394 temporary registers does not force creation of a stack frame.
13395 (VLOGICAL_REGNO_P): Allow vector logical operations in GPRs.
13396 (MODES_TIEABLE_P): Move the VSX tests above the Altivec tests so
13397 that any VSX registers are tieable, even if they are also an
13398 Altivec vector mode.
13399 (r6000_reg_class_enum): Add wm, wr, wv constraints.
13400 (RS6000_BTM_P8_VECTOR): Power8 builtin support.
13401 (RS6000_BTM_CRYPTO): Likewise.
13402 (RS6000_BTM_COMMON): Likewise.
13404 * config/rs6000/rs6000.md (cpu attribute): Add power8.
13405 * config/rs6000/rs6000-opts.h (PROCESSOR_POWER8): Likewise.
13406 (enum rs6000_vector): Add power8 vector support.
13408 2013-05-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
13412 * config/arm/arm.c (any_sibcall_uses_r3): Rename to ..
13413 (any_sibcall_could_use_r3): this and handle indirect calls.
13414 (arm_get_frame_offsets): Rename use of any_sibcall_uses_r3.
13416 2013-05-22 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
13418 * config/rs6000/rs6000.h (MALLOC_ABI_ALIGNMENT): New #define.
13420 2013-05-22 Richard Biener <rguenther@suse.de>
13422 PR middle-end/57349
13423 * profile.c (branch_prob): Do not split blocks that are
13424 abnormally receiving from ECF_RETURNS_TWICE functions.
13426 2013-05-22 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
13428 * recog.c (offsettable_address_addr_space_p): Fix calculation of
13429 address mode. Move pointer mode initialization to the same place.
13431 2013-05-22 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
13433 * read-rtl.c (copy_rtx_for_iterators): Continue applying iterators
13434 while it has any effect.
13436 2013-05-21 Easwaran Raman <eraman@google.com>
13438 PR tree-optimization/57322
13439 * tree-ssa-reassoc.c (build_and_add_sum): If a BB is empty, set the
13440 UID of the statement added to the BB to be 1.
13442 2013-05-21 Jakub Jelinek <jakub@redhat.com>
13444 PR tree-optimization/57331
13445 * tree-vrp.c (simplify_cond_using_ranges): Don't optimize comparison
13446 of conversion from pointer type to integral type with integer.
13448 2013-05-21 Martin Jambor <mjambor@suse.cz>
13451 * ipa-prop.c (ipa_read_node_info): Process param_used and
13452 controlled_uses in the same order as when writing.
13454 2013-05-21 Magnus Granberg <baldrick@free.fr>
13457 * Makefile.in (PLUGIN_HEADERS): Add $(TARGET_H).
13459 2013-05-21 Richard Biener <rguenther@suse.de>
13461 PR tree-optimization/57318
13462 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Do not
13463 estimate stmts with side-effects as likely eliminated.
13465 2013-05-21 Richard Biener <rguenther@suse.de>
13467 PR tree-optimization/57330
13468 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Properly
13469 preserve the call stmts fntype.
13471 2013-05-21 Richard Biener <rguenther@suse.de>
13473 PR tree-optimization/57303
13474 * tree-ssa-sink.c (statement_sink_location): Improve killing
13475 stmt detection and properly handle self-assignments.
13477 2013-05-21 Christian Bruel <christian.bruel@st.com>
13479 * dwarf2out.c (multiple_reg_loc_descriptor): Use dbx_reg_number for
13480 spanning registers. LEAF_REG_REMAP is supported only for contiguous
13481 registers. Set register size out of the PARALLEL loop.
13483 2013-05-20 Oleg Endo <olegendo@gcc.gnu.org>
13486 * config/sh/sh.md (fmasf4): Remove empty constraints strings.
13487 (*fmasf4, *fmasf4_media): New insns.
13489 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
13491 * config/mips/mips.h (BASE_INSN_LENGTH, NOP_INSN_LENGTH): New macros.
13492 * config/mips/mips.c (mips_symbol_insns, mips_address_insns)
13493 (mips_const_insns, mips_split_const_insns, mips_load_store_insns)
13494 (mips_idiv_insns): Update the comments to say that the returned
13495 instruction counts are in units of BASE_INSN_LENGTH.
13496 (mips_adjust_insn_length): Multiply the mips_load_label_num_insns
13497 by BASE_INSN_LENGTH rather than 4. Add the jump separately,
13498 using 2 rather than 4 as the length of indirect MIPS16 and
13499 microMIPS jumps. Use NOP_INSN_LENGTH rather than 4 as the
13500 length of a NOP. Don't divide MIPS16 lengths by 2.
13501 (mips16_split_long_branches): Assume a branch is long if the
13502 length is greater than 4 rather than 8.
13503 * config/mips/mips.md (length): Give MIPS16 lengths directly,
13504 rather than multiplying them by 2. Multiply instruction counts
13505 by BASE_INSN_LENGTH rather than 4.
13506 (*jump_mips16, tls_get_tp_mips16_<mode>)
13507 (*tls_get_tp_mips16_call_<mode>): Divide lengths by 2.
13509 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
13511 * config/mips/mips.md (extended_mips16): Remove branch case.
13512 (length): Remove duplicated extended_mips16 test.
13514 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
13516 * config/mips/t-sde: Don't build 64-bit microMIPS multilibs.
13518 2013-05-18 Richard Sandiford <rdsandiford@googlemail.com>
13520 * recog.h (Recog_data): Rename to...
13521 (recog_data_d): ...this.
13522 (recog_data): Update accordingly.
13523 * recog.c (recog_data): Likewise.
13524 * reload.c (save_recog_data): Likewise.
13525 * config/picochip/picochip.c (picochip_saved_recog_data): Likewise.
13526 (picochip_save_recog_data, picochip_restore_recog_data): Likewise.
13528 2013-05-17 Julian Brown <julian@codesourcery.com>
13530 * gcse.c (compute_ld_motion_mems): If a non-simple MEM is
13531 found in a REG_EQUAL note, invalidate it.
13533 2013-05-17 Easwaran Raman <eraman@google.com>
13535 * tree-ssa-reassoc.c (find_insert_point): New function.
13536 (insert_stmt_after): Likewise.
13537 (get_def_stmt): Likewise.
13538 (ensure_ops_are_available): Likewise.
13539 (not_dominated_by): Likewise.
13540 (rewrite_expr_tree): Do not move statements beyond what is
13541 necessary. Remove call to swap_ops_for_binary_stmt...
13542 (reassociate_bb): ... and move it here.
13543 (build_and_add_sum): Assign UIDs for new statements.
13544 (linearize_expr): Likewise.
13545 (do_reassoc): Renumber gimple statement UIDs.
13547 2013-05-17 Jan Hubicka <jh@suse.cz>
13549 * lto-symtab.c (lto_symtab_merge_cgraph_nodes): Resolve cross module
13551 * cgraph.c (dump_cgraph_node): Do not ice on unresolved alias.
13552 * cgraphunit.c (handle_alias_pairs): Store target of unresolved
13554 (output_weakrefs): Update.
13556 2013-05-17 Po-Chun Chang <pchang9@cs.wisc.edu>
13557 Martin Jambor <mjambor@suse.cz>
13559 PR middle-end/57276
13560 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Break when a
13561 value that corresponds to the given aggval is found in values vector.
13563 2013-05-17 Uros Bizjak <ubizjak@gmail.com>
13565 * config/i386/driver-i386.c (host_detect_local_cpu): Pass mmx, 3dnow,
13566 sse, sse2, sse3, ssse3 and sse4a flags to options.
13568 2013-05-17 David Malcolm <dmalcolm@redhat.com>
13570 * gengtype-state.c: (s_expr_writer): New class, to handle
13571 prettifying of output layout of s-expressions.
13572 (state_writer): New class, to write out gtype.state.
13573 (state_written_type_count): Move this variable into member data of
13575 (s_expr_writer::s_expr_writer): New code: constructor for new class
13576 (state_writer::state_writer(): ditto
13577 (s_expr_writer::write_new_line): New function
13578 (s_expr_writer::write_any_indent): ditto
13579 (s_expr_writer::begin_s_expr): ditto
13580 (s_expr_writer::end_s_expr): ditto
13581 (write_state_fileloc): convert to method of state_writer...
13582 (state_writer:: write_state_fileloc): ...and use methods of
13583 s_expr_writer to write indentation into the gtype.state output file
13584 to visually represent the hierarchical structure of the list
13586 (write_state_fields): ditto, renaming to...
13587 (state_writer::write_state_fields)
13588 (write_state_a_string): ditto, renaming to...
13589 (state_writer::write_state_a_string)
13590 (write_state_string_option): ditto, renaming to...
13591 (state_writer::write_state_string_option)
13592 (write_state_type_option): ditto, renaming to...
13593 (state_writer::write_state_type_option)
13594 (write_state_nested_option): ditto, renaming to...
13595 (state_writer::write_state_nested_option)
13596 (write_state_option): ditto, renaming to...
13597 (state_writer::write_state_option)
13598 (write_state_options): ditto, renaming to...
13599 (state_writer::write_state_options)
13600 (write_state_lang_bitmap): ditto, renaming to...
13601 (state_writer::write_state_lang_bitmap)
13602 (write_state_version): ditto, renaming to...
13603 (state_writer::write_state_version)
13604 (write_state_scalar_type): ditto, renaming to...
13605 (state_writer::write_state_scalar_type)
13606 (write_state_string_type): ditto, renaming to...
13607 (state_writer::write_state_string_type)
13608 (write_state_undefined_type): ditto, renaming to...
13609 (state_writer::write_state_undefined_type)
13610 (write_state_struct_union_type): ditto, renaming to...
13611 (state_writer::write_state_struct_union_type)
13612 (write_state_struct_type): ditto, renaming to...
13613 (state_writer::write_state_struct_type)
13614 (write_state_user_struct_type): ditto, renaming to...
13615 (state_writer::write_state_user_struct_type)
13616 (write_state_lang_struct_type): ditto, renaming to...
13617 (state_writer::write_state_lang_struct_type)
13618 (write_state_param_struct_type): ditto, renaming to...
13619 (state_writer::write_state_param_struct_type)
13620 (write_state_pointer_type): ditto, renaming to...
13621 (state_writer::write_state_pointer_type)
13622 (write_state_array_type): ditto, renaming to...
13623 (state_writer::write_state_array_type)
13624 (write_state_gc_used): ditto, renaming to...
13625 (state_writer::write_state_gc_used)
13626 (write_state_common_type_content): ditto, renaming to...
13627 (state_writer::write_state_common_type_content)
13628 (write_state_type): ditto, renaming to...
13629 (state_writer::write_state_type)
13630 (write_state_pair_list): ditto, renaming to...
13631 (state_writer::write_state_pair_list)
13632 (write_state_pair): ditto, renaming to...
13633 (state_writer::write_state_pair)
13634 (write_state_typedefs): ditto, renaming to...
13635 (state_writer::write_state_typedefs)
13636 (write_state_structures): ditto, renaming to...
13637 (state_writer::write_state_structures)
13638 (write_state_param_structs): ditto, renaming to...
13639 (state_writer::write_state_param_structs)
13640 (write_state_variables): ditto, renaming to...
13641 (state_writer::write_state_variables)
13642 (write_state_srcdir): ditto, renaming to...
13643 (state_writer::write_state_srcdir)
13644 (write_state_files_list): ditto, renaming to...
13645 (state_writer::write_state_files_list)
13646 (write_state_languages): ditto, renaming to...
13647 (state_writer::write_state_languages)
13648 (write_state): create a state_writer instance and use it when
13649 writing out the state file
13651 2013-05-17 Mike Stump <mikestump@comcast.net>
13653 PR rtl-optimization/57304
13654 * web.c (union_match_dups): Ensure that DF_REF_LOC exists before
13655 accessing DF_REF_REAL_LOC.
13657 2013-05-17 Jakub Jelinek <jakub@redhat.com>
13659 PR rtl-optimization/57281
13660 PR rtl-optimization/57300
13661 * config/i386/i386.md (extendsidi2_1 dead reg splitter): Remove.
13662 (extendsidi2_1 peephole2s): Add instead 2 new peephole2s, that undo
13663 what the other splitter did if the registers are dead.
13665 2013-05-17 Richard Biener <rguenther@suse.de>
13667 * tree-ssa-alias.c (stmt_kills_ref_p_1): Properly compare
13670 2013-05-17 Jakub Jelinek <jakub@redhat.com>
13672 * gcc.c (SANITIZER_SPEC): Reject -fsanitize=address -fsanitize=thread
13675 2013-05-17 Marek Polacek <polacek@redhat.com>
13677 * tree-ssa-strlen.c (handle_char_store): Don't invalidate cached
13678 length when doing non-zero store of storing '\0' to '\0'.
13680 2013-05-17 Jakub Jelinek <jakub@redhat.com>
13682 * tree-vect-patterns.c (vect_recog_rotate_pattern): For
13683 vect_external_def oprnd1 with loop_vinfo, try to emit
13684 optional cast, negation and and stmts on the loop preheader
13685 edge instead of into the pattern def seq.
13687 PR tree-optimization/57051
13688 * fold-const.c (const_binop) <case VEC_LSHIFT_EXPR,
13689 case VEC_RSHIFT_EXPR>: Fix BYTES_BIG_ENDIAN handling.
13691 2013-05-16 Nick Clifton <nickc@redhat.com>
13693 * config/rl78/rl78.c (rl78_attribute_table): Add naked.
13694 (rl78_is_naked_func): New function.
13695 (rl78_expand_prologue): Skip prologue generation for naked functions.
13696 (rl78_expand_epilogue): Skip epilogue generation for naked functions.
13697 * doc/extend.texi (naked): Add RL78 to the list of processors
13698 that supports this attribute.
13700 2013-05-16 Jeff Law <law@redhat.com>
13702 * Makefile.in (tree-switch-conversion.o): Depend on $(OPTABS_H).
13704 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
13706 * config/i386/driver-i386.c (host_detect_local_cpu): Determine
13707 cache parameters using detect_caches_amd also for CYRIX,
13708 NSC and TM2 signatures.
13710 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
13711 Dzianis Kahanovich <mahatma@eu.by>
13715 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
13716 VIA/Centaur processors and determine their cache parameters
13717 using detect_caches_amd.
13719 2013-05-16 Teresa Johnson <tejohnson@google.com>
13721 * cfgrtl.c (verify_hot_cold_block_grouping): Return err.
13722 (rtl_verify_edges): New function.
13723 (rtl_verify_bb_insns): Ditto.
13724 (rtl_verify_bb_pointers): Ditto.
13725 (rtl_verify_bb_insn_chain): Ditto.
13726 (rtl_verify_fallthru): Ditto.
13727 (rtl_verify_bb_layout): Ditto.
13728 (rtl_verify_flow_info_1): Outline checks into new functions.
13729 (rtl_verify_flow_info): Ditto.
13731 2013-05-16 Steve Ellcey <sellcey@imgtec.com>
13733 * cfghooks.c (copy_bbs): Add update_dominance argument.
13734 * cfghooks.h (copy_bbs): Update prototype.
13735 * tree-cfg.c (gimple_duplicate_sese_region):
13736 Add update_dominance argument.
13737 * tree-flow.h (gimple_duplicate_sese_region): Update prototype.
13738 * tree-ssa-loop-ch.c (copy_loop_headers): Update
13739 gimple_duplicate_sese_region call.
13740 * tree-vect-loop-manip.c (slpeel_tree_duplicate_loop_to_edge_cfg):
13741 Update copy_bbs call.
13742 * cfgloopmanip.c (duplicate_loop_to_header_edge): Ditto.
13743 * trans-mem.c (ipa_uninstrument_transaction): Ditto.
13745 2013-05-16 Jakub Jelinek <jakub@redhat.com>
13747 * tree-vectorizer.h (NUM_PATTERNS): Increment.
13748 * tree-vect-patterns.c (vect_vect_recog_func_ptrs): Add
13749 vect_recog_rotate_pattern.
13750 (vect_recog_rotate_pattern): New function.
13752 2013-05-16 Jason Merrill <jason@redhat.com>
13754 * Makefile.in (LLINKER): New variable.
13755 (mostlyclean): Remove link mutex.
13756 * configure.ac: Handle --enable-link-mutex.
13757 * lock-and-run.sh: New script.
13759 2013-05-16 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
13762 * config/arm/arm.c (arm_function_ok_for_sibcall): Add check
13765 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
13767 * reorg.c (link_cc0_insns): Wrap in #ifdef HAVE_cc0.
13769 2013-05-16 Greta Yorsh <Greta.Yorsh@arm.com>
13771 * config/arm/arm-protos.h (gen_movmem_ldrd_strd): New declaration.
13772 * config/arm/arm.c (next_consecutive_mem): New function.
13773 (gen_movmem_ldrd_strd): Likewise.
13774 * config/arm/arm.md (movmemqi): Update condition and code.
13775 (unaligned_loaddi, unaligned_storedi): New patterns.
13777 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
13779 * config.gcc: Obsolete *-*-solaris2.9*.
13780 * doc/install.texi (Specific, *-*-solaris2*): Document it.
13782 2013-05-16 Richard Biener <rguenther@suse.de>
13784 * passes.c (init_optimization_passes): Move pass_parallelize_loops
13785 earlier, after GRAPHITE transforms and IV canonicalization.
13787 2013-05-16 Jakub Jelinek <jakub@redhat.com>
13789 * omp-low.c (extract_omp_for_data): For collapsed loops,
13790 if at least one of the loops is known at compile time to
13791 iterate zero times, set count to 0.
13792 (expand_omp_regimplify_p): New function.
13793 (expand_omp_for_generic): For collapsed loops, if at least
13794 one of the loops isn't known to iterate at least once,
13795 add runtime check with setting count to 0.
13796 (expand_omp_for_static_nochunk, expand_omp_for_static_chunk):
13797 For unsigned types if it isn't known at compile time that
13798 the loop will iterate at least once, add runtime check to bypass
13799 the whole loop if initial condition isn't true.
13801 2013-05-16 Nathan Sidwell <nathan@codesourcery.com>
13803 * varasm.c (default_use_anchors_for_symbol_p): Use decl_replaceable_p.
13805 2013-05-16 Marc Glisse <marc.glisse@inria.fr>
13807 PR middle-end/57286
13808 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Disable some
13809 transformations to avoid an infinite loop.
13811 2013-05-16 Marek Polacek <polacek@redhat.com>
13813 * tree-scalar-evolution.c (scev_const_prop): Add more dumps.
13815 2013-05-15 Leif Ekblad <leif@rdos.net>
13817 * config/i386/i386.c (ix86_decompose_address): Use
13818 DEFAULT_TLS_SEG_REG to access TLS segment register.
13819 * config/i386/i386.h (DEFAULT_TLS_SEG_REG): New define.
13820 * config/i386/rdos.h (DEFAULT_TLS_SEG_REG): Ditto.
13821 (TARGET_TLS_DIRECT_SEG_REFS_DEFAULT): Ditto.
13823 2013-05-15 Richard Sandiford <rdsandiford@googlemail.com>
13826 * config/mips/mips.c (mips_function_ok_for_sibcall): Don't allow
13827 sibling calls to functions that would normally be lazily bound,
13828 unless $gp is call-clobbered.
13830 2013-05-15 Uros Bizjak <ubizjak@gmail.com>
13832 * config/i386/i386.c (ix86_option_override_internal): Update
13833 processor_alias_table for missing PTA_PRFCHW and PTA_FXSR flags. Add
13834 PTA_POPCNT to corei7 entry. Do not enable SSE prefetch on
13835 non-SSE 3dNow! targets. Enable TARGET_PRFCHW for TARGET_3DNOW targets.
13836 * config/i386/i386.md (prefetch): Enable for TARGET_PRFCHW instead
13838 (*prefetch_3dnow): Enable for TARGET_PRFCHW only.
13840 2013-05-15 Andreas Schwab <schwab@suse.de>
13842 * config/m68k/m68k.md (*rotlhi3_lowpart, *rotlqi3_lowpart): Name
13843 for rotlhi3+1 and rotlqi3+1, resp. Fix reference to non-existing
13846 2013-05-15 Teresa Johnson <tejohnson@google.com>
13848 * loop-unroll.c (report_unroll_peel): Check decision before
13849 emitting unroll/peel message.
13851 2013-05-15 Teresa Johnson <tejohnson@google.com>
13853 * function.h (has_bb_partition): New rtl_data flag.
13854 (bb_reorder_complete): Ditto.
13855 * cfgcleanup.c (try_crossjump_to_edge): Check for has_bb_partition
13856 instead of flag_reorder_blocks_and_partition.
13857 * cfgrtl.c (verify_hot_cold_block_grouping): Moved from bb-reorder.c,
13858 with some enhancements.
13859 (rtl_verify_flow_info_1): Call verify_hot_cold_block_grouping.
13860 * bb-reorder.c (connect_traces): Check for has_bb_partition
13861 instead of flag_reorder_blocks_and_partition.
13862 (verify_hot_cold_block_grouping): Moved to cfgrtl.c.
13863 (reorder_basic_blocks): Set bb_reorder_complete flag, remove call to
13864 verify_hot_cold_block_grouping.
13865 (partition_hot_cold_basic_blocks): Set has_bb_partition.
13867 2013-05-15 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
13870 * config/arm/predicates.md (call_insn_operand): New predicate.
13871 * config/arm/constraints.md ("Cs", "Ss"): New constraints.
13872 * config/arm/arm.md (*call_insn, *call_value_insn): Match only
13873 if insn is not a tail call.
13874 (*sibcall_insn, *sibcall_value_insn): Adjust for tailcalling through
13876 * config/arm/arm.h (enum reg_class): New caller save register class.
13877 (REG_CLASS_NAMES): Likewise.
13878 (REG_CLASS_CONTENTS): Likewise.
13879 * config/arm/arm.c (arm_function_ok_for_sibcall): Allow tailcalling
13882 2013-05-15 Richard Biener <rguenther@suse.de>
13884 * tree-vect-loop.c (vect_transform_loop): Use MSG_NOTE instead
13885 of MSG_OPTIMIZED_LOCATIONS.
13886 * tree-vect-slp.c (vect_make_slp_decision): Likewise.
13887 (vect_slp_transform_bb): Indicate location in MSG_OPTIMIZED_LOCATIONS
13889 * tree-vectorizer.c (vectorize_loops): Use MSG_NOTE instead
13890 of MSG_OPTIMIZED_LOCATIONS.
13891 (execute_vect_slp): Likewise.
13892 * tree-vect-loop-manip.c (vect_do_peeling_for_loop_bound): Likewise.
13893 (vect_create_cond_for_alias_checks): Likewise.
13894 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Likewise.
13895 (vect_recog_widen_mult_pattern): Likewise.
13896 (vect_recog_widen_sum_pattern): Likewise.
13897 (vect_recog_over_widening_pattern): Likewise.
13898 (vect_recog_widen_shift_pattern): Likewise.
13899 (vect_recog_vector_vector_shift_pattern): Likewise.
13900 (vect_recog_divmod_pattern): Likewise.
13901 (vect_recog_mixed_size_cond_pattern): Likewise.
13902 (vect_recog_bool_pattern): Likewise.
13903 (vect_pattern_recog_1): Likewise.
13905 2013-05-15 Martin Jambor <mjambor@suse.cz>
13907 * ipa-prop.c (ipa_make_edge_direct_to_target): Redirect calls to
13908 non-functions to builtin_unreachable.
13909 * ipa-inline-transform.c (inline_call): Do not assert estimates were
13910 correct when new direct edges were discovered.
13912 2013-05-15 Martin Jambor <mjambor@suse.cz>
13914 * ipa-prop.c (ipa_print_node_jump_functions): Print symbol order in
13915 header, print symbol order instead of node uid, print more information
13916 about indirect edge targets.
13917 (ipa_make_edge_direct_to_target): Print symbol order instead of node
13919 (ipa_make_edge_direct_to_target): Likewise.
13920 (remove_described_reference): Likewise.
13921 (propagate_controlled_uses): Likewise.
13922 (ipa_print_node_params): Also print symbol order.
13923 (ipcp_transform_function): Print symbol order instead of node uids.
13924 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Likewise.
13925 (cgraph_get_create_real_symbol_node): Likewise.
13926 * ipa-cp.c (print_lattice): Likewise.
13927 (print_all_lattices): Likewise.
13928 (determine_versionability): Likewise.
13929 (initialize_node_lattices): Likewise.
13930 (estimate_local_effects): Likewise.
13931 (update_profiling_info): Likewise.
13932 (create_specialized_node): Likewise.
13933 (perhaps_add_new_callers): Likewise.
13934 (decide_about_value): Likewise.
13935 (decide_whether_version_node): Likewise.
13936 (identify_dead_nodes): Likewise.
13937 * ipa-inline-analysis.c (dump_inline_edge_summary): Likewise.
13938 (dump_inline_summary): Likewise.
13939 (estimate_node_size_and_time): Likewise.
13940 (inline_analyze_function): Likewise.
13941 * ipa-inline.c (report_inline_failed_reason): Likewise.
13942 (want_early_inline_function_p): Likewise.
13943 (edge_badness): Likewise.
13944 (update_edge_key): Likewise.
13945 (inline_small_functions): Likewise. Add dumping of order to two other
13947 * ipa-pure-const.c (pure_const_read_summary): Print symbol order
13948 instead of node uids.
13949 (propagate_pure_const): Likewise.
13950 (propagate_pure_const): Likewise.
13951 * ipa-utils.c (dump_cgraph_node_set): Likewise.
13952 * lto-cgraph.c (input_node): Explicitly specify we dump uid.
13953 * lto-symtab.c (lto_cgraph_replace_node): Print symbol order instead
13955 * tree-pretty-print.c (dump_function_header): Likewise.
13956 * tree-sra.c (convert_callers_for_node): Dump in traditional format.
13957 Print symbol order instead of node uids.
13959 2013-05-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
13961 * config/s390/s390.c (s390_register_move_cost): Don't impose the
13962 FPR<->GPR move cost penalty if ldgr/lgdr can be used.
13964 2013-05-15 Richard Biener <rguenther@suse.de>
13966 PR tree-optimization/57275
13967 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Fix
13968 return value for fail to do runtime alias checks for gather loads.
13970 2013-05-15 Jan Hubicka <jh@suse.cz>
13974 * lto-symtab.c (lto_symtab_symbol_p): Add external symbol;
13975 weakrefs are not external.
13976 (lto_symtab_merge_decls): Fix thinko when dealing with
13977 non-lto_symtab decls.
13978 (lto_symtab_merge_cgraph_nodes): Use lto_symtab_symbol_p.
13979 (lto_symtab_prevailing_decl): Get int sync with lto_symtab_symbol_p.
13980 * varpool.c (dump_varpool_node): Dump more flags.
13982 2013-05-15 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
13984 * config/i386/i386.c (processor_alias_table): Add instruction
13985 FSGSBASE for AMD bdver3 architecture.
13987 2013-05-14 Jakub Jelinek <jakub@redhat.com>
13989 * tree.c (warn_deprecated_use): Print file:line using locus color.
13990 * diagnostic.c (diagnostic_report_current_module): Print file:line
13991 and file:line:column using locus color.
13993 2013-05-14 Mike Stump <mikestump@comcast.net>
13995 * gdbinit.in: Add __null.
13997 2013-05-14 Mike Stump <mikestump@comcast.net>
13999 * recog.h: Rename struct recog_data to Recog_data.
14000 * recog.c: Likewise.
14001 * reload.c (can_reload_into): Likewise.
14002 * config/picochip/picochip.c: Likewise.
14004 2013-05-14 Mike Stump <mikestump@comcast.net>
14006 * web.c (union_match_dups): Also check DF_REF_REAL_LOC.
14008 2013-05-14 Steven Bosscher <steven@gcc.gnu.org>
14010 * resource.h (struct resources): Remove unch_memory member.
14011 (CLEAR_RESOURCE): Don't clear unch_memory.
14012 * resource.c (mark_referenced_resources): Don't set it.
14013 (mark_set_resources): Likewise.
14014 (mark_target_live_regs): Don't clear it.
14015 (init_resource_info): Likewise.
14016 * reorg.c (resource_conflicts_p): Don't compare it.
14017 (redundant_insn): Don't set it.
14019 * rtl.h (next_label, skip_consecutive_labels, link_cc0_insns):
14021 * emit-rtl.c (next_label): Remove unused function.
14022 (skip_consecutive_labels, link_cc0_insns): Move to ...
14023 * reorg.c (skip_consecutive_labels, link_cc0_insns): ... here, the
14024 only place where these functions are used, and make them static.
14026 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
14028 * fold-const.c (fold_negate_expr): Handle vectors.
14029 (fold_truth_not_expr): Make it static.
14030 (fold_invert_truthvalue): New static function.
14031 (invert_truthvalue_loc): Handle vectors. Do not call
14032 fold_truth_not_expr directly.
14033 (fold_unary_loc) <BIT_NOT_EXPR>: Handle comparisons.
14034 <TRUTH_NOT_EXPR>: Do not cast to boolean.
14035 (fold_comparison): Handle vector constants.
14036 (fold_binary_loc) <TRUTH_XOR_EXPR>: Remove redundant code.
14037 (fold_ternary_loc) <VEC_COND_EXPR>: Adapt more COND_EXPR optimizations.
14038 * tree.h (fold_truth_not_expr): Remove declaration.
14040 2013-05-14 James Greenhalgh <james.greenhalgh@arm.com>
14042 * config/aarch64/aarch64-simd.md
14043 (aarch64_vcond_internal<mode>): Rename to...
14044 (aarch64_vcond_internal<mode><mode>): ...This, for integer modes.
14045 (aarch64_vcond_internal<VDQF_COND:mode><VDQF:mode>): ...This for
14046 float modes. Clarify all iterator modes.
14047 (vcond<mode><mode>): Use new name for vcond expanders.
14048 (vcond<v_cmp_result><mode>): Likewise.
14049 (vcondu<mode><mode>: Likewise.
14050 * config/aarch64/iterators.md (VDQF_COND): New.
14052 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
14055 * fold-const.c (fold_binary_loc) <shift>: Use an unsigned
14056 variable for the shift amount. Check that we shift by non-negative
14059 2013-05-14 Chung-Lin Tang <cltang@codesourcery.com>
14062 * config/arm/arm.h (EPILOGUE_USES): Only return true
14063 for LR_REGNUM after epilogue_completed.
14065 2013-05-14 Joern Rennecke <joern.rennecke@embecosm.com>
14067 * config/avr/avr.c (avr_encode_section_info): Bail out if the type
14068 is error_mark_node.
14070 2013-05-14 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
14073 * configure.ac (gcc_cv_ld_as_needed): Disable before Solaris 11
14074 and Solaris 11+/x86 with gld.
14075 * configure: Regenerate.
14077 2013-05-14 Jakub Jelinek <jakub@redhat.com>
14079 * expmed.c (expand_shift_1): Canonicalize rotates by
14080 constant bitsize / 2 to bitsize - 1.
14081 * simplify-rtx.c (simplify_binary_operation_1) <case ROTATE,
14082 case ROTATERT>: Likewise.
14085 2013-05-10 Jakub Jelinek <jakub@redhat.com>
14087 * config/i386/i386.md (rotateinv): New code attr.
14088 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
14089 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
14090 roll $31, %eax, etc.
14092 2013-05-14 Richard Biener <rguenther@suse.de>
14094 PR middle-end/57235
14095 * tree-eh.c (sink_clobbers): Give up for successors with
14096 multiple predecessors and no virtual uses.
14098 2013-05-14 Eric Botcazou <ebotcazou@adacore.com>
14100 * config/sparc/sp64-elf.h (CPP_SUBTARGET_SPEC): Delete.
14101 * config/sparc/openbsd64.h (CPP_SUBTARGET_SPEC): Likewise.
14103 2013-05-14 Jakub Jelinek <jakub@redhat.com>
14105 PR middle-end/57251
14106 * expr.c (expand_expr_real_2) <case WIDEN_MULT_EXPR>: Handle
14107 the case when both op0 and op1 have VOIDmode.
14109 2013-05-14 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
14111 * config/rl78/rl78.md(mulsi3_g13): Add additional 'nop' required
14112 in multiply-accumulate mode.
14114 2013-05-13 Guozhi Wei <carrot@google.com>
14116 * dwarf2asm.c (dw2_output_indirect_constant_1): Mark new decl STATIC.
14118 2013-05-13 Kai Tietz <ktietz@redhat.com>
14121 * config/i386/cygming.h (TARGET_PECOFF): Define as true.
14122 * config/i386/i386.h (TARGET_PECOFF): Define by default as false.
14123 (PIC_OFFSET_TABLE_REGNUM): Use TARGET_PECOFF.
14124 * config/i386/i386.c (ix86_option_override_internal): Likewise.
14125 (ix86_expand_prologue): Likewise.
14126 (ix86_expand_split_stack_prologue): Likewise.
14127 (legitimate_pic_address_disp_p): Likewise.
14128 (legitimize_pic_address): Likewise.
14129 (legitimize_tls_address): Likewise.
14130 (legitimize_pe_coff_symbol): Likewise.
14131 (output_pic_addr_const): Likewise.
14132 (construct_plt_address): Likewise.
14133 (ix86_expand_call): Likewise.
14134 (x86_output_mi_thunk): Likewise.
14135 (x86_function_profiler): Likewise.
14137 2013-05-13 Sofiane Naci <sofiane.naci@arm.com>
14139 * config/aarch64/aarch64-simd.md (aarch64_simd_mov<mode>): Group
14140 similar switch cases.
14141 (aarch64_simd_mov): Rename to aarch64_split_simd_mov. Update.
14142 (aarch64_simd_mov_to_<mode>low): Delete.
14143 (aarch64_simd_mov_to_<mode>high): Delete.
14144 (move_lo_quad_<mode>): Add w<-r alternative.
14145 (aarch64_simd_move_hi_quad_<mode>): Likewise.
14146 (aarch64_simd_mov_from_*): Update type attribute.
14147 * config/aarch64/aarch64.c (aarch64_split_simd_move): Refacror switch
14150 2013-05-13 Jan Hubicka <jh@suse.cz>
14152 * mode-switching.c (optimize_mode_switching): Set correct RTL profile.
14153 * config/i386/i386.c (ix86_compute_frame_layout,
14154 ix86_expand_epilogue, emit_i387_cw_initialization,
14155 ix86_expand_vector_move_misalign, ix86_fp_comparison_strategy,
14156 ix86_local_alignment): Fix use of size/speed predicates.
14158 2013-05-13 Jakub Jelinek <jakub@redhat.com>
14160 PR tree-optimization/45216
14161 PR tree-optimization/57157
14162 * tree-ssa-forwprop.c (simplify_rotate): Only recognize
14163 the (-Y) & (B - 1) variant if OP is |.
14164 * expmed.c (expand_shift_1): For rotations by const0_rtx just
14165 return shifted. Use (-op1) & (prec - 1) as other_amount
14166 instead of prec - op1.
14168 2013-05-13 Martin Jambor <mjambor@suse.cz>
14170 PR middle-end/42371
14171 * ipa-prop.h (IPA_UNDESCRIBED_USE): New macro.
14172 (ipa_constant_data): New type.
14173 (ipa_jump_func): Use ipa_constant_data to hold information about
14174 constant jump functions.
14175 (ipa_get_jf_constant): Adjust to jump function type changes.
14176 (ipa_get_jf_constant_rdesc): New function.
14177 (ipa_param_descriptor): New field controlled_uses.
14178 (ipa_get_controlled_uses): New function.
14179 (ipa_set_controlled_uses): Likewise.
14180 * ipa-ref.h (ipa_find_reference): Declare.
14181 * ipa-prop.c (ipa_cst_ref_desc): New type.
14182 (ipa_print_node_jump_functions_for_edge): Adjust for jump function type
14184 (ipa_set_jf_constant): Likewise. Also create reference descriptions.
14185 New parameter cs. Adjust all callers.
14186 (ipa_analyze_params_uses): Detect uncontrolled and controlled uses.
14187 (remove_described_reference): New function.
14188 (jfunc_rdesc_usable): Likewise.
14189 (try_make_edge_direct_simple_call): Decrement controlled use count,
14190 attempt to remove reference if it hits zero.
14191 (combine_controlled_uses_counters): New function.
14192 (propagate_controlled_uses): Likewise.
14193 (ipa_propagate_indirect_call_infos): Call propagate_controlled_uses.
14194 (ipa_edge_duplication_hook): Duplicate reference descriptions.
14195 (ipa_print_node_params): Print described use counter.
14196 (ipa_write_jump_function): Adjust to jump function type changes.
14197 (ipa_read_jump_function): New parameter CS, pass it to
14198 ipa_set_jf_constant. Adjust caller.
14199 (ipa_write_node_info): Stream controlled use count
14200 (ipa_read_node_info): Likewise.
14201 * cgraph.c (cgraph_mark_address_taken_node): Bail out instead of
14203 * ipa-cp.c (ipcp_discover_new_direct_edges): Decrement controlled use
14204 count. Remove cloning-added reference if it reaches zero.
14205 * ipa-ref.c (ipa_find_reference): New function.
14207 2013-05-13 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
14209 * config/i386/i386.c (processor_target_table): Modified default
14210 alignment values for AMD BD and BT architectures.
14212 2013-05-13 Marc Glisse <marc.glisse@inria.fr>
14214 * tree-vect-generic.c (uniform_vector_p): Move ...
14215 * tree.c (uniform_vector_p): ... here.
14216 * tree.h (uniform_vector_p): Declare it.
14217 * fold-const.c (fold_binary_loc) <shift>: Turn the second argument
14220 2013-05-13 Jakub Jelinek <jakub@redhat.com>
14222 PR tree-optimization/57230
14223 * tree-ssa-strlen.c (handle_char_store): Record length for
14224 array store from STRING_CST.
14226 PR tree-optimization/57230
14227 * tree-ssa-strlen.c (handle_char_store): Add missing integer_zerop
14230 2013-05-12 Joern Rennecke <joern.rennecke@embecosm.com>
14232 * config/epiphany/epiphany.c (epiphany_init): Check size of
14233 NUM_MODES_FOR_MODE_SWITCHING.
14234 (epiphany_expand_prologue):
14235 Remove CONFIG_REGNUM initial value handling code.
14236 (epiphany_optimize_mode_switching): Handle EPIPHANY_MSW_ENTITY_CONFIG.
14237 (epiphany_mode_needed, epiphany_mode_entry_exit): Likewise.
14238 (emit_set_fp_mode, epiphany_mode_after): Likewise.
14239 (epiphany_mode_needed) <Handle EPIPHANY_MSW_ENTITY_AND>:
14240 Don't return 1 for FP_MODE_NONE.
14241 * config/epiphany/epiphany.h (NUM_MODES_FOR_MODE_SWITCHING):
14242 Add value for EPIPHANY_MSW_ENTITY_CONFIG.
14243 (EPIPHANY_MSW_ENTITY_CONFIG, EPIPHANY_MSW_ENTITY_NUM): Define.
14244 * config/epiphany/epiphany.md (save_config): New pattern.
14246 2013-05-12 Uros Bizjak <ubizjak@gmail.com>
14248 * config/i386/i386.md (*zero_extendsidi2): Add *x->?r alternative.
14250 2013-05-10 Uros Bizjak <ubizjak@gmail.com>
14252 * config/i386/i386.md (memory): Handle sseishft1.
14253 * config/i386/sse.md (*vec_extractv4si): Remove memory attribute.
14254 (*vec_extractv2di_1): Ditto.
14256 2013-05-10 Vladimir Makarov <vmakarov@redhat.com>
14258 * lra-assigns.c (find_hard_regno_for): Add 1 to the cost of call
14261 2013-05-10 Sebastian Huber <sebastian.huber@embedded-brains.de>
14263 * config/arm/t-rtems-eabi: Remove mthumb/march=armv7 multilib.
14264 Add mthumb/march=armv7-a multilib.
14265 Add mthumb/march=armv7-r multilib.
14266 Add mthumb/march=armv7-a/mfpu=neon/mfloat-abi=hard multilib.
14268 2013-05-10 Ralf Corsépius <ralf.corsepius@rtems.org>
14270 * config/v850/t-rtems: Add more multilibs.
14272 2013-05-10 Richard Biener <rguenther@suse.de>
14274 PR tree-optimization/57214
14275 * tree-ssa-loop-ivcanon.c (propagate_constants_for_unrolling): Do
14276 not propagate from SSA names that occur in abnormal PHI nodes.
14278 2013-05-10 Marc Glisse <marc.glisse@inria.fr>
14280 * stor-layout.c (element_precision): New function.
14281 * machmode.h (element_precision): Declare it.
14282 * tree.c (build_minus_one_cst): New function.
14283 (element_precision): Likewise.
14284 * tree.h (build_minus_one_cst): Declare new function.
14285 (element_precision): Likewise.
14286 * fold-const.c (operand_equal_p): Use element_precision.
14287 (fold_binary_loc): Handle vector types.
14288 * convert.c (convert_to_integer): Use element_precision.
14289 * gimple.c (iterative_hash_canonical_type): Handle complex and vectors
14292 2013-05-10 Richard Sandiford <rdsandiford@googlemail.com>
14294 * config/mips/mips-protos.h (m16_uimm3_b, m16_simm4_1, m16_nsimm4_1)
14295 (m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4, m16_simm8_1)
14296 (m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1, m16_uimm8_4)
14297 (m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
14298 * config/mips/mips.c (m16_check_op, m16_uimm3_b, m16_simm4_1)
14299 (m16_nsimm4_1, m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4)
14300 (m16_simm8_1, m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1)
14301 (m16_uimm8_4, m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
14302 * config/mips/constraints.md (Udb8, Usb5, Usb8, Usd8, Uub8, Uuw5)
14303 (Uuw8): New constraints.
14304 (Usb4): Move into alphabetical order.
14305 * config/mips/predicates.md (db8_operand, sb5_operand, sb8_operand)
14306 (sd8_operand, ub8_operand, uw8_operand): New predicates.
14307 * config/mips/mips.md (*xor<mode>3, *xor<mode>3_mips16): Name
14308 previously unnamed patterns.
14309 (*add<mode>3_mips16, *xor<mode>3_mips16, *<optab>si3_mips16)
14310 (*ashldi3_mips16, *ashrdi3_mips16, *lshrdi3_mips16)
14311 (*slt<u>_<GPR:mode><GPR2:mode>_mips16)
14312 (*sle<u>_<GPR:mode><GPR2:mode>_mips16): Use constraints instead
14313 of set_attr_alternative/if_then_else. Use extended_mips16 instead
14314 of specific lengths.
14316 2013-05-10 Jakub Jelinek <jakub@redhat.com>
14318 * config/i386/i386.md (rotateinv): New code attr.
14319 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
14320 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
14321 roll $31, %eax, etc.
14323 PR tree-optimization/45216
14324 PR tree-optimization/57157
14325 * tree-ssa-forwprop.c (simplify_rotate): New function.
14326 (ssa_forward_propagate_and_combine): Call it.
14328 2013-05-10 Richard Biener <rguenther@suse.de>
14330 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do not
14331 disable peeling when we version for aliasing.
14332 (vector_alignment_reachable_p): Honor explicit user alignment.
14333 (vect_supportable_dr_alignment): Likewise.
14334 * tree-vect-loop-manip.c (vect_can_advance_ivs_p): Use
14335 STMT_VINFO_LOOP_PHI_EVOLUTION_PART instead of recomputing it.
14336 * tree-vect-loop.c (vect_transform_loop): First apply versioning,
14337 then peeling to arrange for the cost-model check to come first.
14339 2013-05-10 Alan Modra <amodra@gmail.com>
14341 * configure.ac (HAVE_AS_TLS): Swap powerpc64 and powerpc cases.
14342 (HAVE_LD_LARGE_TOC): Don't mention AIX in help text.
14343 * configure: Regenerate.
14345 2013-05-10 Alan Modra <amodra@gmail.com>
14348 * varasm.c (default_elf_select_section): Move !DECL_P check..
14349 (get_named_section): ..to here before calling get_section_name.
14351 (default_section_type_flags): Add DECL_P check.
14352 * config/i386/winnt.c (i386_pe_section_type_flags): Likewise.
14353 * config/rs6000/rs6000.c (rs6000_xcoff_section_type_flags): Likewise.
14355 2013-05-09 Joern Rennecke <joern.rennecke@embecosm.com>
14357 * config/epiphany/epiphany.c (epiphany_expand_prologue):
14358 When using gen_stack_adjust_str with a register offset, add a
14359 REG_FRAME_RELATED_EXPR note.
14361 2013-05-09 Uros Bizjak <ubizjak@gmail.com>
14363 * config/i386/sse.md (*vec_extractv4si_0_zext): New pattern.
14364 (*vec_extractv4si_zext_mem): Ditto.
14365 (*vec_extractv2di): Add 0->x and x->x alternatives.
14366 * config/i386/mmx.md (*vec_extractv2si_zext_mem): New pattern.
14367 * config/i386/i386.md (*zero_extendsidi2): Add *Yj->?r alternative.
14369 2013-05-09 Jason Merrill <jason@redhat.com>
14371 N3639 C++1y VLA support
14372 * gimplify.c (gimplify_vla_decl): Don't touch an existing
14375 * tree.c (build_constructor_va): New.
14376 * tree.h: Declare it.
14378 2013-05-09 Martin Jambor <mjambor@suse.cz>
14381 * gimple-fold.c (canonicalize_constructor_val): Call
14382 cgraph_get_create_real_symbol_node instead of cgraph_get_create_node.
14384 2013-05-09 Jan Hubicka <jh@suse.cz>
14385 Richard Biener <rguenther@suse.de>
14388 * symtab.c (symtab_make_decl_local): Do not add private names.
14390 2013-05-09 Jan Hubicka <jh@suse.cz>
14393 * symtab.c (insert_to_assembler_name_hash): Handle clones.
14394 (unlink_from_assembler_name_hash): Likewise.
14395 (symtab_prevail_in_asm_name_hash, symtab_register_node,
14396 symtab_unregister_node, symtab_initialize_asm_name_hash,
14397 change_decl_assembler_name): Update.
14399 2013-05-09 Sofiane Naci <sofiane.naci@arm.com>
14401 * config/aarch64/aarch64.md: New movtf split.
14402 (*movtf_aarch64): Update.
14403 (aarch64_movdi_tilow): Handle TF modes and rename to
14404 aarch64_movdi_<mode>low.
14405 (aarch64_movdi_tihigh): Handle TF modes and rename to
14406 aarch64_movdi_<mode>high
14407 (aarch64_movtihigh_di): Handle TF modes and rename to
14408 aarch64_mov<mode>high_di
14409 (aarch64_movtilow_di): Handle TF modes and rename to
14410 aarch64_mov<mode>low_di
14411 (aarch64_movtilow_tilow): Remove spurious whitespace.
14412 * config/aarch64/aarch64.c (aarch64_split_128bit_move): Handle TFmode
14414 (aarch64_print_operand): Update.
14416 2013-05-09 Alan Modra <amodra@gmail.com>
14418 * configure.ac (HAVE_AS_TLS): Enable tests for powerpcle and
14420 * configure: Regenerate.
14422 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
14424 * config/i386/mmx.md (*vec_extract* splitters): Simplify post-reload
14425 splitter preparation statements.
14426 * config/i386/sse.md (*vec_extract* splitters): Ditto.
14427 (*avx_vperm_broadcast_<mode>): Use adjust_address instead of
14430 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
14432 * gimple-ssa-strength-reduction.c (count_candidates): Change
14433 return value to int.
14434 (analyze_candidates_and_replace): Change type of length to int.
14436 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
14438 * config/i386/sse.md (PEXTR_MODE, PEXTR_MODEx): Remove.
14439 (*vec_extract<mode>): Use VI12_128 mode iterator.
14440 (*vec_extract<mode>_mem): Ditto.
14441 (*vec_extract*_mem splitters): Merge splitters using VI_128 mode
14444 2013-05-08 Diego Novillo <dnovillo@google.com>
14449 2012-08-17 Diego Novillo <dnovillo@google.com>
14452 * configure.ac: Add libintl.h to AC_CHECK_HEADERS list.
14453 * config.in: Regenerate.
14454 * configure: Regenerate.
14455 * intl.h: Always include libintl.h if HAVE_LIBINTL_H is set.
14457 2013-05-08 Jan Hubicka <jh@suse.cz>
14460 * cgraph.c (cgraph_make_node_local_1): Se unique_name.
14461 * cgraph.h (symtab_node_base): Add unique_name.
14462 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
14463 input_overwrite_node, input_varpool_node): Stream unique_name.
14464 * cgraphclones.c (cgraph_create_virtual_clone,
14465 cgraph_function_versioning): Set unique_name.
14466 * ipa.c (function_and_variable_visibility): Set unique_name.
14468 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
14470 * gimple-ssa-strength-reduction.c (find_phi_def): Revert former "fix."
14471 (alloc_cand_and_find_basis): Restrict conditional candidate
14472 processing to CAND_MULTs.
14474 2013-05-08 Jan Hubicka <jh@suse.cz>
14477 lto-symtab.c (lto_symtab_symbol_p): New function.
14478 (lto_symtab_resolve_can_prevail_p, lto_symtab_resolve_symbols,
14479 lto_symtab_resolve_symbols, lto_symtab_merge_decls_2,
14480 lto_symtab_merge_decls_1, lto_symtab_merge_cgraph_nodes_1):
14481 Skip static symbols.
14483 2013-05-08 Paolo Carlini <paolo.carlini@oracle.com>
14485 PR tree-optimization/57200
14486 * tree-ssa-loop-niter.c (do_warn_aggressive_loop_optimizations):
14487 Only call inform if the preceding warning_at returns true.
14489 2013-05-07 Han Shen <shenhan@google.com>
14491 * cfgexpand.c (record_or_union_type_has_array_p): New function.
14492 (expand_used_vars): Add logic handling '-fstack-protector-strong'.
14493 * common.opt (fstack-protector-strong): New option.
14494 * doc/cpp.texi (__SSP_STRONG__): New builtin "__SSP_STRONG__".
14495 * doc/invoke.texi (Optimization Options): Document
14496 "-fstack-protector-strong".
14497 * gcc.c (LINK_SSP_SPEC): Add 'fstack-protector-strong'.
14499 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
14501 * config/mips/mips.c (mips_machine_reorg2): Return 0.
14503 2013-05-07 Vladimir Makarov <vmakarov@redhat.com>
14505 * ira.c (update_equiv_regs): Add insn having equiv memory even if
14506 it is not lhs of the insn.
14507 (setup_reg_equiv): Remove insn having equiv memory which it is not
14509 * lra-constraints.c (process_address): Try to improve generation
14510 code for address base + disp.
14511 (lra_constraints): Make correct the code for checking insn setting
14512 up backward equivalence. Remove insn only if it is in the init
14514 * lra-eliminations.c (update_reg_eliminate): Change return value.
14515 (lra_eliminate): Use the result.
14517 2013-05-07 Uros Bizjak <ubizjak@gmail.com>
14519 * config/i386/sse.md (ssescalarnummask): New mode attribute.
14520 (PEXTR_MODE, PEXTR_MODEx): New mode iterators.
14521 (*vec_extract<mode>): Merge from *sse4_1_pextrb_memory and
14522 *sse4_1_pextrw_memory using PEXTR_MODE mode iterator. Handle
14523 register target operands.
14524 (*vec_extractv8hi_sse2): New pattern.
14525 (*vec_extractv16qi_zext): Rename from *sse4_1_pextrb_<mode>.
14526 (*vec_extractv8hi_zext): Rename from *sse2_pextrw_<mode>.
14527 (*vec_extract<mode>_mem): New insn and split pattern.
14529 2013-05-07 Christophe Lyon <christophe.lyon@linaro.org>
14531 * config/arm/arm.c (arm_asan_shadow_offset): New function.
14532 (TARGET_ASAN_SHADOW_OFFSET): Define.
14533 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Define.
14534 (LINUX_OR_ANDROID_CC): Add ASAN_CC1_SPEC.
14536 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
14538 * gimple-ssa-strength-reduction.c (MAX_INCR_VEC_LEN): New constant.
14539 (incr_vec_index): Return -1 if increment not found.
14540 (create_add_on_incoming_edge): Assert if increment not found.
14541 (record_increment): Limit number of increments recorded.
14542 (all_phi_incrs_profitable): Return false if an increment not found.
14543 (replace_profitable_candidates): Don't process increments that were
14545 (analyze_candidates_and_replace): Limit size of incr_vec.
14547 2013-05-07 Richard Biener <rguenther@suse.de>
14549 * calls.c (special_function_p): setjmp-like functions are leaf.
14550 * builtins.def (BUILT_IN_SETJMP): setjmp is leaf.
14551 * tree-inline.c (update_ssa_across_abnormal_edges): Remove assert.
14553 2013-05-07 Sofiane Naci <sofiane.naci@arm.com>
14555 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): call splitter.
14556 (aarch64_simd_mov<mode>): New expander.
14557 (aarch64_simd_mov_to_<mode>low): New instruction pattern.
14558 (aarch64_simd_mov_to_<mode>high): Likewise.
14559 (aarch64_simd_mov_from_<mode>low): Likewise.
14560 (aarch64_simd_mov_from_<mode>high): Likewise.
14561 (aarch64_dup_lane<mode>): Update.
14562 (aarch64_dup_lanedi): New instruction pattern.
14563 * config/aarch64/aarch64-protos.h (aarch64_split_simd_move): New prototype.
14564 * config/aarch64/aarch64.c (aarch64_split_simd_move): New function.
14566 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
14568 * gimple-ssa-strength-reduction.c (lazy_create_slsr_reg): Remove.
14569 (replace_mult_candidate): Remove unnecessary argument; remove
14570 unnecessary parameter from call to introduce_cast_before_cand.
14571 (replace_unconditional_candidate): Remove unnecessary parameter
14572 from call to replace_mult_candidate.
14573 (replace_conditional_candidate): Likewise.
14574 (insert_initializers): Use make_temp_ssa_name.
14575 (introduce_cast_before_cand): Remove unnecessary argument; use
14576 make_temp_ssa_name.
14577 (replace_one_candidate): Remove unnecessary argument; remove
14578 unnecessary parameter from calls to introduce_cast_before_cand.
14579 (replace_profitable_candidates): Remove unnecessary parameters
14580 from calls to replace_one_candidate.
14582 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
14584 * gimple-ssa-strength-reduction.c (find_phi_def): Don't record a
14585 phi def as possibly hiding a basis for a CAND_ADD whose operands
14586 have been commuted in the analysis.
14587 (alloc_cand_and_find_basis): Add parms to call to find_phi_def.
14589 2013-05-07 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
14591 * config/aarch64/aarch64.md
14592 (cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): Restrict the
14593 shift value between 0-4.
14595 2013-05-07 Richard Biener <rguenther@suse.de>
14597 * double-int.h (rshift): New overload.
14598 * double-int.c (rshift): New function.
14599 * tree-ssa-sccvn.c (copy_reference_ops_from_ref): Optimize.
14600 (create_reference_ops_from_ref): Remove.
14601 (vn_reference_insert): Use shared ops for constructing the
14602 reference and copy it.
14604 2013-05-07 Richard Biener <rguenther@suse.de>
14606 PR middle-end/57190
14607 * tree-eh.c (sink_clobbers): Properly propagate
14608 SSA_NAME_OCCURS_IN_ABNORMAL_PHI.
14610 2013-05-07 Jakub Jelinek <jakub@redhat.com>
14612 PR tree-optimization/57149
14613 * tree-ssa-uninit.c (uninit_undefined_value_p): New inline.
14614 (can_skip_redundant_opnd, compute_uninit_opnds_pos,
14615 collect_phi_def_edges, execute_late_warn_uninitialized): Use
14616 uninit_undefined_value_p instead of ssa_undefined_value_p.
14619 * expr.c (expand_expr_addr_expr_1): Handle COMPOUND_LITERAL_EXPR
14620 for modifier == EXPAND_INITIALIZER.
14622 2013-05-07 Anton Blanchard <anton@samba.org>
14624 * configure.ac (HAVE_LD_LARGE_TOC): Use correct linker emulation
14625 for powerpc64 little endian.
14626 * configure: Regenerate.
14628 2013-05-06 Graham Stott <grahams@btinternet.com>
14630 * expmed.c (init_expmed_rtl): Remove unused fields reg_fld, plus_fld,
14631 mult_fld, sdiv_fld1, udiv_fld1, sdiv_32_fld1, smod_32_fld1,
14632 wide_mult_fld1, wide_lshr_fld1, shift_fld1, shift_mult_fld1,
14633 shift_add_fld1, shift_sub0_fld1, shift_sub1_fld1.
14635 2013-05-06 Graham Stott <grahams@btinternet.com>
14637 * gensupport.c (add_predicate_code): Also exclude SCRATCH from rtx
14638 codes which allow non-lvalues.
14640 2013-05-06 Marc Glisse <marc.glisse@inria.fr>
14642 * tree.c (integer_all_onesp) <COMPLEX_CST>: Test that both
14643 components are all 1s.
14644 (integer_minus_onep): New function.
14645 * tree.h (integer_minus_onep): Declare it.
14646 * fold-const.c (fold_binary_loc) <MULT_EXPR>: Test
14647 integer_minus_onep instead of integer_all_onesp.
14649 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
14652 * config/sh/sh.md (*cmp_div0s_0, *cmp_div0s_1, *movsicc_div0s): Add
14653 variations of these patterns.
14655 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
14657 * config/i386/i386.md (isa): Add x64_sse4 member.
14658 (enabled): Handle x64_sse4.
14659 (*movdi_internal): Add *x->?r alternative to emit pextrq $0,%xmm,%reg
14660 instruction for 64bit SSE4_1 targets. Update insn attributes.
14661 (*movsi_internal): Add *x->?r alternative to emit pextrd $0,%xmm,%reg
14662 instruction for SSE4_1 targets. Update insn attributes.
14663 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
14664 with *sse4_1_pextrd and *sse4_1_pextrq having const_0 selector.
14665 (*vec_extractv2di_1): Merge with *sse4_1_pextrq having
14667 (*vec_extractv4si): Rename from *sse4_1_pextrd.
14668 (*vec_extractv4si_zext): Rename from *sse4_1_pextrd_zext.
14669 (*vec_extract<ssevecmodelower>_0 splitters): Merge splitters together.
14671 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
14674 * config/sh/sh.md (tstsi_t_zero_extract_eq): Use QIHISIDI mode iterator.
14676 2013-05-06 Maxim Kuznetsov <maks.kuznetsov@gmail.com>
14678 * final.c (do_assembler_dialects): Don't handle curly braces and
14679 vertical bar escaped by % as dialect delimiters.
14680 (output_asm_insn): Print curly braces and vertical bar if escaped
14681 by % and ASSEMBLER_DIALECT defined.
14682 * doc/tm.texi.in (ASSEMBLER_DIALECT): Document new standard escapes.
14683 * doc/tm.texi: Regenerated.
14685 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
14687 * config/mips/mips.c: Include tree-pass.h.
14688 (mips_reorg): Split in pre- and post-dbr_schedule parts.
14689 (mips_machine_reorg2): Move mips_reorg post-dbr_schedule parts here.
14690 (pass_mips_machine_reorg2): New machine specific pass.
14691 (insert_pass_mips_machine_reorg2): New pass plugin definition.
14692 (mips_option_override): Register the new pass.
14693 * rtl.h (cleanup_barriers): Remove prototype.
14694 (dbr_schedule): Likewise.
14695 * jump.c (cleanup_barriers): Make static.
14696 * reorg.c (dbr_schedule): Likewise.
14698 2013-05-06 Richard Biener <rguenther@suse.de>
14700 PR tree-optimization/57185
14701 * tree-parloops.c (add_field_for_reduction): Handle anonymous
14702 SSA names properly.
14704 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
14707 * config/i386/i386.c (add_parameter_dependencies): Add dependence
14708 between "first_arg" and "insn", not "last" and "insn".
14710 2013-05-06 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
14712 * gimple-ssa-strength-reduction.c (slsr_process_phi): Re-enable.
14713 (find_candidates_in_block): Re-enable slsr_process_phi.
14714 (create_phi_basis): Fix double counting of candidate adjustment.
14716 2013-05-06 Richard Biener <rguenther@suse.de>
14718 PR middle-end/57147
14719 * tree-cfg.c (gimple_purge_dead_abnormal_call_edges): If
14720 the edge is also fallthru, preserve it and just clear the
14722 * tree-cfgcleanup.c (remove_fallthru_edge): If the edge is
14723 also complex, preserve that and just clear the fallthru flag.
14724 * tree-inline.c (update_ssa_across_abnormal_edges): Also
14725 update virtual operands.
14727 2013-05-06 Alan Modra <amodra@gmail.com>
14729 * config/rs6000/linux.h (DEFAULT_ASM_ENDIAN): Define.
14730 (LINK_OS_LINUX_EMUL): Use ENDIAN_SELECT.
14731 * config/rs6000/linux64.h (DEFAULT_ASM_ENDIAN): Define.
14732 * config/rs6000/sysv4le.h (DEFAULT_ASM_ENDIAN): Define.
14733 (LINK_TARGET_SPEC): Use ENDIAN_SELECT.
14734 * config/rs6000/sysv4.h (DEFAULT_ASM_ENDIAN): Define as -mbig.
14736 2013-05-06 Alan Modra <amodra@gmail.com>
14738 * config/rs6000/sysv4.h (ENDIAN_SELECT): Define, extracted from
14739 (ASM_SPEC): ..here. Emit DEFAULT_ASM_ENDIAN too.
14740 (DEFAULT_ASM_ENDIAN): Define.
14741 (CC1_SPEC, LINK_TARGET_SPEC): Use ENDIAN_SELECT.
14742 * config/rs6000/linux64.h (ASM_SPEC32): Remove endian options.
14743 Update -K PIC clause from sysv4.h.
14744 (ASM_SPEC_COMMON): Use ENDIAN_SELECT.
14745 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Likewise.
14747 2013-05-06 Alan Modra <amodra@gmail.com>
14749 * config/rs6000/rs6000.md (bswapdi 2nd splitter): Don't swap words
14750 twice for little-endian.
14751 (ashrdi3_no_power, ashrdi3): Support little-endian.
14753 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
14756 * config/sh/sh.c (sh_rtx_costs): Handle SMIN and SMAX cases.
14757 * config/sh/sh.md (*clips, uminsi3, *clipu, clipu_one): New insns and
14759 * config/sh/iterators.md (SMIN_SMAX): New code iterator.
14760 * config/sh/predicates.md (arith_reg_or_0_or_1_operand,
14761 clips_min_const_int, clips_max_const_int, clipu_max_const_int):
14764 2013-05-05 Steven Bosscher <steven@gcc.gnu.org>
14765 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
14767 * config.gcc (hppa*-*-*): Remove MASK_BIG_SWITCH from CPU default.
14768 * config/pa/pa.opt: Make mbig-switch a no-op.
14769 * config/pa/pa.h (TARGET_DEFAULT): Remove MASK_BIG_SWITCH.
14770 (CASE_VECTOR_MODE): Always return SImode.
14771 (ASM_OUTPUT_ADDR_VEC_ELT, ASM_OUTPUT_ADDR_DIFF_ELT): Remove code
14772 for the !TARGET_BIG_SWITCH case.
14773 * config/pa/pa-linux.h: Likewise.
14774 * config/pa/pa-openbsd.h: Likewise.
14775 * config/pa/pa-hpux.h: Define TARGET_DEFAULT to 0.
14776 * config/pa/pa.md (short_jump): Remove define_insn.
14777 (casesi): Remove code for the !TARGET_BIG_SWITCH case.
14778 (casesi0): Remove define_insn.
14779 (type): Remove btable_branch.
14780 (pa_combine_type): Likewise.
14781 (in_nullified_branch_delay): Likewise.
14782 (in_call_delay): Likewise.
14783 (define_delay): Likewise.
14784 (define_insn_reservation "Z3"): Likewise.
14785 (define_insn_reservation "Z4"): Likewise.
14786 * config/pa/pa.c (pa_reorg): Remove code for !TARGET_BIG_SWITCH.
14787 (pa_adjust_insn_length): Remove adjustment for btable branches.
14788 * doc/invoke.texi (HPPA Options): Delete documentation for mbig-switch
14791 2013-05-05 Uros Bizjak <ubizjak@gmail.com>
14793 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
14794 from sse2_stored and *sse2_storeq_rex64 using SWI48 mode iterator.
14795 Add m->r,x alternatives.
14796 (*vec_extract<ssevecmodelower>_0 splitters): Merge V2DI and V4SI
14797 splitters using SWI48x mode iterator.
14798 (*vec_extract_v2di_0_sse): Rename from *sse2_storeq. Disable for
14799 TARGET_64BIT. Add m->x alternative.
14800 (*vec_extractv4si_mem): Rename from *vec_ext_v4si_mem.
14801 Add o->x alternative. Enable for TARGET_SSE.
14802 (sse_storeq): Remove expander.
14803 (*vec_extractv2di_1): Enable for TARGET_SSE. Split alternatives
14804 with memory input operand.
14805 (*vec_extractv2di_1 splitter): New.
14806 (*vec_extractv4sf_mem): Rename from *vec_extract_v4sf_mem.
14807 * config/i386/i386.md (ssevecmodelower): New mode attribute.
14809 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
14811 * config/rs6000/rs6000.c (INT_P): Reformat. Delete obsolete comment.
14812 (INT_LOWPART): Delete.
14813 (extract_MB): Adjust.
14814 (extract_ME): Adjust.
14815 (print_operand): Adjust.
14817 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
14819 * config/rs6000/predicates.md (reg_or_add_cint_operand,
14820 reg_or_sub_cint_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
14821 (reg_or_logical_cint_operand, easy_fp_constant,
14822 logical_const_operand): Delete "CONST_DOUBLE" case.
14823 * config/rs6000/rs6000.c (num_insns_constant_wide): Delete
14824 "HOST_BITS_PER_WIDE_INT == 64" test.
14825 (num_insns_constant): Ditto. Delete CONST_DOUBLE DImode/VOIDmode case.
14826 (build_mask64_2_operands): Delete "HOST_BITS_PER_WIDE_INT >= 64" test.
14827 (rs6000_emit_set_const): Delete CONST_DOUBLE case.
14828 (rs6000_emit_set_long_const): Delete "HOST_BITS_PER_WIDE_INT >= 64"
14830 (includes_rldic_lshift_p, includes_rldicr_lshift_p): Delete
14831 CONST_DOUBLE DImode/VOIDmode case.
14832 (INT_P, INT_LOWPART): Delete CONST_DOUBLE case.
14833 (print_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case. Delete
14834 CONST_DOUBLE VOIDmode case.
14835 (output_toc): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
14836 (rs6000_rtx_costs): Delete CONST_DOUBLE DImode/VOIDmode case.
14837 * config/rs6000/rs6000.md (iordi3, xordi3, splitter for these):
14838 Delete CONST_DOUBLE case.
14839 (splitters for mov FMOVE64 const_double): Delete
14840 "HOST_BITS_PER_WIDE_INT == 32" case. Delete
14841 "HOST_BITS_PER_WIDE_INT >= 64" test.
14842 (splitter for mov DI const_int): Delete "HOST_BITS_PER_WIDE_INT == 32"
14844 (mov DI const_double): Delete.
14846 2013-05-04 Jakub Jelinek <jakub@redhat.com>
14848 * combine.c (combine_simplify_rtx) <case SUBREG>: If nonzero_bits
14849 on op shows all bits zero in mode of a lowpart subreg, return zero.
14851 2013-05-03 Michael Meissner <meissner@linux.vnet.ibm.com>
14854 * config/rs6000/rs6000.h (HARD_REGNO_CALLER_SAVE_MODE): Use DFmode
14855 to save TFmode registers and DImode to save TImode registers for
14856 caller save operations.
14857 (HARD_REGNO_CALL_PART_CLOBBERED): TFmode and TDmode do not need to
14858 mark being partially clobbered since they only use the first
14861 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): TFmode
14862 and TDmode only use the upper 64-bits of each VSX register.
14864 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
14866 * gimple-ssa-strength-reduction.c (slsr_process_phi): Disable.
14867 (find_candidates_in_block): Disable slsr_process_phi.
14869 2013-05-03 Guozhi Wei <carrot@google.com>
14871 * coverage.c (coverage_obj_init): Move the construction of gcov
14873 (build_init_ctor): ... here.
14875 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
14877 * gimple-ssa-strength-reduction.c (cand_kind): Add CAND_PHI.
14878 (slsr_cand_d): Redefine def_phi.
14879 (stride_status, phi_adjust_status, count_phis_status): New enums.
14880 (find_phi_def): New.
14881 (find_basis_for_base_expr): New.
14882 (find_basis_for_candidate): Handle hidden bases.
14883 (alloc_cand_and_find_basis): Handle phi candidates.
14884 (slsr_process_phi): New.
14885 (create_mul_ssa_cand): Exclude phi base candidates; use integer_onep.
14886 (create_mul_imm_cand): Likewise.
14887 (create_add_ssa_cand): Exclude phi base candidates.
14888 (create_add_imm_cand): Likewise.
14889 (slsr_process_cast): Likewise.
14890 (slsr_process_copy): Likewise.
14891 (find_candidates_in_block): Handle phi candidates.
14892 (dump_candidate): Likewise.
14893 (unconditional_cands): Delete.
14894 (unconditional_cands_with_known_stride_p): Delete.
14895 (phi_dependent_cand_p): New.
14896 (cand_increment): Handle phi-dependent candidates.
14897 (replace_dependent): Delete.
14898 (replace_mult_candidate): New.
14899 (replace_unconditional_candidate): New.
14900 (incr_vec_index): Move to avoid forward reference.
14901 (create_add_on_incoming_edge): New.
14902 (create_phi_basis): New.
14903 (replace_dependents): Delete.
14904 (replace_conditional_candidate): New.
14905 (phi_add_costs): New.
14906 (replace_uncond_cands_and_profitable_phis): New.
14907 (record_increment): Handle phi adjustments.
14908 (record_phi_increments): New.
14909 (record_increments): Handle phi adjustments.
14910 (phi_incr_cost): New.
14911 (lowest_cost_path): Handle phis.
14912 (total_savings): Likewise.
14913 (analyze_increments): Likewise.
14914 (ncd_with_phi): New.
14915 (ncd_of_cand_and_phis): New.
14916 (nearest_common_dominator_for_cands): Handle phi increments.
14917 (all_phi_incrs_profitable): New.
14918 (replace_profitable_candidates): Handle phi-dependent candidates.
14919 (analyze_candidates_and_replace): Likewise.
14921 2013-05-03 Teresa Johnson <tejohnson@google.com>
14924 * sched-rgn.c (compute_dom_prob_ps): Ensure accumulated probabilities
14925 do not exceed REG_BR_PROB_BASE.
14927 2013-05-03 Jeff Law <law@redhat.com>
14929 PR tree-optimization/57144
14930 * tree-vrp.c (simplify_cond_using_ranges): Verify the constant
14931 operand of the condition will bit into the new type when eliminating
14932 a cast feeding a condition.
14934 2013-05-03 Jakub Jelinek <jakub@redhat.com>
14936 PR rtl-optimization/57130
14937 * combine.c (make_compound_operation) <case SUBREG>: Pass SET instead
14938 of COMPARE as in_code to the recursive call if needed.
14940 2013-05-03 Uros Bizjak <ubizjak@gmail.com>
14942 * config/i386/i386.md (isa): Add x64_sse4_noavx and x64_avx members.
14943 (enabled): Handle new members.
14944 * config/i386/sse.md (*vec_concatv2si): Merge from
14945 *vec_concatv2si_sse2 and vec_concatv2si_sse.
14946 (vec_concatv2di): Merge with *vec_concatv2di_rex64.
14948 2013-05-03 Joern Rennecke <joern.rennecke@embecosm.com>
14950 PR tree-optimization/57027
14951 * tree-ssa-math-opts.c (convert_mult_to_fma): When checking
14952 for fnms opportunity, check we got the prerequisite kind
14953 of tree / gimple before using accessor functions.
14955 2013-05-03 Richard Biener <rguenther@suse.de>
14957 * double-int.h (lshift): New overload without precision
14958 and arith argument.
14959 (operator *=, operator +=, operator -=): Move ...
14960 * double-int.c (operator *=, operator +=, operator -=): ... here
14961 and implement more efficiently.
14962 (mul_double_with_sign): Remove.
14963 (lshift_double): Adjust to take unsinged shift argument, push
14964 dispatching code to callers.
14965 (mul_double_wide_with_sign): Add early out for callers that
14966 are not interested in high parts or overflow.
14967 (lshift): New function.
14968 (lshift, rshift, alshift, arshift, llshift, lrshift): Add
14969 dispatch code here.
14970 (lrotate, rrotate): Use logical shifts.
14971 * expr.c (get_inner_reference): Use lshift.
14972 * fixed-value.c (do_fixed_divide): Likewise.
14973 * tree-dfa.c (get_ref_base_and_extent): Likewise.
14974 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Likewise.
14975 (indirect_refs_may_alias_p): Likewise.
14976 (stmt_kills_ref_p_1): Likewise.
14978 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
14980 * config/aarch64/aarch64-simd.md (simd_fabd): Correct the description.
14982 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
14984 * config/aarch64/aarch64-simd.md (*fabd_scalar<mode>3): Support
14985 scalar form of FABD instruction.
14987 2013-05-02 Vladimir Makarov <vmakarov@redhat.com>
14989 * lra-constraints.c (process_alt_operands): Add checking alt
14990 number to choose the best alternative.
14992 2013-05-02 Richard Biener <rguenther@suse.de>
14994 * tree-eh.c (cleanup_empty_eh_merge_phis): Remove rename_virts
14995 bitmap and its handling.
14996 (pass_cleanup_eh): Set todo_flags_finish to TODO_verify_ssa.
14998 2013-05-02 Richard Biener <rguenther@suse.de>
15000 PR middle-end/57140
15001 * tree-inline.c (copy_loops): Properly handle removed loops.
15002 (copy_cfg_body): Mark destination loops for fixup if source
15003 loops needed fixup.
15005 2013-05-02 Greta Yorsh <Greta.Yorsh@arm.com>
15008 * config/arm/arm.c (arm_expand_epilogue): Check really_return before
15009 generating simple_return for naked functions.
15011 2013-05-02 Martin Jambor <mjambor@suse.cz>
15013 PR middle-end/56988
15014 * ipa-prop.h (ipa_agg_replacement_value): New flag by_ref.
15015 * ipa-cp.c (ipa_get_indirect_edge_target_1): Also check that by_ref
15017 (find_aggregate_values_for_callers_subset): Fill in the by_ref flag of
15018 ipa_agg_replacement_value structures.
15019 (known_aggs_to_agg_replacement_list): Likewise.
15020 * ipa-prop.c (write_agg_replacement_chain): Stream by_ref flag.
15021 (read_agg_replacement_chain): Likewise.
15022 (ipcp_transform_function): Also check that by_ref flags match.
15024 2013-05-02 Richard Biener <rguenther@suse.de>
15026 * graphds.h (struct graph): Add obstack member.
15027 * graphds.c (new_graph): Initialize obstack and allocate
15029 (add_edge): Allocate edge from the obstack.
15030 (free_graph): Free the obstack instead of all edges and vertices.
15032 2013-05-02 Teresa Johnson <tejohnson@google.com>
15034 * loop-unswitch.c (unswitch_loop): Use helper routines with rounding
15036 * cfg.c (update_bb_profile_for_threading): Ditto.
15037 * tree-inline.c (copy_bb): Ditto.
15038 (copy_edges_for_bb): Ditto.
15039 (initialize_cfun): Ditto.
15040 (copy_cfg_body): Ditto.
15041 (expand_call_inline): Ditto.
15042 * ipa-inline-analysis.c (estimate_edge_size_and_time): Ditto.
15043 (estimate_node_size_and_time): Ditto.
15044 (inline_merge_summary): Ditto.
15045 * cgraphclones.c (cgraph_clone_edge): Ditto.
15046 (cgraph_clone_node): Ditto.
15047 * sched-rgn.c (compute_dom_prob_ps): Ditto.
15048 (compute_trg_info): Ditto.
15050 2013-05-02 Ian Bolton <ian.bolton@arm.com>
15052 * config/aarch64/aarch64.md (movsi_aarch64): Only allow to/from
15053 S reg when fp attribute set.
15054 (movdi_aarch64): Only allow to/from D reg when fp attribute set.
15056 2013-05-02 Ian Bolton <ian.bolton@arm.com>
15058 * config/aarch64/aarch64.md (*and_one_cmpl<mode>3_compare0):
15060 (*and_one_cmplsi3_compare0_uxtw): Likewise.
15061 (*and_one_cmpl_<SHIFT:optab><mode>3_compare0): Likewise.
15062 (*and_one_cmpl_<SHIFT:optab>si3_compare0_uxtw): Likewise.
15064 2013-05-02 Richard Biener <rguenther@suse.de>
15066 * tree-scalar-evolution.c (scev_info_hasher): Remove.
15067 (struct instantiate_cache_entry): New type.
15068 (struct instantiate_cache_entry_hasher): New hashtable descriptor.
15069 (struct instantiate_cache_type): New type.
15070 (set_instantiated_value, get_instantiated_value): Remove.
15071 (get_instantiated_value_entry): New function.
15072 (instantiate_scev_name): Use the new cache and adjust.
15073 (instantiate_scev_poly): Adjust.
15074 (instantiate_scev_binary): Likewise.
15075 (instantiate_array_ref): Likewise.
15076 (instantiate_scev_convert): Likewise.
15077 (instantiate_scev_not): Likewise.
15078 (instantiate_scev_3): Likewise.
15079 (instantiate_scev_2): Likewise.
15080 (instantiate_scev_r): Likewise.
15081 (instantiate_scev): Likewise.
15082 (resolve_mixers): Likewise.
15084 2013-05-01 Vladimir Makarov <vmakarov@redhat.com>
15087 * lra-constraints.c (best_small_class_operands_num): Remove.
15088 (process_alt_operands): Remove small_class_operands_num. Take
15089 small classes operands into losers and only if the operand is not
15090 matched. Modify debugging output.
15091 (curr_insn_transform): Remove best_small_class_operands_num.
15094 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
15096 * config/aarch64/aarch64-builtins.c
15097 (aarch64_gimple_fold_builtin.c): Fold more modes for reduc_splus_.
15098 * config/aarch64/aarch64-simd-builtins.def
15099 (reduc_splus_): Add new modes.
15100 (reduc_uplus_): New.
15101 * config/aarch64/aarch64-simd.md (aarch64_addvv4sf): Remove.
15102 (reduc_uplus_v4sf): Likewise.
15103 (reduc_splus_v4sf): Likewise.
15104 (aarch64_addv<mode>): Likewise.
15105 (reduc_uplus_<mode>): Likewise.
15106 (reduc_splus_<mode>): Likewise.
15107 (aarch64_addvv2di): Likewise.
15108 (reduc_uplus_v2di): Likewise.
15109 (reduc_splus_v2di): Likewise.
15110 (aarch64_addvv2si): Likewise.
15111 (reduc_uplus_v2si): Likewise.
15112 (reduc_splus_v2si): Likewise.
15113 (reduc_<sur>plus_<mode>): New.
15114 (reduc_<sur>plus_v2di): Likewise.
15115 (reduc_<sur>plus_v2si): Likewise.
15116 (reduc_<sur>plus_v4sf): Likewise.
15117 (aarch64_addpv4sf): Likewise.
15118 * config/aarch64/arm_neon.h
15119 (vaddv<q>_<s,u,f><8, 16, 32, 64): Rewrite using builtins.
15120 * config/aarch64/iterators.md (unspec): Remove UNSPEC_ADDV,
15121 add UNSPEC_SADDV, UNSPEC_UADDV.
15123 (sur): Add UNSPEC_SADDV, UNSPEC_UADDV.
15125 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
15127 * config/aarch64/arm_neon.h
15128 (v<max,min><nm><q><v>_<sfu><8, 16, 32, 64>): Rewrite using builtins.
15130 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
15132 * config/aarch64/aarch64-builtins
15133 (aarch64_gimple_fold_builtin): Fold reduc_<su><maxmin>_ builtins.
15135 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
15137 * config/aarch64/aarch64-simd-builtins.def
15138 (reduc_smax_): New.
15139 (reduc_smin_): Likewise.
15140 (reduc_umax_): Likewise.
15141 (reduc_umin_): Likewise.
15142 (reduc_smax_nan_): Likewise.
15143 (reduc_smin_nan_): Likewise.
15146 (smax): Update for V2SF, V4SF and V2DF modes.
15149 (smin_nan): Likewise.
15150 * config/aarch64/aarch64-simd.md (<maxmin><mode>3): Rename to...
15151 (<su><maxmin><mode>3): ...This, refactor.
15152 (s<maxmin><mode>3): New.
15153 (<maxmin_uns><mode>3): Likewise.
15154 (reduc_<maxmin_uns>_<mode>): Refactor.
15155 (reduc_<maxmin_uns>_v4sf): Likewise.
15156 (reduc_<maxmin_uns>_v2si): Likewise.
15157 (aarch64_<fmaxmin><mode>: Remove.
15158 * config/aarch64/arm_neon.h (vmax<q>_f<32,64>): Rewrite to use
15160 (vmin<q>_f<32,64>): Likewise.
15161 * config/iterators.md (unspec): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
15163 (su): Add mappings for smax, smin, umax, umin.
15165 (FMAXMINV): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
15166 (FMAXMIN): Rename as...
15167 (FMAXMIN_UNS): ...This.
15169 (fmaxminv): Likewise.
15170 (fmaxmin): Likewise.
15172 (maxmin_uns_op): Likewise.
15174 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
15176 * config/aarch64/arm_neon.h
15177 (vac<ge, gt><sd>_f<32, 64>): Rename to...
15178 (vca<ge, gt><sd>_f<32, 64>): ...this, reimpliment in C.
15179 (vca<ge, gt, lt, le><q>_f<32, 64>): Reimpliment in C.
15181 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
15183 * config/aarch64/aarch64-simd.md (*aarch64_fac<optab><mode>): New.
15184 * config/aarch64/iterators.md (FAC_COMPARISONS): New.
15186 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
15188 * config/aarch64/aarch64-simd.md
15189 (vcond<mode>_internal): Handle special cases for constant masks.
15190 (vcond<mode><mode>): Allow nonmemory_operands for outcome vectors.
15191 (vcondu<mode><mode>): Likewise.
15192 (vcond<v_cmp_result><mode>): New.
15194 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
15196 * config/aarch64/aarch64-builtins.c (BUILTIN_VALLDI): Define.
15197 (aarch64_fold_builtin): Add folding for cm<eq,ge,gt,tst>.
15198 * config/aarch64/aarch64-simd-builtins.def
15199 (cmeq): Update to BUILTIN_VALLDI.
15204 * config/aarch64/arm_neon.h
15205 (vc<eq, lt, le, gt, ge, tst><z><qsd>_<fpsu><8,16,32,64>): Remap
15206 to builtins or C as appropriate.
15208 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
15210 * config/aarch64/aarch64-simd-builtins.def (cmhs): Rename to...
15212 (cmhi): Rename to...
15214 * config/aarch64/aarch64-simd.md
15215 (simd_mode): Add SF.
15216 (aarch64_vcond_internal): Use new names for unsigned comparison insns.
15217 (aarch64_cm<optab><mode>): Rewrite to not use UNSPECs.
15218 * config/aarch64/aarch64.md (*cstore<mode>_neg): Rename to...
15219 (cstore<mode>_neg): ...This.
15220 * config/aarch64/iterators.md
15222 (unspec): Remove UNSPEC_CM<EQ, LE, LT, GE, GT, HS, HI, TST>.
15223 (COMPARISONS): New.
15224 (UCOMPARISONS): Likewise.
15225 (optab): Add missing comparisons.
15231 (VCMP_S): Likewise.
15232 (VCMP_U): Likewise.
15233 (V_cmp_result): Add DF, SF modes.
15234 (v_cmp_result): Likewise.
15236 (vmtype): Likewise.
15237 * config/aarch64/predicates.md (aarch64_reg_or_fp_zero): New.
15239 2013-05-01 Greta Yorsh <Greta.Yorsh@arm.com>
15241 * config/arm/thumb2.md (thumb2_smaxsi3,thumb2_sminsi3): Convert
15242 define_insn to define_insn_and_split.
15243 (thumb32_umaxsi3,thumb2_uminsi3): Likewise.
15244 (thumb2_negdi2,thumb2_abssi2,thumb2_neg_abssi2): Likewise.
15245 (thumb2_mov_scc,thumb2_mov_negscc,thumb2_mov_notscc): Likewise.
15246 (thumb2_movsicc_insn,thumb2_and_scc,thumb2_ior_scc): Likewise.
15247 (thumb2_negscc): Likewise.
15249 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
15251 * config/arm/thumb2.md (thumb2_incscc, thumb2_decscc): Delete.
15253 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
15255 * config/arm/thumb2.md: Remove trailing whitespaces.
15257 2013-04-30 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
15259 * explow.c (plus_constant): Pass "mode" to immed_double_int_const.
15260 Use gen_int_mode rather than GEN_INT.
15262 2013-04-30 H.J. Lu <hongjiu.lu@intel.com>
15264 * value-prof.c (stream_in_histogram_value): Remove the strayed
15267 2013-04-30 Richard Biener <rguenther@suse.de>
15269 PR middle-end/57122
15270 * cfghooks.c (split_edge): Properly check for the loop latch edge.
15272 2013-04-30 Richard Biener <rguenther@suse.de>
15274 PR middle-end/57107
15275 * tree-eh.c (sink_clobbers): Preserve virtual SSA form.
15277 2013-04-30 Andrey Belevantsev <abel@ispras.ru>
15279 PR rtl-optimization/56957
15280 PR rtl-optimization/57105
15281 * sel-sched.c (move_op_orig_expr_found): Remove insn_emitted
15282 variable. Use just INSN_UID for determining whether an insn
15283 should be only disconnected from the insn stream.
15284 * sel-sched-ir.h (EXPR_WAS_CHANGED): Remove.
15286 2013-04-30 Jakub Jelinek <jakub@redhat.com>
15288 PR tree-optimization/57104
15289 * tsan.c (instrument_expr): Don't instrument accesses to
15290 DECL_HARD_REGISTER VAR_DECLs.
15292 2013-04-30 Richard Biener <rguenther@suse.de>
15294 * function.h (loops_for_fn): New inline function.
15295 (set_loops_for_fn): Likewise.
15296 * cfgloop.h (place_new_loop): Add struct function parameter.
15297 (get_loop): Likewise.
15298 (get_loops): Likewise.
15299 (number_of_loops): Likewise.
15300 (fel_next): Adjust.
15301 (fel_init): Likewise.
15302 * cfg.c (get_loop_copy): Adjust.
15303 * cfgloop.c (flow_loops_dump): Likewise.
15304 (record_loop_exits): Likewise.
15305 (verify_loop_structure): Likewise.
15306 * cfgloopanal.c (mark_irreducible_loops): Likewise.
15307 (estimate_reg_pressure_cost): Likewise.
15308 (mark_loop_exit_edges): Likewise.
15309 * cfgloopmanip.c (place_new_loop): Likewise.
15310 (add_loop): Likewise.
15311 (duplicate_loop): Likewise.
15312 * graph.c (draw_cfg_nodes): Likewise.
15313 * graphite-clast-to-gimple.c (translate_clast_user): Likewise.
15314 * graphite-sese-to-poly.c (build_scop_scattering): Likewise.
15315 (extract_affine_chrec): Likewise.
15316 (build_scop_iteration_domain): Likewise.
15317 * graphite.c (graphite_initialize): Likewise.
15318 * ira-build.c (create_loop_tree_nodes): Likewise.
15319 (more_one_region_p): Likewise.
15320 (rebuild_regno_allocno_maps): Likewise.
15321 (mark_loops_for_removal): Likewise.
15322 (mark_all_loops_for_removal): Likewise.
15323 (remove_unnecessary_regions): Likewise.
15324 (ira_build): Likewise.
15325 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
15326 * loop-init.c (fix_loop_structure): Likewise.
15327 (gate_rtl_move_loop_invariants): Likewise.
15328 (gate_rtl_unswitch): Likewise.
15329 (gate_rtl_unroll_and_peel_loops): Likewise.
15330 (rtl_doloop): Likewise.
15331 * lto-streamer-in.c (input_cfg): Likewise.
15332 * lto-streamer-out.c (output_cfg): Likewise.
15333 * modulo-sched.c (sms_schedule): Likewise.
15334 * predict.c (tree_estimate_probability): Likewise.
15335 (tree_estimate_probability_driver): Likewise.
15336 (estimate_loops): Likewise.
15337 * tree-cfg.c (fixup_loop_arrays_after_move): Likewise.
15338 (move_sese_region_to_fn): Likewise.
15339 (debug_loop_num): Likewise.
15340 * tree-chrec.c (chrec_evaluate): Likewise.
15341 (hide_evolution_in_other_loops_than_loop): Likewise.
15342 (chrec_component_in_loop_num): Likewise.
15343 (reset_evolution_in_loop): Likewise.
15344 (evolution_function_is_invariant_rec_p): Likewise.
15345 * tree-if-conv.c (main_tree_if_conversion): Likewise.
15346 * tree-inline.c (copy_loops): Likewise.
15347 (copy_cfg_body): Likewise.
15348 (tree_function_versioning): Likewise.
15349 * tree-loop-distribution.c (rdg_flag_loop_exits): Likewise.
15350 * tree-scalar-evolution.c (chrec_contains_symbols_defined_in_loop):
15352 (add_to_evolution_1): Likewise.
15353 (scev_const_prop): Likewise.
15354 * tree-scalar-evolution.h (get_chrec_loop): Likewise.
15355 * tree-ssa-loop-ch.c (copy_loop_headers): Likewise.
15356 * tree-ssa-loop-im.c (analyze_memory_references): Likewise.
15357 (tree_ssa_lim_initialize): Likewise.
15358 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Likewise.
15359 (verify_loop_closed_ssa): Likewise.
15360 * tree-ssa-loop.c (tree_ssa_loop_init): Likewise.
15361 (tree_ssa_loop_im): Likewise.
15362 (tree_ssa_loop_unswitch): Likewise.
15363 (tree_vectorize): Likewise.
15364 (check_data_deps): Likewise.
15365 (tree_ssa_loop_ivcanon): Likewise.
15366 (tree_ssa_loop_bounds): Likewise.
15367 (tree_complete_unroll): Likewise.
15368 (tree_complete_unroll_inner): Likewise.
15369 (tree_parallelize_loops): Likewise.
15370 (tree_ssa_loop_prefetch): Likewise.
15371 (tree_ssa_loop_ivopts): Likewise.
15372 * tree-ssa.c (execute_update_addresses_taken): Liekwise.
15373 * tree-vectorizer.c (vectorize_loops): Likewise.
15375 2013-04-29 Mike Frysinger <vapier@gentoo.org>
15377 * config/arm/bpabi.h (EABI_LINK_SPEC): Define.
15378 (BPABI_LINK_SPEC): Use new EABI_LINK_SPEC.
15379 * config/arm/linux-eabi.h (LINK_SPEC): Replace BE8_LINK_SPEC
15380 with EABI_LINK_SPEC.
15382 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
15385 * config/i386/i386.md (*zero_extendsidi2): Add "!" to m->?*y
15388 2013-04-29 Vladimir Makarov <vmakarov@redhat.com>
15391 * lra-constraints.c (process_alt_operands): Discourage a bit more
15392 using memory for pseudos. Print cost dump for alternatives.
15393 Modify cost values for conflicts with early clobbers.
15394 (curr_insn_transform): Spill pseudos reassigned to NO_REGS.
15396 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
15399 * config/i386/i386.c (ix86_expand_vec_perm): Validize constant memory.
15401 2013-04-29 Ian Bolton <ian.bolton@arm.com>
15403 * config/aarch64/aarch64.md (movsi_aarch64): Support LDR/STR
15404 from/to S register.
15405 (movdi_aarch64): Support LDR/STR from/to D register.
15407 2013-04-29 Ian Bolton <ian.bolton@arm.com>
15409 * common/config/aarch64/aarch64-common.c: Enable REE pass at O2
15410 or higher by default.
15412 2013-04-29 Richard Biener <rguenther@suse.de>
15414 PR middle-end/57075
15415 * tree-inline.c (copy_edges_for_bb): Still split the bbs,
15416 even if not adding abnormal edges for calls that can make
15419 2013-04-29 Richard Biener <rguenther@suse.de>
15421 PR middle-end/57103
15422 * tree-cfg.c (move_stmt_op): Fix condition under which to update
15424 (move_stmt_r): Remove redundant checking.
15426 2013-04-29 Teresa Johnson <tejohnson@google.com>
15429 * basic-block.h (apply_scale): New function.
15430 (apply_probability): Use apply_scale.
15431 * gimple-streamer-in.c (input_bb): Ditto.
15432 * lto-streamer-in.c (input_cfg): Ditto.
15433 * lto-cgraph.c (merge_profile_summaries): Ditto.
15434 * tree-optimize.c (execute_fixup_cfg): Ditto.
15435 * tree-inline.c (copy_bb): Update comment to use apply_scale.
15436 (copy_edges_for_bb): Ditto.
15437 (copy_cfg_body): Ditto.
15439 2013-04-29 Tom de Vries <tom@codesourcery.com>
15441 * tree-ssa-tail-merge.c (find_same_succ_bb): Skip loop latch bbs.
15442 (replace_block_by): Don't set LOOPS_NEED_FIXUP.
15443 (tail_merge_optimize): Handle current_loops == NULL.
15445 2013-04-26 Jeff Law <law@redhat.com>
15447 * tree-vrp.c (range_fits_type_p): Move to earlier point in file.
15448 (simplify_cond_using_ranges): Generalize code to simplify
15449 COND_EXPRs where one argument is a constant and the other
15450 is an SSA_NAME created by an integral type conversion.
15452 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15454 * config/arm/arm.md (store_minmaxsi): Use only when
15455 optimize_insn_for_size_p.
15457 2013-04-29 Christian Bruel <christian.bruel@st.com>
15460 * sh.md (tstsi_t_zero_extract_eq): Set mode for operand 0.
15462 2013-04-29 Richard Biener <rguenther@suse.de>
15464 PR middle-end/57089
15465 * omp-low.c (expand_omp_taskreg): If the parent function had a broken
15466 loop tree make sure to schedule a fixup for the child as well.
15467 (expand_omp_for_generic): Properly add loops.
15468 (expand_omp_for_static_nochunk): Likewise.
15469 (expand_omp_for_static_chunk): Likewise.
15470 (expand_omp_for): For the degenerate case fixup loops.
15471 (expand_omp_sections): Fix default bb placement in loops.
15472 (expand_omp_atomic_pipeline): Properly add loops.
15474 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15476 * predict.c: Fix typo in comment above #define PROB_VERY_UNLIKELY.
15478 2013-04-29 Tom de Vries <tom@codesourcery.com>
15480 * tree-ssa-tail-merge.c: Update header comment.
15482 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
15484 * config/aarch64/arm_neon.h
15485 (vcvt<sd>_f<32,64>_s<32,64>): Rewrite in C.
15486 (vcvt<q>_f<32,64>_s<32,64>): Rewrite using builtins.
15487 (vcvt_<high_>_f<32,64>_f<32,64>): Likewise.
15488 (vcvt<qsd>_<su><32,64>_f<32,64>): Likewise.
15489 (vcvta<qsd>_<su><32,64>_f<32,64>): Likewise.
15490 (vcvtm<qsd>_<su><32,64>_f<32,64>): Likewise.
15491 (vcvtn<qsd>_<su><32,64>_f<32,64>): Likewise.
15492 (vcvtp<qsd>_<su><32,64>_f<32,64>): Likewise.
15494 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
15496 * config/aarch64/aarch64-simd.md
15497 (<optab><VDQF:mode><fcvt_target>2): New, maps to fix, fixuns.
15498 (<fix_trunc_optab><VDQF:mode><fcvt_target>2): New, maps to
15499 fix_trunc, fixuns_trunc.
15500 (ftrunc<VDQF:mode>2): New.
15501 * config/aarch64/iterators.md (optab): Add fix, fixuns.
15502 (fix_trunc_optab): New.
15504 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
15506 * config/aarch64/aarch64-builtins.c
15507 (aarch64_builtin_vectorized_function): Vectorize over ifloorf,
15508 iceilf, lround, iroundf.
15510 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
15513 * config/i386/i386.h (enum ix86_tune_indices)
15514 <X86_TUNE_INTER_UNIT_MOVES_TO_VEC, X86_TUNE_INTER_UNIT_MOVES_FROM_VEC>:
15515 New, split from X86_TUNE_INTER_UNIT_MOVES.
15516 <X86_TUNE_INTER_UNIT_MOVES>: Remove.
15517 (TARGET_INTER_UNIT_MOVES_TO_VEC): New define.
15518 (TARGET_INTER_UNIT_MOVES_FROM_VEC): Ditto.
15519 (TARGET_INTER_UNIT_MOVES): Remove.
15520 * config/i386/i386.c (initial_ix86_tune_features): Update.
15521 Disable X86_TUNE_INTER_UNIT_MOVES_FROM_VEC for m_ATHLON_K8 only.
15522 (ix86_expand_convert_uns_didf_sse): Use
15523 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
15524 (ix86_expand_vector_init_one_nonzero): Ditto.
15525 (ix86_expand_vector_init_interleave): Ditto.
15526 (inline_secondary_memory_needed): Return true for moves from SSE class
15527 registers for !TARGET_INTER_UNIT_MOVES_FROM_VEC targets and for moves
15528 to SSE class registers for !TARGET_INTER_UNIT_MOVES_TO_VEC targets.
15529 * config/i386/constraints.md (Yi, Ym): Depend on
15530 TARGET_INTER_UNIT_MOVES_TO_VEC.
15531 (Yj, Yn): New constraints.
15532 * config/i386/i386.md (*movdi_internal): Change constraints of
15533 operand 1 from Yi to Yj and from Ym to Yn.
15534 (*movsi_internal): Ditto.
15535 (*movdf_internal): Ditto.
15536 (*movsf_internal): Ditto.
15537 (*float<SWI48x:mode><X87MODEF:mode>2_1): Use
15538 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
15539 (*float<SWI48x:mode><X87MODEF:mode>2_1 splitters): Ditto.
15540 (floatdi<X87MODEF:mode>2_i387_with_xmm): Ditto.
15541 (floatdi<X87MODEF:mode>2_i387_with_xmm splitters): Ditto.
15542 * config/i386/sse.md (movdi_to_sse): Ditto.
15543 (sse2_stored): Change constraint of operand 1 from Yi to Yj.
15544 Use TARGET_INTER_UNIT_MOVES_FROM_VEC instead of
15545 TARGET_INTER_UNIT_MOVES.
15546 (sse_storeq_rex64): Change constraint of operand 1 from Yi to Yj.
15547 (sse_storeq_rex64 splitter): Use TARGET_INTER_UNIT_MOVES_FROM_VEC
15548 instead of TARGET_INTER_UNIT_MOVES.
15549 * config/i386/mmx.md (*mov<mode>_internal): Change constraint of
15550 operand 1 from Yi to Yj and from Ym to Yn.
15552 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
15554 * config/aarch64/aarch64-simd-builtins.def (vec_unpacks_hi_): New.
15555 (float_truncate_hi_): Likewise.
15556 (float_extend_lo_): Likewise.
15557 (float_truncate_lo_): Likewise.
15558 * config/aarch64/aarch64-simd.md (vec_unpacks_lo_v4sf): New.
15559 (aarch64_float_extend_lo_v2df): Likewise.
15560 (vec_unpacks_hi_v4sf): Likewise.
15561 (aarch64_float_truncate_lo_v2sf): Likewise.
15562 (aarch64_float_truncate_hi_v4sf): Likewise.
15563 (vec_pack_trunc_v2df): Likewise.
15564 (vec_pack_trunc_df): Likewise.
15566 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
15568 * config/aarch64/aarch64-builtins.c
15569 (aarch64_fold_builtin): Fold float conversions.
15570 * config/aarch64/aarch64-simd-builtins.def
15571 (floatv2si, floatv4si, floatv2di): New.
15572 (floatunsv2si, floatunsv4si, floatunsv2di): Likewise.
15573 * config/aarch64/aarch64-simd.md
15574 (<optab><fcvt_target><VDQF:mode>2): New, expands to float and floatuns.
15575 * config/aarch64/iterators.md (FLOATUORS): New.
15576 (optab): Add float, floatuns.
15577 (su_optab): Likewise.
15579 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
15581 * config/aarch64/aarch64-builtins.c
15582 (aarch64_builtin_vectorized_function): Use new names for
15584 * config/aarch64/aarch64-simd-builtins.def (fcvtzs): Split as...
15585 (lbtruncv2sf, lbtruncv4sf, lbtruncv2df): ...This.
15586 (fcvtzu): Split as...
15587 (lbtruncuv2sf, lbtruncuv4sf, lbtruncuv2df): ...This.
15588 (fcvtas): Split as...
15589 (lroundv2sf, lroundv4sf, lroundv2df, lroundsf, lrounddf): ...This.
15590 (fcvtau): Split as...
15591 (lrounduv2sf, lrounduv4sf, lrounduv2df, lroundusf, lroundudf): ...This.
15592 (fcvtps): Split as...
15593 (lceilv2sf, lceilv4sf, lceilv2df): ...This.
15594 (fcvtpu): Split as...
15595 (lceiluv2sf, lceiluv4sf, lceiluv2df, lceilusf, lceiludf): ...This.
15596 (fcvtms): Split as...
15597 (lfloorv2sf, lfloorv4sf, lfloorv2df): ...This.
15598 (fcvtmu): Split as...
15599 (lflooruv2sf, lflooruv4sf, lflooruv2df, lfloorusf, lfloorudf): ...This.
15600 (lfrintnv2sf, lfrintnv4sf, lfrintnv2df, lfrintnsf, lfrintndf): New.
15601 (lfrintnuv2sf, lfrintnuv4sf, lfrintnuv2df): Likewise.
15602 (lfrintnusf, lfrintnudf): Likewise.
15603 * config/aarch64/aarch64-simd.md
15604 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Convert to
15606 (aarch64_fcvt<frint_suffix><su><mode>): Remove.
15607 * config/aarch64/iterators.md (FCVT): Include UNSPEC_FRINTN.
15608 (fcvt_pattern): Likewise.
15610 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
15612 * config/aarch64/aarch64-simd.md
15613 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Rename to...
15614 (l<fcvt_pattern><su_optab><VDQF:mode><fcvt_target>2): ... This.
15616 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
15618 * config/aarch64/arm_neon.h (vrndq<a,m,n,p>_f<32, 64>): Rename to...
15619 (vrnd<a,m,n,p>q_f<32, 64>): ...This, implement using builtin.
15620 (vrnd<a,m,n,p>_f32): Implement using builtins.
15621 (vrnd<i,x><q>_f<32, 64>): New.
15623 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
15625 * config/aarch64/aarch64-builtins.c
15626 (aarch64_builtin_vectorized_function): Fold to standard pattern names.
15627 * config/aarch64/aarch64-simd-builtins.def (frintn): New.
15628 (frintz): Rename to...
15630 (frintp): Rename to...
15632 (frintm): Rename to...
15634 (frinti): Rename to...
15635 (nearbyint): ...this.
15636 (frintx): Rename to...
15638 (frinta): Rename to...
15640 * config/aarch64/aarch64-simd.md
15641 (aarch64_frint<frint_suffix><mode>): Delete.
15642 (<frint_pattern><mode>2): Convert to insn.
15643 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRINTN.
15644 * config/aarch64/iterators.md (FRINT): Add UNSPEC_FRINTN.
15645 (frint_pattern): Likewise.
15646 (frint_suffix): Likewise.
15648 2013-04-29 Richard Biener <rguenther@suse.de>
15650 PR tree-optimization/57081
15651 * loop-init.c: Include tree-flow.h.
15652 (loop_optimizer_finalize): Free number of iteration estimates.
15653 * Makefile.in (loop-init.o): Add $(TREE_FLOW_H) dependency.
15655 2013-04-29 Jakub Jelinek <jakub@redhat.com>
15657 PR tree-optimization/57083
15658 * tree-vrp.c (extract_range_from_binary_expr_1): For LSHIFT_EXPR with
15659 non-singleton shift count range, zero extend low_bound for uns case.
15661 * config/i386/predicates.md (general_vector_operand): New predicate.
15662 * config/i386/i386.c (const_vector_equal_evenodd_p): New function.
15663 (ix86_expand_mul_widen_evenodd): Force op1 resp. op2 into register
15664 if they aren't nonimmediate operands. If their original values
15665 satisfy const_vector_equal_evenodd_p, don't shift them.
15666 * config/i386/sse.md (mul<mode>3): Use general_vector_operand
15667 predicates. For the SSE4.1 case force operands[{1,2}] into registers
15668 if not nonimmediate_operand.
15669 (vec_widen_smult_even_v4si): Use nonimmediate_operand predicates
15670 instead of register_operand.
15671 (vec_widen_<s>mult_odd_<mode>): Use general_vector_operand predicates.
15673 2013-04-28 Eric Botcazou <ebotcazou@adacore.com>
15675 * stor-layout.c (finalize_size_functions): Allocate a structure and
15676 reset cfun before dumping the functions.
15678 2013-04-27 Jakub Jelinek <jakub@redhat.com>
15680 * config/i386/i386.c (ix86_expand_call): Make cregs_size unsigned.
15683 * config/i386/i386.c (ix86_expand_mul_widen_evenodd): Don't
15684 use xop_pmacsdqh if uns_p.
15685 * config/i386/sse.md (xop_rotr<mode>3): Fix up computation of
15686 the immediate rotate count.
15688 2013-04-26 Vladimir Makarov <vmakarov@redhat.com>
15690 * rtl.h (struct rtx_def): Add comment for field jump.
15691 (LRA_SUBREG_P): New macro.
15692 * recog.c (register_operand): Check LRA_SUBREG_P.
15693 * lra.c (lra): Add note at the end of RTL code. Align non-empty
15695 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
15696 (lra_final_code_change): Skip subreg change for operators.
15697 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
15698 if there are no operand changes.
15699 * lra-constraints.c (curr_insn_set): New.
15700 (match_reload): Set LRA_SUBREG_P.
15701 (emit_spill_move): Ditto.
15702 (check_and_process_move): Use curr_insn_set. Process only single
15703 set insns. Don't initialize sec_mem_p and change_p.
15704 (simplify_operand_subreg): Use LRA_SUBREG_P.
15705 (reg_in_class_p): New function.
15706 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
15707 of #ifdef. Add code to remove cycling.
15708 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
15709 non-null disp. Reload inner instead of disp when base and index
15710 are null. Try to put lo_sum into register.
15711 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
15712 (check_and_process_move): Move code for move cost check to
15713 simple_move_p. Remove equiv_substitution.
15714 (simple_move_p): New function.
15715 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
15716 curr_insn_set. Call check_and_process_move only for single set
15717 insns. Use the new function. Move call of check_and_process_move
15718 after operand equiv substitution and address process.
15720 2013-04-26 Jakub Jelinek <jakub@redhat.com>
15723 * tree-ssa-uninit.c (compute_uninit_opnds_pos): In functions
15724 with nonlocal goto receivers or returns twice calls, ignore
15725 unininitialized values from abnormal edges to nl goto receiver
15726 or returns twice call.
15728 2013-04-26 Jakub Jelinek <jakub@redhat.com>
15730 PR tree-optimization/57051
15731 * fold-const.c (const_binop): Handle VEC_LSHIFT_EXPR
15732 and VEC_RSHIFT_EXPR if shift count is a multiple of element
15735 2013-04-26 Richard Biener <rguenther@suse.de>
15737 * omp-low.c (finalize_task_copyfn): Do not drop PROP_loops.
15738 (expand_omp_taskreg): Likewise. Mark loops for fixup.
15739 * tree-cfg.c (move_block_to_fn): Remap loop fathers.
15740 (fixup_loop_arrays_after_move): New function.
15741 (move_sese_region_to_fn): Properly outline the loop tree parts
15742 of the SESE region.
15744 2013-04-26 Uros Bizjak <ubizjak@gmail.com>
15746 * config/i386/i386.md (type, unit): Fix long lines.
15748 2013-04-26 Richard Biener <rguenther@suse.de>
15750 * Makefile.in (lto-streamer-in.o): Add $(CFGLOOP_H) dependency.
15751 (lto-streamer-out.o): Likewise.
15752 * cfgloop.c (init_loops_structure): Export, add struct function
15753 argument and adjust.
15754 (flow_loops_find): Adjust.
15755 * cfgloop.h (enum loop_estimation): Add EST_LAST.
15756 (init_loops_structure): Declare.
15757 * lto-streamer-in.c: Include cfgloop.h.
15758 (input_cfg): Input the loop tree.
15759 * lto-streamer-out.c: Include cfgloop.h.
15760 (output_cfg): Output the loop tree.
15761 (output_struct_function_base): Do not drop PROP_loops.
15763 2013-03-26 Richard Biener <rguenther@suse.de>
15765 * tree-cfg.c (execute_build_cfg): Build the loop tree.
15766 (pass_build_cfg): Provide PROP_loops.
15767 (move_sese_region_to_fn): Remove loops that are outlined into fn
15769 * tree-inline.c: Include cfgloop.h.
15770 (initialize_cfun): Do not drop PROP_loops.
15771 (copy_loops): New function.
15772 (copy_cfg_body): Copy loop structure.
15773 (tree_function_versioning): Initialize destination loop tree.
15774 * tree-ssa-loop.c (pass_tree_loop_init): Do not provide PROP_loops.
15775 (pass_parallelize_loops): Do IL verification.
15776 * loop-init.c (loop_optimizer_init): Fixup loops if required.
15777 * tree-optimize.c (execute_fixup_cfg): If we need to cleanup
15778 the CFG make sure we fixup loops as well.
15779 * tree-ssa-tail-merge.c: Include cfgloop.h.
15780 (replace_block_by): When merging loop latches mark loops for fixup.
15781 * lto-streamer-out.c (output_struct_function_base): Drop
15782 PROP_loops for now.
15783 * tree-ssa-phiopt.c: Include tree-scalar-evolution.h.
15784 (tree_ssa_cs_elim): Initialize the loop optimizer and SCEV.
15785 * ipa-split.c: Include cfgloop.h.
15786 (split_function): Add the new return block to the loop tree root.
15787 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Return
15788 whether we have removed the forwarder block.
15789 (merge_phi_nodes): If we removed a forwarder mark loops for fixup.
15790 * cfgloop.h (place_new_loop): Declare.
15791 * cfgloopmanip.c (place_new_loop): Export.
15792 * Makefile.in (asan.o): Add $(CFGLOOP_H) dependency.
15793 (tree-switch-conversion.o): Likewise.
15794 (tree-complex.o): Likewise.
15795 (tree-inline.o): Likewise.
15796 (tree-ssa-tailmerge.o): Likewise.
15797 (ipa-split.o): Likewise.
15798 (tree-ssa-phiopt.o): Add $(SCEV_H) dependency.
15799 (tree-ssa-copy.o): Likewise.
15800 * tree-switch-conversion.c: Include cfgloop.h
15801 (process_switch): If we emit a bit-test cascade, schedule loops
15803 * tree-complex.c: Include cfgloop.h.
15804 (expand_complex_div_wide): Properly add new basic-blocks to loops.
15805 * asan.c: Include cfgloop.h.
15806 (create_cond_insert_point): Properly add new basic-blocks to
15807 loops, schedule loop fixup.
15808 * cfgloop.c (verify_loop_structure): Check that looks are not
15810 * omp-low.c (expand_parallel_call): Properly add new basic-blocks
15812 (expand_omp_for_generic): Likewise.
15813 (expand_omp_sections): Likewise.
15814 (expand_omp_atomic_pipeline): Schedule loops for fixup.
15815 * tree-ssa-copy.c: Include tree-scalar-evolution.h.
15816 (fini_copy_prop): Disable DCE in substitute_and_fold if SCEV
15817 is initialized, not when loops are present.
15818 * tree-parloops.c (parallelize_loops): Remove checking here.
15819 * passes.c (init_optimization_passes): Schedule a copy-propagation
15820 pass before complete unrolling of inner loops.
15822 2013-04-26 Jakub Jelinek <jakub@redhat.com>
15824 * Makefile.in (toplev.o): Depend on diagnostic-color.h.
15825 * diagnostic-color.c (should_colorize): Remove _WIN32 version.
15826 (colorize_init): Add argument to _WIN32 version.
15827 * toplev.c: Include diagnostic-color.h.
15828 (process_options): Default to -fdiagnostics-color=auto if
15829 GCC_COLORS env var is in the environment.
15830 * common.opt (fdiagnostics-color=): Add Var and Init.
15831 * doc/invoke.texi (-fdiagnostics-color=): Document that if GCC_COLORS
15832 env var is in the environment, the default is auto rather than never.
15834 * diagnostic.h (file_name_as_prefix): Add context argument.
15835 * diagnostic.c (file_name_as_prefix): Likewise. Colorize
15836 the string as locus.
15837 * langhooks.c (lhd_print_error_function): Adjust caller.
15839 2013-04-25 Lawrence Crowl <crowl@google.com>
15841 * var-tracking.c (shared_hash_def::htab):
15842 Change type to hash_table. Update dependent calls and types.
15844 2013-04-25 Lawrence Crowl <crowl@google.com>
15846 * Makefile.in: Update as needed below.
15848 * alloc-pool.c (static hash_table <alloc_pool_hasher> alloc_pool_hash):
15849 Move declaration to after the type's method definitons.
15851 * attribs.c (htab_t scoped_attributes::attribute_hash):
15852 Change type to hash_table. Update dependent calls and types.
15854 * bitmap.c (htab_t bitmap_desc_hash):
15855 Change type to hash_table. Update dependent calls and types.
15857 * cselib.c (htab_t cselib_hash_table):
15858 Change type to hash_table. Update dependent calls and types.
15860 * data-streamer.h (struct string_slot): Move to lto-streamer.h.
15861 (hash_string_slot_node): Move implementation into lto-streamer.h
15862 struct string_slot_hasher.
15863 (eq_string_slot_node): Likewise.
15865 * data-streamer-out.c: Update output_block::string_hash_table
15866 dependent calls and types.
15868 * dwarf2cfi.c (htab_t trace_index):
15869 Change type to hash_table. Update dependent calls and types.
15871 * dwarf2out.c (htab_t break_out_includes::cu_hash_table):
15872 Change type to hash_table. Update dependent calls and types.
15873 (htab_t copy_decls_for_unworthy_types::decl_table): Likewise.
15874 (htab_t optimize_external_refs::map): Likewise.
15875 (htab_t output_comp_unit::extern_map): Likewise.
15876 (htab_t output_comdat_type_unit::extern_map): Likewise.
15877 (htab_t output_macinfo::macinfo_htab): Likewise.
15878 (htab_t optimize_location_lists::htab): Likewise.
15879 (htab_t dwarf2out_finish::comdat_type_table): Likewise.
15881 * except.c (htab_t ehspec_hash_type):
15882 Change type to hash_table. Update dependent calls and types.
15883 (assign_filter_values::ttypes): Likewise.
15884 (assign_filter_values::ehspec): Likewise.
15885 (sjlj_assign_call_site_values::ar_hash): Likewise.
15886 (convert_to_eh_region_ranges::ar_hash): Likewise.
15888 * gcse.c (htab_t pre_ldst_table):
15889 Change type to hash_table. Update dependent calls and types.
15891 * ggc-common.c (htab_t saving_htab):
15892 Change type to hash_table. Update dependent calls and types.
15893 (htab_t loc_hash): Likewise.
15894 (htab_t ptr_hash): Likewise.
15895 (call_count): Rename ggc_call_count.
15896 (call_alloc): Rename ggc_call_alloc.
15897 (loc_descriptor): Rename make_loc_descriptor.
15898 (add_statistics): Rename ggc_add_statistics.
15900 * ggc-common.c (saving_htab):
15901 Change type to hash_table. Update dependent calls and types.
15903 * gimple.h (struct gimplify_ctx): Move to gimplify-ctx.h.
15904 (push_gimplify_context): Likewise.
15905 (pop_gimplify_context): Likewise.
15906 (struct gimple_temp_hash_elt): Added.
15907 (struct gimplify_hasher): Likewise.
15908 (struct gimplify_ctx.temp_htab):
15909 Change type to hash_table. Update dependent calls and types.
15911 * gimple-fold.c: Include gimplify-ctx.h.
15913 * gimple-ssa-strength-reduction.c (htab_t base_cand_map):
15914 Change type to hash_table. Update dependent calls and types.
15915 (base_cand_dump_callback): Rename to ssa_base_cand_dump_callback to
15916 avoid potential global name collision.
15918 * gimplify.c: Include gimplify-ctx.h.
15919 (struct gimple_temp_hash_elt): Move to gimplify-ctx.h.
15920 (htab_t gimplify_ctx::temp_htab):
15921 Update dependent calls and types for new type hash_table.
15922 (gimple_tree_hash): Move into gimplify_hasher in gimplify-ctx.h.
15923 (gimple_tree_eq): Move into gimplify_hasher in gimplify-ctx.h.
15925 * gimplify-ctx.h: New.
15926 (struct gimple_temp_hash_elt): Move from gimplify.c.
15927 (class gimplify_hasher): New.
15928 (struct gimplify_ctx): Move from gimple.h.
15929 (htab_t gimplify_ctx::temp_htab):
15930 Change type to hash_table. Update dependent calls and types.
15932 * graphite-clast-to-gimple.c: Include graphite-htab.h.
15933 (htab_t ivs_params::newivs_index):
15934 Change type to hash_table. Update dependent calls and types.
15935 (htab_t ivs_params::params_index): Likewise.
15936 (htab_t print_generated_program::params_index): Likewise.
15937 (htab_t gloog::newivs_index): Likewise.
15938 (htab_t gloog::params_index): Likewise.
15940 * graphite.c: Include graphite-htab.h.
15941 4htab_t graphite_transform_loops::bb_pbb_mapping):
15942 Change type to hash_table. Update dependent calls and types.
15944 * graphite-clast-to-gimple.h: (extern gloog) Move to graphite-htab.h.
15945 (bb_pbb_map_hash): Fold into bb_pbb_htab_type in graphite-htab.h.
15946 (eq_bb_pbb_map): Fold into bb_pbb_htab_type in graphite-htab.h.
15948 * graphite-dependences.c: Include graphite-htab.h.
15949 (loop_is_parallel_p): Change hash table type of parameter.
15951 * graphite-htab.h: New.
15952 (typedef hash_table <bb_pbb_hasher> bb_pbb_htab_type): New.
15953 (extern find_pbb_via_hash): Move from graphite-poly.h.
15954 (extern loop_is_parallel_p): Move from graphite-poly.h.
15955 (extern get_loop_body_pbbs): Move from graphite-poly.h.
15957 * graphite-poly.h (extern find_pbb_via_hash): Move to graphite-htab.h.
15958 (extern loop_is_parallel_p): Move to graphite-htab.h.
15959 (extern get_loop_body_pbbs): Move to graphite-htab.h.
15961 * haifa-sched.c (htab_t delay_htab):
15962 Change type to hash_table. Update dependent calls and types.
15963 (htab_t delay_htab_i2): Likewise.
15965 * ira-color.c (htab_t allocno_hard_regs_htab):
15966 Change type to hash_table. Update dependent calls and types.
15968 * ira-costs.c (htab_t cost_classes_htab):
15969 Change type to hash_table. Update dependent calls and types.
15971 * loop-invariant.c (htab_t merge_identical_invariants::eq):
15972 Change type to hash_table. Update dependent calls and types.
15974 * loop-iv.c (htab_t bivs):
15975 Change type to hash_table. Update dependent calls and types.
15977 * loop-unroll.c (htab_t opt_info::insns_to_split):
15978 Change type to hash_table. Update dependent calls and types.
15979 (htab_t opt_info::insns_with_var_to_expand): Likewise.
15981 * lto-streamer.h (struct string_slot): Move from data-streamer.h
15982 (struct string_slot_hasher): New.
15983 (htab_t output_block::string_hash_table):
15984 Change type to hash_table. Update dependent calls and types.
15986 * lto-streamer-in.c (freeing_string_slot_hasher): New.
15987 (htab_t file_name_hash_table):
15988 Change type to hash_table. Update dependent calls and types.
15990 * lto-streamer-out.c: Update output_block::string_hash_table dependent
15993 * lto-streamer.c (htab_t tree_htab):
15994 Change type to hash_table. Update dependent calls and types.
15996 * omp-low.c: Include gimplify-ctx.h.
15998 * passes.c (htab_t name_to_pass_map):
15999 Change type to hash_table. Update dependent calls and types.
16000 (pass_traverse): Rename to passes_pass_traverse.
16002 * plugin.c (htab_t event_tab):
16003 Change type to hash_table. Update dependent calls and types.
16005 * postreload-gcse.c (htab_t expr_table):
16006 Change type to hash_table. Update dependent calls and types.
16007 (dump_hash_table_entry): Rename dump_expr_hash_table_entry.
16009 * sese.c (debug_rename_map_1): Make extern.
16010 (htab_t copy_bb_and_scalar_dependences::rename_map):
16011 Change type to hash_table. Update dependent calls and types.
16013 * sese.h (extern debug_rename_map): Move to .c file.
16015 * store-motion.c (htab_t store_motion_mems_table):
16016 Change type to hash_table. Update dependent calls and types.
16018 * trans-mem.c (htab_t tm_new_mem_hash):
16019 Change type to hash_table. Update dependent calls and types.
16021 * tree-browser.c (htab_t TB_up_ht):
16022 Change type to hash_table. Update dependent calls and types.
16024 * tree-cfg.c (htab_t discriminator_per_locus):
16025 Change type to hash_table. Update dependent calls and types.
16027 * tree-complex.c: Include tree-hasher.h
16028 (htab_t complex_variable_components):
16029 Change type to hash_table. Update dependent calls and types.
16031 * tree-eh.c (htab_t finally_tree):
16032 Change type to hash_table. Update dependent calls and types.
16034 * tree-flow.h (extern int_tree_map_hash): Moved into tree-hasher
16035 struct int_tree_hasher.
16036 (extern int_tree_map_eq): Likewise.
16037 (uid_decl_map_hash): Removed.
16038 (extern decl_tree_map_eq): Likewise.
16040 * tree-hasher.h: New.
16041 (struct int_tree_hasher): New.
16042 (typedef int_tree_htab_type): New.
16044 * tree-inline.c: Include gimplify-ctx.h.
16046 * tree-mudflap.c: Include gimplify-ctx.h.
16048 * tree-parloops.c: Include tree-hasher.h.
16049 (htab_t eliminate_local_variables_stmt::decl_address):
16050 Change type to hash_table. Update dependent calls and types.
16051 (htab_t separate_decls_in_region::decl_copies): Likewise.
16053 * tree-scalar-evolution.c (htab_t resolve_mixers::cache):
16054 Change type to hash_table. Update dependent calls and types.
16056 * tree-sra.c (candidates):
16057 Change type to hash_table. Update dependent calls and types.
16059 * tree-ssa.c (int_tree_map_eq): Moved into struct int_tree_hasher
16061 (int_tree_map_hash): Likewise.
16063 * tree-ssa-dom.c (htab_t avail_exprs):
16064 Change type to hash_table. Update dependent calls and types.
16066 * tree-ssa-live.c (var_map_base_init::tree_to_index):
16067 Change type to hash_table. Update dependent calls and types.
16069 * tree-ssa-loop-ivopts.c (struct ivopts_data.inv_expr_tab):
16070 Change type to hash_table. Update dependent calls and types.
16072 * tree-ssa-phiopt.c (seen_ssa_names):
16073 Change type to hash_table. Update dependent calls and types.
16075 * tree-ssa-strlen.c (decl_to_stridxlist_htab):
16076 Change type to hash_table. Update dependent calls and types.
16078 * tree-ssa-uncprop.c (equiv):
16079 Change type to hash_table. Update dependent calls and types.
16081 2013-04-25 Jakub Jelinek <jakub@redhat.com>
16083 PR rtl-optimization/57003
16084 * regcprop.c (copyprop_hardreg_forward_1): If ksvd.ignore_set_reg,
16085 call note_stores with kill_clobbered_value callback again after
16086 killing regs_invalidated_by_call.
16088 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
16090 * config/aarch64/aarch64-simd.md
16091 (aarch64_simd_bsl<mode>_internal): Rewrite RTL to not use UNSPEC_BSL.
16092 (aarch64_simd_bsl<mode>): Likewise.
16093 * config/aarch64/iterators.md (unspec): Remove UNSPEC_BSL.
16095 2013-04-25 Marek Polacek <polacek@redhat.com>
16097 PR tree-optimization/57066
16098 * builtins.c (fold_builtin_logb): Return +Inf for -Inf.
16100 2013-04-25 James Greenhalgh <jame.greenhalgh@arm.com>
16102 * config/aarch64/aarch64-simd.md (neg<mode>2): Use VDQ iterator.
16104 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
16106 * config/aarch64/aarch64-builtins.c
16107 (aarch64_fold_builtin): New.
16108 * config/aarch64/aarch64-protos.h (aarch64_fold_builtin): New.
16109 * config/aarch64/aarch64.c (TARGET_FOLD_BUILTIN): Define.
16110 * config/aarch64/aarch64-simd-builtins.def (abs): New.
16111 * config/aarch64/arm_neon.h
16112 (vabs<q>_<f32, 64>): Implement using __builtin_aarch64_fabs.
16114 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
16115 Tejas Belagod <tejas.belagod@arm.com>
16117 * config/aarch64/aarch64-builtins.c
16118 (aarch64_gimple_fold_builtin): New.
16119 * config/aarch64/aarch64-protos.h (aarch64_gimple_fold_builtin): New.
16120 * config/aarch64/aarch64-simd-builtins.def (addv): New.
16121 * config/aarch64/aarch64-simd.md (addpv4sf): New.
16122 (addvv4sf): Update.
16123 * config/aarch64/aarch64.c (TARGET_GIMPLE_FOLD_BUILTIN): Define.
16125 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
16127 * config/aarch64/aarch64.md
16128 (*cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): New pattern.
16130 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
16132 * config/aarch64/aarch64.md (*ngc<mode>): New pattern.
16133 (*ngcsi_uxtw): New pattern.
16135 2013-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
16136 Julian Brown <julian@codesourcery.com>
16138 * config/arm/arm.c (neon_builtin_type_mode): Add T_V4HF.
16139 (TB_DREG): Add T_V4HF.
16140 (v4hf_UP): New macro.
16141 (neon_itype): Add NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
16142 (arm_init_neon_builtins): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
16143 Handle initialisation of V4HF. Adjust initialisation of reinterpret
16145 (arm_expand_neon_builtin): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
16146 (arm_vector_mode_supported_p): Handle V4HF.
16147 (arm_mangle_map): Handle V4HFmode.
16148 * config/arm/arm.h (VALID_NEON_DREG_MODE): Add V4HF.
16149 * config/arm/arm_neon_builtins.def: Add entries for
16150 vcvtv4hfv4sf, vcvtv4sfv4hf.
16151 * config/arm/neon.md (neon_vcvtv4sfv4hf): New pattern.
16152 (neon_vcvtv4hfv4sf): Likewise.
16153 * config/arm/neon-gen.ml: Handle half-precision floating point
16155 * config/arm/neon-testgen.ml: Handle Requires_FP_bit feature.
16156 * config/arm/arm_neon.h: Regenerate.
16157 * config/arm/neon.ml (type elts): Add F16.
16158 (type vectype): Add T_float16x4, T_floatHF.
16159 (type vecmode): Add V4HF.
16160 (type features): Add Requires_FP_bit feature.
16161 (elt_width): Handle F16.
16162 (elt_class): Likewise.
16163 (elt_of_class_width): Likewise.
16164 (mode_of_elt): Refactor.
16165 (type_for_elt): Handle F16, fix error messages.
16166 (vectype_size): Handle T_float16x4.
16167 (vcvt_sh): New function.
16168 (ops): Add entries for vcvt_f16_f32, vcvt_f32_f16.
16169 (string_of_vectype): Handle T_floatHF, T_float16, T_float16x4.
16170 (string_of_mode): Handle V4HF.
16171 * doc/arm-neon-intrinsics.texi: Regenerate.
16173 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
16175 * config/aarch64/aarch64.c (aarch64_print_operand): Fix asm_fprintf
16176 format specifier in 'X' case.
16178 2013-04-25 Alan Modra <amodra@gmail.com>
16181 * config/rs6000/rs6000.md (rotlsi3_internal7): Rename to
16182 rotlsi3_internal7le and condition on !BYTES_BIG_ENDIAN.
16183 (rotlsi3_internal8be): New BYTES_BIG_ENDIAN insn.
16184 Repeat for many other rotate/shift and mask patterns using subregs.
16185 Name lshiftrt insns.
16186 (ashrdisi3_noppc64): Rename to ashrdisi3_noppc64be and condition
16187 on WORDS_BIG_ENDIAN.
16189 2013-04-25 Alan Modra <amodra@gmail.com>
16191 * config.gcc: Support little-endian powerpc-linux targets.
16192 * config/rs6000/linux.h (LINK_OS_LINUX_EMUL): Define.
16193 (LINK_OS_LINUX_SPEC): Define.
16194 * config/rs6000/linuxspe.h (TARGET_DEFAULT):
16195 Preserve MASK_LITTLE_ENDIAN.
16196 * config/rs6000/default64.h (TARGET_DEFAULT): Likewise.
16197 * config/rs6000/linuxaltivec.h (TARGET_DEFAULT): Likewise.
16198 * config/rs6000/linux64.h (OPTION_LITTLE_ENDIAN): Don't zero.
16199 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Define.
16200 (LINK_OS_LINUX_SPEC32, LINK_OS_LINUX_SPEC64): Use above.
16201 * config/rs6000/rs6000.c (output_toc): Don't use .tc for TARGET_ELF.
16202 Correct fp word order for little-endian. Don't shift toc entries
16203 smaller than a word for little-endian.
16204 * config/rs6000/rs6000.md (bswaphi2, bswapsi2 split): Comment.
16205 (bswapdi2 splits): Correct low-part subreg for little-endian.
16206 Remove wrong BYTES_BIG_ENDIAN tests, and rename vars to remove
16207 low/high where such is correct only for be.
16208 * config/rs6000/sysv4.h (SUBTARGET_OVERRIDE_OPTIONS): Allow
16209 little-endian for -mcall-aixdesc.
16211 2013-04-25 Alan Modra <amodra@gmail.com>
16213 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
16214 replace_equiv_address_nv.
16216 2013-04-25 Alan Modra <amodra@gmail.com>
16218 * config/rs6000/rs6000.c (rs6000_emit_set_long_const): Tidy.
16220 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
16223 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
16224 * rtl.h (struct rtx_def): ...
16226 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
16228 PR rtl-optimizations/57046
16229 * lra-constraints (split_reg): Set up lra_risky_transformations_p
16230 for multi-reg splits.
16232 2013-04-24 H.J. Lu <hongjiu.lu@intel.com>
16234 * config/i386/x86-64.h (ASM_SPEC): Support -mx32.
16236 2013-04-24 Sterling Augustine <saugustine@google.com>
16238 * dwarf2out.c (skeleton_debug_str_hash, add_skeleton_AT_string)
16239 (comp_dir_string, debug_str_dwo_section): New.
16240 (DEBUG_STR_DWO_SECTION): Rename to ...
16241 (DEBUG_DWO_STR_SECTION): ... this.
16242 (DEBUG_NORM_STR_SECTION): Delete.
16243 (DEBUG_STR_SECTION, DEBUG_STR_SECTION_FLAGS): Edit definitions.
16244 (DEBUG_STR_DWO_SECTION_FLAGS): New.
16245 (find_AT_string): Move most logic to ...
16246 (find_AT_string_in_table): ... here. New.
16247 (add_top_level_skeleton_die_attrs): Call comp_dir_string and
16248 add_skeleton_AT_string. Delete logic.
16249 (output_skeleton_debug_sections): Remove call to
16250 add_top_level_skeleton_die_attrs.
16251 (add_comp_dir_attribute): Move logic to comp_dir_string.
16252 (dwarf2out_init): Initialize debug_str_dwo_section.
16253 (output_indirect_string): Call find_string_form.
16254 (output_indirect_strings): Rewrite.
16255 (prune_unused_types): Empty skeleton_debug_str_hash.
16256 Call get_skeleton_type_unit and add_top_level_skeleton_die_attrs.
16257 (dwarf2out_finish): Call output_indirect_strings.
16259 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
16261 * doc/cpp.texi: Remove __GXX_EXPERIMENTAL_CXX1Y__.
16263 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
16265 * rtl.h (struct rtx_def): Add comment for field jump.
16266 (LRA_SUBREG_P): New macro.
16267 * recog.c (register_operand): Check LRA_SUBREG_P.
16268 * lra.c (lra): Add note at the end of RTL code. Align non-empty
16270 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
16271 (lra_final_code_change): Skip subreg change for operators.
16272 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
16273 if there are no operand changes.
16274 * lra-constraints.c (curr_insn_set): New.
16275 (match_reload): Set LRA_SUBREG_P.
16276 (emit_spill_move): Ditto.
16277 (check_and_process_move): Use curr_insn_set. Process only single
16278 set insns. Don't initialize sec_mem_p and change_p.
16279 (simplify_operand_subreg): Use LRA_SUBREG_P.
16280 (reg_in_class_p): New function.
16281 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
16282 of #ifdef. Add code to remove cycling.
16283 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
16284 non-null disp. Reload inner instead of disp when base and index
16285 are null. Try to put lo_sum into register.
16286 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
16287 (check_and_process_move): Move code for move cost check to
16288 simple_move_p. Remove equiv_substitution.
16289 (simple_move_p): New function.
16290 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
16291 curr_insn_set. Call check_and_process_move only for single set
16292 insns. Use the new function. Move call of check_and_process_move
16293 after operand equiv substitution and address process.
16295 2013-04-24 James Greenhalgh <james.greenhalgh@arm.com>
16297 * config/aarch64/arm_neon.h (vld1<q>_lane*): Fix constraints.
16298 (vld1<q>_dup_<sufp><8, 16, 32, 64>): Likewise.
16299 (vld1<q>_<sufp><8, 16, 32, 64>): Likewise.
16301 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
16303 * doc/cpp.texi: Document __GXX_EXPERIMENTAL_CXX1Y__.
16305 2013-04-24 Marek Polacek <polacek@redhat.com>
16307 * tree-scalar-evolution.h (analyze_scalar_evolution): Remove.
16308 * tree-scalar-evolution.c (get_exit_conditions_rec): Likewise.
16309 (select_loops_exit_conditions): Likewise.
16310 (number_of_iterations_for_all_loops): Likewise.
16311 (analyze_scalar_evolution_for_all_loop_phi_nodes): Likewise.
16312 (scev_analysis): Likewise.
16314 2013-04-02 Catherine Moore <clm@codesourcery.com>
16315 Chao-ying Fu <fu@mips.com>
16317 * config/mips/micromips.md (jraddiusp): New pattern.
16318 * config/mips/mips.c (mips_expand_epilogue): Use the JRADDIUSP
16319 instruction if possible.
16321 2013-04-24 Alan Modra <amodra@gmail.com>
16323 * config/rs6000/driver-rs6000.c (elf_dcachebsize): Fix comment pasto.
16325 2013-04-24 Julian Brown <julian@codesourcery.com>
16326 Chung-Lin Tang <cltang@codesourcery.com>
16328 * dwarf2out.c (gen_enumeration_type_die): Fix HOST_BITS_PER_WIDE_INT
16329 dependency behavior in enumeration type DIE generation. Add TODO note
16330 to comments about future DW_FORM_sdata/udata re-work of related code.
16332 2013-04-23 Lawrence Crowl <crowl@google.com>
16334 * Makefile.in: Update as needed below.
16336 * hash-table.h (class hash_table):
16337 Correct many methods with parameter types compare_type to the correct
16338 value_type. (Correct code was unlikely to notice the change.)
16339 (hash_table::elements_with_deleted) New.
16340 (class hashtable::iterator): New.
16341 (hashtable::begin()): New.
16342 (hashtable::end()): New.
16343 (FOR_EACH_HASH_TABLE_ELEMENT): New.
16345 * statistics.c (statistics_hashes):
16346 Change type to hash_table. Update dependent calls and types.
16348 * tree-into-ssa.c (var_infos):
16349 Change type to hash_table. Update dependent calls and types.
16351 * tree-ssa-coalesce.c (struct coalesce_list_d.list):
16352 Change type to hash_table. Update dependent calls and types.
16354 * tree-ssa-loop-im.c (struct mem_ref.refs):
16355 Change type to hash_table. Update dependent calls and types.
16357 * tree-ssa-reassoc.c (undistribute_ops_list::ctable):
16358 Change type to hash_table. Update dependent calls and types.
16360 * tree-ssa-sccvn.c (vn_tables_s::nary):
16361 Change type to hash_table. Update dependent calls and types.
16362 (vn_tables_s::phis): Likewise.
16363 (vn_tables_s::references): Likewise.
16365 * tree-ssa-sccvn.h (vn_nary_op_eq): Update parameter and return types.
16366 (vn_reference_eq): Update parameter and return types.
16368 * tree-ssa-structalias.c (pointer_equiv_class_table):
16369 Change type to hash_table. Update dependent calls and types.
16370 (location_equiv_class_table): Likewise.
16372 * tree-vect-data-refs.c: Consequential changes for making
16373 peeling a hash_table.
16375 * tree-vect-loop.c (new_loop_vec_info): Dependent hash_table update.
16376 (destroy_loop_vec_info): Dependent hash_table update.
16378 * tree-vectorizer.h (peeling_htab):
16379 Change type to hash_table. Update dependent calls and types.
16381 2013-04-23 Shiva Chen <shiva0217@gmail.com>
16383 * lra-assigns.c (find_hard_regno_for): Use lra_reg_val_equal_p
16384 to check the register content is equal or not.
16385 * lra-constraints.c (match_reload): Use lra_assign_reg_val
16386 to assign register content record.
16387 * lra-eliminations.c (update_reg_eliminate): Use
16388 lra_update_reg_val_offset to update register content offset.
16389 * lra-int.h (struct lra_reg): Add offset member.
16390 (lra_reg_val_equal_p): New static inline function.
16391 (lra_update_reg_val_offset): New static inline function.
16392 (lra_assign_reg_val): New static inline function.
16393 * lra.c (lra_create_new_reg): Use lra_assign_reg_val
16394 to assign register content record.
16395 (initialize_lra_reg_info_element): Initial offset to zero.
16397 2013-04-23 Catherine Moore <clm@codesourcery.com>
16399 * config/mips/mips.md (*movhi_internal, *movqi_internal): New
16400 operands. Record compression.
16402 2013-04-23 Xinliang David Li <davidxl@google.com>
16404 * cfghhooks.c (dump_bb_for_graph): Support 'slim' graph dump.
16406 2013-04-23 Richard Biener <rguenther@suse.de>
16408 PR middle-end/57036
16409 * tree-inline.c (copy_edges_for_bb): Add can_make_abnormal_goto
16410 parameter, only add abnormal goto edges from the copied body
16411 if the call could perform abnormal gotos.
16412 (copy_cfg_body): Adjust.
16414 2013-04-23 Sofiane Naci <sofiane.naci@arm.com>
16416 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add simd attribute.
16418 2013-04-23 Andreas Schwab <schwab@linux-m68k.org>
16420 * coretypes.h (gimple_stmt_iterator): Add struct to make
16423 2013-04-23 Richard Biener <rguenther@suse.de>
16425 PR tree-optimization/57026
16426 * tree-vrp.c (simplify_conversion_using_ranges): Do not propagate
16427 from SSA names occuring in abnormal PHI nodes.
16429 2013-04-22 Andi Kleen <ak@linux.intel.com>
16431 * lto/lto.c (print_lto_report_1): Fix LTO report names.
16433 2013-04-22 Andi Kleen <ak@linux.intel.com>
16435 * lto/lto.c (print_lto_report_1): Declare early.
16436 (read_cgraph_and_symbols): Call print_lto_report_1 early.
16438 2013-04-22 Andi Kleen <ak@linux.intel.com>
16440 * common.opt (-flto-report-wpa): Add.
16441 * doc/invoke.texi (-flto-report-wpa): Add.
16442 * lto/lto.c (do_whole_program_analysis): Check for lto-report-wpa.
16445 2013-04-22 Xinliang David Li <davidxl@google.com>
16447 * graph.c (draw_cfg_node_succ_edges): Add branch probility as label.
16448 * cfghhooks.c (dump_bb_for_graph): Dump profile count and frquency.
16449 * Makefile.in: New dependency
16451 David Daney <ddaney.cavm@gmail.com>
16453 * configure.ac (gcc_cv_as_micromips_support): Use the
16454 --fatal-warnings option.
16455 * configure: Regenerate.
16457 2013-04-22 Marek Polacek <polacek@redhat.com>
16460 * tsan.c (instrument_expr): Don't instrument expression
16461 in case its size is zero.
16463 2013-04-22 Uros Bizjak <ubizjak@gmail.com>
16467 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
16469 * config/alpha/alpha.c (TARGET_LRA_P): New define.
16471 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
16473 * coretypes.h (gimple_stmt_iterator_d): Forward declare.
16474 (gimple_stmt_iterator): New typedef.
16475 * gimple.h (gimple_stmt_iterator): Rename to...
16476 (gimple_stmt_iterator_d): ... This.
16477 * doc/tm.texi.in (TARGET_FOLD_BUILTIN): Detail restriction that
16478 trees be valid for GIMPLE and GENERIC.
16479 (TARGET_GIMPLE_FOLD_BUILTIN): New.
16480 * gimple-fold.c (gimple_fold_call): Call target hook
16481 gimple_fold_builtin.
16482 * hooks.c (hook_bool_gsiptr_false): New.
16483 * hooks.h (hook_bool_gsiptr_false): New.
16484 * target.def (fold_stmt): New.
16485 * doc/tm.texi: Regenerate.
16487 2013-04-22 Vladimir Makarov <vmakarov@redhat.com>
16490 * lra-eliminations.c (mark_not_eliminable): Prevent elimination of
16491 a set sp if no stack realignment.
16493 2013-04-22 Nick Clifton <nickc@redhat.com>
16495 * config.gcc (tilegx-linux): Extend extra_objs rather than
16497 (tilepro-linux): Likewise.
16499 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
16501 * config/aarch64/aarch64-builtins.c
16503 (CF0, CF1, CF2, CF3, CF4, CF10): New.
16504 (VAR<1-12>): Add MAP parameter.
16505 (BUILTIN_*): Likewise.
16506 * config/aarch64/aarch64-simd-builtins.def: Set MAP parameter.
16507 * config/aarch64/aarch64-simd.md (aarch64_sshl_n<mode>): Remove.
16508 (aarch64_ushl_n<mode>): Likewise.
16509 (aarch64_sshr_n<mode>): Likewise.
16510 (aarch64_ushr_n<mode>): Likewise.
16511 (aarch64_<maxmin><mode>): Likewise.
16512 (aarch64_sqrt<mode>): Likewise.
16513 * config/aarch64/arm_neon.h (vshl<q>_n_*): Use new builtin names.
16514 (vshr<q>_n_*): Likewise.
16516 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
16518 * config/aarch64/aarch64-builtins.c
16519 (aarch64_simd_builtin_type_mode): Handle SF types.
16521 (BUILTIN_GPF): Define.
16522 (aarch64_init_simd_builtins): Handle SF types.
16523 * config/aarch64/aarch64-simd-builtins.def (frecpe): Add support.
16524 (frecps): Likewise.
16525 (frecpx): Likewise.
16526 * config/aarch64/aarch64-simd.md
16527 (simd_types): Update simd_frcp<esx> to simd_frecp<esx>.
16528 (aarch64_frecpe<mode>): New.
16529 (aarch64_frecps<mode>): Likewise.
16530 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRECP<ESX>.
16531 (v8type): Add frecp<esx>.
16532 (aarch64_frecp<FRECP:frecp_suffix><mode>): New.
16533 (aarch64_frecps<mode>): Likewise.
16534 * config/aarch64/iterators.md (FRECP): New.
16535 (frecp_suffix): Likewise.
16536 * config/aarch64/arm_neon.h
16537 (vrecp<esx><qsd>_<fd><32, 64>): Convert to using builtins.
16539 2013-04-22 Christian Bruel <christian.bruel@st.com>
16542 * config/sh/sh.h (enum reg_class): Remove DF_HI_REGS.
16543 (REG_CLASS_NAMES): Idem.
16544 (REG_CLASS_CONTENTS): Idem.
16545 (REGCLASS_HAS_FP_REG): Idem.
16546 * config/sh/sh.c (sh_cannot_change_mode_class): Idem.
16547 (sh_conditional_register_usage): Idem.
16549 2013-04-21 Jeff Law <law@redhat.com>
16551 * tree-ssa-forwprop.c (simplify_conversion_from_bitmask): New function.
16552 (ssa_forward_propagate_and_combine): Use it.
16554 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
16556 * lra.c: Update the flow chart diagram.
16558 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
16560 PR rtl-optimization/56847
16561 * lra-constraints.c (process_alt_operands): Discourage alternative
16562 with non-matche doffsettable memory constraint fro memory with
16565 2013-04-19 Richard Biener <rguenther@suse.de>
16567 PR tree-optimization/56982
16568 * builtins.def (BUILT_IN_LONGJMP): longjmp is not a leaf
16570 * gimplify.c (gimplify_call_expr): Notice special calls.
16571 (gimplify_modify_expr): Likewise.
16572 * tree-cfg.c (make_abnormal_goto_edges): Handle setjmp-like
16573 abnormal control flow receivers.
16574 (call_can_make_abnormal_goto): Handle cfun->calls_setjmp
16575 in the same way as cfun->has_nonlocal_labels.
16576 (gimple_purge_dead_abnormal_call_edges): Likewise.
16577 (stmt_starts_bb_p): Make setjmp-like abnormal control flow
16578 receivers start a basic-block.
16580 2013-04-19 Richard Biener <rguenther@suse.de>
16582 * tree-vectorizer.h (struct _slp_instance): Move load_permutation
16584 (struct _slp_tree): ... here. Make it a vector of unsigned ints.
16585 (SLP_INSTANCE_LOAD_PERMUTATION): Remove.
16586 (SLP_TREE_LOAD_PERMUTATION): Add.
16587 (vect_transform_slp_perm_load): Adjust prototype.
16588 * tree-vect-slp.c (vect_free_slp_tree): Adjust.
16589 (vect_free_slp_instance): Likewise.
16590 (vect_create_new_slp_node): Likewise.
16591 (vect_supported_slp_permutation_p): Remove.
16592 (vect_slp_rearrange_stmts): Adjust.
16593 (vect_supported_load_permutation_p): Likewise. Inline
16594 vect_supported_slp_permutation_p here.
16595 (vect_analyze_slp_instance): Compute load permutations per
16596 slp node instead of per instance.
16597 (vect_get_slp_defs): Adjust.
16598 (vect_transform_slp_perm_load): Likewise.
16599 (vect_schedule_slp_instance): Remove redundant code.
16600 (vect_schedule_slp): Remove hack for PR56270, add it ...
16601 * tree-vect-stmts.c (vectorizable_load): ... here, do not
16602 CSE loads for SLP. Adjust.
16604 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
16606 * config/arm/arm.c (load_multiple_sequence, ldm_stm_operation_p): Fix
16607 spelling in two comments.
16609 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
16612 * config/arm/arm.c (load_multiple_sequence): Require SP
16613 as base register for loads if SP is in the register list.
16615 2013-04-19 Martin Jambor <mjambor@suse.cz>
16617 PR tree-optimization/56718
16618 * ipa-cp.c (ipa_value_from_known_type_jfunc): Moved...
16619 * ipa-prop.c (ipa_binfo_from_known_type_jfunc): ...here, renamed
16620 and made public. Adjusted all callers.
16621 (ipa_intraprocedural_devirtualization): New function.
16622 * ipa-prop.h (ipa_binfo_from_known_type_jfunc): Declare.
16623 (ipa_intraprocedural_devirtualization): Likewise.
16624 * Makefile.in (tree-ssa-pre.o): Add ipa-prop.h to dependencies.
16626 2013-04-19 Richard Biener <rguenther@suse.de>
16628 PR tree-optimization/57000
16629 * tree-ssa-reassoc.c (pass_reassoc): Add TODO_update_ssa_only_virtuals.
16631 2013-04-19 Terry Guo <terry.guo@arm.com>
16633 * config/arm/cortex-m4-fpu.md (cortex_m4_v): Delete cpu unit.
16635 (cortex_m4_v_a, cortex_m4_v_b): ... new cpu units.
16636 (cortex_m4_v, cortex_m4_exa_va, cortex_m4_exb_vb): New reservations.
16637 (cortex_m4_fmacs): Use new reservations.
16638 (cortex_m4_f_load, cortex_m4_f_store): Likewise.
16640 2013-04-18 Vladimir Makarov <vmakarov@redhat.com>
16642 PR rtl-optimization/56999
16643 * lra-coalesce.c (coalescable_pseudo_p): Remove 2nd parameter and
16645 (lra_coalesce): Remove split_origin_bitmap and related code.
16646 * lra.c (lra): Coalesce after undoing inheritance. Recreate live
16647 ranges if necessary.
16649 2013-04-18 Uros Bizjak <ubizjak@gmail.com>
16651 * config/i386/i386.c (x86_64_ms_sysv_extra_clobbered_registers):
16653 (ix86_expand_call): Remove clobbered_registers array and use
16654 x86_64_ms_sysv_extra_clobbered_registers instead.
16655 * config/i386/i386.h (x86_64_ms_sysv_extra_clobbered_registers):
16657 * config/i386/predicates.md (call_rex64_ms_sysv_operation): New
16659 * config/i386/i386.md (*call_rex64_ms_sysv): Use
16660 call_rex64_ms_sysv_operation predicate. Remove explicit clobbers.
16661 (*call_value_rex64_ms_sysv): Ditto.
16663 2013-04-18 Cary Coutant <ccoutant@google.com>
16665 * dwarf2out.c (output_pubnames): Check die_perennial_p of
16666 parent instead of die_mark.
16668 2013-04-18 Diego Novillo <dnovillo@google.com>
16670 * gimple.c (create_gimple_tmp): New.
16671 (get_expr_type): New.
16672 (build_assign): New.
16673 (build_type_cast): New.
16674 * gimple.h (enum ssa_mode): Define.
16675 (gimple_seq_set_location): New.
16676 * asan.c (build_check_stmt): Change some gimple_build_* calls
16677 to use build_assign and build_type_cast.
16679 2013-04-18 Richard Biener <rguenther@suse.de>
16681 * tree-vect-data-refs.c (vect_analyze_group_access): Properly
16682 handle negative step. Remove redundant checks.
16683 (vect_create_data_ref_ptr): Avoid ICEs with non-constant steps.
16684 * tree-vect-stmts.c (vectorizable_load): Instead of asserting
16685 for negative step and grouped loads fail to vectorize.
16687 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
16689 * emit-rtl.c (reset_insn_used_flags): New function.
16690 (reset_all_used_flags): Use it.
16691 (verify_insn_sharing): New function.
16692 (verify_rtl_sharing): Fix verification for SEQUENCEs.
16694 2013-04-18 Jakub Jelinek <jakub@redhat.com>
16696 PR tree-optimization/56984
16697 * tree-vrp.c (register_edge_assert_for_2): For (x >> M) < N
16698 and (x >> M) >= N don't register any assertion if N << M is the
16701 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
16703 * lower-subreg.c (resolve_simple_move): If called self-recursive,
16704 do not delete_insn insns that have not yet been emitted, only
16705 unlink them with remove_insn.
16706 * df-scan.c (df_insn_delete): Revert r197492.
16708 2013-04-17 Steven Bosscher <steven@gcc.gnu.org>
16710 * emit-rtl.c (link_insn_into_chain): Handle chaining of SEQUENCEs.
16711 * reorg.c (emit_delay_sequence): Simplify with emit-rtl API.
16713 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
16715 * config/arm/arm.md (movsicc_insn): Convert define_insn into
16716 define_insn_and_split.
16717 (and_scc,ior_scc,negscc): Likewise.
16718 (cmpsi2_addneg, subsi3_compare): Convert to named patterns.
16720 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
16722 * config/arm/arm.c (use_return_insn): Return 0 for targets that
16723 can benefit from using a sequence of LDRD instructions in epilogue
16724 instead of a single LDM instruction.
16726 2013-04-17 Manuel López-Ibáñez <manu@gcc.gnu.org>
16729 * doc/extend.texi: Fix typo.
16731 2013-04-17 Richard Biener <rguenther@suse.de>
16733 * tree-vect-slp.c (vect_build_slp_tree_1): Split out from ...
16734 (vect_build_slp_tree): ... here.
16735 (vect_build_slp_tree_1): Compute which stmts of the SLP group
16736 match. Remove special-casing of mismatched complex loads.
16737 (vect_build_slp_tree): Based on the result from vect_build_slp_tree_1
16738 re-try the match with swapped commutative operands.
16739 (vect_supported_load_permutation_p): Remove special-casing of
16740 mismatched complex loads.
16741 (vect_analyze_slp_instance): Adjust.
16743 2013-04-17 Richard Biener <rguenther@suse.de>
16745 PR rtl-optimization/56921
16746 * cfgloop.h (struct loop): Add simple_loop_desc member.
16747 (struct niter_desc): Mark with GTY(()).
16748 (simple_loop_desc): Do not use aux field but simple_loop_desc.
16749 * loop-iv.c (get_simple_loop_desc): Likewise.
16750 (free_simple_loop_desc): Likewise.
16753 2013-04-16 Richard Biener <rguenther@suse.de>
16755 PR rtl-optimization/56921
16756 * loop-init.c (pass_rtl_move_loop_invariants): Add
16757 TODO_do_not_ggc_collect to todo_flags_finish.
16758 (pass_rtl_unswitch): Same.
16759 (pass_rtl_unroll_and_peel_loops): Same.
16760 (pass_rtl_doloop): Same.
16762 2013-04-17 Eric Botcazou <ebotcazou@adacore.com>
16764 * tree-ssa-alias.c (nonoverlapping_component_refs_of_decl_p): New.
16765 (decl_refs_may_alias_p): Add REF1 and REF2 parameters.
16766 Use nonoverlapping_component_refs_of_decl_p to disambiguate component
16768 (refs_may_alias_p_1): Adjust call to decl_refs_may_alias_p.
16769 * tree-streamer.c (record_common_node): Adjust reference in comment.
16771 2013-04-17 Terry Guo <terry.guo@arm.com>
16773 * config/arm/cortex-m4.md: Add a new bypass.
16775 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
16777 * config/aarch64/aarch64.md (*adds_<optab><mode>_multp2):
16779 (*subs_<optab><mode>_multp2): New pattern.
16780 (*adds_<optab><ALLX:mode>_<GPI:mode>): New pattern.
16781 (*subs_<optab><ALLX:mode>_<GPI:mode>): New pattern.
16783 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
16785 * config/aarch64/aarch64.md (*adds_mul_imm_<mode>): New pattern.
16786 (*subs_mul_imm_<mode>): New pattern.
16788 2013-04-16 David Edelsohn <dje.gcc@gmail.com>
16791 * config/rs6000/vsx.md (vsx_mov<mode>): Add j->r alternative.
16792 (vsx_movti_64bit): Change j->wa to O->wa. Add n->r alternative.
16793 (vsx_movti_32bit): Change j->wa to O->wa.
16795 2013-04-16 Richard Biener <rguenther@suse.de>
16797 PR rtl-optimization/56921
16798 * loop-init.c (pass_rtl_move_loop_invariants): Add
16799 TODO_do_not_ggc_collect to todo_flags_finish.
16800 (pass_rtl_unswitch): Same.
16801 (pass_rtl_unroll_and_peel_loops): Same.
16802 (pass_rtl_doloop): Same.
16804 2013-04-16 Greta Yorsh <Greta.Yorsh@arm.com>
16806 * config/arm/arm.c (emit_multi_reg_push): New declaration
16807 for an existing function.
16808 (arm_emit_strd_push): New function.
16809 (arm_expand_prologue): Used here.
16810 (arm_emit_ldrd_pop): New function.
16811 (arm_expand_epilogue): Used here.
16812 (arm_get_frame_offsets): Update condition.
16813 (arm_emit_multi_reg_pop): Add a special case for load of a single
16814 register with writeback.
16816 2013-04-16 Uros Bizjak <ubizjak@gmail.com>
16818 * doc/invoke.texi (i386 Option): Reword -mstack-protector-guard
16821 2013-04-16 Richard Biener <rguenther@suse.de>
16823 PR tree-optimization/56756
16824 * tree-ssa-loop-im.c (struct first_mem_ref_loc_1): New functor.
16825 (first_mem_ref_loc): New.
16826 (execute_sm): Place the load temporarily before a previous
16827 access instead of in the latch edge to ensure its SSA dependencies
16828 are defined at points dominating the load.
16830 2013-04-16 Steven Bosscher <steven@gcc.gnu.org>
16832 * cfgrtl.c (cfg_layout_merge_blocks): Revert r184005, implement
16833 correct fix by moving header and footer insn to the footer of
16834 the merged basic block. Clear BB_END of the merged-away block.
16836 PR middle-end/43631
16837 * emit-rtl.c (make_note_raw): New function.
16838 (link_insn_into_chain): New static inline function.
16839 (add_insn): Use it.
16840 (add_insn_before, add_insn_after): Factor insn chain linking code...
16841 (add_insn_before_nobb, add_insn_after_nobb): ...here, new functions
16842 using link_insn_into_chain.
16843 (note_outside_basic_block_p): New helper function for emit_note_after
16844 and emit_note_before.
16845 (emit_note_after): Use nobb variant of add_insn_after if the note
16846 should not be contained in a basic block.
16847 (emit_note_before): Use nobb variant of add_insn_before if the note
16848 should not be contained in a basic block.
16849 (emit_note_copy): Use make_note_raw.
16850 (emit_note): Likewise.
16851 * bb-reorder.c (insert_section_boundary_note): Remove hack to set
16852 BLOCK_FOR_INSN to NULL manually for NOTE_INSN_SWITCH_TEXT_SECTIONS.
16853 * jump.c (cleanup_barriers): Use reorder_insns_nobb to avoid making
16854 the moved barrier the tail of the basic block it follows.
16855 * var-tracking.c (pass_variable_tracking): Add TODO_verify_flow.
16857 2013-04-15 Jakub Jelinek <jakub@redhat.com>
16859 PR tree-optimization/56962
16860 * gimple-ssa-strength-reduction.c (record_increment): Only set
16861 initializer if gimple_assign_rhs_code is {,POINTER_}PLUS_EXPR and
16862 either rhs1 or rhs2 is equal to c->base_expr.
16864 2013-04-15 Richard Biener <rguenther@suse.de>
16866 PR tree-optimization/56933
16867 * tree-vectorizer.h (struct _stmt_vec_info): Remove read_write_dep
16869 (GROUP_READ_WRITE_DEPENDENCE): Remove.
16870 (STMT_VINFO_GROUP_READ_WRITE_DEPENDENCE): Likewise.
16871 * tree-vect-data-refs.c (vect_analyze_group_access): Move
16872 dependence check ...
16873 vect_analyze_data_ref_dependence (vect_analyze_data_ref_dependence):
16875 * tree-vect-stmts.c (new_stmt_vec_info): Do not initialize
16876 GROUP_READ_WRITE_DEPENDENCE.
16878 2013-04-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
16880 * emit-rtl.c (reset_all_used_flags): New function.
16881 (verify_rtl_sharing): Call reset_all_used_flags before and after
16882 performing the checks.
16884 2013-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
16886 * config/arm/arm.c (const_ok_for_dimode_op): Handle AND case.
16887 * config/arm/arm.md (*anddi3_insn): Change to insn_and_split.
16888 * config/arm/constraints.md (De): New constraint.
16889 * config/arm/neon.md (anddi3_neon): Delete.
16890 (neon_vand<mode>): Expand to standard anddi3 pattern.
16891 * config/arm/predicates.md (imm_for_neon_inv_logic_operand):
16892 Move earlier in the file.
16893 (neon_inv_logic_op2): Likewise.
16894 (arm_anddi_operand_neon): New predicate.
16896 2013-04-15 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
16898 * configure.ac (gcc_cv_ld_as_needed): Set
16899 gcc_cv_ld_as_needed_option, gcc_cv_no_as_needed_option.
16900 Use -z ignore, -z record on *-*-solaris2*.
16901 (HAVE_LD_AS_NEEDED): Update comment.
16902 (LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION): Define.
16903 * configure: Regenerate.
16904 * config.in: Regenerate.
16905 * gcc.c (init_gcc_specs) [USE_LD_AS_NEEDED]: Use
16906 LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION.
16907 * config/sol2.h [HAVE_LD_AS_NEEDED] (USE_LD_AS_NEEDED): Define.
16908 * doc/tm.texi.in (USE_LD_AS_NEEDED): Allow for --as-needed
16909 equivalents. Fix markup.
16910 * doc/tm.texi: Regenerate.
16912 2013-04-15 Andrew Hsieh <andrewhsieh.google.com>
16914 * config/i386/i386.opt: New option mstack-protector-guard=.
16915 * config/i386/i386-opts.h: Add enum stack_protector_guard.
16916 * config/i386/i386.h: Define TARGET_SSP_GLOBAL_GUARD and
16917 TARGET_SSP_TLS_GUARD.
16918 * config/i386/i386.c (ix86_option_override_internal): Set
16919 ix86_stack_protector_guard.
16920 * config/i386/i386.md (stack_protect_set): Enable for
16921 TARGET_SSP_TLS_GUARD only.
16922 (stack_protect_set_<mode>): Ditto.
16923 (stack_protect_test): Ditto.
16924 (stack_protect_test_<mode>): Ditto.
16925 * doc/invoke.texi (i386 Option): Document.
16927 2013-04-15 Eric Botcazou <ebotcazou@adacore.com>
16930 * config/sparc/sparc.c (enum sparc_mode_class): Add H_MODE value.
16931 (S_MODES): Set H_MODE bit.
16932 (SF_MODES): Set only S_MODE and SF_MODE bits.
16933 (DF_MODES): Set SF_MODES and only D_MODE and DF_MODE bits.
16934 (sparc_init_modes) <MODE_INT>: Set H_MODE bit for sub-word modes.
16935 <MODE_VECTOR_INT>: Do not set SF_MODE for sub-word modes.
16936 <MODE_FLOAT>: Likewise.
16938 2013-04-15 Joey Ye <joey.ye@arm.com>
16940 * config/arm/arm.c (thumb_far_jump_used_p): Fix typo in comments.
16942 2013-04-15 Joey Ye <joey.ye@arm.com>
16944 * config/arm/arm.c (thumb1_final_prescan_insn): Assert lr save
16946 (thumb_far_jump_used_p): Count instruction size and set
16949 2013-04-14 Eric Botcazou <ebotcazou@adacore.com>
16951 * reorg.c (fill_simple_delay_slots): Reindent block of code.
16952 * resource.c (mark_target_live_regs): Reformat conditional block.
16954 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
16956 * sched-deps.c (deps_analyze_insn): Do not check for EH_REGION insn
16957 notes, they are emitted only just before final.
16958 * sched-int.h: Include insn-attr.h before checking INSN_SCHEDULING.
16960 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
16962 * emit-rtl.c (remove_insn): Do not call df_insn_delete here.
16963 * cfgrtl.c (delete_insn): Call it here instead.
16964 * lra-spills.c (lra_final_code_change): Use delete_insn.
16965 * haifa-sched.c (sched_remove_insn): Likewise.
16966 * sel-sched-ir.c (return_nop_to_pool): Clear INSN_DELETED_P for nops
16967 returning to the nop pool.
16968 (sel_remove_insn): Simplify the only_disconnect case via remove_insn,
16969 use delete_insn for definitive removal. Clear BLOCK_FOR_INSN.
16971 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
16973 * doc/tm.texi.in (LOOP_ALIGN): Remove loop note references.
16974 * doc/tm.texi: Regenerated.
16976 2013-04-12 Uros Bizjak <ubizjak@gmail.com>
16978 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use ANY_QI_REGNO_P in
16981 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
16983 * df-core.c (df_find_def): Compare register numbers.
16984 (df_find_use): Likewise.
16986 2013-04-12 Vladimir Makarov <vmakarov@redhat.com>
16989 * config/i386/i386.c (ix86_hard_regno_mode_ok): Add
16990 lra_in_progress for return.
16992 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
16994 * config/arm/arm.md (mov_scc,mov_negscc,mov_notscc): Convert
16995 define_insn into define_insn_and_split and emit movsicc patterns.
16997 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
16999 * config/arm/arm.c (gen_operands_ldrd_strd): Initialize "base".
17001 2013-04-12 Richard Biener <rguenther@suse.de>
17003 * tree-pass.h (TODO_do_not_ggc_collect): New.
17004 * passes.c (execute_one_ipa_transform_pass): Honor
17005 TODO_do_not_ggc_collect.
17006 (execute_one_pass): Likewise.
17009 2013-04-10 Richard Biener <rguenther@suse.de>
17011 * passes.c (init_optimization_passes): Remove reload pass.
17012 * ira.c (do_reload): Merge into ...
17014 (rest_of_handle_reload): Remove.
17015 (pass_reload): Likewise.
17016 * config/i386/i386.c (ix86_option_override): Refer to ira instead
17017 of reload for vzeroupper pass placement.
17019 2013-04-12 Jakub Jelinek <jakub@redhat.com>
17021 PR tree-optimization/56918
17022 PR tree-optimization/56920
17023 * fold-const.c (int_const_binop_1): Use op1.mul_with_sign (op2, ...)
17024 instead of op1 - op2. Pass 2 * TYPE_PRECISION (type) as second
17025 argument to rshift method. For 2 * HOST_BITS_PER_WIDE_INT precision
17026 use wide_mul_with_sign method.
17028 2013-04-12 Richard Biener <rguenther@suse.de>
17030 * gimple.c (is_gimple_constant): Vector CONSTRUCTORs should
17031 not be considered a gimple constant.
17033 2013-04-12 Marc Glisse <marc.glisse@inria.fr>
17035 * fold-const.c (const_binop): Handle vector shifts by a scalar.
17036 (fold_binary_loc): Call const_binop also for mixed vector-scalar
17039 2013-04-12 Manuel López-Ibáñez <manu@gcc.gnu.org>
17040 Jakub Jelinek <jakub@redhat.com>
17042 * opts.c: Include diagnostic-color.h.
17043 (common_handle_option): Handle OPT_fdiagnostics_color_.
17044 * Makefile.in (OBJS-libcommon): Add diagnostic-color.o.
17045 (diagnostic.o, opts.o, pretty-print.o): Depend on diagnostic-color.h.
17046 (diagnostic-color.o): New.
17047 * common.opt (fdiagnostics-color, fdiagnostics-color=): New options.
17048 (diagnostic_color_rule): New enum.
17049 * dwarf2out.c (gen_producer_string): Don't print -fdiagnostics-color*.
17050 * langhooks.c (lhd_print_error_function): Add %r "locus" and %R around
17051 the location string.
17052 * diagnostic.def: Add 3rd argument to DEFINE_DIAGNOSTIC_KIND macros,
17053 either NULL, or color kind.
17054 * diagnostic-color.c: New file.
17055 * diagnostic-color.h: New file.
17056 * diagnostic-core.h (DEFINE_DIAGNOSTIC_KIND): Adjust macro for 3
17058 * doc/invoke.texi (-fdiagnostics-color): Document.
17059 * pretty-print.h (pp_show_color): Define.
17060 (struct pretty_print_info): Add show_color field.
17061 * diagnostic.c: Include diagnostic-color.h.
17062 (diagnostic_build_prefix): Adjust for 3 argument DEFINE_DIAGNOSTIC_KIND
17063 macros. Colorize error:, warning: etc. strings and also the location
17065 (diagnostic_show_locus): Colorize the caret line.
17066 * pretty-print.c: Include diagnostic-color.h.
17067 (pp_base_format): Handle %r and %R format specifiers. Colorize strings
17068 inside of %< %> quotes or quoted through q format modifier.
17070 2013-04-12 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
17072 * ifcvt.c (end_ifcvt_sequence): Mark a and b for unsharing as well.
17074 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
17076 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Allow NEG
17077 code in CC_NZ mode.
17078 * config/aarch64/aarch64.md (*neg_<shift><mode>3_compare0): New
17081 2013-04-11 Marek Polacek <polacek@redhat.com>
17083 PR tree-optimization/48184
17084 * params.def (PARAM_ALIGN_THRESHOLD): Increase the minimum value to 1.
17086 2013-04-11 Eric Botcazou <ebotcazou@adacore.com>
17088 * stor-layout.c (skip_simple_constant_arithmetic): Move to...
17089 * tree.c (skip_simple_constant_arithmetic): ...here and make public.
17090 (skip_simple_arithmetic): Tidy up.
17091 * tree.h (skip_simple_constant_arithmetic): Declare.
17093 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
17095 * config/aarch64/aarch64.h (REVERSIBLE_CC_MODE): Define.
17097 2013-04-11 Richard Biener <rguenther@suse.de>
17099 * tree-vect-loop.c (get_initial_def_for_induction): Properly
17100 generate vector constants.
17102 2013-04-11 Richard Biener <rguenther@suse.de>
17104 PR tree-optimization/56878
17105 * tree-flow.h (outermost_invariant_loop_for_expr): Declare.
17106 * tree-ssa-loop-ivopts.c (outermost_invariant_loop_for_expr):
17108 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment):
17109 Prefer to align the DR with the most invariant base address.
17111 2013-04-11 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
17113 * opts.c (common_handle_option): Fix formatting and add FALLTHRU
17116 2013-04-11 James Greenhalgh <james.greenhalgh@arm.com>
17118 * config/aarch64/aarch64-simd.md (aarch64_vcond_internal): Fix
17119 floating-point vector comparisons against 0.
17121 2013-04-11 Jakub Jelinek <jakub@redhat.com>
17123 PR tree-optimization/56899
17124 * fold-const.c (extract_muldiv_1): Apply distributive law
17125 only if TYPE_OVERFLOW_WRAPS (ctype).
17127 2013-04-11 Bin Cheng <bin.cheng@arm.com>
17130 * ira-costs.c (scan_one_insn): Check whether the source rtx of
17131 loading has side effect.
17133 2013-04-10 Steven Bosscher <steven@gcc.gnu.org>
17135 * config/sparc/sparc.c: Include tree-pass.h.
17136 (TARGET_MACHINE_DEPENDENT_REORG): Do not redefine.
17137 (sparc_reorg): Rename to sparc_do_work_around_errata. Move to
17138 head of file. Change return type. Split off gate function.
17139 (sparc_gate_work_around_errata): New function.
17140 (pass_work_around_errata): New pass definition.
17141 (insert_pass_work_around_errata) New pass insert definition to
17142 insert pass_work_around_errata just after delayed-branch scheduling.
17143 (sparc_option_override): Insert the pass.
17144 * config/sparc/t-sparc (sparc.o): Add TREE_PASS_H dependence.
17146 2013-04-10 David S. Miller <davem@davemloft.net>
17148 * config/sparc/sparc.h (ASM_CPU_SPEC): Pass -Av8 if -mcpu=supersparc
17149 or -mcpu=hypersparc.
17151 * target.def (cstore_mode): New hook.
17152 * target.h: Include insn-codes.h
17153 * targhooks.c: Likewise.
17154 (default_cstore_mode): New function.
17155 * targhooks.h: Declare it.
17156 * doc/tm.texi.in: New hook slot for TARGET_CSTORE_MODE.
17157 * doc/tm.texi: Rebuild.
17158 * expmed.c (emit_cstore): Obtain cstore boolean result mode using
17159 target hook, rather than inspecting the insn_data.
17160 * config/sparc/sparc.c (sparc_cstore_mode): New function.
17161 (TARGET_CSTORE_MODE): Redefine.
17162 (emit_scc_insn): When TARGET_ARCH64, emit new 64-bit boolean
17164 * config/sparc/predicates.md (cstore_result_operand): New special
17166 * config/sparc/sparc.md (cstoresi4, cstoredi4, cstore<F:mode>4):
17167 Use it for operand 0.
17168 (*seqsi_special): Rewrite using 'P' mode iterator on operand 0.
17169 (*snesi_special): Likewise.
17170 (*snesi_zero): Likewise.
17171 (*seqsi_zero): Likewise.
17172 (*sltu_insn): Likewise.
17173 (*sgeu_insn): Likewise.
17174 (*seqdi_special): Make operand 0 and comparison operation be of
17176 (*snedi_special): Likewise.
17177 (*snedi_special_vis3): Likewise.
17178 (*neg_snesi_zero): Rename to *neg_snesisi_zero.
17179 (*neg_snesi_sign_extend): Rename to *neg_snesidi_zero.
17180 (*snesi_zero_extend): Delete, covered by 'P' mode iterator.
17181 (*neg_seqsi_zero): Rename to *neg_seqsisi_zero.
17182 (*neg_seqsi_sign_extend): Rename to *neg_seqsidi_zero.
17183 (*seqsi_zero_extend): Delete, covered by 'P' mode iterator.
17184 (*sltu_extend_sp64): Likewise.
17185 (*neg_sltu_insn): Rename to *neg_sltusi_insn.
17186 (*neg_sltu_extend_sp64): Rename to *neg_sltudi_insn.
17187 (*sgeu_extend_sp64): Delete, covered by 'P' mode iterator.
17188 (*neg_sgeu_insn): Rename to *neg_sgeusi_insn.
17189 (*neg_sgeu_extend_sp64): Rename to *neg_sgeudi_insn.
17191 2013-04-10 Yufeng Zhang <yufeng.zhang@arm.com>
17193 * config/aarch64/aarch64.c (aarch64_print_extension): New function.
17194 (aarch64_start_file): Use the new function.
17196 2013-04-10 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
17197 Jason Merrill <jason@redhat.com>
17199 * common.opt: Add -gdwarf.
17200 * opts.c (common_handle_option): Handle it.
17201 * gcc.c (ASM_DEBUG_SPEC): Don't expect "-2" for DWARF.
17203 2013-04-10 Richard Biener <rguenther@suse.de>
17205 * passes.c (execute_todo): Do not call ggc_collect conditional here.
17206 (execute_one_ipa_transform_pass): But unconditionally here.
17207 (execute_one_pass): And here.
17208 (init_optimization_passes): Remove reload pass.
17209 * tree-pass.h (TODO_ggc_collect): Remove.
17210 (pass_reload): Likewise.
17211 * ira.c (do_reload): Merge into ...
17213 (rest_of_handle_reload): Remove.
17214 (pass_reload): Likewise.
17215 * config/i386/i386.c (ix86_option_override): Refer to ira instead
17216 of reload for vzeroupper pass placement.
17217 * <everywhere>: Remove TODO_ggc_collect from todo_flags_start
17218 and todo_flags_finish of all passes.
17220 2013-04-10 Richard Biener <rguenther@suse.de>
17222 * tree-vectorizer.h (struct _slp_oprnd_info): Remove
17223 first_const_oprnd field, rename first_def_type to first_op_type.
17224 * tree-vect-slp.c (vect_create_oprnd_info): Adjust.
17225 (vect_get_and_check_slp_defs): Always use the type of the
17226 operand. Allow mixed vect_external_def, vect_constant_def types.
17227 (vect_get_constant_vectors): Handle mixed vect_external_def,
17228 vect_constant_def types.
17230 2013-04-10 Joern Rennecke <joern.rennecke@embecosm.com>
17232 PR tree-optimization/55524
17233 * tree-ssa-math-opts.c
17234 (convert_mult_to_fma): Don't use an fms construct
17235 when we don't have an fms operation, but fnma, and it looks
17236 likely that we'll be able to use the latter.
17238 2013-04-10 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
17240 * cif-code.def (OVERWRITABLE): Correct the comment for overwritable
17242 * ipa-inline.c (can_inline_edge_p): Let dump mechanism report the
17243 inline fail caused by overwritable functions.
17245 2013-04-10 Chung-Ju Wu <jasonwucj@gmail.com>
17247 * combine.c (simplify_compare_const): Use GET_MODE_MASK to filter out
17248 unnecessary bits in the constant power of two case.
17250 2013-04-10 Richard Biener <rguenther@suse.de>
17252 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove
17253 broken code swapping operands.
17254 (vect_build_slp_tree): Do not compute load permutations here.
17255 (vect_analyze_slp_instance): Compute load permutations here,
17256 after building the SLP tree.
17258 2013-04-09 Christian Bruel <christian.bruel@st.com>
17260 * config/sh/sh.md (barrier_align): Use next/prev_active_insn instead
17261 of next/prev_real_insn.
17263 2013-04-09 Jan Hubicka <jh@suse.cz>
17265 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
17266 Drop aliased parameter.
17267 (function_and_variable_visibility): Do not handle alias pairs.
17268 * cgraph.c (varpool_externally_visible_p): Update prototype.
17269 * varpool.c (varpool_add_new_variable): Update.
17271 2013-04-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17273 * config/arm/arm.md (minmax_arithsi_non_canon): New pattern.
17275 2013-04-09 Steven Bosscher <steven@gcc.gnu.org>
17277 * sched-vis.c (print_pattern): Print SEQUENCE of insns as insns.
17279 * config/sparc/sparc.md: Use define_c_enum for "unspec" and "unspecv".
17281 2013-04-09 Marek Polacek <polacek@redhat.com>
17283 PR tree-optimization/48762
17284 * params.def (PARAM_MAX_CSE_INSNS): Increase the minimum value to 1.
17286 2013-04-09 Richard Biener <rguenther@suse.de>
17288 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove code
17290 (vect_build_slp_tree): Likewise.
17291 (vect_analyze_slp_cost_1, vect_analyze_slp_cost): New functions
17292 calculating the cost of a SLP instance.
17293 (vect_analyze_slp_instance): Use it from here, after building
17296 2013-04-09 Jakub Jelinek <jakub@redhat.com>
17298 PR middle-end/56883
17299 * omp-low.c (expand_omp_for_generic, expand_omp_for_static_nochunk,
17300 expand_omp_for_static_chunk): Use simple_p = true in
17301 force_gimple_operand_gsi calls when assigning to addressable decls.
17303 2013-04-09 Jeff Law <law@redhat.com>
17305 * tree-vrp.c (simplify_cond_using_ranges): Simplify test of boolean
17306 when the boolean was created by converting a wider object which
17307 had a boolean range.
17309 2013-04-09 Richard Biener <rguenther@suse.de>
17311 * tree-vectorizer.h (slp_void_p): Remove.
17312 (slp_tree): Typedef before _slp_tree declaration.
17313 (struct _slp_tree): Use a vector of slp_tree as children.
17314 (vect_get_place_in_interleaving_chain): Remove.
17315 * tree-vect-data-refs.c (vect_get_place_in_interleaving_chain):
17317 * tree-vect-slp.c (vect_get_place_in_interleaving_chain): ... here
17319 (vect_free_slp_tree, vect_print_slp_tree, vect_mark_slp_stmts,
17320 vect_mark_slp_stmts_relevant, vect_slp_rearrange_stmts,
17321 vect_detect_hybrid_slp_stmts, vect_slp_analyze_node_operations,
17322 vect_schedule_slp_instance, vect_remove_slp_scalar_calls):
17323 Use slp_node instead of slp_void_p and adjust.
17325 2013-04-09 Richard Biener <rguenther@suse.de>
17327 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Avoid
17328 work that is not necessary.
17330 2013-04-09 Jakub Jelinek <jakub@redhat.com>
17332 PR tree-optimization/56854
17333 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Don't
17334 forward into clobber stmts if it would change MEM_REF lhs into
17337 2013-04-09 Maxim Kuvyrkov <maxim@kugelworks.com>
17339 * tree.c (type_hash_lookup, type_hash_add): Make static.
17340 * tree.h (type_hash_lookup, type_hash_add): Remove global declarations.
17342 2013-04-09 Richard Biener <rguenther@suse.de>
17344 * tree.h (unsave_expr_now): Remove.
17345 * tree-inline.c (mark_local_for_remap_r): Remove.
17346 (unsave_expr_1): Likewise.
17347 (unsave_r): Likewise.
17348 (unsave_expr_now): Likewise.
17349 * tree-ssa-copy.c (replace_exp_1): Use unshare_expr.
17350 (propagate_tree_value): Likewise.
17352 2013-04-08 Steven Bosscher <steven@gcc.gnu.org>
17354 * doc/rtl.texi (sequence): Rewrite documentation to match the
17355 current use of SEQUENCE rtl objects.
17356 * rtl.def (SEQUENCE): Likewise.
17358 * doc/rtl.texi (NOTE_INSN_EH_REGION_BEG, NOTE_INSN_EH_REGION_END):
17359 Update documentation.
17360 (NOTE_INSN_LOOP_BEG, NOTE_INSN_LOOP_END, NOTE_INSN_LOOP_CONT,
17361 NOTE_INSN_LOOP_VTOP): Remove documentation for non-existing notes.
17363 * reg-notes.def (REG_EH_CONTEXT): Remove unused note.
17365 2013-04-08 Teresa Johnson <tejohnson@google.com>
17367 * basic-block.h (GCOV_COMPUTE_SCALE): Define.
17368 * ipa-inline-analysis.c (param_change_prob): Use helper rounding divide
17370 (estimate_edge_size_and_time): Add comment to suggest using rounding
17372 (estimate_node_size_and_time): Ditto.
17373 (remap_edge_change_prob): Use helper rounding divide methods.
17374 * value-prof.c (gimple_divmod_fixed_value_transform): Ditto.
17375 (gimple_mod_pow2_value_transform): Ditto.
17376 (gimple_mod_subtract_transform): Ditto.
17377 (gimple_ic_transform): Ditto.
17378 (gimple_stringops_transform): Ditto.
17379 * stmt.c (conditional_probability): Ditto.
17380 (emit_case_dispatch_table): Ditto.
17381 * lto-cgraph.c (merge_profile_summaries): Ditto.
17382 * tree-optimize.c (execute_fixup_cfg): Ditto.
17383 * cfgcleanup.c (try_forward_edges): Ditto.
17384 * cfgloopmanip.c (scale_loop_profile): Ditto.
17386 (duplicate_loop_to_header_edge): Ditto.
17387 (lv_adjust_loop_entry_edge): Ditto.
17388 * tree-vect-loop.c (vect_transform_loop): Ditto.
17389 * profile.c (compute_branch_probabilities): Ditto.
17390 * cfgbuild.c (compute_outgoing_frequencies): Ditto.
17391 * lto-streamer-in.c (input_cfg): Ditto.
17392 * gimple-streamer-in.c (input_bb): Ditto.
17393 * ipa-cp.c (update_profiling_info): Ditto.
17394 (update_specialized_profile): Ditto.
17395 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge): Ditto.
17396 * cfg.c (update_bb_profile_for_threading): Add comment to suggest using
17398 * sched-rgn.c (compute_dom_prob_ps): Ditto.
17399 (compute_trg_info): Ditto.
17400 * cfgrtl.c (force_nonfallthru_and_redirect): Ditto.
17401 (purge_dead_edges): Ditto.
17402 * loop-unswitch.c (unswitch_loop): Ditto.
17403 * cgraphclones.c (cgraph_clone_edge): Ditto.
17404 (cgraph_clone_node): Ditto.
17405 * tree-inline.c (copy_bb): Ditto.
17406 (copy_edges_for_bb): Ditto.
17407 (initialize_cfun): Ditto.
17408 (copy_cfg_body): Ditto.
17409 (expand_call_inline): Ditto.
17411 2013-04-08 Kai Tietz <ktietz@redhat.com>
17413 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Replaced
17414 TARGET_CYGWIN64 by TARGET_64BIT.
17416 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
17418 * config/epiphany/epiphany.md (GPR_1): New constant.
17419 (define_expand "mov<mode>cc): FAIL if gen_compare_reg returned 0.
17420 * config/epiphany/epiphany.c (gen_compare_reg):
17421 For flag_finite_math_only, avoid swapping operands when r0 and/or r1
17422 is already in place.
17423 Use GPR_0 / GPR_1 instead of 0/1 for r0/r1 register numbers.
17424 Don't require being called during rtl expansion; If y operlaps r0,
17426 (epiphany_compute_frame_size, epiphany_expand_prologue): Use GPR_1.
17427 (epiphany_expand_epilogue): Likewise.
17429 * config/epiphany/epiphany.c (epiphany_select_cc_mode):
17430 Don't use CC_FPmode for ORDERED / UNORDERED.
17431 * config/epiphany/epiphany.md (cmpsf_ord): Make pattern unconditional.
17433 * config/epiphany/constraints.md (CnL): New constraint.
17434 * config/epiphany/epiphany.md (addsi3_i): Add r/r/CnL alternative.
17435 * config/epiphany/predicates.md (add_operand): Allow 1024.
17437 * config/epiphany/epiphany.md (logical_op): New code iterator.
17438 (op_mnc): New code attribute.
17439 (<op_mnc>_f, mov_f, cstoresi4): New patterns.
17440 (mov_f+1, mov_f+2): New peephole2 patterns.
17442 * config/epiphany/epiphany.md (mov_f+2): New peephole2 pattern.
17443 (cstoresi4): Also allow re-use of zero result when doing a NE
17444 comparison to a non-zero operand.
17445 Use (clobber (scratch)) for first insn if the gpr output is not needed.
17447 * config/epiphany/epiphany.md (<insn_opname>v2si3):
17448 Use gen_addsi3_i / gen_subsi3_i.
17450 2013-04-08 Jakub Jelinek <jakub@redhat.com>
17454 * tree-eh.c (optimize_clobbers): Only remove clobbers if bb doesn't
17455 contain anything but clobbers, at most one __builtin_stack_restore,
17456 optionally debug stmts and final resx, and if it has at least one
17457 incoming EH edge. Don't check for SSA_NAME on LHS of a clobber.
17458 (sink_clobbers): Don't check for SSA_NAME on LHS of a clobber.
17459 Instead of moving clobbers with MEM_REF LHS with SSA_NAME address
17460 which isn't defaut definition, remove them.
17461 (unsplit_eh, cleanup_empty_eh): Use single_{pred,succ}_{p,edge}
17462 instead of EDGE_COUNT comparisons or EDGE_{PRED,SUCC}.
17463 * tree-ssa-ccp.c (execute_fold_all_builtins): Remove clobbers
17464 with MEM_REF LHS with SSA_NAME address.
17466 2013-04-08 Jeff Law <law@redhat.com>
17468 * gimple.c (canonicalize_cond_expr_cond): Rewrite x ^ y into x != y.
17470 2013-04-08 Richard Biener <rguenther@suse.de>
17472 * gimple-pretty-print.c (debug_gimple_stmt): Do not print
17474 * tree-vect-loop.c (vect_determine_vectorization_factor): Dump
17475 determined vector type.
17476 (vect_analyze_data_refs): Likewise.
17477 (vect_get_new_vect_var): Adjust.
17478 (vect_create_destination_var): Preserve SSA name versions.
17479 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): Do
17480 not dump anything here.
17482 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
17484 * config/epiphany/epiphany.h (struct GTY (()) machine_function):
17485 Add member lr_slot_known.
17486 * config/epiphany/epiphany.md (reload_insi_ra): Compute lr_slot_offs
17488 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
17489 Remove code that sets lr_slot_offset according to what a previous
17490 version of epiphany_emit_save_restore used to do.
17491 (epiphany_emit_save_restore): When doing an lr save or restore,
17492 set/verify lr_slot_known and lr_slot_offset.
17494 2013-04-08 Xinyu Qi <xyqi@marvell.com>
17497 * config/arm/arm.h (REG_CLASS_CONTENTS): Include IWMMXT_GR_REGS
17500 2013-04-08 Richard Biener <rguenther@suse.de>
17502 * alias.c (find_base_term): Fix thinko in previous change.
17504 2013-04-08 Jakub Jelinek <jakub@redhat.com>
17506 * tree-loop-distribution.c (const_with_all_bytes_same): New function.
17507 (generate_memset_builtin): Only handle integer_all_onesp as -1 val if
17508 TYPE_PRECISION is equal to mode bitsize. Use const_with_all_bytes_same
17509 if possible to compute val.
17510 (classify_partition): Verify CONSTRUCTOR doesn't have any elts.
17511 For QImode integers don't require anything about precision. Use
17512 const_with_all_bytes_same to find out if the constant doesn't have
17513 repeated bytes in it.
17515 2013-04-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
17517 * config/s390/s390.c (s390_expand_insv): Only accept insertions
17520 2013-04-08 Marek Polacek <polacek@redhat.com>
17522 PR rtl-optimization/48182
17523 * params.def (PARAM_MIN_CROSSJUMP_INSNS): Increase the minimum
17526 2013-04-06 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
17529 * config/pa/pa.c (legitimize_pic_address): Before incrementing label
17530 nuses, make sure we have a label.
17532 2013-04-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
17535 * config/rs6000/rs6000.c (rs6000_emit_swdiv_high_precision): Remove.
17536 (rs6000_emit_swdiv_low_precision): Remove.
17537 (rs6000_emit_swdiv): Rewrite to handle between one and four
17538 iterations of Newton-Raphson generally; modify required number of
17539 iterations for some cases.
17540 * config/rs6000/rs6000.h (RS6000_RECIP_HIGH_PRECISION_P): Remove.
17542 2013-04-05 Steven Bosscher <steven@gcc.gnu.org>
17544 * bb-reorder.c (fix_crossing_unconditional_branches): Remove a
17545 set-but-unused variable.
17547 * cgraph.c (cgraph_release_function_body): Clear cfun->cfg to make
17548 basic blocks of released function bodies garbage-collectable.
17550 * ree.c (find_and_remove_re): Do not call df_finish_pass here.
17551 (struct rtl_opt_pass): Add TODO_df_finish.
17553 * rtl.def (DEFINE_SUBST, DEFINE_SUBST_ATTR): Add documentation.
17555 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
17557 * config/arm/constraints.md (q): New constraint.
17558 * config/arm/ldrdstrd.md: New file.
17559 * config/arm/arm.md (ldrdstrd.md) New include.
17560 (arm_movdi): Use "q" instead of "r" constraint
17561 for double-word memory access.
17562 (movdf_soft_insn): Likewise.
17563 * config/arm/vfp.md (movdi_vfp): Likewise.
17564 * config/arm/t-arm (MD_INCLUDES): Add ldrdstrd.md.
17565 * config/arm/arm-protos.h (gen_operands_ldrd_strd): New declaration.
17566 * config/arm/arm.c (gen_operands_ldrd_strd): New function.
17567 (mem_ok_for_ldrd_strd): Likewise.
17568 (output_move_double): Update assertion.
17570 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
17572 * config/arm/arm.md: Comment on splitting Thumb1 patterns.
17574 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
17576 * config/arm/arm.md (arm_smax_insn): Convert define_insn into
17577 define_insn_and_split.
17578 (arm_smin_insn,arm_umaxsi3,arm_uminsi3): Likewise.
17580 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
17582 * config/arm/arm.md (arm_ashldi3_1bit): Convert define_insn into
17583 define_insn_and_split.
17584 (arm_ashrdi3_1bit,arm_lshrdi3_1bit): Likewise.
17585 (shiftsi3_compare): New pattern.
17586 (rrx): New pattern.
17587 * config/arm/unspecs.md (UNSPEC_RRX): New.
17589 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
17591 * config/arm/arm.md (negdi_extendsidi): New pattern.
17592 (negdi_zero_extendsidi): Likewise.
17594 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
17596 * config/arm/arm.md (andsi_iorsi3_notsi): Convert define_insn into
17597 define_insn_and_split.
17598 (arm_negdi2,arm_abssi2,arm_neg_abssi2): Likewise.
17599 (arm_cmpdi_insn,arm_cmpdi_unsigned): Likewise.
17601 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
17603 * config/arm/arm.md (arm_subdi3): Convert define_insn into
17604 define_insn_and_split.
17605 (subdi_di_zesidi,subdi_di_sesidi): Likewise.
17606 (subdi_zesidi_di,subdi_sesidi_di,subdi_zesidi_zesidi): Likewise.
17608 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
17610 * config/arm/arm.md (subsi3_carryin): New pattern.
17611 (subsi3_carryin_const): Likewise.
17612 (subsi3_carryin_compare,subsi3_carryin_compare_const): Likewise.
17613 (subsi3_carryin_shift,rsbsi3_carryin_shift): Likewise.
17615 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
17617 * config/arm/arm.md (incscc,arm_incscc,decscc,arm_decscc): Delete.
17619 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
17621 * config/arm/arm.md (addsi3_carryin_<optab>): Set attribute predicable.
17622 (addsi3_carryin_alt2_<optab>,addsi3_carryin_shift_<optab>): Likewise.
17624 2013-04-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17626 * config/arm/arm.c (arm_expand_builtin): Change fcode
17627 type to unsigned int.
17629 2013-04-05 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
17631 * doc/invoke.texi (ARM Options): Document cortex-a53 support.
17633 2013-04-04 Ian Lance Taylor <iant@google.com>
17635 * doc/standards.texi (Standards): The Go frontend supports the Go 1
17638 2013-04-04 Steven Bosscher <steven@gcc.gnu.org>
17640 PR middle-end/56729
17641 * df-scan.c (df_insn_delete): Disable failing assert.
17643 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17645 * config/arm/arm-protos.h (arm_builtin_vectorized_function):
17646 New function prototype.
17647 * config/arm/arm.c (TARGET_VECTORIZE_BUILTINS): Define.
17648 (TARGET_VECTORIZE_BUILTIN_VECTORIZED_FUNCTION): Likewise.
17649 (arm_builtin_vectorized_function): New function.
17651 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17653 * config/arm/arm_neon_builtins.def: New file.
17654 * config/arm/arm.c (neon_builtin_data): Move contents to
17655 arm_neon_builtins.def.
17656 (enum arm_builtins): Include neon builtin definitions.
17657 (ARM_BUILTIN_NEON_BASE): Move from enum to macro.
17658 * config/arm/t-arm (arm.o): Add dependency on arm_neon_builtins.def.
17660 2013-04-04 Marek Polacek <polacek@redhat.com>
17662 PR tree-optimization/48186
17663 * predict.c (maybe_hot_frequency_p): Return false if
17664 HOT_BB_FREQUENCY_FRACTION is 0.
17665 (cgraph_maybe_hot_edge_p): Likewise.
17667 2013-04-04 Richard Biener <rguenther@suse.de>
17669 PR tree-optimization/56826
17670 * tree-vect-slp.c (vect_build_slp_tree): Compute ncopies
17673 2013-04-04 Richard Biener <rguenther@suse.de>
17675 PR tree-optimization/56213
17676 * tree-vect-data-refs.c (vect_check_strided_load): Remove.
17677 (vect_analyze_data_refs): Allow all non-nested loads as strided loads.
17679 2013-04-04 Richard Biener <rguenther@suse.de>
17681 PR tree-optimization/56837
17682 * tree-loop-distribution.c (classify_partition): For non-zero
17683 values require that the value has the same precision as its
17684 mode to be useful as memset value.
17686 2013-04-03 Nick Clifton <nickc@redhat.com>
17688 * config/v850/v850e3v5.md (fmasf4): Use fmaf.s on E3V5 architectures.
17689 (fmssf4): Use fmsf.s on E3V5 architectures.
17690 (fnmasf4): Use fnmaf.s on E3V5 architectures.
17691 (fnmssf4): Use fnmsf.s on E3V5 architectures.
17693 2013-04-03 Jeff Law <law@redhat.com>
17695 * Makefile.in (lra-constraints.o): Depend on $(OPTABS_H).
17696 (lra-eliminations.o): Likewise.
17698 2013-04-03 Teresa Johnson <tejohnson@google.com>
17700 * gcov-io.c (compute_working_sets): Moved most of body of old
17701 compute_working_sets here from profile.c.
17702 * gcov-io.h (NUM_GCOV_WORKING_SETS): Moved here from profile.c.
17703 (gcov_working_set_t): Moved typedef here from basic-block.h
17704 (compute_working_set): Declare.
17705 * profile.c (NUM_GCOV_WORKING_SETS): Moved to gcov-io.h.
17706 (get_working_sets): Renamed from compute_working_set,
17707 replace most of body with call to new compute_working_sets.
17708 (get_exec_counts): Replace call to compute_working_sets
17709 to get_working_sets.
17710 * profile.h (get_working_sets): Renamed from compute_working_set.
17711 * lto-cgraph.c (input_symtab): Replace call to compute_working_sets
17712 to get_working_sets.
17713 * basic-block.h (gcov_working_set_t): Moved to gcov-io.h.
17714 * gcov-dump.c (dump_working_sets): New function.
17716 2013-04-03 Kenneth Zadeck <zadeck@naturalbridge.com>
17718 * hwint.c (sext_hwi, zext_hwi): New functions.
17719 * hwint.h (HOST_BITS_PER_HALF_WIDE_INT, HOST_HALF_WIDE_INT,
17720 HOST_HALF_WIDE_INT_PRINT, HOST_HALF_WIDE_INT_PRINT_C,
17721 HOST_HALF_WIDE_INT_PRINT_DEC, HOST_HALF_WIDE_INT_PRINT_DEC_C,
17722 HOST_HALF_WIDE_INT_PRINT_UNSIGNED, HOST_HALF_WIDE_INT_PRINT_HEX,
17723 HOST_HALF_WIDE_INT_PRINT_HEX_PURE): New symbols.
17724 (sext_hwi, zext_hwi): New functions.
17726 2013-04-03 Jeff Law <law@redhat.com>
17728 PR tree-optimization/56799
17729 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Bring
17730 back test for widening conversion erroneously dropped in prior change.
17732 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17735 * config/aarch64/aarch64.c (is_jump_table): Use next_active_insn
17736 instead of next_real_insn.
17738 2013-04-03 Marek Polacek <polacek@redhat.com>
17741 * tsan.c (instrument_func_exit): Allow BUILT_IN_RETURN functions.
17743 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17746 * config/arm/arm.c (is_jump_table): Use next_active_insn instead of
17748 (thumb1_output_casesi): Likewise.
17749 (thumb2_output_casesi): Likewise.
17751 2013-04-03 Richard Biener <rguenther@suse.de>
17753 PR tree-optimization/56817
17754 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely):
17756 (tree_unroll_loops_completely_1): ... new function to manually
17757 walk the loop tree, properly defering outer loops of unrolled
17758 loops to later iterations.
17760 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
17762 * tree-vect-stmts.c (vectorizable_store): Accept BIT_FIELD_REF.
17763 (vectorizable_load): Likewise.
17764 * tree-vect-slp.c (vect_build_slp_tree): Likewise.
17765 * tree-vect-data-refs.c (vect_create_data_ref_ptr): Handle VECTOR_TYPE.
17767 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
17769 * tree-flow-inline.h (get_addr_base_and_unit_offset_1): Handle
17772 2013-04-03 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
17774 * config/spu/spu.c (emit_nop_for_insn): Handle JUMP_TABLE_DATA.
17776 2013-04-03 Bin Cheng <bin.cheng@arm.com>
17778 * rtl.h (AUTO_INC_DEC): Fix typo of HAVE_POST_MODIFY_DISP.
17780 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
17782 PR tree-optimization/56790
17783 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Add constant
17786 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
17788 * simplify-rtx.c (simplify_binary_operation_1) <VEC_SELECT>:
17790 (simplify_ternary_operation) <VEC_MERGE>: Use unsigned HOST_WIDE_INT
17791 for masks. Test for side effects. Handle nested VEC_MERGE. Handle
17794 2013-04-03 Jakub Jelinek <jakub@redhat.com>
17797 * tree.h (force_folding_builtin_constant_p): New decl.
17798 * builtins.c (force_folding_builtin_constant_p): New variable.
17799 (fold_builtin_constant_p): Fold immediately also if
17800 force_folding_builtin_constant_p.
17802 2013-04-03 Richard Biener <rguenther@suse.de>
17804 PR tree-optimization/56812
17805 * tree-vect-data-refs.c (vect_slp_analyze_data_ref_dependence):
17806 DRs of the same interleaving chain are independent.
17808 2013-04-02 Jason Merrill <jason@redhat.com>
17810 * gdbinit.in (pbb): Use debug fn.
17812 2013-04-02 Lawrence Crowl <crowl@google.com>
17814 * sese.h (struct ivtype_map_elt_s): Remove unused.
17815 (extern debug_ivtype_map): Remove unused.
17816 (extern eq_ivtype_map_elts): Remove unused.
17817 * sese.c (debug_ivtype_map): Removed unused.
17818 (debug_ivtype_map_1): Removed unused.
17819 (debug_ivtype_elt): Remove unused.
17820 (eq_ivtype_map_elts): Remove unused.
17823 2013-04-02 Kai Tietz <ktietz@redhat.com>
17826 * config/i386/cygming.h (SUB_TARGET_RECORD_STUB): New sub-target macro.
17827 * config/i386/i386-protos.h (i386_pe_record_stub): Add new prototype.
17828 * config/i386/i386.c (legitimize_pe_coff_extern_decl): New static
17830 (legitimize_pe_coff_symbol): Likewise.
17831 (is_imported_p): New helper-function.
17832 (ix86_option_override_internal): Make MEDIUM_PIC the default code-model
17833 for Windows x64 targets.
17834 (ix86_expand_prologue): Optimize for pe-coff targets.
17835 (ix86_expand_split_stack_prologue): Adjust for pe-coff targets.
17836 (legitimate_pic_address_disp_p): Adjust for x64 pe-coff to support
17837 medium/large code-model.
17838 (legitimize_pic_address): Likewise.
17839 (legitimize_tls_address): Likewise.
17840 (ix86_expand_call): Likewise.
17841 (x86_output_mi_thunk): Likewise.
17842 (get_dllimport_decl): Add new beimport argument.
17843 (construct_plt_address): Don't assert for x64 pe-coff targets.
17844 * config/i386/i386.h (PIC_OFFSET_TABLE_REGNUM): Adjust for x64 pe-coff
17846 (SYMBOL_FLAG_STUBVAR): New macro.
17847 (SYMBOL_REF_STUBVAR_P): Likewise.
17848 * config/i386/winnt.c (stub_list): New structure.
17849 (stub_head): New local variable.
17850 (i386_pe_record_stub): New function.
17851 (i386_pe_file_end): Emit refptr-stubs.
17853 2013-04-02 Jakub Jelinek <jakub@redhat.com>
17855 PR rtl-optimization/56745
17856 * ifcvt.c (cond_exec_find_if_block): Don't try to optimize
17857 if then_bb has no successors and else_bb is EXIT_BLOCK_PTR.
17860 * tree-ssa-alias.c (stmt_kills_ref_p_1): If base != ref->base
17861 and both of them are MEM_REFs, just compare first argument for
17862 equality and attempt to deal even with differing offsets.
17865 * tree-cfg.c (verify_gimple_assign_single): Allow lhs
17866 of gimple_clobber_p to be MEM_REF.
17867 * gimplify.c (gimplify_modify_expr): Gimplify *to_p of
17868 an assignment from TREE_CLOBBER_P. Allow it to be MEM_REF
17869 after gimplification.
17870 * asan.c (get_mem_ref_of_assignment): Don't instrument
17871 gimple_clobber_p stmts.
17872 * tree-ssa-dse.c (dse_optimize_stmt): Allow DSE of
17873 gimple_clobber_p stmt if they have MEM_REF lhs and
17874 are dead because of another gimple_clobber_p stmt.
17875 * tree-ssa-live.c (clear_unused_block_pointer): Treat
17876 gimple_clobber_p stmts like debug stmts.
17877 (remove_unused_locals): Remove clobbers with MEM_REF lhs
17878 that refer to unused VAR_DECLs or uninitialized values.
17879 * tree-sra.c (sra_ipa_reset_debug_stmts): Also remove
17880 gimple_clobber_p stmts if they refer to removed parameters.
17881 (get_repl_default_def_ssa_name, sra_ipa_modify_expr): Fix up
17884 2013-04-02 Uros Bizjak <ubizjak@gmail.com>
17886 * config/i386/i386.md (*testqi_ext_3): Merge with *testqi_ext_3_rex64
17887 using SWI48 mode attribute.
17889 2013-04-02 Wei Mi <wmi@google.com>
17891 * config/i386/i386.c (ix86_rtx_costs): Set proper rtx cost for
17892 ashl<mode>3_mask, *<shift_insn><mode>3_mask and
17893 *<rotate_insn><mode>3_mask in i386.md.
17895 2013-04-02 Alexander Ivchenko <alexander.ivchenko@intel.com>
17897 * config.gcc (arm*-*-linux-*): Remove duplicate t-linux-android.
17899 2013-04-02 Richard Biener <rguenther@suse.de>
17901 PR tree-optimization/56778
17902 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence):
17903 Runtime alias tests are not supported for gather loads.
17904 * tree-vect-loop-manip.c (vect_loop_versioning): Insert
17905 stmts referenced from SSA operands before updating SSA form.
17907 2013-04-02 Ian Caulfield <ian.caulfield@arm.com>
17908 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
17910 * config/arm/arm-arches.def (armv8-a): Default to cortex-a53.
17911 * config/arm/t-arm (MD_INCLUDES): Depend on cortex-a53.md.
17912 * config/arm/cortex-a53.md: New file.
17913 * config/arm/bpabi.h (BE8_LINK_SPEC): Handle cortex-a53.
17914 * config/arm/arm.md (generic_sched, generic_vfp): Handle cortex-a53.
17915 * config/arm/arm.c (arm_issue_rate): Likewise.
17916 * config/arm/arm-tune.md: Regenerate
17917 * config/arm/arm-tables.opt: Regenerate.
17918 * config/arm/arm-cores.def: Add cortex-a53.
17920 2013-04-02 Zhenqiang Chen <zhenqiang.chen@arm.com>
17922 * config/arm/uclinux-elf.h: Add %L to LINK_GCC_C_SEQUENCE_SPEC for
17925 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
17927 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add variants for
17928 scalar load/store operations using B/H registers.
17929 (*zero_extend<SHORT:mode><GPI:mode>2_aarch64): Likewise.
17931 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
17933 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add alternatives for
17935 * config/aarch64/aarch64.c
17936 (aarch64_simd_scalar_immediate_valid_for_move): New.
17937 * config/aarch64/aarch64-protos.h
17938 (aarch64_simd_scalar_immediate_valid_for_move): New.
17939 * config/aarch64/constraints.md (Dh, Dq): New.
17940 * config/aarch64/iterators.md (hq): New.
17942 2013-04-02 Eric Botcazou <ebotcazou@adacore.com>
17944 * reorg.c (get_branch_condition): Deal with conditional returns.
17945 (fill_simple_delay_slots): Remove dead code dealing with jumps.
17947 2013-04-01 Wei Mi <wmi@google.com>
17949 * config/i386/i386.md (*ashl<mode>3_mask): Rewrite as define_insn.
17950 Truncate operand 2 using %b asm operand modifier.
17951 (*<shift_insn><mode>3_mask): Ditto.
17952 (*<rotate_insn><mode>3_mask): Ditto.
17954 2013-04-01 Steven Bosscher <steven@gcc.gnu.org>
17956 PR middle-end/56798
17957 * cfgbuild.c (inside_basic_block_p): Restore check broken at r197234.
17959 2013-03-31 Kaz Kojima <kkojima@gcc.gnu.org>
17961 * config/sh/sh.md (casesi_worker_1): Use next_active_insn instead
17963 (casesi_worker_2, casesi_shift_media, casesi_load_media): Likewise.
17965 2013-03-30 Lawrence Crowl <crowl@google.com>
17967 * dse.c (clear_alias_sets): Remove never set.
17968 (disqualified_clear_alias_sets): Remove never set.
17969 (clear_alias_mode_pool): Remove never set.
17970 (dse_step0): Remove condition that is never true.
17971 (canon_address): Remove condition that is never true.
17972 (dse_step7): Remove condition that is never true.
17973 (rest_of_handle_dse): Remove condition that is never true.
17974 (rest_of_handle_dse::did_global): Remove never read from above.
17975 (dse_step2_spill): Remove never called from above.
17976 (dse_step5_spill): Remove never called from above.
17978 2013-03-30 Steven Bosscher <steven@gcc.gnu.org>
17980 * doc/md.texi (Standard Names) <casesi>: Update documentation for
17981 JUMP_TABLE_DATA changes.
17982 * doc/tm.texi.in (Dispatch Tables) <ASM_OUTPUT_CASE_LABEL>: Likewise.
17983 * doc/rtl.texi (Flags) <INSN_DELETED_P, SCHED_GROUP_P>: Likewise.
17984 (Insns) <jump_table_data>: New entry.
17985 * doc/tm.texi: Regenerate.
17987 * cfgrtl.c (fixup_reorder_chain): Do not emit barriers to BB_FOOTER.
17989 * postreload-gcse.c (bb_has_well_behaved_predecessors): Correct test
17990 for table jump at the end of a basic block using tablejump_p.
17991 * targhooks.c (default_invalid_within_doloop): Likewise.
17992 * config/rs6000/rs6000.c (TARGET_INVALID_WITHIN_DOLOOP): Remove
17993 target hook implementation that is identical to the default hook.
17994 (rs6000_invalid_within_doloop): Remove.
17996 * bb-reorder.c (fix_crossing_unconditional_branches): Remove set but
17997 unused variable from tablejump_p call.
17999 * rtl.def (JUMP_TABLE_DATA): New RTX_INSN object.
18000 * rtl.h (RTX_PREV, RTX_NEXT): Adjust for new JUMP_TABLE_DATA.
18001 (INSN_DELETED_P): Likewise.
18002 (emit_jump_table_data): New prototype.
18003 * gengtype.c (adjust_field_rtx_def): Handle JUMP_TABLE_DATA fields
18004 after 4th as unused.
18005 * print-rtl.c (print_rtl): Handle JUMP_TABLE_DATA.
18006 * sched-vis.c (print_insn): Likewise.
18007 * emit-rtl.c (active_insn_p): Consider JUMP_TABLE_DATA an active
18008 insn for compatibility with back ends that use next_active_insn to
18009 identify jump table data.
18010 (set_insn_deleted): Remove no longer useful JUMP_TABLE_DATA_P check.
18011 (remove_insn): Likewise.
18012 (emit_insn): Do not accept JUMP_TABLE_DATA objects in insn chains
18014 (emit_debug_insn, emit_jump_insn, emit_call_insn, emit_label): Idem.
18015 (emit_jump_table_data): New function.
18017 * cfgbuild.c (inside_basic_block_p): A JUMP_INSN is always inside a
18018 basic block, a JUMP_TABLE_DATA never is.
18019 (control_flow_insn_p): JUMP_TABLE_DATA is not a control flow insn.
18020 * cfgrtl.c (duplicate_insn_chain): Split handling of JUMP_TABLE_DATA
18021 off from code handling real insns.
18022 * final.c (get_attr_length_1): Simplify for JUMP_INSNs.
18023 * function.c (instantiate_virtual_regs): Remove JUMP_TABLE_DATA_P
18024 test, now redundant because JUMP_TABLE_DATA is not an INSN_P insn.
18025 * gcse.c (insert_insn_end_basic_block): Likewise, JUMP_TABLE_DATA_P
18026 is not a NONDEBUG_INSN_P.
18027 * ira-costs.c (scan_one_insn): Likewise.
18028 * jump.c (mark_all_labels): Likewise.
18029 (mark_jump_label_1): Likewise.
18030 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
18031 * lra.c (get_insn_freq): Expect all insns reaching here to be in
18033 (check_rtl): Remove JUMP_TABLE_DATA_P test, not a NONDEBUG_INSN_P insn.
18034 * predict.c (expensive_function_p): Use FOR_BB_INSNS.
18035 * reload1.c (calculate_needs_all_insns): Call set_label_offsets for
18036 JUMP_TABLE_DATA_P insns.
18037 (calculate_elim_costs_all_insns): Likewise.
18038 (set_label_offsets): Recurse on the PATTERN of JUMP_TABLE_DATA insns.
18039 (elimination_costs_in_insn): Remove redundant JUMP_TABLE_DATA_P test.
18040 (delete_output_reload): Code style fixups.
18041 * reorg.c (dbr_schedule): Move JUMP_TABLE_DATA_P up to avoid setting
18042 insn flags on this non-insn.
18043 * sched-rgn.c (add_branch_dependences): Treat JUMP_TABLE_DATA insns
18044 as scheduling barriers, for pre-change compatibility.
18045 * stmt.c (emit_case_dispatch_table): Emit jump table data not as
18046 JUMP_INSN objects but instead as JUMP_TABLE_DATA objects.
18048 * config/alpha/alpha.c (alpha_does_function_need_gp): Remove
18049 redundant JUMP_TABLE_DATA_P test.
18050 * config/arm/arm.c (thumb_far_jump_used_p): Likewise.
18051 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
18052 (frv_for_each_packet): Likewise.
18053 * config/i386/i386.c (min_insn_size): Likewise.
18054 (ix86_avoid_jump_mispredicts): Likewise.
18055 * config/m32r/m32r.c (m32r_is_insn): Likewise.
18056 * config/mep/mep.c (mep_reorg_erepeat): Likewise.
18057 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
18058 (mips16_insn_length): Robustify.
18059 (mips_has_long_branch_p): Remove redundant JUMP_TABLE_DATA_P test.
18060 (mips16_split_long_branches): Likewise.
18061 * config/pa/pa.c (pa_combine_instructions): Likewise.
18062 * config/rs6000/rs6000.c (get_next_active_insn): Treat
18063 JUMP_TABLE_DATA objects as active insns, like in active_insn_p.
18064 * config/s390/s390.c (s390_chunkify_start): Treat JUMP_TABLE_DATA
18065 as contributing to pool range lengths.
18066 * config/sh/sh.c (find_barrier): Restore check for ADDR_DIFF_VEC.
18067 Remove redundant JUMP_TABLE_DATA_P test.
18068 (sh_loop_align): Likewise.
18069 (split_branches): Likewise.
18070 (sh_insn_length_adjustment): Likewise.
18071 * config/spu/spu.c (get_branch_target): Likewise.
18073 2013-03-29 Jan Hubicka <jh@suse.cz>
18075 * lto-cgraph.c (output_profile_summary, input_profile_summary): Use
18076 gcov streaming; stream hot bb threshold to ltrans.
18077 * predict.c (get_hot_bb_threshold): Break out from ....
18078 (maybe_hot_count_p): ... here.
18079 (set_hot_bb_threshold): New function.
18080 * lto-section-in.c (lto_section_name): Add profile.
18081 * profile.h (get_hot_bb_threshold, set_hot_bb_threshold): Declare.
18082 * ipa.c: Include hash-table.h, tree-inline.h, profile.h, lto-streamer.h
18083 and data-streamer.h
18084 (histogram_entry): New structure.
18085 (histogram, histogram_pool): New global vars.
18086 (histogram_hash): New structure.
18087 (histogram_hash::hash): New method.
18088 (histogram_hash::equal): Likewise.
18089 (account_time_size): New function.
18090 (cmp_counts): New function.
18091 (dump_histogram): New function.
18092 (ipa_profile_generate_summary): New function.
18093 (ipa_profile_write_summary): New function.
18094 (ipa_profile_read_summary): New function.
18095 (ipa_profile): Decide on threshold.
18096 (pass_ipa_profile): Add ipa_profile_write_summary and
18097 ipa_profile_read_summary.
18098 * Makefile.in (ipa.o): Update dependencies.
18099 * lto-streamer.h (LTO_section_ipa_profile): New section.
18101 2013-03-29 Gabriel Dos Reis <gdr@integrable-solutions.net>
18103 * tree.h (VAR_P): New.
18105 2013-03-29 Paolo Carlini <paolo.carlini@oracle.com>
18108 * doc/invoke.texi ([-fwhole-program]): Fix typo.
18110 2013-03-29 Steven Bosscher <steven@gcc.gnu.org>
18112 * cfgbuild.c (inside_basic_block_p): Use JUMP_TABLE_DATA_P in lieu
18113 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
18114 (control_flow_insn_p): Likewise.
18115 * cfgrtl.c (duplicate_insn_chain): Likewise.
18116 * final.c (get_attr_length_1): Likewise.
18117 (shorten_branches): Likewise.
18118 (final_scan_insn): Likewise.
18119 * function.c (instantiate_virtual_regs): Likewise.
18120 * gcse.c (insert_insn_end_basic_block): Likewise.
18121 * ira-costs.c (scan_one_insn): Likewise.
18122 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
18123 * lra.c (check_rtl): Likewise.
18124 * reload1.c (elimination_costs_in_insn): Likewise.
18125 * reorg.c (follow_jumps): Likewise.
18127 * config/arm/arm.c (is_jump_table): Use JUMP_TABLE_DATA_P in lieu
18128 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
18129 (thumb_far_jump_used_p): Likewise.
18130 * config/bfin/bfin.c (workaround_rts_anomaly): Likewise.
18131 (workaround_speculation): Likewise.
18132 (add_sched_insns_for_speculation): Likewise.
18133 * config/c6x/c6x.c (reorg_emit_nops): Likewise.
18134 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
18135 (frv_for_each_packet): Likewise.
18136 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Likewise.
18137 * config/ia64/ia64.c (emit_all_insn_group_barriers): Likewise.
18138 (final_emit_insn_group_barriers): Likewise.
18139 * config/m32r/m32r.c (m32r_is_insn): Likewise.
18140 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
18141 (mips16_insn_length): Likewise.
18142 * config/pa/pa.c (pa_reorg): Likewise.
18143 (pa_combine_instructions): Likewise.
18144 * config/rs6000/rs6000.c (rs6000_invalid_within_doloop): Likewise.
18145 * config/sh/sh.c (fixup_addr_diff_vecs): Likewise.
18146 (sh_reorg): Likewise.
18147 (split_branches): Likewise.
18148 * config/spu/spu.c (get_branch_target): Likewise.
18150 * config/s390/s390.c (s390_chunkify_start): Simplify logic using
18153 2013-03-29 Kirill Yukhin <kirill.yukhin@intel.com>
18155 * config/i386/avx2intrin.h (_mm256_broadcastsi128_si256):
18156 Fix declaration name.
18158 2013-03-28 Lawrence Crowl <crowl@google.com>
18160 * graphds.h (struct graph.indicies): Remove unused.
18161 * graphite-poly.h (struct graph.original_pddrs): Remove unused.
18162 (SCOP_ORIGINAL_PDDRS): Remove unused.
18163 * sese.h (extern insert_loop_close_phis): Removed unused.
18164 (extern insert_guard_phis): Removed unused.
18165 (extern ivtype_map_elt_info): Removed unused.
18166 (new_ivtype_map_elt): Removed unused.
18167 * sese.c (ivtype_map_elt_info): Removed unused.
18169 2013-03-28 Lawrence Crowl <crowl@google.com>
18171 * Makefile.in: Add several missing include dependences.
18173 (test-dump.o): New. This object is not added to any executable,
18174 but is present for ad-hoc testing.
18176 (debug (const bitmap_head_def &)): New.
18177 (debug (const bitmap_head_def *)): New.
18179 (extern debug (const bitmap_head_def &)): New.
18180 (extern debug (const bitmap_head_def *)): New.
18182 (debug (edge_def &)): New.
18183 (debug (edge_def *)): New.
18185 (debug (basic_block_def &)): New.
18186 (debug (basic_block_def *)): New.
18188 (dump_node (const_tree, int, FILE *)): Correct source file.
18190 (debug (die_struct &)): New.
18191 (debug (die_struct *)): New.
18193 (extern debug (die_struct &)): New.
18194 (extern debug (die_struct *)): New.
18195 * gimple-pretty-print.c
18196 (debug (gimple_statement_d &)): New.
18197 (debug (gimple_statement_d *)): New.
18198 * gimple-pretty-print.h
18199 (extern debug (gimple_statement_d &)): New.
18200 (extern debug (gimple_statement_d *)): New.
18202 (debug (ira_allocno_copy &)): New.
18203 (debug (ira_allocno_copy *)): New.
18204 (debug (ira_allocno &)): New.
18205 (debug (ira_allocno *)): New.
18207 (extern debug (ira_allocno_copy &)): New.
18208 (extern debug (ira_allocno_copy *)): New.
18209 (extern debug (ira_allocno &)): New.
18210 (extern debug (ira_allocno *)): New.
18212 (debug (live_range &)): New.
18213 (debug (live_range *)): New.
18215 (debug (lra_live_range &)): New.
18216 (debug (lra_live_range *)): New.
18218 (debug (lra_live_range &)): New.
18219 (debug (lra_live_range *)): New.
18221 (debug (omega_pb_d &)): New.
18222 (debug (omega_pb_d *)): New.
18224 (extern debug (omega_pb_d &)): New.
18225 (extern debug (omega_pb_d *)): New.
18227 (debug (const rtx_def &)): New.
18228 (debug (const rtx_def *)): New.
18230 (debug_tree (tree): Move within file.
18231 (debug_raw (const tree_node &)): New.
18232 (debug_raw (const tree_node *)): New.
18233 (dump_tree_via_hooks (const tree_node *, int)): New.
18234 (debug (const tree_node &)): New.
18235 (debug (const tree_node *)): New.
18236 (debug_verbose (const tree_node &)): New.
18237 (debug_verbose (const tree_node *)): New.
18238 (debug_head (const tree_node &)): New.
18239 (debug_head (const tree_node *)): New.
18240 (debug_body (const tree_node &)): New.
18241 (debug_body (const tree_node *)): New.
18242 (debug_vec_tree (tree): Move and reimplement in terms of dump.
18243 (debug (vec<tree, va_gc> &)): New.
18244 (debug (vec<tree, va_gc> *)): New.
18246 (extern debug (const rtx_def &)): New.
18247 (extern debug (const rtx_def *)): New.
18249 (debug_raw (simple_bitmap_def &)): New.
18250 (debug_raw (simple_bitmap_def *)): New.
18251 (debug (simple_bitmap_def &)): New.
18252 (debug (simple_bitmap_def *)): New.
18254 (extern debug (simple_bitmap_def &)): New.
18255 (extern debug (simple_bitmap_def *)): New.
18256 (extern debug_raw (simple_bitmap_def &)): New.
18257 (extern debug_raw (simple_bitmap_def *)): New.
18259 (debug (vinsn_def &)): New.
18260 (debug (vinsn_def *)): New.
18261 (debug_verbose (vinsn_def &)): New.
18262 (debug_verbose (vinsn_def *)): New.
18263 (debug (expr_def &)): New.
18264 (debug (expr_def *)): New.
18265 (debug_verbose (expr_def &)): New.
18266 (debug_verbose (expr_def *)): New.
18267 (debug (vec<rtx> &)): New.
18268 (debug (vec<rtx> *)): New.
18270 (extern debug (vinsn_def &)): New.
18271 (extern debug (vinsn_def *)): New.
18272 (extern debug_verbose (vinsn_def &)): New.
18273 (extern debug_verbose (vinsn_def *)): New.
18274 (extern debug (expr_def &)): New.
18275 (extern debug (expr_def *)): New.
18276 (extern debug_verbose (expr_def &)): New.
18277 (extern debug_verbose (expr_def *)): New.
18278 (extern debug (vec<rtx> &)): New.
18279 (extern debug (vec<rtx> *)): New.
18281 (_list_iter_cond_expr): Make inline instead of static.
18283 (debug (sreal &)): New.
18284 (debug (sreal *)): New.
18286 (extern debug (sreal &)): New.
18287 (extern debug (sreal *)): New.
18289 (extern debug_raw (const tree_node &)): New.
18290 (extern debug_raw (const tree_node *)): New.
18291 (extern debug (const tree_node &)): New.
18292 (extern debug (const tree_node *)): New.
18293 (extern debug_verbose (const tree_node &)): New.
18294 (extern debug_verbose (const tree_node *)): New.
18295 (extern debug_head (const tree_node &)): New.
18296 (extern debug_head (const tree_node *)): New.
18297 (extern debug_body (const tree_node &)): New.
18298 (extern debug_body (const tree_node *)): New.
18299 (extern debug (vec<tree, va_gc> &)): New.
18300 (extern debug (vec<tree, va_gc> *)): New.
18302 (debug (struct loop &)): New.
18303 (debug (struct loop *)): New.
18304 (debug_verbose (struct loop &)): New.
18305 (debug_verbose (struct loop *)): New.
18306 * tree-dump.c: Add header dependence.
18308 (extern debug (struct loop &)): New.
18309 (extern debug (struct loop *)): New.
18310 (extern debug_verbose (struct loop &)): New.
18311 (extern debug_verbose (struct loop *)): New.
18313 (debug (data_reference &)): New.
18314 (debug (data_reference *)): New.
18315 (debug (vec<data_reference_p> &)): New.
18316 (debug (vec<data_reference_p> *)): New.
18317 (debug (vec<ddr_p> &)): New.
18318 (debug (vec<ddr_p> *)): New.
18320 (extern debug (data_reference &)): New.
18321 (extern debug (data_reference *)): New.
18322 (extern debug (vec<data_reference_p> &)): New.
18323 (extern debug (vec<data_reference_p> *)): New.
18324 (extern debug (vec<ddr_p> &)): New.
18325 (extern debug (vec<ddr_p> *)): New.
18327 (debug (pt_solution &)): New.
18328 (debug (pt_solution *)): New.
18330 (extern debug (pt_solution &)): New.
18331 (extern debug (pt_solution *)): New.
18333 (debug (_var_map &)): New.
18334 (debug (_var_map *)): New.
18335 (debug (tree_live_info_d &)): New.
18336 (debug (tree_live_info_d *)): New.
18338 (extern debug (_var_map &)): New.
18339 (extern debug (_var_map *)): New.
18340 (extern debug (tree_live_info_d &)): New.
18341 (extern debug (tree_live_info_d *)): New.
18343 2013-03-28 Jan Hubicka <jh@suse.cz>
18345 * lto-cgraph.c (merge_profile_summaries): Fix overflows.
18347 2013-03-28 Ian Bolton <ian.bolton@arm.com>
18349 * config/aarch64/aarch64.md (aarch64_can_eliminate): Keep frame
18350 record only when desired or required.
18352 2013-03-28 Uros Bizjak <ubizjak@gmail.com>
18354 * config/i386/i386.md (*vec_extract2vdi_1): Merge with
18355 *vec_extractv2di_1_rex64. Use x64 isa attribute.
18357 2013-03-28 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
18359 * config/aarch64/aarch64.md (*and<mode>3_compare0): New pattern.
18360 (*andsi3_compare0_uxtw): New pattern.
18361 (*and_<SHIFT:optab><mode>3_compare0): New pattern.
18362 (*and_<SHIFT:optab>si3_compare0_uxtw): New pattern.
18364 2013-03-28 Jan Hubicka <jh@suse.cz>
18366 * data-streamer-in.c (streamer_read_gcov_count): New function.
18367 * gimple-streamer-out.c: Include value-prof.h.
18368 (output_gimple_stmt): Output histogram.
18369 (output_bb): Use streamer_write_gcov_count.
18370 * value-prof.c: Include data-streamer.h
18371 (dump_histogram_value): Add HIST_TYPE_MAX.
18372 (stream_out_histogram_value): New function.
18373 (stream_in_histogram_value): New function.
18374 * value-prof.h (enum hist_type): Add HIST_TYPE_MAX.
18375 (stream_out_histogram_value, stream_in_histogram_value): Declare.
18376 * data-streamer-out.c (streamer_write_gcov_count): New function.
18377 (streamer_write_gcov_count_stream): New function.
18378 * lto-cgraph.c (lto_output_edge): Update counter streaming.
18379 (lto_output_node): Likewise.
18380 (input_node, input_edge): Likewise.
18381 * lto-streamer-out.c (output_cfg): Update streaming.
18382 * lto-streamer-in.c (input_cfg): Likewise.
18383 * data-streamer.h (streamer_write_gcov_count,
18384 streamer_write_gcov_count_stream, streamer_read_gcov_count): Declare.
18385 * gimple-streamer-in.c: Include value-prof.h
18386 (input_gimple_stmt): Input histograms.
18387 (input_bb): Update profile streaming.
18389 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
18391 * genmodes.c (emit_max_int): New function.
18392 (emit_insn_modes_h): Added call to emit_max_function.
18393 * doc/rtl.texi (MAX_BITSIZE_MODE_ANY_INT, MAX_BITSIZE_MODE_ANY_MODE):
18395 * machmode.def: Fixed comment.
18397 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
18399 * combine.c (try_combine): Removed useless assert.
18400 * cselib.c (rtx_equal_for_cselib_1): Removed unnecessary parens.
18402 2013-03-28 Marek Polacek <polacek@redhat.com>
18403 Richard Biener <rguenther@suse.de>
18405 PR tree-optimization/56695
18406 * tree-vect-stmts.c (vectorizable_condition): Unconditionally
18407 build signed result of a vector comparison.
18408 * tree-cfg.c (verify_gimple_comparison): Check that a result
18409 of a vector comparison has signed type.
18411 2013-03-28 Richard Biener <rguenther@suse.de>
18413 PR tree-optimization/37021
18414 * tree-vect-slp.c (vect_build_slp_tree): When not unrolling
18415 do not restrict gaps between groups.
18416 * tree-vect-stmts.c (vectorizable_load): Properly account for
18417 a gap between groups.
18419 2013-03-28 Eric Botcazou <ebotcazou@adacore.com>
18421 * toplev.c (process_options): Do not disable -fomit-frame-pointer on a
18422 general basis if unwind info is requested and ACCUMULATE_OUTGOING_ARGS
18425 2013-03-27 Gerald Pfeifer <gerald@pfeifer.com>
18427 * doc/invoke.texi (AVR Options): Tweak link for AVR-LibC user manual.
18428 * doc/extend.texi (Named Address Spaces): Ditto.
18429 (Variable Attributes): Ditto.
18431 2013-03-27 Kai Tietz <ktietz@redhat.com>
18433 * config.build: Add support for cygwin x64 target.
18434 * config.gcc: Likewise.
18435 * config.host: Likewise.
18436 * configure.ac: Likewise
18437 * configure: Regenerated.
18439 2013-03-27 Kai Tietz <ktietz@redhat.com>
18441 * config/i386/cygwin-stdint.h: Add support for cygwin x64 target.
18442 * config/i386/t-cygwin-w64: New file.
18443 * config/i386/cygwin-w64.h: New file.
18444 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Extend
18445 and add support for x64-cygwin target.
18446 (CPP_SPEC): Likewise.
18447 (CXX_WRAP_SPEC_LIST): Undefine before define.
18448 (LIBGCJ_SONAME): Use 15 as version.
18450 2013-03-27 Richard Biener <rguenther@suse.de>
18452 PR tree-optimization/56716
18453 * tree-ssa-structalias.c (perform_var_substitution): Adjust
18454 dumping for ref nodes.
18456 2013-03-27 Martin Jambor <mjambor@suse.cz>
18458 PR tree-optimization/55334
18459 * ipa-cp.c (initialize_node_lattices): Allow IPA-CP through and to
18460 restricted pointers to arrays.
18462 2013-03-27 Gabriel Dos Reis <gdr@integrable-solutions.net>
18464 * Makefile.in (.SUFFIXES): Add .cc.
18465 (.c.o): Apply same recipe for implicit rule .cc.o.
18467 2013-03-27 Richard Biener <rguenther@suse.de>
18469 PR tree-optimization/37021
18470 * tree-vect-data-refs.c (vect_check_strided_load): Allow
18471 REALPART/IMAGPART_EXPRs around the supported refs.
18472 * tree-ssa-structalias.c (find_func_aliases): Assume that
18473 floating-point values are not used to transfer pointers.
18475 2013-03-27 Alexander Ivchenko <alexander.ivchenko@intel.com>
18477 * target.def (TARGET_HAS_IFUNC_P): New target hook.
18478 * doc/tm.texi.in (TARGET_HAS_IFUNC_P): New.
18479 * doc/tm.texi: Regenerate.
18480 * targhooks.h (default_has_ifunc_p): New.
18481 * targhooks.c (default_has_ifunc_p): Ditto.
18482 * config/linux-protos.h: New file.
18483 * config/linux-android.h (TARGET_HAS_IFUNC_P): Using version of this
18484 hook for linux which disables support of indirect functions in android.
18485 * config/linux-android.c: New file.
18486 * config/t-linux-android.c: Ditto.
18487 * config.gcc: Added new object file linux-android.o.
18488 * config/i386/i386.c (ix86_get_function_versions_dispatcher):
18489 Using TARGET_HAS_IFUNC hook instead of HAVE_GNU_INDIRECT_FUNCTION.
18490 * varasm.c (do_assemble_alias): Likewise.
18491 * configure.ac: Define HAVE_GNU_INDIRECT_FUNCTION as zero if the target
18492 doesn't support indirect functions.
18493 * configure: Regenerate.
18495 2013-03-27 Bin Cheng <bin.cheng@arm.com>
18498 * config/arm/arm.c (thumb1_rtx_costs, thumb1_size_rtx_costs): Fix
18499 rtx costs for SET/ASHIFT/ASHIFTRT/LSHIFTRT/ROTATERT patterns with
18502 2013-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
18504 * config/s390/s390.h (TARGET_FLT_EVAL_METHOD): Define.
18506 2013-03-27 Terry Guo <terry.guo@arm.com>
18508 * config/arm/arm-cores.def: Added core cortex-r7.
18509 * config/arm/arm-tune.md: Regenerated.
18510 * config/arm/arm-tables.opt: Regenerated.
18511 * doc/invoke.texi: Added entry for core cortex-r7.
18513 2013-03-27 Walter Lee <walt@tilera.com>
18515 * config/tilegx/tilegx.c (tilegx_expand_prologue): Avoid
18516 double-decrement of next_scratch_regno.
18518 2013-03-27 Walter Lee <walt@tilera.com>
18520 * config/tilegx/tilegx.md (insn_v1mulu): Fix predicates on
18522 (insn_v1mulus): Ditto.
18523 (insn_v2muls): Ditto.
18525 2013-03-27 Walter Lee <walt@tilera.com>
18527 * config/tilegx/tilegx.h (ASM_OUTPUT_ADDR_VEC_ELT): Delete extra tab.
18528 (ASM_OUTPUT_ADDR_DIFF_ELT): Ditto.
18530 2013-03-27 Walter Lee <walt@tilera.com>
18532 * config/tilegx/tilegx.md (*sibcall_insn): Fix type atribute for jr.
18533 (*sibcall_value): Ditto.
18535 2013-03-27 Walter Lee <walt@tilera.com>
18537 * config/tilegx/tilegx.md (insn_mnz_<mode>): Replaced by ...
18538 (insn_mnz_v8qi): ... this ...
18539 (insn_mnz_v4hi): ... and this. Replace (const_int 0) with the
18541 (insn_v<n>mnz): Replaced by ...
18542 (insn_v1mnz): ... this ...
18543 (insn_v2mnz): ... and this. Replace (const_int 0) with the vector
18545 (insn_mz_<mode>): Replaced by ...
18546 (insn_mz_v8qi): ... this ...
18547 (insn_mz_v4hi): ... and this. Replace (const_int 0) with the
18549 (insn_v<n>mz): Replaced by ...
18550 (insn_v1mz): ... this ...
18551 (insn_v2mz): ... and this. Replace (const_int 0) with the vector
18554 2013-03-26 Eric Botcazou <ebotcazou@adacore.com>
18556 * doc/invoke.texi (SPARC options): Remove -mlittle-endian.
18558 2013-03-26 Roland McGrath <mcgrathr@google.com>
18560 * config/arm/arm.c (arm_print_operand: case 'w'): Use fputs rather
18561 than fprintf with a non-constant, non-format string.
18563 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
18565 * config/i386/i386.md (*cmpqi_ext_1): Merge with *cmpqi_ext_1_rex64
18566 using nox64 isa attribute. Use nonimmediate_x86nomem_operand as
18567 operand 0 predicate.
18568 (*cmpqi_ext_3): Merge with *cmpqi_ext_3_rex64 using nox64 isa
18569 attribute. Use general_x64nomem_operand as operand 1 predicate.
18570 (*movqi_extv_1): Merge with *movqi_extv_1_rex64 using nox64 isa
18571 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
18572 (*movqi_extzv_2): Merge with *movqi_extzv_2_rex64 using nox64 isa
18573 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
18574 (mov<mode>_insv_1): Remove expander. Merge insn with
18575 movsi_insv_1 using SWI48 mode iterator and nox64 isa attribute.
18576 Use general_x64nomem_operand as operand 1 predicate.
18577 (addqi_ext_1): Merge with *addqi_ext_1_rex64 using nox64 isa attribute.
18578 (*testqi_ext_1): Merge with *testqi_ext_1_rex64 using nox64 isa
18579 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
18580 (*andqi_ext_1): Merge with *andqi_ext_1_rex64 using nox64 isa
18581 attribute. Use nonimmediate_x64nomem_operand as operand 2 predicate.
18582 (*<code>qi_ext_1): Merge with *<code>qi_ext_1_rex64 using nox64 isa
18583 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
18584 (*xorqi_cc_ext_1): Merge with *xorqi_cc_ext_1_rex64 using nox64
18585 isa attribute. Use general_x64nomem_operand as operand 2 predicate.
18586 * config/i386/predicates.md (nonimmediate_x64nomem_operand): New.
18587 (general_x64nomem_operand): Ditto.
18589 2013-03-26 Sebastian Huber <sebastian.huber@embedded-brains.de>
18591 * config/rtems.opt: Add -pthread option.
18593 2013-03-26 Richard Biener <rguenther@suse.de>
18595 * alias.c (find_base_term): Avoid redundant and not used recursion.
18596 (base_alias_check): Get the initial base term from the caller.
18597 (true_dependence_1): Compute and pass base terms to base_alias_check.
18598 (write_dependence_p): Likewise.
18599 (may_alias_p): Likewise.
18601 2013-03-26 Sofiane Naci <sofiane.naci@arm.com>
18603 * config/aarch64/aarch64.c (aarch64_classify_address): Support
18604 PC-relative load in SI modes and above only.
18606 2013-03-26 Xinyu Qi <xyqi@marvell.com>
18608 * config/arm/arm.h (FIRST_IWMMXT_GR_REGNUM): Add comment.
18609 * config/arm/iwmmxt.md (WCGR0): Update.
18610 (WCGR1, WCGR2, WCGR3): Likewise.
18612 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
18614 * config/i386/i386.md (*movdfcc_1): Merge with *movdfcc_1_rex64.
18615 Use x64 and nox64 isa attributes.
18617 2013-03-26 Richard Biener <rguenther@suse.de>
18619 * emit-rtl.c (set_mem_attributes_minus_bitpos): Remove
18620 alignment computations and rely on get_object_alignment_1
18621 for the !TYPE_P case.
18622 Commonize DECL/COMPONENT_REF handling in the ARRAY_REF path.
18624 2013-03-26 Walter Lee <walt@tilera.com>
18626 * config/tilegx/tilegx.h (PROFILE_BEFORE_PROLOGUE): Define.
18627 * config/tilegx/tilepro.h (PROFILE_BEFORE_PROLOGUE): Define.
18629 2013-03-25 Jeff Law <law@redhat.com>
18631 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Add missing
18632 check for INTEGRAL_TYPE_P that was missing due to checking in
18633 wrong version of prior patch.
18635 2013-03-25 Walter Lee <walt@tilera.com>
18637 * config/tilegx/tilegx-builtins.h (enum tilegx_builtin): Add
18638 TILEGX_INSN_SHUFFLEBYTES1.
18639 * config/tilegx/tilegx.c (tilegx_builtin_info): Add entry for
18641 (tilegx_builtins): Ditto.
18642 * config/tilegx/tilegx.md (insn_shufflebytes1): New pattern.
18644 2013-03-25 Walter Lee <walt@tilera.com>
18646 * config/tilegx/tilegx.md (floatsisf2): New pattern.
18647 (floatunssisf2): New pattern.
18648 (floatsidf2): New pattern.
18649 (floatunssidf2): New pattern.
18651 2013-03-25 Walter Lee <walt@tilera.com>
18653 * config/tilegx/tilegx.c (expand_set_cint64_one_inst): Inline
18654 tests for constraint J, K, N, P.
18656 2013-03-25 Walter Lee <walt@tilera.com>
18658 * config/tilegx/tilegx.c (tilegx_asm_preferred_eh_data_format):
18659 Use indirect/pcrel encoding.
18660 * config/tilepro/tilepro.c (tilepro_asm_preferred_eh_data_format):
18663 2013-03-25 Steve Ellcey <sellcey@mips.com>
18665 * config/mips/mmips-cpus.def (74kc, 74kf2_1, 74kf, 74kf, 74kf1_1,
18666 74kfx, 74kx, 74kf3_2): Add PTF_AVOID_IMADD.
18667 * config/mips/mips.c (mips_option_override): Set IMADD default.
18668 * config/mips/mips.h (PTF_AVOID_IMADD): New.
18669 (ISA_HAS_MADD_MSUB): Remove MIPS16 check.
18670 (GENERATE_MADD_MSUB): Remove TUNE_74K check, add MIPS16 check.
18671 * config/mips/mips.md (mimadd): New flag for integer madd/msub.
18672 * doc/invoke.texi (-mimadd/-mno-imadd): New.
18674 2013-03-25 Jeff Law <law@redhat.com>
18676 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Rework
18677 slightly to avoid creating and folding useless trees. Simplify
18678 slightly by restricting to INTEGER_CSTs and using int_fits_type_p.
18680 2013-03-25 Uros Bizjak <ubizjak@gmail.com>
18682 * config/i386/i386.md (*zero_extendsidi2): Merge with
18683 *zero_extendsidi2_rex64. Use x64 and nox64 isa attributes.
18684 * config/i386/predicates.md (x86_64_zext_operand): Rename from
18685 x86_64_zext_general_operand. Use nonimmediate_operand on 32bit
18686 targets. Clarify comment.
18688 2013-03-25 Martin Jambor <mjambor@suse.cz>
18690 * ipa-prop.c (ipa_write_jump_function): Stream simple and aritmetic
18691 pass-through jump functions differently.
18692 (ipa_read_jump_function): Likewise. Also use setter functions to set
18695 2013-03-25 Martin Jambor <mjambor@suse.cz>
18697 * ipa-cp.c (ipa_get_indirect_edge_target): Renamed to
18698 ipa_get_indirect_edge_target_1, added parameter agg_reps and ability to
18700 (ipa_get_indirect_edge_target): New function.
18701 (devirtualization_time_bonus): New parameter known_aggs, pass it to
18702 ipa_get_indirect_edge_target. Update all callers.
18703 (ipcp_discover_new_direct_edges): New parameter aggvals. Pass it to
18704 ipa_get_indirect_edge_target_1 instead of calling
18705 ipa_get_indirect_edge_target.
18706 (create_specialized_node): Pass aggvlas to
18707 ipcp_discover_new_direct_edges.
18709 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18711 * config/arm/arm.md (f_sels, f_seld): New types.
18712 (*cmov<mode>): New pattern.
18713 * config/arm/predicates.md (arm_vsel_comparison_operator): New
18716 2013-03-25 Kai Tietz <ktietz@redhat.com>
18718 * config/i386/xm-mingw32.h (__USE_MINGW_ANSI_STDIO): Enable
18719 POSIX-printf for mingw-hosted builds.
18721 2013-03-25 Richard Biener <rguenther@suse.de>
18723 PR middle-end/56694
18724 * tree-eh.c (lower_eh_must_not_throw): Strip BLOCKs from the
18725 must-not-throw stmt location.
18727 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18729 * config/arm/arm.c (arm_emit_load_exclusive): Add acq parameter.
18730 Emit load-acquire versions when acq is true.
18731 (arm_emit_store_exclusive): Add rel parameter.
18732 Emit store-release versions when rel is true.
18733 (arm_split_compare_and_swap): Use acquire-release instructions
18735 of barriers when appropriate.
18736 (arm_split_atomic_op): Likewise.
18737 * config/arm/arm.h (TARGET_HAVE_LDACQ): New macro.
18738 * config/arm/unspecs.md (VUNSPEC_LAX): New unspec.
18739 (VUNSPEC_SLX): Likewise.
18740 (VUNSPEC_LDA): Likewise.
18741 (VUNSPEC_STL): Likewise.
18742 * config/arm/sync.md (atomic_load<mode>): New pattern.
18743 (atomic_store<mode>): Likewise.
18744 (arm_load_acquire_exclusive<mode>): Likewise.
18745 (arm_load_acquire_exclusivesi): Likewise.
18746 (arm_load_acquire_exclusivedi): Likewise.
18747 (arm_store_release_exclusive<mode>): Likewise.
18749 2013-03-25 Catherine Moore <clm@codesourcery.com>
18751 * config/mips/constraints.md (u, Udb7 Uead, Uean, Uesp, Uib3,
18752 Uuw6, Usb4, ZS, ZT, ZU, ZV, ZW): New constraints.
18753 * config/mip/predicates.md (lwsp_swsp_operand,
18754 lw16_sw16_operand, lhu16_sh16_operand, lbu16_operand,
18755 sb16_operand, db4_operand, db7_operand, ib3_operand,
18756 sb4_operand, ub4_operand, uh4_operand, uw4_operand,
18757 uw5_operand, uw6_operand, addiur2_operand, addiusp_operand,
18758 andi16_operand): New predicates.
18759 * config/mips/mips.md (compression): New attribute.
18760 (enabled): New attribute.
18761 (length): Consider compression in computing length.
18762 (shift_compression): New code attribute.
18763 (*add<mode>3): New operands. Record compression.
18764 (sub<mode>3): Likewise.
18765 (one_cmpl<mode>2): Likewise.
18766 (*and<mode>3): Likewise.
18767 (*ior<mode>3): Likewise.
18768 (unnamed pattern for xor): Likewise.
18769 (*zero_extend<SHORT:mode><GPR:mode>2): Likewise.
18770 (*<optab><mode>3): Likewise.
18771 (*mov<mode>_internal: Likewise.
18772 * config/mips/mips-protos.h (mips_signed_immediate_p): New.
18773 (mips_unsigned_immediate_p): New.
18774 (umips_lwsp_swsp_address_p): New.
18775 (m16_based_address_p): New.
18776 * config/mips/mips-protos.h (mips_signed_immediate_p): New prototype.
18777 (mips_unsigned_immediate_p): New prototype.
18778 (lwsp_swsp_address_p): New prototype.
18779 (m16_based_address_p): New prototype.
18780 * config/mips/mips.c (mips_unsigned_immediate_p): New function.
18781 (mips_signed_immediate_p): New function.
18782 (m16_based_address_p): New function.
18783 (lwsp_swsp_address_p): New function.
18784 (mips_print_operand_punctuation): Recognize short delay slot insns
18785 for microMIPS.add<mode>3"
18787 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18790 * config/arm/iterators.md (v_cmp_result): New mode attribute.
18791 * config/arm/neon.md (vcond<mode><mode>): Handle unordered cases.
18793 2013-03-25 Richard Biener <rguenther@suse.de>
18795 PR tree-optimization/56689
18796 * tree-vrp.c (execute_vrp): Mark loops for fixup if we removed
18799 2013-03-25 Richard Biener <rguenther@suse.de>
18801 * tree-ssa-loop-im.c (struct mem_ref): Use bitmap_head instead
18803 (memory_references): Likewise.
18804 (outermost_indep_loop, mem_ref_alloc, mark_ref_stored,
18805 gather_mem_refs_stmt, record_dep_loop, ref_indep_loop_p_1,
18806 ref_indep_loop_p_2, find_refs_for_sm): Adjust.
18807 (gather_mem_refs_in_loops): Fold into ...
18808 (analyze_memory_references): ... this. Move initialization
18809 to tree_ssa_lim_initialize.
18810 (fill_always_executed_in): Rename to ...
18811 (fill_always_executed_in_1): ... this.
18812 (fill_always_executed_in): Move contains_call computation to
18813 this new function from ...
18814 (tree_ssa_lim_initialize): ... here.
18815 (tree_ssa_lim): Call fill_always_executed_in.
18817 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
18819 * postreload.c (reload_combine): Fix code detecting returns.
18821 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
18823 * function.c (emit_use_return_register_into_block): On cc0 targets,
18824 do not emit the sequence between cc0 setter and user.
18826 2013-03-25 Kai Tietz <ktietz@redhat.com>
18828 * config/i386/predicates.md (local_symbolic_operand): Interpret
18829 dll-imported symbols as none-local.
18831 2013-03-25 Richard Biener <rguenther@suse.de>
18833 * tree-ssa-loop-im.c (struct depend): Remove.
18834 (struct lim_aux_data): Make depends a vec of gimples.
18835 (free_lim_aux_data): Adjust.
18836 (add_dependency): Likewise.
18837 (set_level): Likewise.
18839 2013-03-25 Richard Biener <rguenther@suse.de>
18841 PR middle-end/56434
18842 * calls.c (expand_call): Use MALLOC_ABI_ALIGNMENT to annotate
18843 the pointer returned by calls with ECF_MALLOC set.
18845 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
18847 * config/i386/mmx.md (mov<mode>): Add ?!Ym,r and r,?!Ym alternatives.
18849 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
18851 * config/i386/mmx.md (mov<mode>): Merge with movv2sf expander
18852 using MMXMODE mode iterator.
18853 (*move<mode>_internal): Merge with *movv2sf_internal and
18854 *movv2sf_internal_rex64 using MMXMODE mode iterator.
18856 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
18858 * gcse.c (oprs_unchanged_p): Respect flag_gcse_lm.
18859 (record_last_mem_set_info): Likewise.
18861 * df-core.c (rest_of_handle_df_initialize): Use XCNEWVEC instead
18862 of XNEWVEC followed by memset.
18863 (df_worklist_dataflow): Use XNEWVEC instead of xmalloc with a cast.
18865 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
18867 * config/avr/avr.c, config/bfin/bfin.c, config/c6x/c6x.c,
18868 config/epiphany/epiphany.c, config/frv/frv.c, config/ia64/ia64.c,
18869 config/iq2000/iq2000.c, config/mcore/mcore.c, config/mep/mep.c,
18870 config/mmix/mmix.c, config/pa/pa.c, config/rs6000/rs6000.c,
18871 config/s390/s390.c, config/sparc/sparc.c, config/spu/spu.c,
18872 config/stormy16/stormy16.c, config/v850/v850.c, config/xtensa/xtensa.c,
18873 dwarf2out.c, hw-doloop.c, resource.c, rtl.h : Where applicable, use
18874 the predicates NOTE_P, NONJUMP_INSN_P, JUMP_P, CALL_P, LABEL_P, and
18875 BARRIER_P instead of GET_CODE.
18877 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
18879 * config/sparc/sparc.c (sparc_emit_probe_stack_range): Fix small
18880 inaccuracy in the probing code.
18882 * config/sparc/sparc.md (ctrapsi4): Add predicate for operand #3.
18883 (ctrapdi4): Likewise.
18885 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
18887 * calls.c (expand_call): Add missing guard to code handling return
18888 of non-BLKmode structures in MSB.
18889 * function.c (expand_function_end): Likewise.
18891 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
18893 * combine.c (try_combine): Adjust comment. Do not add the set of
18894 insn #0 if the destination indirectly is set or dies in insn #2.
18895 Tidy up code to distribute a new note.
18897 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
18899 * config/i386/i386.md (*movdi_internal): Set prefix_rex attribute
18900 also for alternatives 16 and 17.
18902 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
18904 * config/i386/sse.md (*mov<mode>_internal): Merge with
18905 *mov<mode>_internal_rex64. Use x64 and nox64 isa attributes.
18906 Emit insn template depending on type attribute. Use
18907 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
18908 movd instead of movq mnemonic for interunit moves. Rewrite mode
18909 attribute calculation. Remove unit attribute calculation.
18910 Set prefix attribute to maybe_vex for sselog1 and ssemov types.
18911 Set prefix_data16 attribute for DImode ssemov types.
18912 Use Ym instead of y for SSE-MMX conversion alternatives.
18913 Reorder operand constraints.
18915 2013-03-22 Steven Bosscher <steven@gcc.gnu.org>
18917 * df.h (df_insn_delete): Adjust prototype.
18918 * emit-rtl.c (remove_insn): Pass a basic block to df_insn_delete
18919 and let it decide whether mark the basic block dirty.
18920 (set_insn_deleted): Only pass INSN_P insns to df_insn_delete.
18921 * df-scan.c (df_insn_info_delete): New helper function, split
18922 off from df_insn_delete.
18923 (df_scan_free_bb_info): Use it.
18924 (df_insn_rescan, df_insn_rescan_all, df_process_deferred_rescans):
18926 (df_insn_delete): Likewise. Take insn rtx as argument. Verify
18927 that the insn is actually an insn and it has a non-NULL basic block.
18928 Do not mark basic block dirty if only deleting a DEBUG_INSN.
18930 2013-03-22 Richard Biener <rguenther@suse.de>
18932 * tree-ssa-loop-im.c (struct mem_ref): Remove indep_ref and
18934 (mem_ref_alloc): Do not allocate them.
18935 (refs_independent_p): Do not query or maintain a cache.
18937 2013-03-22 Richard Biener <rguenther@suse.de>
18939 * tree-ssa-loop-im.c (memory_references): Drop all_refs_in_loop.
18940 (gather_mem_refs_in_loops): Do not compute it.
18941 (analyze_memory_references): Do not allocate it.
18942 (tree_ssa_lim_finalize): Do not free it.
18943 (for_all_locs_in_loop): Do not query all_refs_in_loop.
18945 2013-03-22 Richard Biener <rguenther@suse.de>
18947 * is-a.h (as_a): Use gcc_checking_assert.
18949 2013-03-22 Ian Bolton <ian.bolton@arm.com>
18951 * config/aarch64/aarch64.c (aarch64_print_operand): New
18952 format specifier for printing a constant in hex.
18953 * config/aarch64/aarch64.md (insv_imm<mode>): Use the X
18954 format specifier for printing second operand.
18956 2013-03-22 Richard Biener <rguenther@suse.de>
18958 * tree-ssa-loop-im.c (memory_references): Add refs_stored_in_loop
18960 (gather_mem_refs_in_loops): Perform store accumulation here.
18961 (create_vop_ref_mapping_loop): Remove.
18962 (create_vop_ref_mapping): Likewise.
18963 (analyze_memory_references): Initialize refs_stored_in_loop.
18964 (LOOP_DEP_BIT): New define to map to bits in (in)dep_loop bitmaps.
18965 (record_indep_loop): Remove.
18966 (record_dep_loop): New function.
18967 (ref_indep_loop_p_1): Adjust to only walk over references
18968 in the loop, not its subloops.
18969 (ref_indep_loop_p): Rename to ...
18970 (ref_indep_loop_p_2): ... this and recurse over the loop tree,
18971 maintaining a more fine-grained cache.
18972 (ref_indep_loop_p): Wrap ref_indep_loop_p_2.
18973 (tree_ssa_lim_finalize): Free refs_stored_in_loop.
18975 2013-03-22 Richard Biener <rguenther@suse.de>
18977 * tree-ssa-loop-im.c (struct mem_ref_locs): Remove.
18978 (struct mem_ref): Make accesses_in_loop a vec of a vec of
18979 aggregate mem_ref_loc.
18980 (free_mem_ref_locs): Inline into ...
18981 (memref_free): ... this and adjust.
18982 (mem_ref_alloc): Adjust.
18983 (mem_ref_locs_alloc): Remove.
18984 (record_mem_ref_loc): Adjust.
18985 (get_all_locs_in_loop): Rewrite into ...
18986 (for_all_locs_in_loop): ... this iterator.
18987 (rewrite_mem_ref_loc): New functor.
18988 (rewrite_mem_refs): Use for_all_locs_in_loop.
18989 (sm_set_flag_if_changed): New functor.
18990 (execute_sm_if_changed_flag_set): Use for_all_locs_in_loop.
18991 (ref_always_accessed): New functor.
18992 (ref_always_accessed_p): Use for_all_locs_in_loop.
18994 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
18996 * tree-pass.h (PROP_gimple_lvec): New.
18997 * passes.c (dump_properties): Handle PROP_gimple_lvec.
18998 (init_optimization_passes): Move pass_lower_vector.
18999 * tree-vect-generic.c (gate_expand_vector_operations_ssa): Test
19001 (pass_lower_vector): Provide PROP_gimple_lvec.
19002 (pass_lower_vector_ssa): Likewise.
19003 * cfgexpand.c (pass_expand): Require PROP_gimple_lvec.
19005 2013-03-21 Mark Wielaard <mjw@redhat.com>
19007 * dwarf2out.c (size_of_aranges): Skip DECL_IGNORED_P functions.
19009 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
19011 * config/i386/i386.md (*movdi_internal): Disparage slightly
19012 all MMX moves to/from memory. Use Yi instead of x for SSE-MMX
19013 conversion alternatives.
19015 2013-03-21 Jakub Jelinek <jakub@redhat.com>
19017 PR middle-end/48087
19018 * diagnostic.def (DK_WERROR): New kind.
19019 * diagnostic.h (werrorcount): Define.
19020 * diagnostic.c (diagnostic_report_diagnostic): For DK_WARNING
19021 promoted to DK_ERROR, increment DK_WERROR counter instead of
19023 * toplev.c (toplev_main): Call print_ignored_options even if
19024 just werrorcount is non-zero. Exit with FATAL_EXIT_CODE
19025 even if just werrorcount is non-zero.
19028 * dwarf2out.c (tree_add_const_value_attribute): Call ggc_free (array)
19030 (resolve_one_addr): Fail if referenced STRING_CST hasn't been written.
19031 (string_cst_pool_decl): New function.
19032 (optimize_one_addr_into_implicit_ptr): New function.
19033 (resolve_addr_in_expr): Optimize DWARF location expression
19034 DW_OP_addr DW_OP_stack_value where DW_OP_addr refers to some variable
19035 which doesn't live in memory, but has DW_AT_location or
19036 DW_AT_const_value, or refers to a string literal, into
19037 DW_OP_GNU_implicit_pointer.
19038 (optimize_location_into_implicit_ptr): New function.
19039 (resolve_addr): If removing DW_AT_location of a variable because
19040 it was DW_OP_addr of address of the variable, but the variable doesn't
19041 live in memory, try to emit const value attribute for the initializer.
19043 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
19045 * tree.h (VECTOR_TYPE_P): New macro.
19046 (VECTOR_INTEGER_TYPE_P, VECTOR_FLOAT_TYPE_P, FLOAT_TYPE_P,
19047 TYPE_MODE): Use it.
19048 * fold-const.c (fold_cond_expr_with_comparison): Use build_zero_cst.
19049 VEC_COND_EXPR cannot be lvalues.
19050 (fold_ternary_loc) <VEC_COND_EXPR>: Merge with the COND_EXPR case.
19052 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
19054 * simplify-rtx.c (simplify_binary_operation_1) <VEC_CONCAT>:
19055 Restrict the transformation to equal modes.
19057 2013-03-21 Richard Biener <rguenther@suse.de>
19059 PR tree-optimization/39326
19060 * tree-ssa-loop-im.c (UNANALYZABLE_MEM_ID): New define.
19061 (MEM_ANALYZABLE): Adjust.
19062 (record_mem_ref_loc): Move bitmap ops ...
19063 (gather_mem_refs_stmt): ... here. Use the shared mem-ref for
19064 unanalyzable refs, do not record locations for it.
19065 (analyze_memory_references): Allocate ref zero as shared
19067 (refs_independent_p): Do not test for unanalyzed mems here.
19068 (ref_indep_loop_p_1): Special-case disambiguation against
19069 the unanalyzed ref.
19070 (ref_indep_loop_p): Assert we are not queried for the unanalyzed mem.
19072 2013-03-21 Christophe Lyon <christophe.lyon@linaro.org>
19074 * config/arm/arm-protos.h (tune_params): Add
19075 prefer_neon_for_64bits field.
19076 * config/arm/arm.c (prefer_neon_for_64bits): New variable.
19077 (arm_slowmul_tune): Default prefer_neon_for_64bits to false.
19078 (arm_fastmul_tune, arm_strongarm_tune, arm_xscale_tune): Ditto.
19079 (arm_9e_tune, arm_v6t2_tune, arm_cortex_tune): Ditto.
19080 (arm_cortex_a15_tune, arm_cortex_a5_tune): Ditto.
19081 (arm_cortex_a9_tune, arm_v6m_tune, arm_fa726te_tune): Ditto.
19082 (arm_option_override): Handle -mneon-for-64bits new option.
19083 * config/arm/arm.h (TARGET_PREFER_NEON_64BITS): New macro.
19084 (prefer_neon_for_64bits): Declare new variable.
19085 * config/arm/arm.md (arch): Rename neon_onlya8 and neon_nota8 to
19086 avoid_neon_for_64bits and neon_for_64bits. Remove onlya8 and nota8.
19087 (arch_enabled): Handle new arch types. Remove support for onlya8
19089 (one_cmpldi2): Use new arch names.
19090 (zero_extend<mode>di2, extend<mode>di2): Ditto.
19091 * config/arm/arm.opt (mneon-for-64bits): Add option.
19092 * config/arm/neon.md (adddi3_neon, subdi3_neon, iordi3_neon)
19093 (anddi3_neon, xordi3_neon, ashldi3_neon, <shift>di3_neon): Use
19094 neon_for_64bits instead of nota8 and avoid_neon_for_64bits instead
19096 * doc/invoke.texi (-mneon-for-64bits): Document.
19098 2013-03-21 Richard Biener <rguenther@suse.de>
19100 PR tree-optimization/39326
19101 * tree-ssa-loop-im.c (bb_loop_postorder): New global static.
19102 (sort_bbs_in_loop_postorder_cmp): New function.
19103 (gather_mem_refs_in_loops): Assign mem-ref IDs in loop postorder.
19105 2013-03-21 Richard Biener <rguenther@suse.de>
19107 * tree-vect-data-refs.c (vect_update_interleaving_chain): Remove.
19108 (vect_insert_into_interleaving_chain): Likewise.
19109 (vect_drs_dependent_in_basic_block): Inline ...
19110 (vect_slp_analyze_data_ref_dependence): ... here. New function,
19112 (vect_analyze_data_ref_dependence): ... here. Simplify.
19113 (vect_check_interleaving): Simplify.
19114 (vect_analyze_data_ref_dependences): Likewise. Split out ...
19115 (vect_slp_analyze_data_ref_dependences): ... this new function.
19116 (dr_group_sort_cmp): New function.
19117 (vect_analyze_data_ref_accesses): Compute data-reference groups
19118 here instead of in vect_analyze_data_ref_dependence. Use
19119 a more efficient algorithm.
19120 * tree-vect-slp.c (vect_slp_analyze_bb_1): Use
19121 vect_slp_analyze_data_ref_dependences. Call
19122 vect_analyze_data_ref_accesses earlier.
19123 * tree-vect-loop.c (vect_analyze_loop_2): Likewise.
19124 * tree-vectorizer.h (vect_analyze_data_ref_dependences): Adjust.
19125 (vect_slp_analyze_data_ref_dependences): New prototype.
19127 2013-03-21 Richard Biener <rguenther@suse.de>
19129 * tree-ssa-loop-im.c (can_sm_ref_p): Do not test whether
19130 ref is stored in the loop.
19131 (find_refs_for_sm): Walk only over all stores.
19132 (store_motion_loop): Allocate from lim_bitmap_obstack.
19133 (store_motion): Likewise.
19135 2013-03-21 Richard Biener <rguenther@suse.de>
19137 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge):
19138 Update virtual SSA form.
19140 2013-03-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
19142 * configure.ac (gcc_cv_ld_eh_frame_ciev3): New test.
19143 * configure: Regenerate.
19144 * config.in: Regenerate.
19145 * config/sol2.c (solaris_override_options): Only enforce DWARF 2
19146 if !HAVE_LD_EH_FRAME_CIEV3.
19148 2013-03-21 Richard Biener <rguenther@suse.de>
19150 * tree-cfg.c (verify_expr_no_block): New function.
19151 (verify_expr_location_1): Verify that neither DECL_DEBUG_EXPR
19152 nor DECL_VALUE_EXPR have locations with associated blocks.
19153 * tree-ssa-live.c (clear_unused_block_pointer_1): Remove.
19154 (clear_unused_block_pointer): Remove code dealing with
19155 blocks in DECL_DEBUG_EXPR locations.
19157 2013-03-21 Richard Biener <rguenther@suse.de>
19159 * tree.h (DECL_DEBUG_EXPR_IS_FROM): Rename to ...
19160 (DECL_HAS_DEBUG_EXPR_P): ... this. Guard properly.
19161 * tree.c (copy_node_stat): Do not copy DECL_HAS_DEBUG_EXPR_P.
19162 * dwarf2out.c (add_var_loc_to_decl): Use DECL_HAS_DEBUG_EXPR_P
19163 instead of DECL_DEBUG_EXPR_IS_FROM.
19164 * gimplify.c (gimplify_modify_expr): Likewise.
19165 * tree-cfg.c (verify_expr_location_1): Likewise.
19166 * tree-complex.c (create_one_component_var): Likewise.
19167 * tree-sra.c (create_access_replacement): Likewise.
19168 * tree-ssa-live.c (clear_unused_block_pointer_1): Likewise.
19169 (clear_unused_block_pointer): Likewise.
19170 * tree-streamer-in.c (unpack_ts_decl_common_value_fields): Likewise.
19171 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Likewise.
19172 * var-tracking.c (var_debug_decl): Likewise.
19173 (track_expr_p): Likewise.
19174 * tree-inline.c (add_local_variables): Likewise. Set
19175 DECL_HAS_DEBUG_EXPR_P after copying it.
19176 * tree-diagnostic.c (default_tree_printer): Use DECL_HAS_DEBUG_EXPR_P
19177 instead of DECL_DEBUG_EXPR_IS_FROM. Guard properly.
19179 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
19182 * configure.ac (HAVE_AS_IX86_INTERUNIT_MOVQ): New test.
19183 * configure: Regenerate.
19184 * config.in: Regenerate.
19185 * config/i386/i386.md (*movdf_internal): Use
19186 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
19187 movd instead of movq mnemonic for interunit moves.
19188 (*movdi_internal): Ditto.
19190 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
19192 * config/aarch64/aarch64-simd.md (simd_fabd): New Attribute.
19193 (abd<mode>_3): New pattern.
19194 (aba<mode>_3): New pattern.
19195 (fabd<mode>_3): New pattern.
19197 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
19199 * config/aarch64/aarch64-elf.h (REGISTER_PREFIX): Remove.
19200 * config/aarch64/aarch64.c (aarch64_print_operand): Remove all
19201 occurrence of REGISTER_PREFIX as its empty string.
19203 2013-03-20 Jeff Law <law@redhat.com>
19205 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Record
19206 addititional equivalences for equality comparisons between an SSA_NAME
19207 and a constant where the SSA_NAME was set from a widening conversion.
19209 2013-03-20 Walter Lee <walt@tilera.com>
19211 * config/tilegx/sync.md (atomic_test_and_set): New pattern.
19213 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
19215 * config/i386/i386.md (*movoi_internal_avx): Emit insn template
19216 depending on type attribute.
19217 (*movti_internal): Ditto.
19218 (*movtf_internal): Ditto.
19219 (*movxf_internal): Ditto.
19220 (*movdf_internal): Ditto.
19221 (*movsf_internal): Ditto.
19223 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
19225 * config/i386/i386.md (*movti_internal): Set prefix attribute to
19226 maybe_vex for sselog1 and ssemov types.
19227 (*movdi_internal): Reorder operand constraints.
19228 (*movsi_internal): Ditto. Set prefix attribute to
19229 maybe_vex for sselog1 and ssemov types.
19230 (*movtf_internal): Set prefix attribute to maybe_vex
19231 for sselog1 and ssemov types.
19232 (*movdf_internal): Ditto. Set prefix_data16 attribute for
19233 DImode ssemov types. Reorder operand constraints.
19234 (*movsf_internal): Set type of alternatives 3,4 to imov. Set prefix
19235 attribute to maybe_vex for sselog1 and ssemov types. Set prefix_data16
19236 attribute for SImode ssemov types. Reorder operand constraints.
19238 2013-03-20 Martin Jambor <mjambor@suse.cz>
19240 * params.def (PARAM_IPA_CP_ARRAY_INDEX_HINT_BONUS): New parameter.
19241 * ipa-cp.c (hint_time_bonus): Add abonus for known array indices.
19243 2013-03-20 Pat Haugen <pthaugen@us.ibm.com>
19245 * config/rs6000/predicates.md (indexed_address, update_address_mem
19246 update_indexed_address_mem): New predicates.
19247 * config/rs6000/vsx.md (vsx_extract_<mode>_zero): Set correct "type"
19248 attribute for load/store instructions.
19249 * config/rs6000/dfp.md (movsd_store): Likewise.
19250 (movsd_load): Likewise.
19251 * config/rs6000/rs6000.md (zero_extend<mode>di2_internal1): Likewise.
19252 (unnamed HI->DI extend define_insn): Likewise.
19253 (unnamed SI->DI extend define_insn): Likewise.
19254 (unnamed QI->SI extend define_insn): Likewise.
19255 (unnamed QI->HI extend define_insn): Likewise.
19256 (unnamed HI->SI extend define_insn): Likewise.
19257 (unnamed HI->SI extend define_insn): Likewise.
19258 (extendsfdf2_fpr): Likewise.
19259 (movsi_internal1): Likewise.
19260 (movsi_internal1_single): Likewise.
19261 (movhi_internal): Likewise.
19262 (movqi_internal): Likewise.
19263 (movcc_internal1): Correct mnemonic for stw insn. Set correct "type"
19264 attribute for load/store instructions.
19265 (mov<mode>_hardfloat): Set correct "type" attribute for load/store
19267 (mov<mode>_softfloat): Likewise.
19268 (mov<mode>_hardfloat32): Likewise.
19269 (mov<mode>_hardfloat64): Likewise.
19270 (mov<mode>_softfloat64): Likewise.
19271 (movdi_internal32): Likewise.
19272 (movdi_internal64): Likewise.
19273 (probe_stack_<mode>): Likewise.
19275 2013-03-20 Michael Meissner <meissner@linux.vnet.ibm.com>
19277 * config/rs6000/vector.md (VEC_R): Add 32-bit integer, binary
19278 floating point, and decimal floating point to reload iterator.
19280 * config/rs6000/constraints.md (wl constraint): New constraints to
19281 return FLOAT_REGS if certain options are used to reduce the number
19282 of separate patterns that exist in the file.
19283 (wx constraint): Likewise.
19284 (wz constraint): Likewise.
19286 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
19287 -mdebug=reg, print wg, wl, wx, and wz constraints.
19288 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
19289 Initialize the reload functions for 64-bit binary/decimal floating
19291 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
19292 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
19293 create the buffer on the stack to overcome not having a 32-bit
19295 (rs6000_emit_move): Likewise.
19296 (rs6000_secondary_memory_needed_rtx): Likewise.
19297 (rs6000_alloc_sdmode_stack_slot): Likewise.
19298 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
19299 via xxlxor, just like DFmode 0.0.
19301 * config/rs6000/rs6000.h (TARGET_NO_SDMODE_STACK): New macro,
19302 define as 1 if we are running on a power7 or newer.
19303 (enum r6000_reg_class_enum): Add new constraints.
19305 * config/rs6000/dfp.md (movsd): Delete, combine with binary
19306 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
19307 with other moves by using conditional constraits (wg). Use LFIWZX
19308 and STFIWX for loading SDmode on power7. Use xxlxor to create 0.0f.
19309 (movsd splitter): Likewise.
19310 (movsd_hardfloat): Likewise.
19311 (movsd_softfloat): Likewise.
19313 * config/rs6000/rs6000.md (FMOVE32): New iterators to combine
19314 binary and decimal floating point moves.
19315 (fmove_ok): New attributes to combine binary and decimal floating
19316 point moves, and to combine power6x (mfpgpr) moves along normal
19318 (real_value_to_target): Likewise.
19319 (f32_lr): Likewise.
19320 (f32_lm): Likewise.
19321 (f32_li): Likewise.
19322 (f32_sr): Likewise.
19323 (f32_sm): Likewise.
19324 (f32_si): Likewise.
19325 (movsf): Combine binary and decimal floating point moves. Combine
19326 power6x (mfpgpr) moves with other moves by using conditional
19327 constraits (wg). Use LFIWZX and STFIWX for loading SDmode on power7.
19328 (mov<mode> for SFmode/SDmode); Likewise.
19329 (SFmode/SDmode splitters): Likewise.
19330 (movsf_hardfloat): Likewise.
19331 (mov<mode>_hardfloat for SFmode/SDmode): Likewise.
19332 (movsf_softfloat): Likewise.
19333 (mov<mode>_softfloat for SFmode/SDmode): Likewise.
19335 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wl,
19336 wx and wz constraints.
19338 * config/rs6000/constraints.md (wg constraint): New constraint to
19339 return FLOAT_REGS if -mmfpgpr (power6x) was used.
19341 * config/rs6000/rs6000.h (enum r6000_reg_class_enum): Add wg
19344 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
19345 -mdebug=reg, print wg, wl, wx, and wz constraints.
19346 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
19347 Initialize the reload functions for 64-bit binary/decimal floating
19349 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
19350 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
19351 create the buffer on the stack to overcome not having a 32-bit
19353 (rs6000_emit_move): Likewise.
19354 (rs6000_secondary_memory_needed_rtx): Likewise.
19355 (rs6000_alloc_sdmode_stack_slot): Likewise.
19356 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
19357 via xxlxor, just like DFmode 0.0.
19359 * config/rs6000/dfp.md (movdd): Delete, combine with binary
19360 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
19361 with other moves by using conditional constraits (wg). Use LFIWZX
19362 and STFIWX for loading SDmode on power7.
19363 (movdd splitters): Likewise.
19364 (movdd_hardfloat32): Likewise.
19365 (movdd_softfloat32): Likewise.
19366 (movdd_hardfloat64_mfpgpr): Likewise.
19367 (movdd_hardfloat64): Likewise.
19368 (movdd_softfloat64): Likewise.
19370 * config/rs6000/rs6000.md (FMOVE64): New iterators to combine
19371 64-bit binary and decimal floating point moves.
19372 (FMOVE64X): Likewise.
19373 (movdf): Combine 64-bit binary and decimal floating point moves.
19374 Combine power6x (mfpgpr) moves with other moves by using
19375 conditional constraits (wg).
19376 (mov<mode> for DFmode/DDmode): Likewise.
19377 (DFmode/DDmode splitters): Likewise.
19378 (movdf_hardfloat32): Likewise.
19379 (mov<mode>_hardfloat32 for DFmode/DDmode): Likewise.
19380 (movdf_softfloat32): Likewise.
19381 (movdf_hardfloat64_mfpgpr): Likewise.
19382 (movdf_hardfloat64): Likewise.
19383 (mov<mode>_hardfloat64 for DFmode/DDmode): Likewise.
19384 (movdf_softfloat64): Likewise.
19385 (mov<mode>_softfloat64 for DFmode/DDmode): Likewise.
19386 (reload_<mode>_load): Move to later in the file so they aren't in
19387 the middle of the floating point move insns.
19388 (reload_<mode>_store): Likewise.
19390 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wg
19393 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print out wg
19394 constraint if -mdebug=reg.
19395 (rs6000_initi_hard_regno_mode_ok): Enable wg constraint if -mfpgpr.
19396 Enable using dd reload support if needed.
19398 * config/rs6000/dfp.md (movtd): Delete, combine with 128-bit
19399 binary and decimal floating point moves in rs6000.md.
19400 (movtd_internal): Likewise.
19402 * config/rs6000/rs6000.md (FMOVE128): Combine 128-bit binary and
19403 decimal floating point moves.
19405 (movtf_internal): Likewise.
19406 (mov<mode>_internal, TDmode/TFmode): Likewise.
19407 (movtf_softfloat): Likewise.
19408 (mov<mode>_softfloat, TDmode/TFmode): Likewise.
19410 * config/rs6000/rs6000.md (movdi_mfpgpr): Delete, combine with
19411 movdi_internal64, using wg constraint for move direct operations.
19412 (movdi_internal64): Likewise.
19414 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print
19415 MODES_TIEABLE_P for selected modes. Print the numerical value of
19416 the various virtual registers. Use GPR/FPR first/last values,
19417 instead of hard coding the register numbers. Print which modes
19418 have reload functions registered.
19419 (rs6000_option_override_internal): If -mdebug=reg, trace the options
19420 settings before/after setting cpu, target and subtarget settings.
19421 (rs6000_secondary_reload_trace): Improve the RTL dump for -mdebug=addr
19422 and for secondary reload failures in rs6000_secondary_reload_inner.
19423 (rs6000_secondary_reload_fail): Likewise.
19424 (rs6000_secondary_reload_inner): Likewise.
19426 * config/rs6000/rs6000.md (FIRST_GPR_REGNO): Add convenience
19427 macros for first/last GPR and FPR registers.
19428 (LAST_GPR_REGNO): Likewise.
19429 (FIRST_FPR_REGNO): Likewise.
19430 (LAST_FPR_REGNO): Likewise.
19432 * config/rs6000/vector.md (mul<mode>3): Use the combined macro
19433 VECTOR_UNIT_ALTIVEC_OR_VSX_P instead of separate calls to
19434 VECTOR_UNIT_ALTIVEC_P and VECTOR_UNIT_VSX_P.
19435 (vcond<mode><mode>): Likewise.
19436 (vcondu<mode><mode>): Likewise.
19437 (vector_gtu<mode>): Likewise.
19438 (vector_gte<mode>): Likewise.
19439 (xor<mode>3): Don't allow logical operations on TImode in 32-bit
19440 to prevent the compiler from converting DImode operations to TImode.
19441 (ior<mode>3): Likewise.
19442 (and<mode>3): Likewise.
19443 (one_cmpl<mode>2): Likewise.
19444 (nor<mode>3): Likewise.
19445 (andc<mode>3): Likewise.
19447 * config/rs6000/constraints.md (wt constraint): New constraint
19448 that returns VSX_REGS if TImode is allowed in VSX registers.
19450 * config/rs6000/predicates.md (easy_fp_constant): 0.0f is an easy
19451 constant under VSX.
19453 * config/rs6000/rs6000-modes.def (PTImode): Define, PTImode is
19454 similar to TImode, but it is restricted to being in the GPRs.
19456 * config/rs6000/rs6000.opt (-mvsx-timode): New switch to allow
19457 TImode to occupy a single VSX register.
19459 * config/rs6000/rs6000-cpus.def (ISA_2_6_MASKS_SERVER): Default to
19460 -mvsx-timode for power7/power8.
19461 (power7 cpu): Likewise.
19462 (power8 cpu): Likewise.
19464 * config/rs6000/rs6000.c (rs6000_hard_regno_nregs_internal): Make
19465 sure that TFmode/TDmode take up two registers if they are ever
19466 allowed in the upper VSX registers.
19467 (rs6000_hard_regno_mode_ok): If -mvsx-timode, allow TImode in VSX
19469 (rs6000_init_hard_regno_mode_ok): Likewise.
19470 (rs6000_debug_reg_global): Add debugging for PTImode and wt
19471 constraint. Print if LRA is turned on.
19472 (rs6000_option_override_internal): Give an error if -mvsx-timode
19473 and VSX is not enabled.
19474 (invalid_e500_subreg): Handle PTImode, restricting it to GPRs. If
19475 -mvsx-timode, restrict TImode to reg+reg addressing, and PTImode
19476 to reg+offset addressing. Use PTImode when checking offset
19477 addresses for validity.
19478 (reg_offset_addressing_ok_p): Likewise.
19479 (rs6000_legitimate_offset_address_p): Likewise.
19480 (rs6000_legitimize_address): Likewise.
19481 (rs6000_legitimize_reload_address): Likewise.
19482 (rs6000_legitimate_address_p): Likewise.
19483 (rs6000_eliminate_indexed_memrefs): Likewise.
19484 (rs6000_emit_move): Likewise.
19485 (rs6000_secondary_reload): Likewise.
19486 (rs6000_secondary_reload_inner): Handle PTImode. Allow 64-bit
19487 reloads to fpr registers to continue to use reg+offset addressing,
19488 but 64-bit reloads to altivec registers need reg+reg addressing.
19489 Drop test for PRE_MODIFY, since VSX loads/stores no longer support
19490 it. Treat LO_SUM like a PLUS operation.
19491 (rs6000_secondary_reload_class): If type is 64-bit, prefer to use
19492 FLOAT_REGS instead of VSX_RGS to allow use of reg+offset addressing.
19493 (rs6000_cannot_change_mode_class): Do not allow TImode in VSX
19494 registers to share a register with a smaller sized type, since VSX
19495 puts scalars in the upper 64-bits.
19496 (print_operand): Add support for PTImode.
19497 (rs6000_register_move_cost): Use VECTOR_MEM_VSX_P instead of
19498 VECTOR_UNIT_VSX_P to catch types that can be loaded in VSX
19499 registers, but don't have arithmetic support.
19500 (rs6000_memory_move_cost): Add test for VSX.
19501 (rs6000_opt_masks): Add -mvsx-timode.
19503 * config/rs6000/vsx.md (VSm): Change to use 64-bit aligned moves
19506 (VSr): Use wt constraint for TImode.
19507 (VSv): Drop TImode support.
19508 (vsx_movti): Delete, replace with versions for 32-bit and 64-bit.
19509 (vsx_movti_64bit): Likewise.
19510 (vsx_movti_32bit): Likewise.
19511 (vec_store_<mode>): Use VSX iterator instead of vector iterator.
19512 (vsx_and<mode>3): Delete use of '?' constraint on inputs, just put
19513 one '?' on the appropriate output constraint. Do not allow TImode
19514 logical operations on 32-bit systems.
19515 (vsx_ior<mode>3): Likewise.
19516 (vsx_xor<mode>3): Likewise.
19517 (vsx_one_cmpl<mode>2): Likewise.
19518 (vsx_nor<mode>3): Likewise.
19519 (vsx_andc<mode>3): Likewise.
19520 (vsx_concat_<mode>): Likewise.
19521 (vsx_xxpermdi_<mode>): Fix thinko for non V2DF/V2DI modes.
19523 * config/rs6000/rs6000.h (MASK_VSX_TIMODE): Map from
19524 OPTION_MASK_VSX_TIMODE.
19525 (enum rs6000_reg_class_enum): Add RS6000_CONSTRAINT_wt.
19526 (STACK_SAVEAREA_MODE): Use PTImode instead of TImode.
19528 * config/rs6000/rs6000.md (INT mode attribute): Add PTImode.
19529 (TI2 iterator): New iterator for TImode, PTImode.
19530 (wd mode attribute): Add values for vector types.
19531 (movti_string): Replace TI move operations with operations for TImode
19532 and PTImode. Add support for TImode being allowed in VSX registers.
19533 (mov<mode>_string, TImode/PTImode): Likewise.
19534 (movti_ppc64): Likewise.
19535 (mov<mode>_ppc64, TImode/PTImode): Likewise.
19536 (TI mode splitters): Likewise.
19538 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wt
19541 2013-03-20 Marc Glisse <marc.glisse@inria.fr>
19543 PR tree-optimization/56355
19544 * fold-const.c (tree_binary_nonnegative_warnv_p) <MULT_EXPR>:
19545 Also handle integers with undefined overflow.
19547 2013-03-20 Catherine Moore <clm@codesourcery.com>
19548 Maciej W. Rozycki <macro@codesourcery.com>
19549 Tom de Vries <tom@codesourcery.com>
19550 Nathan Sidwell <nathan@codesourcery.com>
19551 Iain Sandoe <iain@codesourcery.com>
19552 Nathan Froyd <froydnj@codesourcery.com>
19553 Chao-ying Fu <fu@mips.com>
19555 * doc/extend.texi: (micromips, nomicromips, nocompression):
19556 Document new function attributes.
19557 * doc/invoke.texi (minterlink-compressed, mmicromips,
19558 m14k, m14ke, m14kec): Document new options.
19559 (minterlink-mips16): Update documentation.
19560 * doc/md.texi (ZC, ZD): Document new constraints.
19561 * configure.ac (gcc_cv_as_micromips): Check if linker
19562 supports the .set micromips directive.
19563 * configure: Regenerate.
19564 * config.in: Regenerate.
19565 * config/mips/mips-tables.opt: Regenerate.
19566 * config/mips/micromips.md: New file.
19567 * constraints.md (ZC, ZD): New constraints.
19568 * config/mips/predicates.md (movep_src_register): New predicate.
19569 (movep_src_operand): New predicate.
19570 (non_volatile_mem_operand): New predicate.
19571 * config/mips/mips.md (multimem): New type.
19572 (length): Differentiate between 17-bit and 18-bit branch offsets.
19573 (MOVEP1, MOVEP2): New mode iterator.
19574 (mov_<load>l): Use ZC constraint.
19575 (mov_<load>r): Likewise.
19576 (mov_<store>l): Likewise.
19577 (mov_<store>r): Likewise.
19578 (*branch_equality<mode>_inverted): Add microMIPS support.
19579 (*branch_equality<mode>): Likewise.
19580 (*jump_absolute): Likewise.
19581 (indirect_jump_<mode>): Likewise.
19582 (tablejump_<mode>): Likewise.
19583 (<optab>_internal): Likewise.
19584 (sibcall_internal): Likewise.
19585 (sibcall_value_internal): Likewise.
19586 (prefetch): Use constraint ZD.
19587 * config/mips/mips.opt (minterlink-compressed): New option.
19588 (minterlink-mips16): Now an alias for minterlink-compressed.
19589 (mmicromips): New option.
19590 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
19591 (compare_and_swap_12): Likewise.
19592 (sync_add<mode>): Likewise.
19593 (sync_<optab>_12): Likewise.
19594 (sync_old_<optab>_12): Likewise.
19595 (sync_new_<optab>_12): Likewise.
19596 (sync_nand_12): Likewise.
19597 (sync_old_nand_12): Likewise.
19598 (sync_new_nand_12): Likewise.
19599 (sync_sub<mode>): Likewise.
19600 (sync_old_add<mode>): Likewise.
19601 (sync_old_sub<mode>): Likewise.
19602 (sync_new_add<mode>): Likewise.
19603 (sync_new_sub<mode>): Likewise.
19604 (sync_<optab><mode>): Likewise.
19605 (sync_old_<optab><mode>): Likewise.
19606 (sync_new_<optab><mode>): Likewise.
19607 (sync_nand<mode>): Likewise.
19608 (sync_old_nand<mode>): Likewise.
19609 (sync_new_nand<mode>): Likewise.
19610 (sync_lock_test_and_set<mode>): Likewise.
19611 (test_and_set_12): Likewise.
19612 (atomic_compare_and_swap<mode>): Likewise.
19613 (atomic_exchange<mode>_llsc): Likewise.
19614 (atomic_fetch_add<mode>_llsc): Likewise.
19615 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
19616 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
19617 (umips_save_restore_pattern_p): Likewise.
19618 (umips_load_store_pair_p): Likewise.
19619 (umips_output_load_store_pair): Likewise.
19620 (umips_movep_target_p): Likewise.
19621 (umips_12bit_offset_address_p): Likewise.
19622 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
19623 (mips_base_mips16): Rename this...
19624 (mips_base_compression_flags): ...to this. Update all uses.
19625 (mips_attribute_table): Add micromips, nomicromips and nocompression.
19626 (mips_mips16_decl_p): Delete.
19627 (mips_nomips16_decl_p): Delete.
19628 (mips_get_compress_on_flags): New function.
19629 (mips_get_compress_off_flags): New function.
19630 (mips_get_compress_mode): New function.
19631 (mips_get_compress_on_name): New function.
19632 (mips_get_compress_off_name): New function.
19633 (mips_insert_attributes): Support multiple compression types.
19634 (mips_merge_decl_attributes): Likewise.
19635 (umips_12bit_offset_address_p): New function.
19636 (mips_start_function_definition): Emit .set micromips directive.
19637 (mips_call_may_need_jalx_p): New function.
19638 (mips_function_ok_for_sibcall): Add microMIPS support.
19639 (mips_print_operand_punctuation): Support short delay slots and
19641 (umips_swm_mask, umips_swm_encoding): New.
19642 (umips_build_save_restore): New function.
19643 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
19644 (was_mips16_p): Remove.
19645 (old_compression_mode): New.
19646 (mips_set_compression_mode): New function.
19647 (mips_set_current_function): Add microMIPS support.
19648 (mips_option_override): Likewise.
19649 (umips_save_restore_pattern_p): New function.
19650 (umips_output_save_restore): New function.
19651 (umips_load_store_pair_p_1): New function.
19652 (umips_load_store_pair_p): New function.
19653 (umips_output_load_store_pair_1): New function.
19654 (umips_output_load_store_pair): New function.
19655 (umips_movep_target_p) New function.
19656 (mips_prepare_pch_save): Add microMIPS support.
19657 * config/mips/mips.h (TARGET_COMPRESSION): New.
19658 (TARGET_CPU_CPP_BUILTINS): Update macro
19659 to use new compression flags and to support microMIPS.
19660 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
19661 (MIPS_ARCH_FLOAT_SPEC): Likewise.
19662 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
19663 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
19664 (ASM_SPEC): Support mmicromips and mno-micromips.
19665 (M16STORE_REG_P): New macro.
19666 (MIPS_CALL): Support TARGET_MICROMIPS.
19667 (MICROMIPS_J): New macro.
19668 (mips_base_mips16): Rename this...
19669 (mips_base_compression_flags): ...to this.
19670 (UMIPS_12BIT_OFFSET_P): New macro.
19671 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
19672 (MULTILIB_DIRNAMES): Likewise.
19673 2013-03-20 Richard Biener <rguenther@suse.de>
19675 PR tree-optimization/56661
19676 * tree-ssa-sccvn.c (visit_use): Only value-number calls if
19677 the result does not have to be distinct.
19679 2013-03-20 Richard Biener <rguenther@suse.de>
19681 * tree-inline.c (copy_tree_body_r): Sync MEM_REF code with
19684 2013-03-20 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
19685 Steven Bosscher <steven@gcc.gnu.org>
19687 PR rtl-optimization/56605
19688 * loop-iv.c (implies_p): Handle equal RTXs and subregs.
19690 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
19693 * config/i386/i386.md (*movdi_internal): Handle broken assemblers
19694 that require movd instead of movq.
19696 2013-03-20 Richard Biener <rguenther@suse.de>
19698 * tree-ssa-structalias.c (struct variable_info): Add pointer
19699 to the first field of an aggregate with sub-vars. Make
19700 this and the pointer to the next subfield its ID.
19701 (vi_next): New function.
19702 (nothing_id, anything_id, readonly_id, escaped_id, nonlocal_id,
19703 storedanything_id, integer_id): Increment by one.
19704 (new_var_info, get_call_vi, lookup_call_clobber_vi,
19705 get_call_clobber_vi): Adjust.
19706 (solution_set_expand): Simplify and speedup.
19707 (solution_set_add): Inline into ...
19708 (set_union_with_increment): ... this. Adjust accordingly.
19709 (do_sd_constraint): Likewise.
19710 (do_ds_constraint): Likewise.
19711 (do_complex_constraint): Simplify.
19712 (build_pred_graph): Adjust.
19713 (solve_graph): Likewise. Simplify and speedup.
19714 (get_constraint_for_ssa_var, get_constraint_for_ptr_offset,
19715 get_constraint_for_component_ref, get_constraint_for_1,
19716 first_vi_for_offset, first_or_preceding_vi_for_offset,
19717 create_function_info_for, create_variable_info_for_1,
19718 create_variable_info_for, intra_create_variable_infos): Adjust.
19719 (init_base_vars): Push NULL for ID zero.
19720 (compute_points_to_sets): Adjust.
19722 2013-03-20 Richard Biener <rguenther@suse.de>
19724 * cfgloop.c (verify_loop_structure): Streamline and avoid
19725 ICEing on corrupt loop tree.
19726 * graph.c (draw_cfg_nodes_for_loop): Avoid ICEing on corrupt
19729 2013-03-20 Richard Biener <rguenther@suse.de>
19731 * tree-vect-loop-manip.c (slpeel_can_duplicate_loop_p): Do not
19732 check whether an SSA update is needed.
19734 2013-03-20 Richard Sandiford <rdsandiford@googlemail.com>
19736 * config/mips/constraints.md (T): Rename to...
19740 * config/mips/mips.md (*movdi_64bit, *movdi_64bit_mips16)
19741 (*mov<mode>_internal, *mov<mode>_mips16): Update accordingly.
19743 2013-03-19 Ian Bolton <ian.bolton@arm.com>
19745 * config/aarch64/aarch64.md (*sub<mode>3_carryin): New pattern.
19746 (*subsi3_carryin_uxtw): Likewise.
19748 2013-03-19 Ian Bolton <ian.bolton@arm.com>
19750 * config/aarch64/aarch64.md (*ror<mode>3_insn): New pattern.
19751 (*rorsi3_insn_uxtw): Likewise.
19753 2013-03-19 Ian Bolton <ian.bolton@arm.com>
19755 * config/aarch64/aarch64.md (*extr<mode>5_insn): New pattern.
19756 (*extrsi5_insn_uxtw): Likewise.
19758 2013-03-19 Richard Biener <rguenther@suse.de>
19760 PR tree-optimization/56273
19761 * passes.c (init_optimization_passes): Move second VRP after DOM.
19763 2013-03-19 Uros Bizjak <ubizjak@gmail.com>
19765 * config/i386/i386.md (*movti_internal): Merge from
19766 *movti_internal_rex64 and *movti_internal_sse. Use x64 isa attribute.
19767 (*movdi_internal): Merge with *movdi_internal_rex64. Use x64 and
19768 nox64 isa attributes.
19770 2013-03-18 Richard Biener <rguenther@suse.de>
19772 * tree-ssa-structalias.c (find): Use gcc_checking_assert.
19774 (merge_node_constraints): Likewise.
19775 (build_succ_graph): Likewise.
19776 (valid_graph_edge): Inline into single caller.
19777 (unify_nodes): Likewise. Use bitmap_set_bit return value
19779 (scc_visit): Fix formatting and variable use.
19780 (do_sd_constraint): Use gcc_checking_assert.
19781 (do_ds_constraint): Likewise.
19782 (do_complex_constraint): Likewise.
19783 (condense_visit): Likewise. Cleanup.
19784 (dump_pred_graph): New function.
19785 (perform_var_substitution): Dump the pred-graph before
19786 variable substitution.
19787 (find_equivalent_node): Use gcc_checking_assert.
19788 (rewrite_constraints): Guard checking loop with ENABLE_CHECKING.
19790 2013-03-18 Richard Biener <rguenther@suse.de>
19792 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
19793 Remove cond_expr_stmt_list argument and do not gimplify the
19795 (vect_loop_versioning): Adjust.
19796 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref):
19797 Cleanup to use less temporaries.
19798 (vect_create_data_ref_ptr): Cleanup.
19800 2013-03-18 Jakub Jelinek <jakub@redhat.com>
19802 PR tree-optimization/56635
19803 * fold-const.c (operand_equal_p): For MEM_REF and TARGET_MEM_REF,
19804 require types_compatible_p types.
19806 2013-03-18 Nick Clifton <nickc@redhat.com>
19808 * config/stormy16/stormy16.c (xstormy16_expand_prologue): Remove
19809 spurious backslash.
19811 * config/mn10300/mn10300.c (mn10300_get_live_callee_saved_regs):
19812 Add missing line to comment describing function.
19814 2013-03-18 Richard Biener <rguenther@suse.de>
19816 PR tree-optimization/56210
19817 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
19818 Handle string / character search functions.
19819 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
19821 2013-03-18 Richard Biener <rguenther@suse.de>
19823 PR middle-end/56483
19824 * cfgexpand.c (expand_gimple_cond): Inline gimple_cond_single_var_p
19825 and implement properly.
19826 * gimple.h (gimple_cond_single_var_p): Remove.
19828 2013-03-18 Richard Biener <rguenther@suse.de>
19830 * tree-data-ref.h (find_data_references_in_loop): Declare.
19831 * tree-data-ref.c (get_references_in_stmt): Use a stack
19832 vector pre-allocated in the callers.
19833 (find_data_references_in_stmt): Adjust.
19834 (graphite_find_data_references_in_stmt): Likewise.
19835 (create_rdg_vertices): Likewise.
19836 (find_data_references_in_loop): Export.
19837 * tree-vect-data-refs.c (vect_analyze_data_ref_dependences):
19838 Compute dependences here...
19839 (vect_analyze_data_refs): ...not here. When we encounter
19840 a non-vectorizable data reference in basic-block vectorization
19841 truncate the data reference vector. Do not bother to
19842 fixup data-dependence information for gather loads.
19843 * tree-vect-slp.c (vect_slp_analyze_bb_1): Check the number
19844 of data references, as reported.
19846 2013-03-18 Richard Biener <rguenther@suse.de>
19848 PR tree-optimization/3713
19849 * tree-ssa-sccvn.c (visit_copy): Simplify. Always propagate
19850 has_constants and expr.
19851 (stmt_has_constants): Properly valueize SSA names when deciding
19852 whether the stmt has constants.
19854 2013-03-18 Richard Biener <rguenther@suse.de>
19856 * tree-ssa-loop-manip.c (find_uses_to_rename): Do not scan the
19857 whole function when there is nothing to do.
19858 * tree-ssa-loop.c (pass_vectorize): Remove TODO_update_ssa.
19859 * tree-vectorizer.c (vectorize_loops): Update virtual and
19860 loop-closed SSA once.
19861 * tree-vect-loop.c (vect_transform_loop): Do not update SSA here.
19863 2013-03-18 Richard Biener <rguenther@suse.de>
19865 PR middle-end/56113
19866 * domwalk.c (bb_postorder): New global static.
19867 (cmp_bb_postorder): New function.
19868 (walk_dominator_tree): Replace scheme imposing an order for
19869 visiting dominator sons by one sorting them at the time they
19870 are pushed on the stack.
19872 2013-03-18 Richard Biener <rguenther@suse.de>
19874 PR tree-optimization/39326
19875 * tree-ssa-loop-im.c (refs_independent_p): Exploit symmetry.
19876 (struct mem_ref): Replace mem member with ao_ref typed member.
19877 (MEM_ANALYZABLE): Adjust.
19878 (memref_eq): Likewise.
19879 (mem_ref_alloc): Likewise.
19880 (gather_mem_refs_stmt): Likewise.
19881 (mem_refs_may_alias_p): Use the ao_ref to query the alias oracle.
19882 (execute_sm_if_changed_flag_set): Adjust.
19883 (execute_sm): Likewise.
19884 (ref_always_accessed_p): Likewise.
19885 (refs_independent_p): Likewise.
19886 (can_sm_ref_p): Likewise.
19888 2013-03-18 Jakub Jelinek <jakub@redhat.com>
19891 * tree.c (tree_int_cst_min_precision): For integer_zerop (value)
19892 return 1 even for !unsignedp.
19894 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
19896 * config/i386/i386.md (isa): Add x64 and nox64.
19897 (enabled): Define x64 for TARGET_64BIT and nox64 for !TARGET_64BIT.
19898 (*pushtf): Enable *roF alternative for x64 isa only.
19899 (*pushxf): Merge with *pushxf_nointeger. Use Yx*r constraint. Set
19900 mode attribute of integer alternatives to DImode for TARGET_64BIT.
19901 (*pushdf): Merge with *pushdf_rex64. Use x64 and nox64 isa attributes.
19902 (*movtf_internal): Merge from *movtf_internal_rex64 and
19903 *movtf_internal_sse. Use x64 and nox64 isa attributes.
19904 (*movxf_internal): Merge with *movxf_internal_rex64. Use x64 and
19905 nox64 isa attributes.
19906 (*movdf_internal): Merge with *movdf_internal_rex64. Use x64 and
19907 nox64 isa attributes.
19908 * config/i386/constraints.md (Yd): Do not set for TARGET_64BIT.
19910 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
19912 * config/alpha/alpha.c (TARGET_LRA_P): New define.
19914 2013-03-17 Jakub Jelinek <jakub@redhat.com>
19917 * config/arm/arm.h (REG_CLASS_NAMES): Add "SFP_REG" and "AFP_REG"
19918 class names. Remove trailing comma after "ALL_REGS".
19920 2013-03-16 Jan Hubicka <jh@suse.cz>
19922 * cgraph.h (cgraph_get_create_real_symbol_node): Declare.
19923 * cgraph.c (cgraph_get_create_real_symbol_node): New function.
19924 * cgrpahbuild.c: Use cgraph_get_create_real_symbol_node instead
19925 of cgraph_get_create_node.
19926 * ipa-prop.c (ipa_make_edge_direct_to_target): Likewise.
19928 2013-03-16 Jason Merrill <jason@redhat.com>
19931 * dwarf2out.c (gen_generic_params_dies): Indicate default arguments
19932 with DW_AT_default_value.
19934 2013-03-16 Jakub Jelinek <jakub@redhat.com>
19936 * BASE-VER: Set to 4.9.0.
19938 2013-03-14 Andi Kleen <ak@linux.intel.com>
19941 * doc/extend.texi: Document __ATOMIC_HLE_ACQUIRE,
19942 __ATOMIC_HLE_RELEASE. Document __builtin_ia32 TSX intrincs.
19943 Document _x* TSX intrinsics.
19945 2013-03-14 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
19946 David Holsgrove <david.holsgrove@xilinx.com>
19948 * configure.ac: Add MicroBlaze TLS support detection.
19949 * configure: Regenerate.
19950 * config/microblaze/microblaze-protos.h
19951 (microblaze_cannot_force_const_mem, microblaze_tls_referenced_p,
19952 symbol_mentioned_p, label_mentioned_p): Add prototypes.
19953 * config/microblaze/microblaze.c (microblaze_address_type): Add
19954 ADDRESS_TLS and tls_reloc address types.
19955 (microblaze_address_info): Add tls_reloc.
19956 (TARGET_HAVE_TLS): Define.
19957 (get_tls_get_addr, microblaze_tls_symbol_p, microblaze_tls_operand_p_1,
19958 microblaze_tls_referenced_p, microblaze_cannot_force_const_mem,
19959 symbol_mentioned_p, label_mentioned_p, tls_mentioned_p,
19960 load_tls_operand, microblaze_call_tls_get_addr,
19961 microblaze_legitimize_tls_address): New functions.
19962 (microblaze_classify_unspec): Handle UNSPEC_TLS.
19963 (get_base_reg): Use microblaze_tls_symbol_p.
19964 (microblaze_classify_address): Handle TLS.
19965 (microblaze_legitimate_pic_operand): Use symbol_mentioned_p,
19966 label_mentioned_p and microblaze_tls_referenced_p.
19967 (microblaze_legitimize_address): Handle TLS.
19968 (microblaze_address_insns): Handle ADDRESS_TLS.
19969 (pic_address_needs_scratch): Handle TLS.
19970 (print_operand_address): Handle TLS.
19971 (microblaze_expand_prologue): Check TLS_NEEDS_GOT.
19972 (microblaze_expand_move): Handle TLS.
19973 (microblaze_legitimate_constant_p): Check
19974 microblaze_cannot_force_const_mem and microblaze_tls_symbol_p.
19975 (TARGET_CANNOT_FORCE_CONST_MEM): Define.
19976 * config/microblaze/microblaze.h (TLS_NEEDS_GOT): Define
19977 (PIC_OFFSET_TABLE_REGNUM): Set.
19978 * config/microblaze/linux.h (TLS_NEEDS_GOT): Define.
19979 * config/microblaze/microblaze.md (UNSPEC_TLS): Define.
19980 (addsi3, movsi_internal2, movdf_internal): Update constraints
19981 * config/microblaze/predicates.md (arith_plus_operand): Define
19982 (move_operand): Redefine as move_src_operand,
19983 check microblaze_tls_referenced_p.
19985 2013-03-14 Ian Bolton <ian.bolton@arm.com>
19987 * config/aarch64/aarch64.md: (*and<mode>3nr_compare0): Use CC_NZ.
19988 (*and_<SHIFT:optab><mode>3nr_compare0): Likewise.
19990 2013-03-14 Ian Bolton <ian.bolton@arm.com>
19992 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return correct
19995 2013-03-14 Jakub Jelinek <jakub@redhat.com>
19997 PR tree-optimization/53265
19998 * common.opt (Waggressive-loop-optimizations): New option.
19999 * tree-ssa-loop-niter.c: Include tree-pass.h.
20000 (do_warn_aggressive_loop_optimizations): New function.
20001 (record_estimate): Call it. Don't add !is_exit bounds to loop->bounds
20002 if number_of_latch_executions returned constant.
20003 (estimate_numbers_of_iterations_loop): Call number_of_latch_executions
20004 early. If number_of_latch_executions returned constant, set
20005 nb_iterations_upper_bound back to it.
20006 * cfgloop.h (struct loop): Add warned_aggressive_loop_optimizations
20008 * Makefile.in (tree-ssa-loop-niter.o): Depend on $(TREE_PASS_H).
20009 * doc/invoke.texi (-Wno-aggressive-loop-optimizations): Document.
20011 * config/aarch64/t-aarch64-linux (MULTARCH_DIRNAME): Remove.
20012 (MULTILIB_OSDIRNAMES): Set.
20013 * genmultilib: If defaultosdirname doesn't start with :: , set
20014 defaultosdirname2 instead, clear it and emit two . multilib_raw
20015 entries instead of just one.
20017 2013-03-14 Kaz Kojima <kkojima@gcc.gnu.org>
20019 * config/sh/linux.h (TARGET_DEFAULT): Remove MASK_USERMODE.
20020 (SUBTARGET_OVERRIDE_OPTIONS): Set TARGET_USERMODE as default.
20021 * config/sh/netbsd-elf.h (TARGET_DEFAULT): Remove MASK_USERMODE.
20022 (SUBTARGET_OVERRIDE_OPTIONS): New.
20024 2013-03-13 Oleg Endo <olegendo@gcc.gnu.org>
20027 * config/sh/sh.opt (FPU_SINGLE_ONLY): New mask.
20028 (musermode): Convert to Var(TARGET_USERMODE).
20029 * config/sh/sh.h (SELECT_SH2A_SINGLE_ONLY, SELECT_SH4_SINGLE_ONLY,
20030 MASK_ARCH): Add MASK_FPU_SINGLE_ONLY.
20031 * config/sh/sh.c (sh_option_override): Use
20032 TARGET_FPU_DOUBLE || TARGET_FPU_SINGLE_ONLY for call-fp case.
20033 * config/sh/sh.md (udivsi3_i1, divsi3_i1): Remove ! TARGET_SH4
20035 (udivsi3_i4, divsi3_i4): Use TARGET_FPU_DOUBLE condition instead of
20037 (udivsi3_i4_single, divsi3_i4_single): Use
20038 TARGET_FPU_SINGLE_ONLY || TARGET_FPU_DOUBLE instead of TARGET_HARD_SH4.
20040 2013-03-13 Dave Korn <dave.korn.cygwin@....>
20042 * config/i386/cygwin.h (SHARED_LIBGCC_SPEC): Make shared libgcc the
20045 2013-03-13 Richard Biener <rguenther@suse.de>
20047 PR tree-optimization/56608
20048 * tree-vect-slp.c (vect_schedule_slp): Do not remove scalar
20049 calls when vectorizing basic-blocks.
20051 2013-03-13 Jakub Jelinek <jakub@redhat.com>
20054 * config.gcc: On arm, mips, sh and sparc add vxworks-dummy.h to
20057 2013-03-12 Jakub Jelinek <jakub@redhat.com>
20059 * doc/invoke.texi (-Waddr-space-convert): Move into the table earlier.
20061 2013-03-11 Jan Hubicka <jh@suse.cz>
20064 * lto-streamer-out.c (output_symbol_p): Skip references from
20065 constructors of external variables.
20067 2013-03-11 Jan Hubicka <jh@suse.cz>
20069 PR middle-end/56571
20070 * valtrack.c (cleanup_auto_inc_dec): Unshare clobbers originating
20072 * emit-rtl.c (verify_rtx_sharing): Likewise.
20073 (copy_insn_1): Likewise.
20074 * rtl.c (copy_rtx): Likewise.
20076 2013-03-11 Georg-Johann Lay <avr@gjlay.de>
20079 * config/avr/avr.c (avr_print_operand): Add space after '%c' in
20080 output_operand_lossage message.
20082 2013-03-11 Richard Earnshaw <rearnsha@arm.com>
20085 * arm.c (shift_op): Validate RTL pattern on the fly.
20086 (arm_print_operand, case 'S'): Don't use shift_operator to validate
20089 2013-03-10 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
20092 * config/pa/pa.md (call_value): Check for calls to powf and direct to
20093 new call patterns that clobber %fr12.
20094 (call_val_powf, call_val_powf_pic, call_val_powf_64bit): New insn,
20095 split and postreload patterns.
20096 * config/pa/pa.c (pa_conditional_register_usage): Revert marking
20097 registers %fr12 and %fr12R as call used.
20099 2013-03-09 Steven Bosscher <steven@gcc.gnu.org>
20101 * dse.c (delete_dead_store_insn): Respect TDF_DETAILS.
20102 (canon_address, record_store, replace_read, check_mem_read_rtx,
20103 scan_insn, dse_step1, dse_step2_init, dse_step2_spill,
20104 dse_step4, dse_step5_nospill, dse_step5_spill, dse_step6,
20105 rest_of_handle_dse): Likewise.
20107 2013-03-09 Richard Sandiford <rdsandiford@googlemail.com>
20109 PR middle-end/56524
20110 * tree.h (tree_optimization_option): Rename target_optabs to optabs.
20112 (TREE_OPTIMIZATION_OPTABS): Update after previous field change.
20113 (TREE_OPTIMIZATION_BASE_OPTABS): New macro.
20114 (save_optabs_if_changed): Replace with...
20115 (init_tree_optimization_optabs): ...this.
20116 * optabs.c (save_optabs_if_changed): Rename to...
20117 (init_tree_optimization_optabs): ...this. Take the optimization node
20118 as argument. Do nothing if the base optabs are already correct.
20119 Reuse the existing TREE_OPTIMIZATION_OPTABS memory if we need
20120 to recompute optabs.
20121 * function.h (function): Remove optabs field.
20122 * function.c (invoke_set_current_function_hook): Call
20123 init_tree_optimization_optabs. Use the result to initialize
20126 2013-02-27 Aldy Hernandez <aldyh@redhat.com>
20128 * trans-mem.c (expand_transaction): Do not set PR_INSTRUMENTEDCODE
20129 if GTMA_HAS_NO_INSTRUMENTATION.
20130 (generate_tm_state): Keep GTMA_HAS_NO_INSTRUMENTATION bit.
20131 (ipa_tm_transform_transaction): Set GTMA_HAS_NO_INSTRUMENTATION.
20132 * gimple.h (GTMA_HAS_NO_INSTRUMENTATION): Define.
20133 * gimple-pretty-print.c (dump_gimple_transaction): Handle
20134 GTMA_HAS_NO_INSTRUMENTATION.
20136 2013-03-08 Jakub Jelinek <jakub@redhat.com>
20138 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Don't link against
20141 2013-03-08 Marek Polacek <polacek@redhat.com>
20142 Jakub Jelinek <jakub@redhat.com>
20144 PR tree-optimization/56478
20145 * predict.c (is_comparison_with_loop_invariant_p): Change the
20146 type of loop_step to tree.
20147 (predict_loops): Adjust.
20148 (predict_iv_comparison): Perform the computations on double_ints.
20150 2013-03-08 Richard Biener <rguenther@suse.de>
20152 PR tree-optimization/56570
20153 * tree-cfg.c (verify_expr_location_1): Verify locations for
20155 * tree-sra.c (create_access_replacement): Strip locations
20156 from DECL_DEBUG_EXPRs.
20158 2013-03-08 Richard Biener <rguenther@suse.de>
20160 * tree-inline.c (expand_call_inline): Do not associate
20161 a BLOCK with the location in BLOCK_SOURCE_LOCATION.
20162 * tree-cfg.c (verify_location): Verify BLOCK_SOURCE_LOCATION.
20164 2013-03-08 Richard Biener <rguenther@suse.de>
20166 * tree-ssa-ter.c (is_replaceable_p): Do not TER across location
20167 or block changes with -Og. Fix for location / block encoding
20168 changes and PHI arguments with locations.
20170 2013-03-07 Steven Bosscher <steven@gcc.gnu.org>
20172 * bitmap.c (struct bitmap_descriptor_d): Use unsigned HOST_WIDEST_INT
20174 (struct output_info): Likewise.
20175 (register_overhead): Remove bad gcc_assert.
20176 (bitmap_find_bit): If there is only a single bitmap element, do not
20177 count a miss as a search.
20178 (print_statistics): Update for counter type changes.
20179 (dump_bitmap_statistics): Likewise. Print headers such that they
20180 are properly lined up with the printed counters.
20182 2013-03-07 Jakub Jelinek <jakub@redhat.com>
20184 PR tree-optimization/56559
20185 * tree-ssa-reassoc.c (zero_one_operation): When looking at rhs2,
20186 check that it has only a single use.
20188 2013-03-07 Richard Biener <rguenther@suse.de>
20190 * doc/invoke.texi (fwhole-program): Discourage use in combination
20193 2013-03-06 Jakub Jelinek <jakub@redhat.com>
20195 * config/arm/t-arm (TM_H, OPTIONS_H_EXTRA): Add arm-cores.def.
20197 PR tree-optimization/56539
20198 * tree-tailcall.c (adjust_return_value_with_ops): Use GSI_SAME_STMT
20199 instead of GSI_CONTINUE_LINKING as last argument to
20200 force_gimple_operand_gsi. Adjust function comment.
20202 * config/aarch64/t-aarch64 (TM_H, OPTIONS_H_EXTRA): Add
20205 PR middle-end/56548
20206 * expr.c (expand_cond_expr_using_cmove): When expanding cmove in
20207 promoted mode, convert the result back to the original mode.
20209 2013-03-06 Richard Biener <rguenther@suse.de>
20211 PR middle-end/56294
20212 * tree-into-ssa.c (insert_phi_nodes_for): Add dumping.
20213 (insert_updated_phi_nodes_compare_uids): New function.
20214 (update_ssa): Sort symbols_to_rename after UID before
20215 traversing it to insert PHI nodes.
20217 2013-03-06 Richard Biener <rguenther@suse.de>
20219 PR middle-end/50494
20220 * tree-vect-data-refs.c (vect_can_force_dr_alignment_p):
20221 Do not adjust alignment of DECL_IN_CONSTANT_POOL decls.
20224 2013-02-13 Richard Biener <rguenther@suse.de>
20227 * varasm.c (output_constant_def_1): Get the decl representing
20228 the constant as argument.
20229 (output_constant_def): Wrap output_constant_def_1.
20230 (make_decl_rtl): Use output_constant_def_1 with the decl
20231 representing the constant.
20232 (build_constant_desc): Optionally re-use a decl already
20233 representing the constant.
20234 (tree_output_constant_def): Adjust.
20236 2013-03-06 Joey Ye <joey.ye@arm.com>
20239 * gcc.c (convert_white_space): New function.
20240 (main): Handles white space in function name.
20242 2013-03-06 Oleg Endo <olegendo@gcc.gnu.org>
20245 * config/sh/sh.c (sh_option_override): Check for TARGET_DYNSHIFT
20246 instead of TARGET_SH2 for call-table case. Do not set sh_div_strategy
20247 to SH_DIV_CALL_TABLE for TARGET_SH2.
20248 * config.gcc (sh_multilibs): Add m2 and m2a to sh*-*-linux* multilib
20250 * doc/invoke.texi (SH options): Document mdiv= call-div1, call-fp,
20251 call-table options.
20253 2013-03-05 Sterling Augustine <saugustine@google.com>
20254 Cary Coutant <ccoutant@google.com>
20257 * dwarf2out.c (resolve_addr): Don't call
20258 remove_loc_list_addr_table_entries a second time for the same
20261 2013-03-05 Jakub Jelinek <jakub@redhat.com>
20264 * cfgexpand.c (expand_debug_parm_decl): Call copy_rtx on incoming.
20265 (avoid_complex_debug_insns): New function.
20266 (expand_debug_locations): Call it.
20268 PR rtl-optimization/56484
20269 * ifcvt.c (noce_process_if_block): If else_bb is NULL, avoid extending
20270 lifetimes of hard registers on small register class machines.
20272 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
20274 * config/microblaze/microblaze-protos.h: Rename
20275 microblaze_is_interrupt_handler to microblaze_is_interrupt_variant.
20276 * config/microblaze/microblaze.c (microblaze_attribute_table): Add
20278 (microblaze_fast_interrupt_function_p): New function.
20279 (microblaze_is_interrupt_handler): Rename to
20280 microblaze_is_interrupt_variant and add fast_interrupt check.
20281 (microblaze_must_save_register): Use microblaze_is_interrupt_variant.
20282 (save_restore_insns): Likewise.
20283 (compute_frame_size): Likewise.
20284 (microblaze_function_prologue): Add FAST_INTERRUPT_NAME.
20285 (microblaze_globalize_label): Likewise.
20286 * config/microblaze/microblaze.h: Define FAST_INTERRUPT_NAME.
20287 * config/microblaze/microblaze.md: Use wrapper
20288 microblaze_is_interrupt_variant.
20290 2013-03-05 Kai Tietz <ktietz@redhat.com>
20292 * sdbout.c (sdbout_one_type): Switch to current function's section
20293 supporting cold/hot.
20295 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
20297 * doc/invoke.texi (MicroBlaze): Add -mbig-endian, -mlittle-endian,
20300 2013-03-05 Jakub Jelinek <jakub@redhat.com>
20302 PR middle-end/56461
20303 * ggc-common.c (gt_pch_save): For ENABLE_VALGRIND_CHECKING,
20304 if VALGRIND_GET_VBITS is defined, temporarily make object
20305 memory all defined, and restore previous valgrind addressability
20306 and definability afterwards. Free this_object at the end.
20308 PR middle-end/56461
20309 * lra.c (lra): Call lra_clear_live_ranges if live_p,
20310 right before calling lra_create_live_ranges, also call it
20311 when clearing live_p. Only call lra_clear_live_ranges
20312 at the end if live_p.
20314 PR middle-end/56461
20315 * sched-deps.c (delete_dep_node): Free DEP_REPLACE.
20317 2013-03-05 Richard Biener <rguenther@suse.de>
20319 PR tree-optimization/56521
20320 * tree-ssa-sccvn.c (set_value_id_for_result): Always initialize
20323 2013-03-05 Steven Bosscher <steven@gcc.gnu.org>
20326 * except.h (remove_unreachable_eh_regions): New prototype.
20327 * except.c (remove_eh_handler_splicer): New function, split out
20328 of remove_eh_handler.
20329 (remove_eh_handler): Use remove_eh_handler_splicer. Add comment
20330 warning about running it on many EH regions one at a time.
20331 (remove_unreachable_eh_regions_worker): New function, walk the
20332 EH tree in depth-first order and remove non-marked regions.
20333 (remove_unreachable_eh_regions): New function.
20334 * tree-eh.c (mark_reachable_handlers): New function, split out
20335 from remove_unreachable_handlers.
20336 (remove_unreachable_handlers): Use mark_reachable_handlers and
20337 remove_unreachable_eh_regions.
20338 (remove_unreachable_handlers_no_lp): Use mark_reachable_handlers
20339 and remove_unreachable_eh_regions.
20341 2013-03-05 Richard Biener <rguenther@suse.de>
20343 PR middle-end/56525
20344 * loop-init.c (fix_loop_structure): Remove loops in two stages,
20345 not freeing them until the end.
20347 2013-03-05 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
20349 * config/s390/s390.h: Define DWARF2_ASM_LINE_DEBUG_INFO.
20351 2013-03-05 Richard Biener <rguenther@suse.de>
20353 PR tree-optimization/56270
20354 * tree-vect-slp.c (vect_schedule_slp): Clear vectorized stmts
20355 of loads after scheduling an SLP instance.
20357 2013-03-05 Jakub Jelinek <jakub@redhat.com>
20359 * Makefile.in (dg_target_exps): Add aarch64.exp, epiphany.exp and
20361 (check_gcc_parallelize): Run guality.exp as a separate job from
20362 vect.exp with unsorted.exp and $(dg_target_exps) separately from
20363 struct-layout-1.exp with stackalign.exp.
20365 * alias.c (init_alias_analysis): Clear reg_known_equiv_p bitmap.
20367 PR middle-end/56461
20368 * tree-vect-slp.c (vect_supported_load_permutation_p): Free
20369 load_index sbitmap even if some bit in it isn't set.
20371 PR middle-end/56461
20372 * tree-ssa-loop-niter.c (bb_queue): Remove typedef.
20373 (discover_iteration_bound_by_body_walk): Change queues to
20374 vec<vec<basic_block> > and queue to vec<basic_block>. Fix up
20375 spelling in comment. Call safe_push on queues[bound_index] directly.
20376 Release queues[queue_index] in every iteration unconditionally.
20377 Release bounds vector.
20379 PR middle-end/56461
20380 * tree-vect-stmts.c (free_stmt_vec_info_vec): Call
20381 free_stmt_vec_info on any left-over stmt_vec_info in the vector.
20382 * tree-vect-loop.c (vect_create_epilog_for_reduction): Release
20385 2013-03-05 Richard Biener <rguenther@suse.de>
20388 * tree-inline.c (remap_blocks_to_null): New function.
20389 (expand_call_inline): When expanding a call stmt without
20390 an associated block inline remap all callee blocks to NULL.
20392 2013-03-05 Jakub Jelinek <jakub@redhat.com>
20394 PR rtl-optimization/56494
20395 * simplify-rtx.c (simplify_truncation): If C is narrower than A,
20396 optimize (truncate:A (subreg:B (truncate:C X) 0)) into
20397 (subreg:A (truncate:C X) 0) instead of (truncate:A X).
20399 PR middle-end/56461
20400 * sel-sched-ir.c (free_sched_pools): Release
20401 succs_info_pool.stack[succs_info_pool.max_top] vectors too
20402 if succs_info_pool.max_top isn't -1.
20405 * opts.c (opts_obstack, opts_concat): Moved to...
20406 * opts-common.c (opts_obstack, opts_concat): ... here.
20408 2013-03-04 Jakub Jelinek <jakub@redhat.com>
20410 PR middle-end/56461
20411 * diagnostic.c (diagnostic_append_note): Save and restore old prefix.
20413 2013-03-04 Martin Jambor <mjambor@suse.cz>
20415 * tree-dfa.c (get_or_create_ssa_default_def): Use parameter fn in
20416 all appropriate places.
20418 2013-01-04 Eric Botcazou <ebotcazou@adacore.com>
20420 PR tree-optimization/56424
20421 * ipa-split.c (split_function): Do not set the RSO flag if result is
20422 not by reference and its type is a register type.
20424 2013-03-04 David Holsgrove <david.holsgrove@xilinx.com>
20426 * config/microblaze/microblaze.c (microblaze_valid_pic_const): New
20427 (microblaze_legitimate_pic_operand): Likewise
20428 * config/microblaze/microblaze.h (LEGITIMATE_PIC_OPERAND_P): calls
20429 new function microblaze_legitimate_pic_operand
20430 * config/microblaze/microblaze-protos.h
20431 (microblaze_legitimate_pic_operand): Declare.
20433 2013-03-04 Edgar E. Iglesias <edgar.iglesias@gmail.com>
20435 * config/microblaze/predicates.md (call_insn_simple_operand):
20436 New predicate for supported rtx code types.
20437 * config/microblaze/microblaze.md (call_internal1): Use
20438 call_insn_simple_operand predicate.
20440 2013-03-04 Jakub Jelinek <jakub@redhat.com>
20442 PR middle-end/56461
20443 * tree-loop-distribution.c (ldist_gen): Call partition_free after each
20444 partitions.ordered_remove.
20446 PR middle-end/56461
20447 * tree-vect-stmts.c (vectorizable_conversion): Don't call
20448 vec_oprnds0.create (1) for modifier == NONE.
20450 PR middle-end/56461
20451 * tree-vect-stmts.c (vectorizable_shift): Don't call create methods
20452 on vec_oprnds0 or vec_oprnds1 before loop, only call it on
20453 vec_oprnds1 right before pushing anything to it for
20456 PR middle-end/56461
20457 * tree-vect-loop.c (destroy_loop_vec_info): For !clean_stmts, just
20458 set nbbs to 0 instead of having separate code path.
20459 (vect_analyze_loop_form): Call destroy_loop_vec_info with true
20460 instead of false as last argument if returning NULL.
20462 2013-03-03 Sandra Loosemore <sandra@codesourcery.com>
20464 * target.def (TARGET_OPTION_VALID_ATTRIBUTE_P): Update comments;
20465 the attribute is now called "target" instead of "option".
20466 (TARGET_OPTION_PRAGMA_PARSE): Likewise, for the pragma.
20467 * doc/tm.texi.in (Target Attributes): Likewise document the correct
20468 attribute/pragma name for TARGET_OPTION_VALID_P and
20469 TARGET_OPTION_PRAGMA_PARSE. Also copy-edit and correct markup.
20470 * doc/tm.texi: Regenerated.
20472 2013-03-02 David Holsgrove <david.holsgrove@xilinx.com>
20474 * config/microblaze/microblaze.c:
20475 Check mcpu, pcmp requirement and set TARGET_REORDER to 0 if not met.
20476 * config/microblaze/microblaze.h: Add -mxl-reorder to
20478 * config/microblaze/microblaze.md: New bswapsi2 and bswaphi2.
20479 instructions emitted if TARGET_REORDER.
20480 * config/microblaze/microblaze.opt: New option -mxl-reorder set to 1
20481 or 0 for -m/-mno case, but initialises as 2 to detect default use case
20484 2013-03-01 Xinliang David Li <davidxl@google.com>
20486 * tree-ssa-uninit.c (compute_control_dep_chain): Limit post-dom
20489 2013-03-01 Jakub Jelinek <jakub@redhat.com>
20491 PR middle-end/56461
20492 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Release path
20493 vector even when returning true. Fix up function comment formatting.
20495 PR middle-end/56461
20496 * ira-build.c (ira_loop_nodes_count): New variable.
20497 (create_loop_tree_nodes): Initialize it.
20498 (finish_loop_tree_nodes): Use it instead of looking at current_loops.
20500 PR middle-end/56461
20501 * tree-vect-data-refs.c (vect_permute_store_chain): Avoid using copy
20502 method on dr_chain and result_chain.
20503 * tree-vect-stmts.c (vectorizable_store): Only call
20504 result_chain.create if j == 0.
20506 PR middle-end/56461
20507 * tree-vect-stmts.c (vect_create_vectorized_promotion_stmts): Call
20508 vec_oprnds0->release (); rather than vec_oprnds0->truncate (0)
20509 before overwriting it.
20511 2013-03-01 Tobias Burnus <burnus@net-b.de>
20513 * doc/extended.texi (C Extensions): Change order in @menu
20515 (Other MIPS Built-in Functions): Move last MIPS entry before
20516 "picoChip Built-in Functions".
20517 (SH Built-in Functions): Move after RX Built-in Functions.
20518 * doc/gcc.texi (Introduction): Change order in @menu
20520 * doc/md.texi (Constraints): Ditto.
20521 * gty.texi (Type Information): Ditto.
20522 (User-provided marking routines for template types): Make
20524 * doc/invoke.texi (AArch64 Options): Move before
20525 "Adapteva Epiphany Options".
20527 2013-02-28 Konstantin Serebryany <konstantin.s.serebryany@gmail.com>
20528 Jakub Jelinek <jakub@redhat.com>
20531 * asan.c (gate_asan): Lookup no_sanitize_address instead of
20532 no_address_safety_analysis attribute.
20533 * doc/extend.texi (no_address_safety_attribute): Rename to
20534 no_sanitize_address attribute, mention no_address_safety_analysis
20535 attribute as deprecated alias.
20537 2013-02-28 Jakub Jelinek <jakub@redhat.com>
20539 PR middle-end/56461
20540 * tree-vectorizer.h (vect_get_slp_defs): Change 3rd argument
20541 type to vec<vec<tree> > *.
20542 * tree-vect-slp.c (vect_get_slp_defs): Likewise. Change vec_defs
20543 to be vec<tree> instead of vec<tree> *, set vec_defs
20544 to vNULL and call vec_defs.create (number_of_vects), adjust other
20546 * tree-vect-stmts.c (vect_get_vec_defs, vectorizable_call,
20547 vectorizable_condition): Adjust vect_get_slp_defs callers.
20549 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
20551 * config/aarch64/aarch64.c
20552 (aarch64_float_const_representable): Remove unused variable.
20554 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
20556 * config/aarch64/aarch64.c (aarch64_mangle_type): Make static.
20558 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
20560 * config/aarch64/aarch64-builtins.c
20561 (aarch64_init_simd_builtins): Make static.
20563 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
20565 * config/aarch64/aarch64.c
20566 (aarch64_simd_make_constant): Make static.
20568 2013-02-28 Martin Jambor <mjambor@suse.cz>
20570 * tree-sra.c (load_assign_lhs_subreplacements): Do not put replacements
20571 with no initialization to the RHS of debug statements.
20573 2013-02-28 Martin Jambor <mjambor@suse.cz>
20575 PR tree-optimization/56294
20576 * tree-sra.c (analyze_access_subtree): Create replacement declarations.
20578 (get_access_replacement): Do not call create_access_replacement.
20579 Assert a replacement exists.
20580 (get_repl_default_def_ssa_name): Create the replacement declaration
20583 2013-02-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
20585 * config/arm/arm.c (arm_output_mi_thunk): Call final_start_function and
20586 final_end_function.
20588 2013-02-28 Marek Polacek <polacek@redhat.com>
20590 PR rtl-optimization/56466
20591 * loop-unroll.c (unroll_and_peel_loops): Call fix_loop_structure
20592 if we're changing a loop.
20593 (peel_loops_completely): Likewise.
20595 2013-02-28 Paolo Carlini <paolo.carlini@oracle.com>
20598 * doc/invoke.texi ([-Wctor-dtor-privacy]): Complete.
20600 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
20603 * config/avr/avr.c (avr_init_builtins): Use 'n' instead of empty
20604 macro parameters with: FX_FTYPE_FX, FX_FTYPE_FX_INT, INT_FTYPE_FX,
20605 INTX_FTYPE_FX, FX_FTYPE_INTX.
20606 * config/avr/builtins.def: Adjust respective DEF_BUILTIN.
20608 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
20610 * avr/avr-mcus.def (ata5272, ata5505, attiny1634, ata6285)
20611 (ata6286, atmega8a, atmega48pa, ata5790, ata5790n, ata5795)
20612 (atmega164pa, atmega165pa, atmega168pa, atmega16hva, atmega16hvb)
20613 (atmega16hvbrevb, atmega16m1, atmega16u4, atmega26hvg, atmega32a)
20614 (atmega32a, atmega3250pa, atmega3290pa, atmega32c1, atmega32m1)
20615 (atmega32u4, atmega32u6, atmega64a, atmega6490a, atmega6490p)
20616 (atmega64c1, atmega64m1, atmega64rfa2, atmega64rfr2, atmega32hvb)
20617 (atmega32hvbrevb, atmega16hva2, atmega48hvf, at90pwm161)
20618 (atmega128a, atmega1284, atmxt112sl, atmxt224, atmxt224e)
20619 (atmxt336s, atxmega16a4u, atxmega16c4, atxmega32a4u, atxmega32c4)
20620 (atxmega32e5, atxmega64a3u, atxmega64a4u, atxmega64b1, atxmega64b3)
20621 (atxmega64c3, atxmega64d4, atxmega128a3u, atxmega128b1)
20622 (atxmega128b3, atxmega128c3, atxmega128d4, atmxt540s, atmxt540sreva)
20623 (atxmega192a3u, atxmega192c3, atxmega256a3u, atxmega256c3)
20624 (atxmega384c3, atxmega384d3, atxmega128a4u): New AVR_MCU.
20625 (avrxmega6): Increase max flash segments from 5 to 6.
20626 * config/avr/t-multilib: Regenerate.
20627 * config/avr/avr-tables.opt: Regenerate.
20628 * doc/avr-mmcu.texi: Regenerate.
20630 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
20632 * config/avr/avr.h (device_to_arch): Rename to device_to_ld.
20633 (avr_device_to_arch): Rename to avr_device_to_ld.
20634 (avr_device_to_as): New prototype.
20635 (EXTRA_SPEC_FUNCTIONS): Add device_to_as.
20636 (ASM_SPEC): Use device_to_as to get -mmcu= and -mno-skip-bug=.
20637 * config/avr/driver-avr.c (avr_device_to_as): New.
20638 (avr_device_to_arch): Rename to avr_device_to_ld.
20640 2013-02-27 Jakub Jelinek <jakub@redhat.com>
20642 PR middle-end/56461
20643 * tree-vect-data-refs.c (vect_permute_load_chain): Avoid using copy
20644 method on dr_chain and result_chain.
20646 PR middle-end/56461
20647 * tree-ssa-loop-niter.c (maybe_lower_iteration_bound): Call
20648 pointer_set_destroy on not_executed_last_iteration.
20650 PR middle-end/56461
20651 * tree-vect-loop.c (vectorizable_reduction): Release vect_defs vector.
20653 PR middle-end/56461
20654 * ipa-pure-const.c (propagate): Use FOR_EACH_FUNCTION instead of
20655 FOR_EACH_DEFINED_FUNCTION when freeing state.
20657 PR middle-end/56461
20658 * df-scan.c (df_insn_delete): Use df_scan_free_mws_vec before
20660 (df_insn_rescan_debug_internal): Use df_scan_free_mws_vec before
20663 PR middle-end/56461
20664 * ipa-cp.c (decide_whether_version_node): Call vec_free on
20665 known_aggs[i].items and release known_aggs vector.
20667 PR middle-end/56461
20668 * ipa-reference.c (propagate): Free node_info even for alias nodes.
20670 2013-02-27 Edgar E. Iglesias <edgar.iglesias@gmail.com>
20672 * config/microblaze/microblaze.c (microblaze_emit_compare):
20673 Use xor for EQ/NE comparisions.
20674 * config/microblaze/microblaze.md (cstoresf4): Add constraints
20675 (cbranchsf4): Adjust operator to comparison_operator.
20677 2013-02-27 Jakub Jelinek <jakub@redhat.com>
20679 PR middle-end/56461
20680 * tree-flow.h (edge_var_map_vector): Change into va_heap, vl_embed
20682 * tree-ssa.c (redirect_edge_var_map_add): Use vec_safe_reserve and
20683 vec_safe_push, always update *slot.
20684 (redirect_edge_var_map_clear): Use vec_free.
20685 (redirect_edge_var_map_dup): Use vec_safe_copy and vec_safe_reserve.
20686 (free_var_map_entry): Use vec_free.
20687 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Use
20688 FOR_EACH_VEC_SAFE_ELT instead of FOR_EACH_VEC_ELT.
20690 2013-02-27 Andrey Belevantsev <abel@ispras.ru>
20692 PR middle-end/45472
20693 * sel-sched-ir.c (merge_expr): Also change vinsn of merged expr
20694 when the may_trap_p bit of the exprs being merged differs.
20695 Reorder tests for speculativeness in the logical and operator.
20697 2013-02-27 Jakub Jelinek <jakub@redhat.com>
20699 * incpath.c (add_standard_paths): Use reconcat instead of concat
20700 where appropriate and avoid leaking memory.
20702 * opts.h: Include obstack.h.
20703 (opts_concat): New prototype.
20704 (opts_obstack): New declaration.
20705 * opts.c (opts_concat): New function.
20706 (opts_obstack): New variable.
20707 (init_options_struct): Call gcc_init_obstack on opts_obstack.
20708 (finish_options): Use opts_concat instead of concat
20709 and XOBNEWVEC instead of XNEWVEC.
20710 * opts-common.c (generate_canonical_option, decode_cmdline_option,
20711 generate_option): Likewise.
20712 * Makefile.in (OPTS_H): Depend on $(OBSTACK_H).
20713 * lto-wrapper.c (main): Call gcc_init_obstack on opts_obstack.
20716 * stmt.c (expand_switch_as_decision_tree_p): If flag_pic
20717 and ASM_OUTPUT_ADDR_DIFF_ELT isn't defined, return true.
20719 2013-02-26 Jakub Jelinek <jakub@redhat.com>
20721 PR middle-end/56461
20722 * lra-spills.c (lra_spill): Free spill_hard_reg at the end.
20724 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
20726 * config/arm/arm.c (const_ok_for_dimode_op): Back out last change.
20727 (arm_block_move_unaligned_straight): Likewise.
20728 (arm_adjust_block_mem): Likewise.
20730 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
20733 * config/lm32/lm32.c (gen_int_relational): Remove unused variables
20734 temp, cond and label.
20735 * config/lm32/lm32.md (ashlsi3): Remove unused variable one.
20738 * config/c6x/c6x.c (dbx_register_map): Change to unsigned.
20739 * config/c6x/c6x.h (dbx_register_map): Update declaration.
20742 * config/cr16/cr16-protos.h: Move end of RTX_CODE guard below end
20743 of prologue/epilogue functions.
20746 * config/tilegx/tilegx.c (tilegx_expand_prologue):
20747 Remove unused variable cfa_offset.
20748 * config/tilepro/tilepro.c (tilepro_expand_prologue): Likewise.
20751 * config/mn10300/mn10300.c (mn10300_expand_epilogue): Avoid offset
20752 type promotion to unsigned.
20755 * config/arm/arm.c (const_ok_for_dimode_op): Make code consistent
20756 for HOST_WIDE_INT of 32 bit / same size as int.
20757 (arm_block_move_unaligned_straight): Likewise.
20758 (arm_adjust_block_mem): Likewise.
20761 * config/mep/t-mep (mep-pragma.o): Use ALL_COMPILERFLAGS instead of
20764 2013-02-26 Marek Polacek <polacek@redhat.com>
20766 PR tree-optimization/56426
20767 * tree-ssa-loop.c (tree_ssa_loop_init): Always call scev_initialize.
20769 2013-02-26 Richard Biener <rguenther@suse.de>
20772 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc): Remove
20773 unused variable loops.
20775 2013-02-26 Jakub Jelinek <jakub@redhat.com>
20777 PR tree-optimization/56448
20778 * fold-const.c (operand_equal_p) <case tcc_reference>: Don't look at
20779 TREE_SIDE_EFFECTS if flags contain OEP_CONSTANT_ADDRESS_OF.
20780 Clear OEP_CONSTANT_ADDRESS_OF from flags before recursing on second or
20781 later operands of the references, or even first operand for
20782 INDIRECT_REF, TARGET_MEM_REF or MEM_REF.
20784 PR tree-optimization/56443
20785 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): For
20786 overaligned types, pass TYPE_UNSIGNED (scalar_type) as second argument
20787 to type_for_mode langhook.
20789 2013-02-25 Matt Turner <mattst88@gmail.com>
20791 * doc/invoke.texi: Document r4700.
20793 2013-02-25 Richard Biener <rguenther@suse.de>
20795 PR tree-optimization/56175
20796 * tree-ssa-forwprop.c (hoist_conversion_for_bitop_p): New predicate,
20798 (simplify_bitwise_binary): ... here. Also guard the conversion
20799 of (type) X op CST to (type) (X op ((type-x) CST)) with it.
20801 2013-02-25 Catherine Moore <clm@codesourcery.com>
20804 2013-02-24 Catherine Moore <clm@codesourcery.com>
20805 Maciej W. Rozycki <macro@codesourcery.com>
20806 Tom de Vries <tom@codesourcery.com>
20807 Nathan Sidwell <nathan@codesourcery.com>
20808 Iain Sandoe <iain@codesourcery.com>
20809 Nathan Froyd <froydnj@codesourcery.com>
20810 Chao-ying Fu <fu@mips.com>
20812 * doc/extend.texi: (micromips, nomicromips, nocompression):
20813 Document new function attributes.
20814 * doc/invoke.texi (minterlink-compressed, mmicromips,
20815 m14k, m14ke, m14kec): Document new options.
20816 (minterlink-mips16): Update documentation.
20817 * doc/md.texi (ZC, ZD): Document new constraints.
20818 * configure.ac (gcc_cv_as_micromips): Check if linker
20819 supports the .set micromips directive.
20820 * configure: Regenerate.
20821 * config.in: Regenerate.
20822 * config/mips/mips-tables.opt: Regenerate.
20823 * config/mips/micromips.md: New file.
20824 * constraints.md (ZC, AD): New constraints.
20825 * config/mips/predicates.md (movep_src_register): New predicate.
20826 (movep_src_operand): New predicate.
20827 (non_volatile_mem_operand): New predicate.
20828 * config/mips/mips.md (multimem): New type.
20829 (length): Differentiate between 17-bit and 18-bit branch offsets.
20830 (MOVEP1, MOVEP2): New mode iterator.
20831 (mov_<load>l): Use ZC constraint.
20832 (mov_<load>r): Likewise.
20833 (mov_<store>l): Likewise.
20834 (mov_<store>r): Likewise.
20835 (*branch_equality<mode>_inverted): Add microMIPS support.
20836 (*branch_equality<mode>): Likewise.
20837 (*jump_absolute): Likewise.
20838 (indirect_jump_<mode>): Likewise.
20839 (tablejump_<mode>): Likewise.
20840 (<optab>_internal): Likewise.
20841 (sibcall_internal): Likewise.
20842 (sibcall_value_internal): Likewise.
20843 (prefetch): Use constraint ZD.
20844 * config/mips/mips.opt (minterlink-compressed): New option.
20845 (minterlink-mips16): Now an alias for minterlink-compressed.
20846 (mmicromips): New option.
20847 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
20848 (compare_and_swap_12): Likewise.
20849 (sync_add<mode>): Likewise.
20850 (sync_<optab>_12): Likewise.
20851 (sync_old_<optab>_12): Likewise.
20852 (sync_new_<optab>_12): Likewise.
20853 (sync_nand_12): Likewise.
20854 (sync_old_nand_12): Likewise.
20855 (sync_new_nand_12): Likewise.
20856 (sync_sub<mode>): Likewise.
20857 (sync_old_add<mode>): Likewise.
20858 (sync_old_sub<mode>): Likewise.
20859 (sync_new_add<mode>): Likewise.
20860 (sync_new_sub<mode>): Likewise.
20861 (sync_<optab><mode>): Likewise.
20862 (sync_old_<optab><mode>): Likewise.
20863 (sync_new_<optab><mode>): Likewise.
20864 (sync_nand<mode>): Likewise.
20865 (sync_old_nand<mode>): Likewise.
20866 (sync_new_nand<mode>): Likewise.
20867 (sync_lock_test_and_set<mode>): Likewise.
20868 (test_and_set_12): Likewise.
20869 (atomic_compare_and_swap<mode>): Likewise.
20870 (atomic_exchange<mode>_llsc): Likewise.
20871 (atomic_fetch_add<mode>_llsc): Likewise.
20872 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
20873 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
20874 (umips_save_restore_pattern_p): Likewise.
20875 (umips_load_store_pair_p): Likewise.
20876 (umips_output_load_store_pair): Likewise.
20877 (umips_movep_target_p): Likewise.
20878 (umips_12bit_offset_address_p): Likewise.
20879 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
20880 (mips_base_mips16): Rename this...
20881 (mips_base_compression_flags): ...to this. Update all uses.
20882 (mips_attribute_table): Add micromips, nomicromips and nocompression.
20883 (mips_mips16_decl_p): Delete.
20884 (mips_nomips16_decl_p): Delete.
20885 (mips_get_compress_on_flags): New function.
20886 (mips_get_compress_off_flags): New function.
20887 (mips_get_compress_mode): New function.
20888 (mips_get_compress_on_name): New function.
20889 (mips_get_compress_off_name): New function.
20890 (mips_insert_attributes): Support multiple compression types.
20891 (mips_merge_decl_attributes): Likewise.
20892 (umips_12bit_offset_address_p): New function.
20893 (mips_start_function_definition): Emit .set micromips directive.
20894 (mips_call_may_need_jalx_p): New function.
20895 (mips_function_ok_for_sibcall): Add microMIPS support.
20896 (mips_print_operand_punctuation): Support short delay slots and
20898 (umips_swm_mask, umips_swm_encoding): New.
20899 (umips_build_save_restore): New function.
20900 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
20901 (was_mips16_p): Remove.
20902 (old_compression_mode): New.
20903 (mips_set_compression_mode): New function.
20904 (mips_set_current_function): Add microMIPS support.
20905 (mips_option_override): Likewise.
20906 (umips_save_restore_pattern_p): New function.
20907 (umips_output_save_restore): New function.
20908 (umips_load_store_pair_p_1): New function.
20909 (umips_load_store_pair_p): New function.
20910 (umips_output_load_store_pair_1): New function.
20911 (umips_output_load_store_pair): New function.
20912 (umips_movep_target_p) New function.
20913 (mips_prepare_pch_save): Add microMIPS support.
20914 * config/mips/mips.h (TARGET_COMPRESSION): New.
20915 (TARGET_CPU_CPP_BUILTINS): Update macro
20916 to use new compression flags and to support microMIPS.
20917 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
20918 (MIPS_ARCH_FLOAT_SPEC): Likewise.
20919 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
20920 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
20921 (ASM_SPEC): Support mmicromips and mno-micromips.
20922 (M16STORE_REG_P): New macro.
20923 (MIPS_CALL): Support TARGET_MICROMIPS.
20924 (MICROMIPS_J): New macro.
20925 (mips_base_mips16): Rename this...
20926 (mips_base_compression_flags): ...to this.
20927 (UMIPS_12BIT_OFFSET_P): New macro.
20928 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
20929 (MULTILIB_DIRNAMES): Likewise.
20931 2013-02-25 Tom de Vries <tom@codesourcery.com>
20933 PR rtl-optimization/56131
20934 * insn-notes.def (INSN_NOTE_BASIC_BLOCK): Update comment.
20935 * cfgrtl.c (delete_insn): Don't reorder NOTE_INSN_DELETED_LABEL and
20936 NOTE_INSN_BASIC_BLOCK if BLOCK_FOR_INSN == NULL.
20938 2013-02-25 Tobias Burnus <burnus@net-b.de>
20940 * doc/invoke.texi (-fsanitize=): Move from optimization
20941 to debugging options.
20943 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
20945 * sched-deps.c (sched_analyze_insn): Fix typo in comment.
20947 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
20948 Alexander Monakov <amonakov@ispras.ru>
20950 PR middle-end/56077
20951 * sched-deps.c (sched_analyze_insn): When reg_pending_barrier,
20952 flush pending lists also on non-jumps. Adjust comment.
20954 2013-02-24 Catherine Moore <clm@codesourcery.com>
20955 Maciej W. Rozycki <macro@codesourcery.com>
20956 Tom de Vries <tom@codesourcery.com>
20957 Nathan Sidwell <nathan@codesourcery.com>
20958 Iain Sandoe <iain@codesourcery.com>
20959 Nathan Froyd <froydnj@codesourcery.com>
20960 Chao-ying Fu <fu@mips.com>
20962 * doc/extend.texi: (micromips, nomicromips, nocompression):
20963 Document new function attributes.
20964 * doc/invoke.texi (minterlink-compressed, mmicromips,
20965 m14k, m14ke, m14kec): Document new options.
20966 (minterlink-mips16): Update documentation.
20967 * doc/md.texi (ZC, ZD): Document new constraints.
20968 * configure.ac (gcc_cv_as_micromips): Check if linker
20969 supports the .set micromips directive.
20970 * configure: Regenerate.
20971 * config.in: Regenerate.
20972 * config/mips/mips-tables.opt: Regenerate.
20973 * config/mips/micromips.md: New file.
20974 * constraints.md (ZC, AD): New constraints.
20975 * config/mips/predicates.md (movep_src_register): New predicate.
20976 (movep_src_operand): New predicate.
20977 (non_volatile_mem_operand): New predicate.
20978 * config/mips/mips.md (multimem): New type.
20979 (length): Differentiate between 17-bit and 18-bit branch offsets.
20980 (MOVEP1, MOVEP2): New mode iterator.
20981 (mov_<load>l): Use ZC constraint.
20982 (mov_<load>r): Likewise.
20983 (mov_<store>l): Likewise.
20984 (mov_<store>r): Likewise.
20985 (*branch_equality<mode>_inverted): Add microMIPS support.
20986 (*branch_equality<mode>): Likewise.
20987 (*jump_absolute): Likewise.
20988 (indirect_jump_<mode>): Likewise.
20989 (tablejump_<mode>): Likewise.
20990 (<optab>_internal): Likewise.
20991 (sibcall_internal): Likewise.
20992 (sibcall_value_internal): Likewise.
20993 (prefetch): Use constraint ZD.
20994 * config/mips/mips.opt (minterlink-compressed): New option.
20995 (minterlink-mips16): Now an alias for minterlink-compressed.
20996 (mmicromips): New option.
20997 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
20998 (compare_and_swap_12): Likewise.
20999 (sync_add<mode>): Likewise.
21000 (sync_<optab>_12): Likewise.
21001 (sync_old_<optab>_12): Likewise.
21002 (sync_new_<optab>_12): Likewise.
21003 (sync_nand_12): Likewise.
21004 (sync_old_nand_12): Likewise.
21005 (sync_new_nand_12): Likewise.
21006 (sync_sub<mode>): Likewise.
21007 (sync_old_add<mode>): Likewise.
21008 (sync_old_sub<mode>): Likewise.
21009 (sync_new_add<mode>): Likewise.
21010 (sync_new_sub<mode>): Likewise.
21011 (sync_<optab><mode>): Likewise.
21012 (sync_old_<optab><mode>): Likewise.
21013 (sync_new_<optab><mode>): Likewise.
21014 (sync_nand<mode>): Likewise.
21015 (sync_old_nand<mode>): Likewise.
21016 (sync_new_nand<mode>): Likewise.
21017 (sync_lock_test_and_set<mode>): Likewise.
21018 (test_and_set_12): Likewise.
21019 (atomic_compare_and_swap<mode>): Likewise.
21020 (atomic_exchange<mode>_llsc): Likewise.
21021 (atomic_fetch_add<mode>_llsc): Likewise.
21022 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
21023 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
21024 (umips_save_restore_pattern_p): Likewise.
21025 (umips_load_store_pair_p): Likewise.
21026 (umips_output_load_store_pair): Likewise.
21027 (umips_movep_target_p): Likewise.
21028 (umips_12bit_offset_address_p): Likewise.
21029 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
21030 (mips_base_mips16): Rename this...
21031 (mips_base_compression_flags): ...to this. Update all uses.
21032 (mips_attribute_table): Add micromips, nomicromips and nocompression.
21033 (mips_mips16_decl_p): Delete.
21034 (mips_nomips16_decl_p): Delete.
21035 (mips_get_compress_on_flags): New function.
21036 (mips_get_compress_off_flags): New function.
21037 (mips_get_compress_mode): New function.
21038 (mips_get_compress_on_name): New function.
21039 (mips_get_compress_off_name): New function.
21040 (mips_insert_attributes): Support multiple compression types.
21041 (mips_merge_decl_attributes): Likewise.
21042 (umips_12bit_offset_address_p): New function.
21043 (mips_start_function_definition): Emit .set micromips directive.
21044 (mips_call_may_need_jalx_p): New function.
21045 (mips_function_ok_for_sibcall): Add microMIPS support.
21046 (mips_print_operand_punctuation): Support short delay slots and
21048 (umips_swm_mask, umips_swm_encoding): New.
21049 (umips_build_save_restore): New function.
21050 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
21051 (was_mips16_p): Remove.
21052 (old_compression_mode): New.
21053 (mips_set_compression_mode): New function.
21054 (mips_set_current_function): Add microMIPS support.
21055 (mips_option_override): Likewise.
21056 (umips_save_restore_pattern_p): New function.
21057 (umips_output_save_restore): New function.
21058 (umips_load_store_pair_p_1): New function.
21059 (umips_load_store_pair_p): New function.
21060 (umips_output_load_store_pair_1): New function.
21061 (umips_output_load_store_pair): New function.
21062 (umips_movep_target_p) New function.
21063 (mips_prepare_pch_save): Add microMIPS support.
21064 * config/mips/mips.h (TARGET_COMPRESSION): New.
21065 (TARGET_CPU_CPP_BUILTINS): Update macro
21066 to use new compression flags and to support microMIPS.
21067 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
21068 (MIPS_ARCH_FLOAT_SPEC): Likewise.
21069 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
21070 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
21071 (ASM_SPEC): Support mmicromips and mno-micromips.
21072 (M16STORE_REG_P): New macro.
21073 (MIPS_CALL): Support TARGET_MICROMIPS.
21074 (MICROMIPS_J): New macro.
21075 (mips_base_mips16): Rename this...
21076 (mips_base_compression_flags): ...to this.
21077 (UMIPS_12BIT_OFFSET_P): New macro.
21078 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
21079 (MULTILIB_DIRNAMES): Likewise.
21081 2013-02-24 Jakub Jelinek <jakub@redhat.com>
21084 * target-globals.c (save_target_globals): For init_reg_sets and
21085 target_reinit remporarily set this_fn_optabs to this_target_optabs.
21087 2013-02-22 James Grennahlgh <james.greenhalgh@arm.com>
21089 * config/aarch64/aarch64-simd-builtins.def: Add copyright header.
21090 * config/aarch64/t-aarch64
21091 (aarch64-builtins.o): Depend on aarch64-simd-builtins.def.
21093 2013-02-22 Vladimir Makarov <vmakarov@redhat.com>
21095 PR inline-asm/56148
21096 * lra-constraints.c (process_alt_operands): Reload operand
21097 conflicting with earlier clobber only if no more other conflicting
21100 2013-02-22 Jakub Jelinek <jakub@redhat.com>
21103 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Link in libasan_preinit.o
21104 if not linking a shared library.
21106 2013-02-22 Seth LaForge <sethml@google.com>
21108 * config.gcc (arm*-*-eabi*): Treat arm*eb as big-endian.
21110 2013-02-22 Greta Yorsh <Greta.Yorsh@arm.com>
21112 * config/arm/arm.md (split for extendsidi): Update condition.
21113 (zero_extend<mode>di2,extend<mode>di2): Add an alternative.
21114 * config/arm/iterators.md (qhs_extenddi_cstr): Likewise.
21115 (qhs_zextenddi_cstr): Likewise.
21117 2013-02-21 Jakub Jelinek <jakub@redhat.com>
21119 PR middle-end/56420
21120 * expmed.c (EXACT_POWER_OF_2_OR_ZERO_P): Do subtraction in uhwi, to
21121 avoid signed wrapping.
21122 (expand_mult): Handle properly multiplication by
21123 ((dword_type) -1) << (BITS_PER_WORD - 1). Improve multiplication by
21124 ((dword_type) 1) << (BITS_PER_WORD - 1). Avoid undefined behavior
21125 in the compiler if coeff is HOST_WIDE_INT_MIN.
21126 (expand_divmod): Don't make ext_op1 static, change it's type to
21127 uhwi. Avoid undefined behavior in -INTVAL (op1).
21129 PR rtl-optimization/50339
21130 * lower-subreg.h (struct lower_subreg_choices): Add splitting_ashiftrt
21132 * lower-subreg.c (compute_splitting_shift): Handle ASHIFTRT.
21133 (compute_costs): Call compute_splitting_shift also for ASHIFTRT
21134 into splitting_ashiftrt field.
21135 (find_decomposable_shift_zext, resolve_shift_zext): Handle also
21137 (dump_choices): Fix up printing LSHIFTRT choices, print ASHIFTRT
21140 2013-02-20 Aldy Hernandez <aldyh@redhat.com>
21142 PR middle-end/56108
21143 * trans-mem.c (execute_tm_mark): Do not expand transactions that
21144 are sure to go irrevocable.
21146 2013-02-21 Hans-Peter Nilsson <hp@axis.com>
21148 * doc/rtl.texi (vec_concat, vec_duplicate): Mention that
21149 scalars are valid operands.
21151 2013-02-21 Martin Jambor <mjambor@suse.cz>
21153 PR tree-optimization/56310
21154 * ipa-cp.c (agg_replacements_to_vector): New parameter index, copy
21155 only matching indices and non-negative final offsets.
21156 (intersect_aggregates_with_edge): Pass src_idx to
21157 agg_replacements_to_vector. Pass src_idx insstead of index to
21158 intersect_with_agg_replacements.
21160 2013-02-21 Martin Jambor <mjambor@suse.cz>
21162 * ipa-cp.c (good_cloning_opportunity_p): Dump the real threshold
21163 instead of hard-wired defaults.
21165 2013-02-21 Maciej W. Rozycki <macro@codesourcery.com>
21167 * doc/invoke.texi (MIPS Options): Update documentation of the
21168 floating-point multiply-accumulate instruction restrictions.
21170 2013-02-21 Kostya Serebryany <kcc@google.com>
21172 * config/i386/i386.c (ix86_asan_shadow_offset): Use 0x7fff8000 as
21173 asan_shadow_offset on x86_64 linux.
21175 2013-02-21 Richard Biener <rguenther@suse.de>
21177 PR tree-optimization/56415
21179 2013-02-11 Richard Biener <rguenther@suse.de>
21181 PR tree-optimization/56273
21182 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
21185 2013-02-21 Jakub Jelinek <jakub@redhat.com>
21188 * doc/invoke.texi (-fdump-rtl-pro_and_epilogue): Use @item
21191 PR inline-asm/56405
21192 * expr.c (expand_expr_real_1) <case TARGET_MEM_REF, MEM_REF>: Don't
21193 use movmisalign or extract_bit_field for EXPAND_MEMORY modifier.
21195 2013-02-20 Jan Hubicka <jh@suse.cz>
21197 PR tree-optimization/56265
21198 * ipa-prop.c (ipa_make_edge_direct_to_target): Fixup callgraph
21199 when target is referenced for first time.
21201 2013-02-20 Richard Biener <rguenther@suse.de>
21203 * tree-call-cdce.c (tree_call_cdce): Do not remove unused locals.
21204 * tree-ssa-forwprop.c (ssa_forward_propagate_and_combine): Likewise.
21205 * tree-ssa-dce.c (perform_tree_ssa_dce): Likewise.
21206 * tree-ssa-copyrename.c (copy_rename_partition_coalesce): Do
21207 not return anything.
21208 (rename_ssa_copies): Do not remove unused locals.
21209 * tree-ssa-ccp.c (do_ssa_ccp): Likewise.
21210 * tree-ssanames.c (pass_release_ssa_names): Remove unused locals first.
21211 * passes.c (execute_function_todo): Do not schedule unused locals
21212 removal if cleanup_tree_cfg did something.
21213 * tree-ssa-live.c (remove_unused_locals): Dump statistics
21214 about the number of removed locals.
21216 2013-02-20 Richard Biener <rguenther@suse.de>
21218 PR tree-optimization/56398
21219 * tree-vect-loop-manip.c (adjust_debug_stmts): Skip SSA default defs.
21221 2013-02-20 Martin Jambor <mjambor@suse.cz>
21223 PR tree-optimization/55334
21224 * ipa-cp.c (initialize_node_lattices): Disable IPA-CP through and to
21225 restricted pointers to arrays.
21227 2013-02-20 Richard Biener <rguenther@suse.de>
21228 Jakub Jelinek <jakub@redhat.com>
21230 PR tree-optimization/56396
21231 * tree-ssa-ccp.c (n_const_val): New static variable.
21232 (get_value): Return NULL for SSA names we don't have a lattice
21234 (ccp_initialize): Initialize n_const_val.
21235 * tree-ssa-copy.c (n_copy_of): New static variable.
21236 (init_copy_prop): Initialize n_copy_of.
21237 (get_value): Return NULL_TREE for SSA names we don't have a
21240 2013-02-20 Martin Jambor <mjambor@suse.cz>
21242 * ipa-cp.c (initialize_node_lattices): Fix dumping condition.
21244 2013-02-20 Richard Biener <rguenther@suse.de>
21246 * genpreds.c (write_lookup_constraint): Do not compare first
21247 letter of the constraint again.
21249 2013-02-20 Richard Biener <rguenther@suse.de>
21251 * tree-ssa-loop-ivopts.c (alloc_use_cost_map): Use bitmap_count_bits
21253 (get_use_iv_cost): Terminate hashtable walk when coming across
21256 2013-02-20 Igor Zamyatin <igor.zamyatin@intel.com>
21258 * config/i386/i386.c (initial_ix86_tune_features): Turn on fp
21259 reassociation for avx2 targets.
21261 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
21263 * config/microblaze/microblaze.c: microblaze_has_clz = 0
21264 Add version check for v8.10.a to enable microblaze_has_clz
21265 * config/microblaze/microblaze.h: Add TARGET_HAS_CLZ as combined
21266 version and TARGET_PATTERN_COMPARE check
21267 * config/microblaze/microblaze.md: New clzsi2 instruction
21269 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
21271 * config/microblaze/microblaze.md (call_value_intern): Check symbol is
21272 function before branching.
21274 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
21276 * sel-sched-dump.c (dump_insn_rtx_flags): Explicitly set
21278 (dump_insn_rtx_1): Pass PATTERN (insn) to str_pattern_slim.
21280 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
21282 PR middle-end/55889
21283 * sel-sched.c: Include ira.h.
21284 (implicit_clobber_conflict_p): New function.
21285 (moveup_expr): Use it.
21286 * Makefile.in (sel-sched.o): Depend on ira.h.
21288 2013-02-19 Richard Biener <rguenther@suse.de>
21290 PR tree-optimization/56384
21291 * tree-ssa-sccvn.h (struct vn_phi_s): Add type member.
21292 (vn_hash_type): Split out from ...
21293 (vn_hash_constant_with_type): ... here.
21294 * tree-ssa-sccvn.c (vn_phi_compute_hash): Use vn_hash_type.
21295 (vn_phi_eq): Compare types from vn_phi_s structure.
21296 (vn_phi_lookup): Populate vn_phi_s type.
21297 (vn_phi_insert): Likewise.
21299 2013-02-19 Jakub Jelinek <jakub@redhat.com>
21301 PR tree-optimization/56350
21302 * tree-vect-loop.c (vectorizable_reduction): If orig_stmt, return false
21303 if haven't found reduction or nested cycle operand, rather than
21304 asserting we must find it.
21306 PR tree-optimization/56381
21307 * tree-ssa-pre.c (create_expression_by_pieces): Fix up last argument
21310 2013-02-18 Aldy Hernandez <aldyh@redhat.com>
21311 Jakub Jelinek <jakub@redhat.com>
21314 * genopinit.c (raw_optab_handler): Use this_fn_optabs.
21315 (swap_optab_enable): Same.
21316 (init_all_optabs): Use argument instead of global.
21317 * tree.h (struct tree_optimization_option): New field target_optabs.
21318 * expr.h (init_all_optabs): Add argument to prototype.
21319 (TREE_OPTIMIZATION_OPTABS): New.
21320 (save_optabs_if_changed): Protoize.
21321 * optabs.h: Declare this_fn_optabs.
21322 * optabs.c (save_optabs_if_changed): New.
21323 Declare this_fn_optabs.
21324 (init_optabs): Add argument to init_all_optabs() call.
21325 * function.c (invoke_set_current_function_hook): Handle per
21327 * function.h (struct function): New field optabs.
21328 * config/mips/mips.c (mips_set_mips16_mode): Handle when
21329 optimization_current_node has changed.
21330 * target-globals.h (save_target_globals_default_opts): Protoize.
21331 * target-globals.c (save_target_globals_default_opts): New.
21333 2013-02-18 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
21336 * config/pa/pa.c (pa_conditional_register_usage): On HP-UX, mark
21337 registers %fr12 and %fr12R as call used.
21340 * config/pa/predicates.md (base14_operand): Except for BLKmode, QImode
21341 and HImode, require all displacements to be an integer multiple of
21343 * config/pa/pa.c (pa_legitimate_address_p): For REG+BASE addresses,
21344 only allow QImode and HImode when reload is in progress and strict is
21345 true. Likewise for symbolic addresses. Use base14_operand to check
21346 displacements in REG+BASE addresses.
21348 2013-02-18 Richard Biener <rguenther@suse.de>
21350 PR tree-optimization/56366
21351 * tree-vect-loop.c (get_initial_def_for_induction): Properly
21352 handle sign-conversion of outer-loop initial induction value.
21354 2013-02-18 Richard Biener <rguenther@suse.de>
21356 PR middle-end/56349
21357 * cfghooks.c (merge_blocks): If we merge a latch into another
21358 block adjust references to it.
21359 * cfgloop.c (flow_loops_find): Reset latch before recomputing it.
21360 (verify_loop_structure): Verify that a recorded latch is in fact
21363 2013-02-18 Richard Biener <rguenther@suse.de>
21365 PR tree-optimization/56321
21366 * tree-ssa-reassoc.c (propagate_op_to_single_use): Properly
21367 order SSA name release and virtual operand unlinking.
21369 2013-02-17 Edgar E. Iglesias <edgar.iglesias@gmail.com>
21371 * config/microblaze/microblaze.md (save_stack_block): Define.
21372 (restore_stack_block): Likewise.
21374 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
21376 * config/microblaze/linux.h (TARGET_SUPPORTS_PIC): Define as 1.
21377 * config/microblaze/microblaze.h (TARGET_SUPPORTS_PIC): Define as 1.
21378 * config/microblaze/microblaze.c (microblaze_option_override):
21379 Bail out early for PIC modes when target does not support PIC.
21381 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
21383 * config/microblaze/microblaze.c (microblaze_asm_trampoline_template):
21384 Replace with a microblaze version.
21385 (microblaze_trampoline_init): Adapt for microblaze.
21386 * config/microblaze/microblaze.h (TRAMPOLINE_SIZE): Adapt for
21389 2013-02-16 Jakub Jelinek <jakub@redhat.com>
21390 Dodji Seketeli <dodji@redhat.com>
21393 * asan.c (get_mem_refs_of_builtin_call): White space and style cleanup.
21394 (instrument_mem_region_access): Do not forget to always put
21395 instrumentation of the of 'base' and 'base + len' in a "if (len !=
21396 0) statement, even for cases where either 'base' or 'base + len'
21397 are not instrumented -- because they have been previously
21398 instrumented. Simplify the logic by putting all the statements
21399 instrument 'base + len' inside a sequence, and then insert that
21400 sequence right before the current insertion point. Then, to
21401 instrument 'base + len', just get an iterator on that statement.
21402 And do not forget to update the pointer to iterator the function
21403 received as argument.
21405 2013-02-15 Vladimir Makarov <vmakarov@redhat.com>
21407 PR rtl-optimization/56348
21408 * lra-assigns.c (reload_pseudo_compare_func): Prefer bigger pseudos.
21410 2013-02-15 Steven Bosscher <steven@gcc.gnu.org>
21412 * graph.c (start_graph_dump): Print dumpfile base as digraph label.
21413 (clean_graph_dump_file): Pass base to start_graph_dump.
21415 2013-02-14 Richard Henderson <rth@redhat.com>
21418 * lower-subreg.c (simple_move): Check dest mode instead of src mode.
21420 2013-02-14 Steven Bosscher <steven@gcc.gnu.org>
21422 * collect2-aix.h: Define F_LOADONLY.
21424 2013-02-14 Richard Biener <rguenther@suse.de>
21427 * varasm.c (output_constant_def_1): Get the decl representing
21428 the constant as argument.
21429 (output_constant_def): Wrap output_constant_def_1.
21430 (make_decl_rtl): Use output_constant_def_1 with the decl
21431 representing the constant.
21432 (build_constant_desc): Optionally re-use a decl already
21433 representing the constant.
21434 (tree_output_constant_def): Adjust.
21436 2013-02-14 Dodji Seketeli <dodji@redhat.com>
21439 * asan.c (instrument_builtin_call): Really put the length of the
21440 second source argument into src1_len.
21442 2013-02-13 Jakub Jelinek <jakub@redhat.com>
21444 * asan.c (create_cond_insert_point): Add create_then_fallthru_edge
21445 argument. If it is false, don't create edge from then_bb to
21447 (insert_if_then_before_iter): Pass true to it.
21448 (build_check_stmt): Pass false to it.
21449 (transform_statements): Flush hash table only on extended basic
21450 block boundaries, rather than at the beginning of every bb.
21451 Don't flush hash table on nonfreeing_call_p calls.
21452 * tree-flow.h (nonfreeing_call_p): New prototype.
21453 * tree-ssa-phiopt.c (nonfreeing_call_p): No longer static.
21455 2013-02-13 David S. Miller <davem@davemloft.net>
21457 * expmed.c (expand_shift_1): Only strip scalar integer subregs.
21459 2013-02-13 Vladimir Makarov <vmakarov@redhat.com>
21462 * ira.c (max_regno_before_ira): Move from ...
21464 (fix_reg_equiv_init): Use max_regno_before_ira instead of
21467 2013-02-13 Jakub Jelinek <jakub@redhat.com>
21469 * config/i386/i386.c (ix86_asan_shadow_offset): Revert last change.
21471 2013-02-13 Richard Biener <rguenther@suse.de>
21474 * gimple-streamer-out.c (output_gimple_stmt): Undo wrapping
21475 globals in MEM_REFs.
21477 2013-02-13 Richard Biener <rguenther@suse.de>
21479 * loop-init.c (loop_optimizer_init): Clear loop state when
21480 re-initializing preserved loops.
21481 * loop-unswitch.c (unswitch_single_loop): Return whether
21482 we unswitched the loop. Do not verify loop state here.
21483 (unswitch_loops): When we unswitched a loop discover new loops.
21485 2013-02-13 Kostya Serebryany <kcc@google.com>
21487 * config/i386/i386.c: Use 0x7fff8000 as asan_shadow_offset
21489 * sanitizer.def: Rename __asan_init to __asan_init_v1.
21491 2013-02-12 Dodji Seketeli <dodji@redhat.com>
21493 Avoid instrumenting duplicated memory access in the same basic block
21494 * Makefile.in (asan.o): Add new dependency on hash-table.h
21495 * asan.c (struct asan_mem_ref, struct mem_ref_hasher): New types.
21496 (asan_mem_ref_init, asan_mem_ref_get_end, get_mem_ref_hash_table)
21497 (has_stmt_been_instrumented_p, empty_mem_ref_hash_table)
21498 (free_mem_ref_resources, has_mem_ref_been_instrumented)
21499 (has_stmt_been_instrumented_p, update_mem_ref_hash_table)
21500 (get_mem_ref_of_assignment): New functions.
21501 (get_mem_refs_of_builtin_call): Extract from
21502 instrument_builtin_call and tweak a little bit to make it fit with
21504 (instrument_builtin_call): Use the new
21505 get_mem_refs_of_builtin_call. Use gimple_call_builtin_p instead
21506 of is_gimple_builtin_call.
21507 (instrument_derefs, instrument_mem_region_access): Insert the
21508 instrumented memory reference into the hash table.
21509 (maybe_instrument_assignment): Renamed instrument_assignment into
21510 this, and change it to advance the iterator when instrumentation
21511 actually happened and return true in that case. This makes it
21512 homogeneous with maybe_instrument_assignment, and thus give a
21513 chance to callers to be more 'regular'.
21514 (transform_statements): Clear the memory reference hash table
21515 whenever we enter a new BB, when we cross a function call, or when
21516 we are done transforming statements. Use
21517 maybe_instrument_assignment instead of instrumentation. No more
21518 need to special case maybe_instrument_assignment and advance the
21519 iterator after calling it; it's now handled just like
21520 maybe_instrument_call. Update comment.
21522 2013-02-13 Richard Biener <rguenther@suse.de>
21524 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc):
21525 Fix loop discovery code.
21527 2013-02-12 Vladimir Makarov <vmakarov@redhat.com>
21529 PR inline-asm/56148
21530 * lra-constraints.c (process_alt_operands): Match early clobber
21531 operand with itself. Check conflicts with earlyclobber only if
21532 the operand is not reloaded. Prefer to reload conflicting operand
21533 if earlyclobber and matching operands are the same.
21535 2013-02-12 Richard Biener <rguenther@suse.de>
21538 * lto-streamer-out.c (write_symbol): Do not output symbols
21539 for hard register variables.
21541 2013-02-12 Georg-Johann Lay <avr@gjlay.de>
21544 * config/avr/avr-dimode.md (umulsidi3, mulsidi3): New expanders.
21545 (umulsidi3_insn, mulsidi3_insn): New insns.
21547 2013-02-12 Christophe Lyon <christophe.lyon@linaro.org>
21549 * config/arm/arm-protos.h (struct cpu_vec_costs): New struct type.
21550 (struct tune_params): Add vec_costs field.
21551 * config/arm/arm.c (arm_builtin_vectorization_cost)
21552 (arm_add_stmt_cost): New functions.
21553 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST)
21554 (TARGET_VECTORIZE_ADD_STMT_COST): Define.
21555 (arm_default_vec_cost): New struct of type cpu_vec_costs.
21556 (arm_slowmul_tune, arm_fastmul_tune, arm_strongarm_tune)
21557 (arm_xscale_tune, arm_9e_tune, arm_v6t2_tune, arm_cortex_tune)
21558 (arm_cortex_a15_tune, arm_cortex_a5_tune, arm_cortex_a9_tune)
21559 (arm_v6m_tune, arm_fa726te_tune): Define new vec_costs field.
21561 2013-02-12 Richard Biener <rguenther@suse.de>
21564 * gimple-streamer-in.c (input_gimple_stmt): Strip MEM_REFs off
21565 decls again if possible.
21567 2013-02-12 Richard Biener <rguenther@suse.de>
21569 PR middle-end/56288
21570 * tree-ssa.c (verify_ssa_name): Fix check, move
21571 SSA_NAME_IN_FREE_LIST check up.
21573 2013-02-12 Jakub Jelinek <jakub@redhat.com>
21574 Steven Bosscher <steven@gcc.gnu.org>
21576 PR rtl-optimization/56151
21577 * optabs.c (add_equal_note): Don't return 0 if target is a MEM,
21578 equal to op0 or op1, and last_insn pattern is CODE operation
21579 with MEM dest and one of the operands matches that MEM.
21581 2013-02-11 Sriraman Tallam <tmsriramgoogle.com>
21583 * doc/extend.texi: Document Function Multiversioning and "default"
21584 parameter string to target attribute.
21585 * config/i386/i386.c (get_builtin_code_for_version): Return 0 if
21586 target attribute parameter is "default".
21587 (ix86_compare_version_priority): Remove checks for target attribute.
21588 (ix86_mangle_function_version_assembler_name): Change error to sorry.
21589 Remove check for target attribute equal to NULL. Add assert.
21590 (ix86_generate_version_dispatcher_body): Change error to sorry.
21592 2013-02-11 Iain Sandoe <iain@codesourcery.com>
21593 Jack Howarth <howarth@bromo.med.uc.edu>
21594 Patrick Marlier <patrick.marlier@gmail.com>
21597 * config/darwin.h: Replace ENDFILE_SPEC with TM_DESTRUCTOR and
21598 define ENDFILE_SPEC as TM_DESTRUCTOR.
21599 * config/i386/darwin.h (ENDFILE_SPEC): Use TM_DESTRUCTOR.
21601 2013-02-11 Alexander Potapenko <glider@google.com>
21602 Jack Howarth <howarth@bromo.med.uc.edu>
21603 Jakub Jelinek <jakub@redhat.com>
21606 * config/darwin.c (cdtor_record): Rename ctor_record.
21607 (sort_cdtor_records): Rename sort_ctor_records.
21608 (finalize_dtors): New routine to sort destructors by
21609 priority before use in assemble_integer.
21610 (machopic_asm_out_destructor): Use finalize_dtors if needed.
21612 2013-02-11 Uros Bizjak <ubizjak@gmail.com>
21614 PR rtl-optimization/56275
21615 * simplify-rtx.c (avoid_constant_pool_reference): Check that
21616 offset is non-negative and less than cmode size before
21617 calling simplify_subreg.
21619 2013-02-11 Richard Biener <rguenther@suse.de>
21621 PR tree-optimization/56264
21622 * cfgloop.h (fix_loop_structure): Adjust prototype.
21623 * loop-init.c (fix_loop_structure): Return the number of
21624 newly discovered loops.
21625 * tree-cfgcleanup.c (repair_loop_structures): When new loops
21626 are discovered, do a full loop-closed SSA rewrite.
21628 2013-02-11 Richard Biener <rguenther@suse.de>
21630 PR tree-optimization/56273
21631 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
21633 (check_array_ref): Fix missing newline in dumps.
21634 (search_for_addr_array): Likewise.
21636 2013-02-09 David Edelsohn <dje.gcc@gmail.com>
21638 * config/rs6000/aix61.h (OS_MISSING_ALTIVEC): Undefine.
21640 2013-02-09 Jakub Jelinek <jakub@redhat.com>
21643 * config/rs6000/rs6000.h (ASSEMBLER_DIALECT): Define.
21645 2013-02-08 Vladimir Makarov <vmakarov@redhat.com>
21647 PR rtl-optimization/56246
21648 * lra-constraints.c (simplify_operand_subreg): Try to reuse
21650 * lra.c (lra): Clear lra_optional_reload_pseudos only when all
21651 constraints are satisfied.
21653 2013-02-08 Jeff Law <law@redhat.com>
21656 * emit-rtl.c (reg_is_parm_p): New function.
21657 * regs.h (reg_is_parm_p): New prototype.
21658 * ira-conflicts.c (ira_build_conflicts): Allow parameters in
21659 callee-clobbered registers.
21661 2013-02-08 Michael Meissner <meissner@linux.vnet.ibm.com>
21664 * config/rs6000/rs6000.c (rs6000_builtin_vectorized_libmass):
21665 If there is no implicit builtin declaration, just return NULL.
21667 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
21669 * config/i386/sse.md (FMAMODEM): New mode iterator.
21670 (fma<mode>4, fms<mode>4, fnma<mode>4, fnms<mode>4): Use FMAMODEM
21671 mode iterator. Do not use TARGET_SSE_MATH in insn constraint.
21673 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
21675 * config/i386/gnu-user.h (TARGET_CAN_SPLIT_STACK): Define only
21676 when HAVE_GAS_CFI_PERSONALITY_DIRECTIVE is set.
21677 * config/i386/gnu-user64.h (TARGET_CAN_SPLIT_STACK): Ditto.
21679 2013-02-08 Edgar E. Iglesias <edgar.iglesias@gmail.com>
21681 * config.gcc (microblaze*-linux*): Add TARGET_BIG_ENDIAN_DEFAULT.
21682 (microblaze*-*-elf): Likewise.
21683 * config/microblaze/linux.h: Add -mbig-endian / -mlittle-endian to
21685 * config/microblaze/microblaze-c.c: Add builtin defines for
21686 _LITTLE_ENDIAN and _BIG_ENDIAN.
21687 * config/microblaze/microblaze.h: Add TARGET_ENDIAN_DEFAULT and
21688 add to TARGET_DEFAULT flags.
21689 Expand ASM_SPEC and LINK_SPEC.
21690 Update BYTES_BIG_ENDIAN and WORDS_BIG_ENDIAN.
21691 * config/microblaze/microblaze.md: Update extendsidi2 and
21692 movdi_internal instructions to use low-order / high-order reg
21694 * config/microblaze/microblaze.opt: Add mbig-endian and mlittle-endian
21695 options and inversemask / mask of LITTLE_ENDIAN.
21696 * config/microblaze/t-microblaze: Expand multilib options to
21697 include mlittle-endian (le) and update exceptions patterns.
21699 2013-02-08 Jakub Jelinek <jakub@redhat.com>
21701 PR rtl-optimization/56195
21702 * lra-constraints.c (get_reload_reg): Don't reuse regs
21703 if they have smaller mode than requested, if they have
21704 wider mode than requested, try to return a SUBREG.
21706 PR tree-optimization/56250
21707 * fold-const.c (extract_muldiv_1) <case NEGATE_EXPR>: Don't optimize
21708 if type is unsigned and code isn't MULT_EXPR.
21710 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
21712 PR tree-optimization/56064
21713 * fixed-value.c (fixed_from_double_int): Sign/zero extend payload
21714 bits according to mode.
21715 * fixed-value.h (fixed_from_double_int)
21716 (const_fixed_from_double_int): Adjust comments.
21718 2013-02-08 Richard Biener <rguenther@suse.de>
21721 * lto-streamer.h (struct data_in): Remove current_file, current_line
21722 and current_col members.
21723 * lto-streamer-out.c (lto_output_location): Stream changed bits
21724 en-block for efficiency.
21725 * lto-streamer-in.c (clear_line_info): Remove.
21726 (lto_input_location): Cache current file, line and column
21727 globally via local statics. Read changed bits en-block.
21728 (input_function): Do not call clear_line_info.
21729 (lto_read_body): Likewise.
21730 (lto_input_toplevel_asms): Likewise.
21732 2013-02-08 Michael Matz <matz@suse.de>
21734 PR tree-optimization/52448
21735 * tree-ssa-phiopt.c (struct name_to_bb): Add phase member.
21736 (nt_call_phase): New static.
21737 (add_or_mark_expr): Only mark accesses with newer phase than any
21739 (nonfreeing_call_p): New.
21740 (nt_init_block): Update nt_call_phase, mark blocks as visited.
21741 (nt_fini_block): Keep blocks marked as visited.
21742 (get_non_trapping): Initialize nt_call_phase, and reset aux pointer.
21744 2013-02-08 Richard Biener <rguenther@suse.de>
21746 * ira.c (ira): Free broken dominator information.
21748 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
21750 * config/i386/i386.c (ix86_spill_class): Use INTEGER_CLASS_P macro.
21752 2013-02-08 Marek Polacek <polacek@redhat.com>
21754 * cfgloop.c (verify_loop_structure): Add more checking of headers.
21756 2013-02-08 Richard Biener <rguenther@suse.de>
21758 PR middle-end/56181
21759 * cfgloop.h (flow_loops_find): Adjust.
21760 (bb_loop_header_p): Declare.
21761 * cfgloop.c (bb_loop_header_p): New function split out from ...
21762 (flow_loops_find): ... here. Adjust function signature,
21763 support incremental loop structure update.
21764 (verify_loop_structure): Cleanup. Verify a loop is a loop.
21765 * cfgloopmanip.c (fix_loop_structure): Move ...
21766 * loop-init.c (fix_loop_structure): ... here.
21767 (apply_loop_flags): Split out from ...
21768 (loop_optimizer_init): ... here.
21769 (fix_loop_structure): Use apply_loop_flags. Use flow_loops_find
21770 in incremental mode, only remove dead loops here.
21772 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
21775 * config/avr/avr.md (unspec) <UNSPEC_ROUND>: Add.
21776 * config/avr/avr-fixed.md (ALL4QA, ALL124QA): New mode iterators.
21777 (round<mode>3, round<mode>3_const): New expanders for fixed-mode.
21778 (*round<mode>3.libgcc): New insns for fixed-modes.
21779 * config/avr/builtins.def (ABSxx): Use a non-NULL LIBNAME.
21780 (ROUNDxx, COUNTLSxx, BITSxx, xxBITS): New DEF_BUILTINs.
21781 (ROUNDFX, COUNTLSFX, ABSFX): New DEF_BUILTINs.
21782 * config/avr/stdfix.h (absFX, bitsFX, FXbits): Remove inline
21783 implementations. Define to __builtin_avr_absFX,
21784 __builtin_avr_bitsFX, __builtin_avr_FXbits, respectively.
21785 (roundFX, countlsFX): Define to __builtin_avr_roundFX,
21786 __builtin_avr_countlsFX, respectively.
21787 * config/avr/avr-c.c (target.h): Include it.
21788 (enum avr_builtin_id): New enum.
21789 (avr_resolve_overloaded_builtin): New static function.
21790 (avr_register_target_pragmas): Use it to set
21791 targetm.resolve_overloaded_builtin.
21792 * config/avr/avr.c (avr_init_builtins): Supply myriads of local
21793 tree nodes used by DEF_BUILTIN.
21794 (avr_expand_builtin) <AVR_BUILTIN_ROUNDxx>: Sanity-check them.
21795 (avr_fold_builtin) <AVR_BUILTIN_BITSxx>: Fold to VIEW_COVERT_EXPR.
21796 <AVR_BUILTIN_xxBITS>: Same.
21798 2013-02-08 Richard Biener <rguenther@suse.de>
21800 * cfgloop.c (verify_loop_structure): Properly handle
21801 a loop exiting to another loop header.
21802 * ira-int.h (ira_loops): Remove.
21803 * ira.c (ira_loops): Remove.
21804 (ira): Use loop_optimizer_init and loop_optimizer_finalize.
21805 (do_reload): Use loop_optimizer_finalize.
21806 * ira-build.c (create_loop_tree_nodes): Use get_loops and
21807 number_of_loops to access the loop tree.
21808 (more_one_region_p): Likewise.
21809 (finish_loop_tree_nodes): Likewise.
21810 (rebuild_regno_allocno_maps): Likewise.
21811 (mark_loops_for_removal): Likewise.
21812 (mark_all_loops_for_removal): Likewise.
21813 (remove_unnecessary_regions): Likewise.
21814 (ira_build): Likewise.
21815 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
21817 2013-02-08 Richard Biener <rguenther@suse.de>
21819 * Makefile.in (tree-tailcall.o): Add $(CFGLOOP_H) dependency.
21820 * ipa-pure-const.c (analyze_function): Avoid calling
21821 mark_irreducible_loops twice.
21822 * tree-tailcall.c (tree_optimize_tail_calls_1): Mark loops for fixup.
21824 2013-02-07 David S. Miller <davem@davemloft.net>
21826 * dwarf2out.c (based_loc_descr): Perform leaf register remapping
21828 * var-tracking.c (vt_add_function_parameter): Test the presence of
21829 HAVE_window_save properly and do not remap argument registers when
21830 we have a leaf function.
21832 2013-02-07 Uros Bizjak <ubizjak@gmail.com>
21835 * ggc-page.c (ggc_print_statistics): Use HOST_LONG_LONG_FORMAT
21837 * config/i386/i386.c (ix86_print_operand): Ditto.
21839 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
21841 * lra-constraints.c (process_alt_operands): Fix recently added comment.
21843 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
21845 PR rtl-optimization/56225
21846 * lra-constraints.c (process_alt_operands): Check that reload hard
21847 reg can hold value for strict_low_part.
21849 2013-02-07 Jakub Jelinek <jakub@redhat.com>
21852 * dwarf2out.c (dwarf2_debug_hooks): Set end_function hook to
21853 dwarf2out_end_function.
21854 (in_first_function_p, maybe_at_text_label_p,
21855 first_loclabel_num_not_at_text_label): New variables.
21856 (dwarf2out_var_location): In the first function find out
21857 lowest loclabel_num N where .LVLN is known not to be equal to .Ltext0.
21858 (find_empty_loc_ranges_at_text_label, dwarf2out_end_function): New
21861 2013-02-07 Eric Botcazou <ebotcazou@adacore.com>
21863 PR rtl-optimization/56178
21864 * cse.c (cse_insn): Do not create a REG_EQUAL note if the source is a
21865 SUBREG of a register. Tidy up related block of code.
21866 * fwprop.c (forward_propagate_and_simplify): Do not create a REG_EQUAL
21867 note if the source is a register or a SUBREG of a register.
21869 2013-02-07 Jakub Jelinek <jakub@redhat.com>
21872 * config/rs6000/rs6000.md (ptrm): New mode attr.
21873 (call_indirect_aix<ptrsize>, call_indirect_aix<ptrsize>_nor11,
21874 call_value_indirect_aix<pttrsize>,
21875 call_value_indirect_aix<pttrsize>_nor11): Use <ptrm> instead of
21878 2013-02-07 Michael Haubenwallner <michael.haubenwallner@salomon.at>
21880 * collect2.c (main): Set aix64_flag for -G and -bsvr4 too, disable
21881 if -bnortl. Convert to strcmp and strncmp.
21883 2013-02-07 Alan Modra <amodra@gmail.com>
21886 * config/rs6000/rs6000.c (mem_operand_gpr): Check that LO_SUM
21887 addresses won't wrap when offsetting.
21888 (rs6000_secondary_reload): Provide secondary reloads needed for
21889 wrapping LO_SUM addresses.
21891 2013-02-06 Thomas Schwinge <thomas@codesourcery.com>
21893 * config/gnu.h (GNU_USER_TARGET_OS_CPP_BUILTINS): Never define
21894 MACH, just __MACH__.
21896 2013-02-06 Richard Biener <rguenther@suse.de>
21898 * tracer.c (tracer): Mark loops with LOOPS_NEED_FIXUP
21899 instead of calling fix_loop_structure.
21901 2013-02-06 Jakub Jelinek <jakub@redhat.com>
21903 PR middle-end/56217
21904 * omp-low.c (use_pointer_for_field): Return false if
21905 lower_send_shared_vars doesn't generate any copy-out code.
21907 2013-02-06 Tom de Vries <tom@codesourcery.com>
21909 PR rtl-optimization/56131
21910 * cfgrtl.c (delete_insn): Use NOTE_BASIC_BLOCK instead of BLOCK_FOR_INSN
21911 to get the bb of a NOTE_INSN_BASIC_BLOCK. Handle the case that the bb
21912 of the label is NULL. Add comment.
21914 2013-02-05 Jakub Jelinek <jakub@redhat.com>
21916 * tree.h (struct tree_decl_with_vis): Remove thread_local field.
21919 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Define.
21920 (STATIC_LIBTSAN_LIBS): Likewise.
21921 * gcc.c (ADD_STATIC_LIBTSAN_LIBS, LIBTSAN_EARLY_SPEC): Define.
21922 (LIBTSAN_SPEC): Add ADD_STATIC_LIBTSAN_LIBS, if LIBTSAN_EARLY_SPEC
21923 is defined, don't add anything else beyond that.
21924 (SANITIZER_EARLY_SPEC, SANITIZER_SPEC): Define.
21925 (LINK_COMMAND_SPEC): Use them.
21927 PR tree-optimization/56205
21928 * tree-stdarg.c (check_all_va_list_escapes): Return true if
21929 there are any PHI nodes that set non-va_list_escape_vars SSA_NAME
21930 and some va_list_escape_vars SSA_NAME appears in some PHI argument.
21932 2013-02-05 Richard Biener <rguenther@suse.de>
21934 PR tree-optimization/53342
21935 PR tree-optimization/53185
21936 * tree-vectorizer.h (vect_check_strided_load): Remove.
21937 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do
21938 not disallow peeling for vectorized strided loads.
21939 (vect_check_strided_load): Make static and simplify.
21940 (vect_analyze_data_refs): Adjust.
21941 * tree-vect-stmts.c (vectorizable_load): Handle peeled loops
21942 correctly when vectorizing strided loads.
21944 2013-02-05 Richard Biener <rguenther@suse.de>
21946 * doc/install.texi: Refer to ISL, not PPL.
21948 2013-02-05 Jan Hubicka <jh@suse.cz>
21950 PR tree-optimization/55789
21951 * params.def (PARAM_EARLY_INLINER_MAX_ITERATIONS): Drop to 1.
21953 2013-02-05 Jan Hubicka <jh@suse.cz>
21955 PR tree-optimization/55789
21956 * cgraphclones.c (cgraph_remove_node_and_inline_clones): Remove
21957 the dead call anyway.
21959 2013-02-05 Eric Botcazou <ebotcazou@adacore.com>
21962 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Add missing guard.
21964 2013-02-04 Alexander Potapenko <glider@google.com>
21965 Jack Howarth <howarth@bromo.med.uc.edu>
21966 Jakub Jelinek <jakub@redhat.com>
21969 * config/darwin.c (sort_ctor_records): Stabilized qsort
21970 on constructor priority by using original position.
21971 (finalize_ctors): New routine to sort constructors by
21972 priority before use in assemble_integer.
21973 (machopic_asm_out_constructor): Use finalize_ctors if needed.
21975 2013-02-04 Jakub Jelinek <jakub@redhat.com>
21978 * config/i386/winnt.c (i386_pe_assemble_visibility): Don't warn
21979 about visibility on artificial decls.
21980 * config/sol2.c (solaris_assemble_visibility): Likewise.
21982 2013-02-04 Kai Tietz <ktietz@redhat.com>
21985 * config/i386/i386.c (function_value_ms_64): Add additional valtype
21986 argument and improve checking of return-argument types for 16-byte
21988 (ix86_function_value_1): Add additional valtype argument on call
21989 of function_value_64.
21990 (return_in_memory_ms_64): Sync 16-byte sized mode handling with
21991 handling infunction_value_64 function.
21993 2013-02-04 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
21995 * reload.c (subst_reloads): Fix DEBUG_RELOAD build issue.
21997 2013-02-04 Richard Biener <rguenther@suse.de>
21999 PR tree-optimization/56188
22000 * tree-ssa-structalias.c (label_visit): Consider case with
22001 initially non-empty points-to set.
22002 (perform_var_substitution): Dump node mapping and clean up.
22004 2013-02-04 Richard Guenther <rguenther@suse.de>
22007 * lto-symtab.c (lto_symtab_merge_decls_1): Make non-builtin
22008 node prevail as last resort.
22009 (lto_symtab_merge_decls): Remove guard on LTRANS here.
22010 (lto_symtab_prevailing_decl): Builtins are their own prevailing decl.
22012 2013-02-04 Richard Biener <rguenther@suse.de>
22014 PR tree-optimization/56113
22015 * tree-ssa-structalias.c (equiv_class_lookup, equiv_class_add):
22017 (equiv_class_lookup_or_add): ... this.
22018 (label_visit): Adjust and fix error in previous patch.
22019 (perform_var_substitution): Adjust.
22021 2013-02-03 Oleg Endo <olegendo@gcc.gnu.org>
22023 * config/sh/divtab.c: Fix formatting and comments throughout the file.
22024 * config/sh/sh4-300.md: Likewise.
22025 * config/sh/sh4a.md: Likewise.
22026 * config/sh/constraints.md: Likewise.
22027 * config/sh/sh.md: Likewise.
22028 * config/sh/netbsd-elf.h: Likewise.
22029 * config/sh/predicates.md: Likewise.
22030 * config/sh/sh-protos.h: Likewise.
22031 * config/sh/ushmedia.h: Likewise.
22032 * config/sh/linux.h: Likewise.
22033 * config/sh/sh.c: Likewise.
22034 * config/sh/superh.h: Likewise.
22035 * config/sh/elf.h: Likewise.
22036 * config/sh/sh4.md: Likewise.
22037 * config/sh/sh.h: Likewise.
22039 2013-02-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
22041 * config/pa/constraints.md: Adjust unused letters. Change "T"
22042 constraint to match_test floating_point_store_memory_operand().
22043 * config/pa/predicates.md (reg_plus_base_memory_operand): New.
22044 (base14_operand): New.
22045 (floating_point_store_memory_operand): New.
22046 (integer_store_memory_operand): Revise to use base14_operand and
22047 reg_plus_base_memory_operand.
22048 (move_dest_operand): Allow symbolic_memory_operands.
22049 (symbolic_memory_operand): Check for LO_SOM.
22050 (symbolic_operand): Change default case to break.
22051 * config/pa/pa.md: Remove unamed DFmode and SFmode patterns to force
22052 CONST_DOUBLE values to be reloaded by putting them into memory when
22053 the destination is a floating point register.
22054 (movdf): Remove code to handle CONST_DOUBLE.
22056 (reload_indf_r1): New.
22057 (reload_insf_r1): New.
22058 Consistently use "Q" and "T" constraints with integer and floating
22059 point move instructions, respectively.
22060 (movdi): Remove FAIL.
22061 Change predicate for source operand unamed DImode move from
22062 general_operand to move_src_operand.
22063 (umulsidi3): Change predicate for destination operand to
22065 Likewise for similar unamed patterns.
22066 * config/pa/pa-protos.h (pa_legitimize_reload_address): Declare.
22067 * config/pa/pa.c (pa_symbolic_expression_p): Remove extra parenthesis.
22068 (hppa_legitimize_address): Simplify mask calculation.
22069 (pa_emit_move_sequence): Revised handling of secondary reloads from
22070 REG+D addresses for floating point loads and stores. Directly handle
22071 loading CONST0_RTX (mode) to a floating point register.
22072 (pa_secondary_reload): Handle reloading DF and SFmode constant values
22073 to floating point registers. Don't restrict secondary reloads to
22074 floating point registers to integer modes. Revise some comments and
22076 (TARGET_LEGITIMATE_ADDRESS_P): Define.
22077 (pa_legitimate_address_p): New.
22078 (pa_legitimize_reload_address): New.
22079 * config/pa/pa.h (STRICT_REG_OK_FOR_INDEX_P): New.
22080 (STRICT_REG_OK_FOR_BASE_P): New.
22081 (GO_IF_LEGITIMATE_ADDRESS): Delete. Update some related comments.
22082 (LEGITIMIZE_RELOAD_ADDRESS): Revise to use pa_legitimize_reload_address.
22084 2013-02-03 David Edelsohn <dje.gcc@gmail.com>
22085 Andrew Dixie <andrewd@gentrack.com>
22087 * collect2.c (GCC_CHECK_HDR): Do not scan objects with F_LOADONLY
22090 2013-02-03 Richard Sandiford <rdsandiford@googlemail.com>
22092 * expmed.c (extract_bit_field_1): Pass the full width of the
22093 structure to get_best_reg_extraction_insn.
22095 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
22098 * configure.ac (use_cxa_atexit): Add AIX.
22099 * configure: Regenerate.
22101 * config/rs6000/aix61.h (STARTFILE_SPEC): Add crtcxa.o.
22103 2013-02-01 Jakub Jelinek <jakub@redhat.com>
22106 * final.c (need_profile_function): New variable.
22107 (final_start_function): Drop ATTRIBUTE_UNUSED from first argument.
22108 If first of NOTE_INSN_BASIC_BLOCK or NOTE_INSN_FUNCTION_BEG
22109 is only preceeded by NOTE_INSN_VAR_LOCATION or NOTE_INSN_DELETED
22110 notes, targetm.asm_out.function_prologue doesn't emit anything,
22111 HAVE_prologue and profiler should be emitted before prologue,
22112 set need_profile_function instead of emitting it.
22113 (final_scan_insn): If need_profile_function, emit
22114 profile_function on the first NOTE_INSN_BASIC_BLOCK or
22115 NOTE_INSN_FUNCTION_BEG note.
22117 2013-02-01 Richard Henderson <rth@redhat.com>
22119 * config/rs6000/rs6000.md (smulditi3): New.
22122 * config/alpha/alpha.md (umulditi3): New.
22124 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
22126 * config/rs6000/xcoff.h (ASM_OUTPUT_ALIGNED_COMMON): Use floor_log2.
22127 (ASM_OUTPUT_ALIGNED_LOCAL): New.
22129 2013-02-01 Richard Biener <rguenther@suse.de>
22131 PR tree-optimization/56113
22132 * tree-ssa-structalias.c (label_visit): Reduce work for
22133 single-predecessor nodes.
22135 2013-02-01 Eric Botcazou <ebotcazou@adacore.com>
22137 * fold-const.c (make_range_step) <TRUTH_NOT_EXPR>: Bail out if the
22138 range isn't testing for zero.
22140 2013-01-31 Steven Bosscher <steven@gcc.gnu.org>
22142 PR middle-end/56113
22143 * fwprop.c (fwprop_init): Set up loops without CFG modifications.
22145 2013-01-31 Hiroyuki Ono <hiroyuki.ono.jc@renesas.com>
22146 Nick Clifton <nickc@redhat.com>
22148 * config/v850/constraints.md (Q): Define as a memory constraint.
22149 * config/v850/predicates.md (label_ref_operand): New predicate.
22150 (e3v5_shift_operand): New predicate.
22151 (ior_operator): New predicate.
22152 * config/v850/t-v850: Add e3v5 multilib.
22153 * config/v850/v850-protos.h (v850_adjust_insn_length): Prototype.
22154 (v850_gen_movdi): Prototype.
22155 * config/v850/v850.c: Add support for e3v5 architecture.
22156 Rename all uses of TARGET_V850E || TARGET_V850E2_ALL to
22158 (construct_save_jarl): Add e3v5 long JARL support.
22159 (v850_adjust_insn_length): New function. Adjust length of call
22160 insns when using e3v5 instructions.
22161 (v850_gen_movdi): New function: Generate instructions to move a
22163 * config/v850/v850.h (TARGET_CPU_v850e3v5): Define.
22164 (CPP_SPEC): Define __v850e3v5__ as appropriate.
22165 (TARGET_USE_FPU): Enable for e3v5.
22166 (CONST_OK_FOR_W): New macro.
22167 (ADJUST_INSN_LENGTH): Define.
22168 * config/v850/v850.md (UNSPEC_LOOP): Define.
22169 (attr cpu): Add v850e3v5.
22170 Rename all uses of TARGET_V850E2 to TARGET_V850E2V3_UP.
22171 (movdi): New pattern.
22172 (movdi_internal): New pattern.
22173 (cbranchsf4): Conditionalize on TARGET_USE_FPU.
22174 (cbranchdf4): Conditionalize on TARGET_USE_FPU.
22175 (cstoresf4): Likewise.
22176 (cstoredf4): Likewise.
22177 (insv): New pattern.
22178 (rotlso3_a): New pattern.
22179 (rotlsi3_b): New pattern
22180 (rotlsi3_v850e3v5): New pattern.
22181 (doloop_begin): New pattern.
22182 (fix_loop_counter): New pattern.
22183 (doloop_end): New pattern.
22184 (branch_normal): Add e3v5 long branch support.
22185 (branch_invert): Likewise.
22186 (branch_z_normal): Likewise.
22187 (branch_z_invert): Likewise.
22188 (branch_nz_normal): Likewise.
22189 (branch_nz_invert): Likewise.
22190 (call_internal_short): Add e3v5 register-indirect JARL support.
22191 (call_internal_long): Likewise.
22192 (call_value_internal_short): Likewise.
22193 (call_value_internal_long): Likewise.
22194 * config/v850/v850.opt (mv850e3v5, mv850e2v4): New options.
22195 (mloop): New option.
22196 * config.gcc: Add support for configuring v840e3v5 target.
22197 * doc/invoke.texi: Document new v850 specific command line options.
22199 2013-01-31 Paul Koning <ni1d@arrl.net>
22203 * dwarf2out.c (prune_unused_types_mark): Mark all of parent's
22204 children if parent is a class.
22205 (prune_unused_types_prune): Don't add DW_AT_declaration.
22207 2013-01-31 Richard Biener <rguenther@suse.de>
22209 PR tree-optimization/56157
22210 * tree-vect-slp.c (vect_get_slp_defs): More thoroughly try to
22211 match up operand with SLP child.
22213 2013-01-31 Jason Merrill <jason@redhat.com>
22216 * dwarf2out.c (gen_struct_or_union_type_die): Always schedule template
22217 parameters the first time.
22218 (gen_scheduled_generic_parms_dies): Check completeness here.
22220 2013-01-31 Richard Biener <rguenther@suse.de>
22222 PR middle-end/53073
22223 * common.opt (faggressive-loop-optimizations): New flag,
22224 enabled by default.
22225 * doc/invoke.texi (faggressive-loop-optimizations): Document.
22226 * tree-ssa-loop-niter.c (estimate_numbers_of_iterations_loop): Guard
22227 infer_loop_bounds_from_undefined by it.
22229 2013-01-31 Richard Biener <rguenther@suse.de>
22231 PR tree-optimization/56150
22232 * tree-ssa-loop-manip.c (find_uses_to_rename_stmt): Do not
22233 visit virtual operands.
22234 (find_uses_to_rename_bb): Likewise.
22236 2013-01-31 Richard Biener <rguenther@suse.de>
22238 PR tree-optimization/56150
22239 * tree-ssa-tail-merge.c (gimple_equal_p): Properly handle
22240 mixed store non-store stmts.
22242 2013-01-30 Jakub Jelinek <jakub@redhat.com>
22245 * gcc.c (LIBASAN_SPEC): Define just to ADD_STATIC_LIBASAN_LIBS if
22246 LIBASAN_EARLY_SPEC is defined.
22247 (LIBASAN_EARLY_SPEC): Define to empty string if not already defined.
22248 (LINK_COMMAND_SPEC): Add LIBASAN_EARLY_SPEC for -fsanitize=address,
22250 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Define.
22253 * config/i386/i386.c (ix86_valid_target_attribute_inner_p): Diagnose
22254 invalid args instead of ICEing on it.
22255 (ix86_valid_target_attribute_tree): Return error_mark_node if
22256 ix86_valid_target_attribute_inner_p failed.
22257 (ix86_valid_target_attribute_p): Return false only if
22258 ix86_valid_target_attribute_tree returned error_mark_node. Allow
22259 target("default") attribute.
22260 (sorted_attr_string): Change argument from const char * to tree,
22261 merge in all target attribute arguments rather than just one.
22262 Formatting fix. Use XNEWVEC instead of xmalloc and XDELETEVEC
22263 instead of free. Avoid using strcat.
22264 (ix86_mangle_function_version_assembler_name): Mangle
22265 target("default") as if no target attribute is present. Adjust
22266 sorted_attr_string caller. Avoid leaking memory. Use XNEWVEC
22267 instead of xmalloc and XDELETEVEC instead of free.
22268 (ix86_function_versions): Don't return true if one of the decls
22269 doesn't have target attribute. If they don't and one of the decls
22270 is DECL_FUNCTION_VERSIONED, report an error. Adjust
22271 sorted_attr_string caller. Use XDELETEVEC instead of free.
22272 (ix86_supports_function_versions): Remove.
22273 (make_name): Fix up formatting.
22274 (make_dispatcher_decl): Remove resolver_name and its initialization.
22275 Avoid leaking memory.
22276 (is_function_default_version): Return true if there is
22277 target("default") attribute rather than no target attribute at all.
22278 (make_resolver_func): Avoid leaking memory.
22279 (ix86_generate_version_dispatcher_body): Likewise.
22280 (TARGET_OPTION_SUPPORTS_FUNCTION_VERSIONS): Remove.
22281 * target.def (supports_function_versions): Remove.
22282 * doc/tm.texi.in (SUPPORTS_FUNCTION_VERSIONS): Remove.
22283 * doc/tm.texi: Regenerated.
22285 2013-01-30 Vladimir Makarov <vmakarov@redhat.com>
22287 PR rtl-optimization/56144
22288 * lra-constraints.c (get_reload_reg): Don't reuse reload pseudo
22289 for values with side effects.
22291 2013-01-30 Richard Biener <rguenther@suse.de>
22293 * sparseset.h (sparseset_bit_p): Use gcc_checking_assert.
22294 (sparseset_pop): Likewise.
22295 * cfganal.c (compute_idf): Likewise. Increase work-stack size
22296 to be able to use quick_push in the worker loop.
22298 2013-01-30 Marek Polacek <polacek@redhat.com>
22300 * cfgcleanup.c (cleanup_cfg): Don't mark affected BBs.
22302 2013-01-30 Richard Biener <rguenther@suse.de>
22305 * lto-symtab.c (lto_symtab_merge_decls_1): Guard DECL_BUILT_IN check.
22307 2013-01-30 Georg-Johann Lay <avr@gjlay.de>
22309 PR tree-optimization/56064
22310 * fixed-value.c (fixed_from_double_int): New function.
22311 * fixed-value.h (fixed_from_double_int): New prototype.
22312 (const_fixed_from_double_int): New static inline function.
22313 * fold-const.c (native_interpret_fixed): New static function.
22314 (native_interpret_expr) <FIXED_POINT_TYPE>: Use it.
22315 (can_native_interpret_type_p) <FIXED_POINT_TYPE>: Return true.
22316 (native_encode_fixed): New static function.
22317 (native_encode_expr) <FIXED_CST>: Use it.
22318 (native_interpret_int): Move double_int worker code to...
22319 * double-int.c (double_int::from_buffer): ...this new static method.
22320 * double-int.h (double_int::from_buffer): Prototype it.
22322 2013-01-30 Richard Biener <rguenther@suse.de>
22324 * tree-ssa-structalias.c (final_solutions, final_solutions_obstack):
22325 New pointer-map and obstack.
22326 (init_alias_vars): Allocate pointer-map and obstack.
22327 (delete_points_to_sets): Free them.
22328 (find_what_var_points_to): Cache result.
22329 (find_what_p_points_to): Adjust for changed interface of
22330 find_what_var_points_to.
22331 (compute_points_to_sets): Likewise.
22332 (ipa_pta_execute): Likewise.
22334 2013-01-30 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
22336 * configure.ac (HAVE_AS_SPARC_NOBITS): New test.
22337 * configure: Regenerate.
22338 * config.in: Regenerate.
22339 * config/sparc/sparc.c (sparc_solaris_elf_asm_named_section): Emit
22340 #nobits/#progbits if supported.
22342 2013-01-29 Oleg Endo <olegendo@gcc.gnu.org>
22345 * config/sh/sh.md (bclr_m2a, bset_m2a, bst_m2a, bld_m2a, bldsign_m2a,
22346 bld_reg, *bld_regqi, band_m2a, bandreg_m2a, bor_m2a, borreg_m2a,
22347 bxor_m2a, bxorreg_m2a): Add satisfies_constraint_K03 condition.
22349 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
22351 * config/arm/cortex-a7.md (cortex_a7_neon, cortex_a7_all): Remove.
22352 (cortex_a7_idiv): Use cortex_a7_both instead of cortex_a7_all.
22354 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
22356 * config/arm/arm.c (cortexa7_younger): Return true for TYPE_CALL.
22357 * config/arm/cortex-a7.md (cortex_a7_call): Update required units.
22359 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
22361 * config/arm/arm-protos.h (arm_mac_accumulator_is_result): New
22363 * config/arm/arm.c (arm_mac_accumulator_is_result): New function.
22364 * config/arm/cortex-a7.md: New bypasses using
22365 arm_mac_accumulator_is_result.
22367 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
22369 * config/arm/cortex-a7.md (cortex_a7_neon_mul): New reservation.
22370 (cortex_a7_neon_mla): Likewise.
22371 (cortex_a7_fpfmad): New reservation.
22372 (cortex_a7_fpmacs): Use ffmas and update required units.
22373 (cortex_a7_fpmuld): Update required units and latency.
22374 (cortex_a7_fpmacd): Likewise.
22375 (cortex_a7_fdivs, cortex_a7_fdivd): Likewise.
22376 (cortex_a7_neon). Likewise.
22377 (bypass) Update participating units.
22379 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
22381 * config/arm/arm.md (type): Add ffmas and ffmad to "type" attribute.
22382 * config/arm/vfp.md (fma,fmsub,fnmsub,fnmadd): Change type
22384 * config/arm/vfp11.md (vfp_farith): Use ffmas.
22385 (vfp_fmul): Use ffmad.
22386 * config/arm/cortex-r4f.md (cortex_r4_fmacs): Use ffmas.
22387 (cortex_r4_fmacd): Use ffmad.
22388 * config/arm/cortex-m4-fpu.md (cortex_m4_fmacs): Use ffmas.
22389 * config/arm/cortex-a9.md (cortex_a9_fmacs): Use ffmas.
22390 (cortex_a9_fmacd): Use ffmad.
22391 * config/arm/cortex-a8-neon.md (cortex_a8_vfp_macs): Use ffmas.
22392 (cortex_a8_vfp_macd): Use ffmad.
22393 * config/arm/cortex-a5.md (cortex_a5_fpmacs): Use ffmas.
22394 (cortex_a5_fpmacd): Use ffmad.
22395 * config/arm/cortex-a15-neon.md (cortex_a15_vfp_macs) Use ffmas.
22396 (cortex_a15_vfp_macd): Use ffmad.
22397 * config/arm/arm1020e.md (v10_fmul): Use ffmas and ffmad.
22399 2013-01-29 Jason Merrill <jason@redhat.com>
22402 * varasm.c (default_assemble_visibility): Don't warn about
22403 visibility on artificial decls.
22405 2013-01-29 Richard Biener <rguenther@suse.de>
22407 PR tree-optimization/56113
22408 * tree-ssa-structalias.c (equiv_class_lookup): Also return
22410 (label_visit): Free duplicate bitmaps and record the leader instead.
22411 (perform_var_substitution): Adjust.
22413 2013-01-29 Richard Biener <rguenther@suse.de>
22415 PR tree-optimization/55270
22416 * tree-ssa-dom.c (eliminate_degenerate_phis): If we changed
22417 the CFG, schedule loops for fixup.
22419 2013-01-29 Nick Clifton <nickc@redhat.com>
22421 * config/rl78/rl78.c (rl78_regno_mode_code_ok_for_base_p): Allow
22424 2013-01-28 Leif Ekblad <leif@rdos.net>
22426 * config.gcc (i[34567]86-*-rdos*, x86_64-*-rdos*): New targets.
22427 * config/i386/i386.h (TARGET_RDOS): New macro.
22428 (DEFAULT_LARGE_SECTION_THRESHOLD): New macro.
22429 * config/i386/i386.c (ix86_option_override_internal): For 64bit
22430 TARGET_RDOS, set ix86_cmodel to CM_MEDIUM_PIC and flag_pic to 1.
22431 * config/i386/i386.opt (mlarge-data-threshold): Initialize to
22432 DEFAULT_LARGE_SECTION_THRESHOLD.
22433 * config/i386/i386.md (R14_REG, R15_REG): New constants.
22434 * config/i386/rdos.h: New file.
22435 * config/i386/rdos64.h: New file.
22437 2013-01-28 Bernd Schmidt <bernds@codesourcery.com>
22440 * reload.c (find_valid_class_1): Use in_hard_reg_set_p instead of
22443 2013-01-28 Jakub Jelinek <jakub@redhat.com>
22445 PR rtl-optimization/56117
22446 * sched-deps.c (sched_analyze_2) <case PREFETCH>: For use_cselib
22447 call cselib_lookup_from_insn on the MEM before calling
22448 add_insn_mem_dependence.
22450 2013-01-28 Richard Biener <rguenther@suse.de>
22452 * tree-inline.c (remap_gimple_stmt): Do not assing a BLOCK
22453 to a stmt that didn't have one.
22454 (copy_phis_for_bb): Likewise for PHI arguments.
22455 (copy_debug_stmt): Likewise for debug stmts.
22457 2013-01-28 Richard Biener <rguenther@suse.de>
22459 PR tree-optimization/56034
22460 * tree-loop-distribution.c (enum partition_kind): Add PKIND_REDUCTION.
22461 (partition_builtin_p): Adjust.
22462 (generate_code_for_partition): Handle PKIND_REDUCTION. Assert
22463 it is the last partition.
22464 (rdg_flag_uses): Check SSA_NAME_IS_DEFAULT_DEF before looking
22465 up the vertex for the definition.
22466 (classify_partition): Classify whether a partition is a
22467 PKIND_REDUCTION, thus has uses outside of the loop.
22468 (ldist_gen): Inherit PKIND_REDUCTION when merging partitions.
22469 Merge all PKIND_REDUCTION partitions into the last partition.
22470 (tree_loop_distribution): Seed partitions from reductions as well.
22472 2013-01-28 Jakub Jelinek <jakub@redhat.com>
22474 PR tree-optimization/56125
22475 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Don't optimize
22476 pow(x,c) into sqrt(x) * powi(x, n/2) or
22477 1.0 / (sqrt(x) * powi(x, abs(n/2))) if c is an integer or when
22478 optimizing for size.
22479 Don't optimize pow(x,c) into powi(x, n/3) * powi(cbrt(x), n%3) or
22480 1.0 / (powi(x, abs(n)/3) * powi(cbrt(x), abs(n)%3)) if 2c is an
22483 PR tree-optimization/56094
22484 * gimplify.c (force_gimple_operand_1): Temporarily set input_location
22485 to UNKNOWN_LOCATION while gimplifying expr.
22487 2013-01-27 Uros Bizjak <ubizjak@gmail.com>
22490 * config/i386/i386.md (*movabs<mode>_1): Add square brackets around
22491 operand 0 in movabs insn template for -masm=intel asm alternative.
22492 (*movabs<mode>_2): Ditto for operand 1.
22494 2013-01-26 David Holsgrove <david.holsgrove@xilinx.com>
22497 * config.gcc (microblaze*-linux*): Add tmake_file to allow building
22500 2013-01-26 Edgar E. Iglesias <edgar.iglesias@gmail.com>
22502 * config.gcc (microblaze*-*-*): Rename microblaze*-*-elf, update
22505 2013-01-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
22507 * config/aarch64/aarch64.c (TARGET_FIXED_CONDITION_CODE_REGS):
22508 Undef to avoid warning.
22510 2013-01-25 Michael Haubenwallner <michael.haubenwallner@salomon.at>
22512 * configure.ac (gcc_cv_ld_static_dynamic): Define for AIX native ld.
22513 * configure: Regenerate.
22515 2013-01-25 Jakub Jelinek <jakub@redhat.com>
22517 PR tree-optimization/56098
22518 * tree-ssa-phiopt.c (nt_init_block): Don't call add_or_mark_expr
22519 for stmts with volatile ops.
22520 (cond_store_replacement): Don't optimize if assign has volatile ops.
22521 (cond_if_else_store_replacement_1): Don't optimize if either
22522 then_assign or else_assign have volatile ops.
22523 (hoist_adjacent_loads): Don't optimize if either def1 or def2 have
22526 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
22528 * doc/invoke.texi (AVR Built-in Macros): Document __XMEGA__.
22530 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
22532 * doc/extend.texi (Example of asm with clobbered asm reg): Fix
22533 missing ':' in asm example.
22535 2013-01-25 Tejas Belagod <tejas.belagod@arm.com>
22537 * config/aarch64/aarch64-simd-builtins.def: Separate sq<r>dmulh_lane
22538 entries into lane and laneq entries.
22539 * config/aarch64/aarch64-simd.md (aarch64_sq<r>dmulh_lane<mode>):
22540 Remove AdvSIMD scalar modes.
22541 (aarch64_sq<r>dmulh_laneq<mode>): New.
22542 (aarch64_sq<r>dmulh_lane<mode>): New RTL pattern for Scalar AdvSIMD
22544 * config/aarch64/arm_neon.h: Fix all the vq<r>dmulh_lane* intrinsics'
22545 builtin implementations to relfect changes in RTL in aarch64-simd.md.
22546 * config/aarch64/iterators.md (VCOND): New.
22549 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
22552 * config/avr/builtins.def (DEF_BUILTIN): Add LIBNAME argument.
22553 Add NULL LIBNAME argument to existing definitions.
22554 (ABSHR, ABSR, ABSLR, ABSLLR, ABSHK, ABSK, ABSLK, ABSLLK): New.
22555 * config/avr/avr-c.c (DEF_BUILTIN): Add LIBNAME argument.
22556 * config/avr/avr.c (DEF_BUILTIN): Same.
22557 (avr_init_builtins): Pass down LIBNAME to add_builtin_function.
22558 (avr_expand_builtin): Expand to a vanilla call if a libgcc
22559 implementation is available (DECL_ASSEMBLER_NAME is set).
22560 (avr_fold_absfx): New static function.
22561 (avr_fold_builtin): Use it to handle: AVR_BUILTIN_ABSHR,
22562 AVR_BUILTIN_ABSR, AVR_BUILTIN_ABSLR, AVR_BUILTIN_ABSLLR,
22563 AVR_BUILTIN_ABSHK, AVR_BUILTIN_ABSK, AVR_BUILTIN_ABSLK,
22564 AVR_BUILTIN_ABSLLK.
22565 * config/avr/stdfix.h (abshr, absr, abslr, absllr)
22566 (abshk, absk, abslk, absllk): Provide as static inline functions.
22568 2013-01-25 Marek Polacek <polacek@redhat.com>
22570 PR tree-optimization/56035
22571 * cfgloopmanip.c (fix_loop_structure): Remove redundant condition.
22573 2012-01-24 Uros Bizjak <ubizjak@gmail.com>
22575 * config/i386/i386.md (*movti_internal_rex64): Add (o,e) alternative.
22576 (*movtf_internal_rex64): Add (!o,C) alternative
22577 (*movxf_internal_rex64): Ditto.
22578 (*movdf_internal_rex64): Add (?r,C) and (?m,C) alternatives.
22580 2013-01-24 Shenghou Ma <minux.ma@gmail.com>
22582 * doc/invoke.texi: fix typo.
22583 * doc/objc.texi: fix typo.
22585 2013-01-24 Richard Sandiford <rdsandiford@googlemail.com>
22587 * config/mips/mips.md (*and<mode>3_mips16): Use the "W" constraint
22588 for the first two alternatives.
22590 2013-01-24 Diego Novillo <dnovillo@google.com>
22592 * Makefile.in (GGC): Remove. Replace all instances with ggc-page.o.
22593 (ggc-zone.o): Remove.
22594 * configure.ac: Remove option --with-gc.
22595 * configure: Re-generate.
22596 * doc/install.texi: Remove documentation for --with-gc.
22597 * gengtype.c (write_enum_defn): Remove. Update all users.
22598 (write_Types_process_field): Remove generation of gt_e_* argument.
22599 (output_type_enum): Remove. Update all users.
22600 (write_enum_defn): Remove. Update all users.
22601 (enum alloc_zone): Remove. Update all users.
22602 (write_splay_tree_allocator_def): Remove generation of gt_e_* argument.
22603 * ggc-common.c (ggc_splay_alloc): Remove first argument.
22604 Update all callers.
22605 (struct ptr_data): Remove field TYPE. Update all users.
22606 (gt_pch_note_object): Remove argument TYPE. Update all users.
22607 * ggc-internal.h (ggc_pch_alloc_object): Remove last argument.
22609 * ggc-none.c (ggc_alloc_typed_stat): Remove.
22610 (struct alloc_zone): Remove.
22611 (ggc_internal_alloc_zone_stat): Remove.
22612 (ggc_internal_cleared_alloc_zone_stat): Remove.
22613 * ggc-page.c (ggc_alloc_typed_stat): Remove.
22614 (ggc_pch_count_object): Remove last argument. Update all users.
22615 (ggc_pch_alloc_object): Remove last argument. Update all users.
22616 (struct alloc_zone): Remove.
22617 * ggc-zone.c: Remove.
22618 * ggc.h (gt_pch_note_object): Remove last argument. Update all users.
22619 (struct alloc_zone): Remove.
22620 (ggc_alloc_typed_stat): Remove.
22621 (ggc_alloc_typed): Remove.
22622 (ggc_splay_alloc): Remove first argument.
22623 (rtl_zone): Remove. Update all users.
22624 (tree_zone): Remove. Update all users.
22625 (tree_id_zone): Remove. Update all users.
22626 (ggc_internal_zone_alloc_stat): Remove. Update all users.
22627 (ggc_internal_zone_cleared_alloc_stat): Remove. Update all users.
22628 (ggc_internal_zone_vec_alloc_stat): Remove. Update all users.
22629 * tree-ssanames.c: Remove references to zone allocator in comments.
22631 2013-01-24 Georg-Johann Lay <avr@gjlay.de>
22633 * config/avr/avr.c (avr_out_fract): Make register numbers that
22634 might be outside of source operand signed.
22636 2013-01-24 Uros Bizjak <ubizjak@gmail.com>
22638 * config/i386/constraints.md (Yf): New constraint.
22639 * config/i386/i386.md (*movdf_internal_rex64): Use Yf*f instead
22640 of f constraint to conditionaly disable x87 register preferences.
22641 (*movdf_internal): Ditto.
22642 (*movsf_internal): Ditto.
22644 2013-01-24 Steven Bosscher <steven@gcc.gnu.org>
22646 PR inline-asm/55934
22647 * lra-assigns.c (assign_by_spills): Throw away the pattern of asms
22648 that have operands with impossible constraints.
22649 Add a FIXME for a speed-up opportunity.
22650 * lra-constraints.c (process_alt_operands): Verify that a class
22651 selected from constraints on asms is valid for the operand mode.
22652 (curr_insn_transform): Remove incorrect comment.
22654 2013-01-23 David Edelsohn <dje.gcc@gmail.com>
22656 * config/rs6000/rs6000.c (rs6000_delegitimize_address): Check that
22657 TOC operand is a valid symbol ref in the constant pool.
22659 2013-01-23 Edgar E. Iglesias <edgar.iglesias@gmail.com>
22661 * config/microblaze/linux.h: Add TARGET_OS_CPP_BUILTINS
22663 2013-01-23 Georg-Johann Lay <avr@gjlay.de>
22666 * config/avr/stdfix.h: New file.
22667 * t-avr (stdfix-gcc.h): New rule to build it.
22668 (EXTRA_HEADERS): Set it to install stdfix.h, stdfix-gcc.h.
22670 2013-01-23 Kostya Serebryany <kcc@google.com>
22672 * config/darwin.h: remove dependency on
22673 CoreFoundation (asan on Mac OS).
22675 2013-01-23 Jakub Jelinek <jakub@redhat.com>
22678 * config/arm/arm.md (cbranchdi4, cstoredi4): Use s_register_operand
22679 instead of cmpdi_operand for first comparison operand.
22680 Don't assert that comparison operands aren't both constants.
22682 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
22684 * doc/install.texi (Downloading the Source): Update references to
22685 downloading separate components.
22687 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
22689 * doc/extend.texi (__int128): Improve grammar.
22691 2013-01-22 Uros Bizjak <ubizjak@gmail.com>
22694 * config/i386/i386.md (*movti_internal_rex64): Change (o,riF)
22695 alternative to (o,r).
22696 (*movdi_internal_rex64): Remove (!o,n) alternative.
22697 (DImode immediate->memory splitter): Remove.
22698 (DImode immediate->memory peephole2): Remove.
22699 (movtf): Enable for TARGET_64BIT || TARGET_SSE.
22700 (*movtf_internal_rex64): Rename from *movtf_internal. Change (!o,F*r)
22701 alternative to (!o,*r).
22702 (*movtf_internal_sse): New pattern.
22703 (*movxf_internal_rex64): New pattern.
22704 (*movxf_internal): Disable for TARGET_64BIT.
22705 (*movdf_internal_rex64): Remove (!o,F) alternative.
22707 2013-01-22 Jakub Jelinek <jakub@redhat.com>
22709 PR middle-end/56074
22710 * dumpfile.c (dump_loc): Only print loc if LOCATION_LOCUS (loc)
22711 isn't UNKNOWN_LOCATION nor BUILTINS_LOCATION.
22712 * tree-vect-loop-manip.c (find_loop_location): Also ignore
22713 stmt locations where LOCATION_LOCUS of the stmt location is
22714 UNKNOWN_LOCATION or BUILTINS_LOCATION.
22717 * config/i386/i386.md (UNSPEC_STOS): New.
22718 (strset_singleop, *strsetdi_rex_1, *strsetsi_1, *strsethi_1,
22719 *strsetqi_1): Add UNSPEC_STOS.
22721 2013-01-22 Paolo Carlini <paolo.carlini@oracle.com>
22724 * doc/invoke.texi: Remove left over -Wsynth example.
22726 2013-01-21 Jakub Jelinek <jakub@redhat.com>
22728 PR tree-optimization/56051
22729 * fold-const.c (fold_binary_loc): Don't fold
22730 X < (cast) (1 << Y) into (X >> Y) != 0 if cast is either
22731 a narrowing conversion, or widening conversion from signed
22734 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
22736 PR rtl-optimization/56023
22737 * haifa-sched.c (fix_inter_tick): Do not update ticks of instructions,
22738 dependent on debug instruction.
22740 2013-01-21 Martin Jambor <mjambor@suse.cz>
22742 PR middle-end/56022
22743 * function.c (allocate_struct_function): Call
22744 invoke_set_current_function_hook earlier.
22746 2013-01-21 Jakub Jelinek <jakub@redhat.com>
22748 * reload1.c (init_reload): Only initialize reload_obstack
22749 during the first call.
22751 2013-01-21 Marek Polacek <polacek@redhat.com>
22753 * cfgloop.c (verify_loop_structure): Fix up grammar.
22755 2013-01-21 Yi-Hsiu Hsu <ahsu@marvell.com>
22757 * config/arm/marvell-pj4.md (pj4_shift_conds, pj4_alu_shift,
22758 pj4_alu_shift_conds, pj4_shift): Handle simple_alu_shift.
22760 2013-01-21 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
22763 * config/arm/marvell-pj4.md: Update copyright year.
22764 Fix up use of alu to alu_reg and simple_alu_imm.
22766 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
22768 * config/i386/i386.md (enabled): Do not disable fma4 for TARGET_FMA.
22770 2013-01-20 Vladimir Makarov <vmakarov@redhat.com>
22773 * lra-constraints.c (curr_insn_transform): Don't reuse original
22774 insn for secondary memory move when memory mode should be different.
22776 2013-01-20 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
22778 * config/pa/pa.md (atomic_loaddi, atomic_loaddi_1, atomic_storedi,
22779 atomic_storedi_1): New patterns.
22781 2013-01-20 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
22783 btver2 pipeline descriptions.
22784 * config/i386/i386.c: Enable CPU_BTVER2 to use btver2 pipeline
22786 * config/i386/i386.md (btver2_decode): New type attributes.
22787 * config/i386/sse.md (btver2_decode, btver2_sse_attr): New
22789 * config/i386/btver2.md: New file describing btver2 pipelines.
22791 2013-01-19 Andrew Pinski <apinski@cavium.com>
22793 PR tree-optimization/52631
22794 * tree-ssa-sccvn (visit_use): Before looking up the original
22795 statement, try looking up the simplified expression.
22797 2013-01-19 Anthony Green <green@moxielogic.com>
22799 * config/moxie/moxie.c (moxie_expand_prologue): Set
22800 current_function_static_stack_size.
22802 2013-01-18 Jakub Jelinek <jakub@redhat.com>
22804 PR tree-optimization/56029
22805 * tree-phinodes.c (reserve_phi_args_for_new_edge): Set
22806 gimple_phi_arg_location for the new arg to UNKNOWN_LOCATION.
22808 2013-01-18 Sharad Singhai <singhai@google.com>
22810 PR tree-optimization/55995
22811 * dumpfile.c (dump_loc): Print location only if available.
22812 * tree-vectorizer.c (increase_alignment): Intialize vect_location.
22814 2013-01-18 Vladimir Makarov <vmakarov@redhat.com>
22817 * lra-constraints.c (curr_insn_transform): Reuse original insn for
22818 secondary memory move.
22819 (inherit_reload_reg): Use rclass instead of cl for
22820 check_secondary_memory_needed_p.
22822 2013-01-18 Jakub Jelinek <jakub@redhat.com>
22824 PR middle-end/56015
22825 * expr.c (expand_expr_real_2) <case COMPLEX_EXPR>: Handle
22826 the case where writing real complex part of target modifies op1.
22828 2013-01-18 James Greenhalgh <james.greenhalgh@arm.com>
22830 * config/aarch64/aarch64-simd.md
22831 (aarch64_vcond_internal<mode>): Handle unordered cases.
22832 * config/aarch64/iterators.md (v_cmp_result): New.
22834 2013-01-18 Yi-Hsiu Hsu <ahsu@marvell.com>
22835 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
22837 * config/arm/marvell-pj4.md: New file.
22838 * config/arm/arm.c (arm_issue_rate): Add marvell_pj4.
22839 * config/arm/arm.md (generic_sched): Add marvell_pj4.
22840 (generic_vfp): Likewise.
22841 * config/arm/arm-cores.def: Add marvell-pj4.
22842 * config/arm/arm-tune.md: Regenerate.
22843 * config/arm/arm-tables.opt: Regenerate.
22844 * config/arm/bpabi.h (BE8_LINK_SPEC): Add marvell_pj4.
22845 * doc/invoke.texi: Document marvell-pj4.
22847 2013-01-18 Tejas Belagod <tejas.belagod@arm.com>
22849 * config/aarch64/arm_neon.h: Map scalar types to standard types.
22851 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
22856 * var-tracking.c (negative_power_of_two_p): New.
22857 (global_get_addr_cache, local_get_addr_cache): New.
22858 (get_addr_from_global_cache, get_addr_from_local_cache): New.
22859 (vt_canonicalize_addr): Rewrite using the above. Adjust the
22861 (vt_stack_offset_p): Remove.
22862 (vt_canon_true_dep): Always canonicalize loc's address.
22863 (clobber_overlapping_mems): Make sure we have a MEM.
22864 (local_get_addr_clear_given_value): New.
22865 (val_reset): Clear local cached entries.
22866 (compute_bb_dataflow): Create and release the local cache.
22867 Disable duplicate MEMs clobbering.
22868 (emit_notes_in_bb): Clobber MEMs likewise.
22869 (vt_emit_notes): Create and release the local cache.
22870 (vt_initialize, vt_finalize): Create and release the global
22871 cache, respectively.
22872 * alias.c (rtx_equal_for_memref_p): Compare operands of ENTRY_VALUEs.
22874 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
22876 PR libmudflap/53359
22877 * tree-mudflap.c (mudflap_finish_file): Skip deferred decls
22878 not found in the symtab.
22880 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
22883 PR rtl-optimization/55547
22884 PR rtl-optimization/53827
22887 * alias.c (offset_overlap_p): New, factored out of...
22888 (memrefs_conflict_p): ... this. Use absolute sizes. Retain
22889 the conservative special case for symbolic constants. Don't
22890 adjust zero sizes on alignment.
22892 2013-01-18 Bernd Schmidt <bernds@codesourcery.com>
22894 PR rtl-optimization/52573
22895 * regrename.c (build_def_use): Ignore REG_DEAD notes if there is a
22896 REG_UNUSED for the same register.
22898 2013-01-17 Richard Biener <rguenther@suse.de>
22899 Marek Polacek <polacek@redhat.com>
22901 PR rtl-optimization/55833
22902 * loop-unswitch.c (unswitch_loops): Move loop verification...
22903 (unswitch_single_loop): ...here. Call mark_irreducible_loops.
22904 * cfgloopmanip.c (fix_loop_placement): Add IRRED_INVALIDATED parameter.
22905 Set it to true when we're removing a loop from hierarchy tree in
22906 an irreducible region.
22907 (fix_bb_placements): Adjust caller.
22908 (fix_loop_placements): Likewise.
22910 2013-01-17 Georg-Johann Lay <avr@gjlay.de>
22912 * config/avr/builtins.def (DEF_BUILTIN): Factor out
22913 "__builtin_avr_" from NAME, turn NAME to an uppercase identifier.
22914 Factor out 'CODE_FOR_' from ICODE, use 'nothing' instead of '-1'.
22915 Remove ID. Adjust comments.
22916 * config/avr/avr-c.c (avr_builtin_name): Remove.
22917 (avr_cpu_cpp_builtins): Use DEF_BUILTIN instead of for-loop.
22918 * config/avr/avr.c (avr_tolower): New static function.
22919 (DEF_BUILTIN): Remove parameter ID. Prefix ICODE by 'CODE_FOR_'.
22920 Stringify NAME, prefix it with "__builtin_avr_" and lowercase it.
22921 (avr_expand_builtin): Assert insn_code != CODE_FOR_nothing for
22924 2013-01-17 Jan Hubicka <jh@suse.cz>
22926 PR tree-optimization/55273
22927 * loop-iv.c (iv_number_of_iterations): Consider zero iteration case.
22929 2013-01-17 Uros Bizjak <ubizjak@gmail.com>
22932 * config/i386/sync.md (atomic_store<mode>): Always generate SWImode
22933 store through atomic_store<mode>_1.
22934 (atomic_store<mode>_1): Macroize insn using SWI mode iterator.
22936 2013-01-17 Martin Jambor <mjambor@suse.cz>
22938 PR tree-optimizations/55264
22939 * ipa-inline-transform.c (can_remove_node_now_p_1): Never return true
22940 for virtual methods.
22941 * ipa.c (symtab_remove_unreachable_nodes): Never return true for
22942 virtual methods before inlining is over.
22943 * cgraph.h (cgraph_only_called_directly_or_aliased_p): Return false for
22945 * cgraphclones.c (cgraph_create_virtual_clone): Mark clones as
22948 2013-01-16 Vladimir Makarov <vmakarov@redhat.com>
22950 PR rtl-optimization/56005
22951 * sched-deps.c (sched_analyze_2): Check deps->readonly for adding
22952 pending reads for prefetch.
22954 2013-01-16 Ian Bolton <ian.bolton@arm.com>
22956 * config/aarch64/aarch64.md
22957 (*cstoresi_neg_uxtw): New pattern.
22958 (*cmovsi_insn_uxtw): New pattern.
22959 (*<optab>si3_uxtw): New pattern.
22960 (*<LOGICAL:optab>_<SHIFT:optab>si3_uxtw): New pattern.
22961 (*<optab>si3_insn_uxtw): New pattern.
22962 (*bswapsi2_uxtw): New pattern.
22964 2013-01-16 Richard Biener <rguenther@suse.de>
22966 * tree-inline.c (tree_function_versioning): Remove set but
22967 never used variable.
22969 2013-01-16 Richard Biener <rguenther@suse.de>
22971 PR tree-optimization/55964
22972 * tree-flow.h (rename_variables_in_loop): Remove.
22973 (rename_variables_in_bb): Likewise.
22974 * tree-loop-distribution.c (update_phis_for_loop_copy): Remove.
22975 (copy_loop_before): Adjust and delete update-ssa status.
22976 * tree-vect-loop-manip.c (rename_variables_in_bb): Make static.
22977 (rename_variables_in_bb): Likewise. Properly walk over predecessors.
22978 (rename_variables_in_loop): Remove.
22979 (slpeel_update_phis_for_duplicate_loop): Likewise.
22980 (slpeel_tree_duplicate_loop_to_edge_cfg): Handle nested loops,
22981 use available cfg machinery instead of duplicating it.
22982 Update PHI nodes and perform poor-mans SSA update here.
22983 (slpeel_tree_peel_loop_to_edge): Adjust.
22985 2013-01-16 Richard Biener <rguenther@suse.de>
22987 PR tree-optimization/54767
22988 PR tree-optimization/53465
22989 * tree-vrp.c (vrp_meet_1): Revert original fix for PR53465.
22990 (vrp_visit_phi_node): For PHI arguments coming via backedges
22991 drop all symbolical range information.
22992 (execute_vrp): Compute backedges.
22994 2013-01-16 Richard Biener <rguenther@suse.de>
22996 * doc/install.texi: Update CLooG and ISL requirements to
22999 2013-01-16 Christian Bruel <christian.bruel@st.com>
23002 * config/sh/sh.c (sh_expand_prologue): Postpone new_stack mem symbol.
23003 (broken_move): Handle UNSPECV_SP_SWITCH_B.
23004 * config/sh/sh.md (sp_switch_1): Use set (reg:SI SP_REG).
23006 2013-01-16 DJ Delorie <dj@redhat.com>
23008 * config/sh/sh.md (UNSPECV_SP_SWITCH_B): New.
23009 (UNSPECV_SP_SWITCH_E): New.
23010 (sp_switch_1): Change to an unspec.
23011 (sp_switch_2): Change to an unspec. Don't use post-inc when we
23014 2013-01-16 Uros Bizjak <ubizjak@gmail.com>
23016 * emit-rtl.c (need_atomic_barrier_p): Mask memory model argument
23017 with MEMMODEL_MASK before comparing with MEMMODEL_* memory types.
23018 * optabs.c (maybe_emit_sync_lock_test_and_set): Ditto.
23019 (expand_mem_thread_fence): Ditto.
23020 (expand_mem_signal_fence): Ditto.
23021 (expand_atomic_load): Ditto.
23022 (expand_atomic_store): Ditto.
23024 2013-01-16 Alexandre Oliva <aoliva@redhat.com>
23026 PR rtl-optimization/55547
23027 PR rtl-optimization/53827
23030 * alias.c (memrefs_conflict_p): Set sizes to negative after
23033 2013-01-15 Jakub Jelinek <jakub@redhat.com>
23036 * function.c (thread_prologue_and_epilogue_insns): Always
23037 add crtl->drap_reg to set_up_by_prologue.set, even if
23038 stack_realign_drap is false.
23040 2013-01-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
23042 * config/vax/vax.md (add<mode>3, sub<mode>3, mul<mode>3, div<mode>3,
23043 and<mode>3, *and<mode>_const_int, ior<mode>3, xor<mode>3, ashrsi3,
23044 *call): Fix indention.
23046 2013-01-15 Tom de Vries <tom@codesourcery.com>
23049 * optabs.c (widen_operand): Use gen_lowpart instead of gen_rtx_SUBREG.
23052 2013-01-15 Vladimir Makarov <vmakarov@redhat.com>
23054 PR rtl-optimization/55153
23055 * sched-deps.c (sched_analyze_2): Add pending reads for prefetch.
23057 2013-01-15 Martin Jambor <mjambor@suse.cz>
23059 PR tree-optimization/55920
23060 * tree-sra.c (analyze_access_subtree): Do not mark non-removable
23061 accesses as grp_to_be_debug_replaced.
23063 2013-01-15 Jakub Jelinek <jakub@redhat.com>
23065 PR tree-optimization/55920
23066 * tree-sra.c (sra_modify_assign): If for lacc->grp_to_be_debug_replaced
23067 there is non-useless type conversion needed from debug rhs to lhs,
23068 use build_debug_ref_for_model and/or VIEW_CONVERT_EXPR.
23070 2013-01-15 Joseph Myers <joseph@codesourcery.com>
23071 Mikael Pettersson <mikpe@it.uu.se>
23074 * config/arm/arm.h (ADDR_VEC_ALIGN): Align SImode jump tables for
23076 (ASM_OUTPUT_CASE_LABEL): Remove.
23077 (ASM_OUTPUT_BEFORE_CASE_LABEL): Define to empty.
23078 * final.c (shorten_branches): Update alignment of labels before
23079 jump tables if CASE_VECTOR_SHORTEN_MODE.
23081 2013-01-15 Richard Biener <rguenther@suse.de>
23084 * system.h: Do not include gmp.h for building host tools.
23086 2013-01-15 Richard Biener <rguenther@suse.de>
23088 PR middle-end/55882
23089 * emit-rtl.c (set_mem_attributes_minus_bitpos): Correctly
23090 account for bitpos when computing alignment.
23092 2013-01-15 Vladimir Yakovlev <vladimir.b.yakovlev@intel.com>
23094 * config/i386/i386-c.c (ix86_target_macros_internal): New case.
23095 (ix86_target_macros_internal): Likewise.
23097 * config/i386/i386.c (m_CORE2I7): Removed.
23098 (m_CORE_HASWELL): New macro.
23099 (m_CORE_ALL): Likewise.
23100 (initial_ix86_tune_features): m_CORE2I7 is replaced by m_CORE_ALL.
23101 (initial_ix86_arch_features): Likewise.
23102 (processor_target_table): Initializations for Core avx2.
23103 (cpu_names): New names "core-avx2".
23104 (ix86_option_override_internal): Changed PROCESSOR_COREI7 by
23105 PROCESSOR_CORE_HASWELL.
23106 (ix86_issue_rate): New case.
23107 (ia32_multipass_dfa_lookahead): Likewise.
23108 (ix86_sched_init_global): Likewise.
23110 * config/i386/i386.h (TARGET_HASWELL): New macro.
23111 (target_cpu_default): New TARGET_CPU_DEFAULT_haswell.
23112 (processor_type): New PROCESSOR_HASWELL.
23114 2013-01-15 Jakub Jelinek <jakub@redhat.com>
23116 PR tree-optimization/55955
23117 * tree-vect-loop.c (vectorizable_reduction): Give up early on
23118 *SHIFT_EXPR and *ROTATE_EXPR codes.
23120 PR tree-optimization/48766
23121 * opts.c (common_handle_option): For -fwrapv disable -ftrapv, for
23122 -ftrapv disable -fwrapv.
23124 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
23127 * config/avr/avr-c.c (avr_cpu_cpp_builtins): Define __FLASH
23128 etc. to 1 and not to __flash.
23129 Use LL suffix for __INT24_MAX__ with -mint8.
23130 Use ULL suffix for __UINT24_MAX__ with -mint8.
23132 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
23134 * config/avr/avr-arch.h
23135 (struct base_arch_s): Use typedef avr_arch_t instead.
23136 (struct arch_info_s): Use typedef avr_arch_info_t instead.
23137 (struct mcu_type_s): Use typedef avr_mcu_t instead.
23138 * config/avr/avr.c: Same.
23139 * config/avr/avr-devices.c: Same.
23140 * config/avr/driver-avr.c: Same.
23141 * config/avr/gen-avr-mmcu-texi.c: Same.
23142 * config/avr/avr-mcus.def: Adjust comment.
23144 2013-01-14 Tejas Belagod <tejas.belagod@arm.com>
23146 * config/aarch64/aarch64-simd.md (*aarch64_simd_ld1r<mode>): New.
23147 * config/aarch64/iterators.md (VALLDI): New.
23149 2013-01-14 Uros Bizjak <ubizjak@gmail.com>
23150 Andi Kleen <ak@linux.intel.com>
23153 * config/i386/sync.md (atomic_store<mode>_1): New pattern.
23154 (atomic_store<mode>): Call atomic_store<mode>_1 for IX86_HLE_RELEASE
23157 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
23159 * config/avr/avr-stdint.h: Remove trailing blanks.
23160 * config/avr/avr-log.h: Same.
23161 * config/avr/avr-arch.h: Same.
23162 * config/avr/avr-devices.c: Same.
23163 * config/avr/avr-dimode.md: Same.
23164 * config/avr/predicates.md: Same.
23165 * config/avr/avr-c.c: Same. And fix typo.
23167 * config/avr/avr-protos.h: Same. And:
23168 (function_arg_regno_p): Rename to avr_function_arg_regno_p.
23169 (init_cumulative_args): Rename to avr_init_cumulative_args.
23170 (expand_prologue): Rename to avr_expand_prologue.
23171 (expand_epilogue): Rename to avr_expand_epilogue.
23172 (adjust_insn_length): Rename to avr_adjust_insn_length.
23173 (notice_update_cc): Rename to avr_notice_update_cc.
23174 (final_prescan_insn): Rename to avr_final_prescan_insn.
23175 * config/avr/avr.c: Same.
23176 * config/avr/avr.h: Same.
23177 * config/avr/avr.md: Remove trailing blanks.
23178 (prologue): Use avr_expand_prologue.
23179 (epilogue, sibcall_epilogue): Use avr_expand_epilogue.
23181 2013-01-14 Richard Biener <rguenther@suse.de>
23183 * tree-cfg.c (verify_expr_location, verify_expr_location_1,
23184 verify_location, collect_subblocks): New functions.
23185 (verify_gimple_in_cfg): Verify that locations only reference
23186 BLOCKs in the functions BLOCK tree.
23188 2013-01-14 Richard Biener <rguenther@suse.de>
23190 * tree-cfgcleanup.c (remove_forwarder_block): Unshare propagated
23192 * graphite-sese-to-poly.c (insert_out_of_ssa_copy): Properly
23194 (insert_out_of_ssa_copy_on_edge): Likewise.
23195 (rewrite_close_phi_out_of_ssa): Likewise.
23196 * tree-ssa.c (insert_debug_temp_for_var_def): Properly unshare
23198 * tree-ssa-pre.c (insert_into_preds_of_block): Properly unshare
23199 propagated constants.
23200 * tree-cfg.c (tree_node_can_be_shared): Handled component-refs
23203 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
23205 * config/avr/avr-modes.def: Add GPL copyright notice.
23207 2013-01-13 Uros Bizjak <ubizjak@gmail.com>
23209 * config/i386/sync.md (mem_thread_fence): Mask operands[0] with
23210 MEMMODEL_MASK to determine memory model.
23211 (atomic_store<mode>): Ditto from operands[2].
23212 * config/i386/i386.c (ix86_memmodel_check): Declare "strong" as bool.
23214 2013-01-13 Jakub Jelinek <jakub@redhat.com>
23217 * gimple-fold.c (get_symbol_constant_value): Call unshare_expr.
23218 (fold_gimple_assign): Don't call unshare_expr here.
23219 (fold_ctor_reference): Call unshare_expr.
23221 2013-01-13 Terry Guo <terry.guo@arm.com>
23223 * Makefile.in (s-mlib): New argument MULTILIB_REUSE.
23224 * doc/fragments.texi: Document MULTILIB_REUSE.
23225 * gcc.c (multilib_reuse): New internal spec.
23226 (set_multilib_dir): Also search multilib from multilib_reuse.
23227 * genmultilib (tmpmultilib3): Refactor code.
23228 (tmpmultilib4): Ditto.
23229 (multilib_reuse): New multilib argument.
23231 2013-01-13 Richard Sandiford <rdsandiford@googlemail.com>
23233 * Makefile.in: Update copyright.
23235 2013-01-12 Tom de Vries <tom@codesourcery.com>
23237 PR middle-end/55890
23238 * calls.c (expand_call): Check if arg_nr is valid.
23240 2013-01-11 Michael Meissner <meissner@linux.vnet.ibm.com>
23242 * doc/extend.texi (X86 Built-in Functions): Add whitespace in
23243 __builtin_ia32_paddb256 and __builtin_ia32_pavgb256
23244 documentation. Add missing '__' in front of
23245 __builtin_ia32_packssdw256.
23247 2013-01-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
23250 * config/s390/s390.c (s390_preferred_reload_class): Do not return
23251 NO_REGS for larl operands.
23252 (s390_reload_larl_operand): Use s390_load_address instead of
23255 2013-01-11 Richard Biener <rguenther@suse.de>
23257 * tree-cfg.c (verify_node_sharing_1): Split out from ...
23258 (verify_node_sharing): ... here.
23259 (verify_gimple_in_cfg): Use verify_node_sharing_1 for walk_tree.
23261 2013-01-11 Eric Botcazou <ebotcazou@adacore.com>
23263 * configure.ac (Tree checking): Set TREECHECKING to yes if enabled.
23264 Substitute TREECHECKING.
23265 * configure: Regenerate.
23266 * Makefile.in (TREECHECKING): New.
23268 2013-01-11 Richard Guenther <rguenther@suse.de>
23270 PR tree-optimization/44061
23271 * tree-vrp.c (extract_range_basic): Compute zero as
23272 value-range for __builtin_constant_p of function parameters.
23274 2013-01-10 Richard Sandiford <rdsandiford@googlemail.com>
23276 Update copyright years.
23278 2013-01-10 Vladimir Makarov <vmakarov@redhat.com>
23280 PR rtl-optimization/55672
23281 * lra-eliminations.c (mark_not_eliminable): Permit addition with
23282 const to be eliminable.
23284 2013-01-10 David Edelsohn <dje.gcc@gmail.com>
23286 * configure.ac (HAVE_AS_TLS): Add check for powerpc-ibm-aix.
23287 * configure: Regenerate.
23289 2013-01-10 Richard Biener <rguenther@suse.de>
23291 * builtins.c (expand_builtin_init_trampoline): Use set_mem_attributes.
23293 2013-01-10 Richard Biener <rguenther@suse.de>
23296 * tree-into-ssa.c (rewrite_add_phi_arguments): Do not set
23297 locations for virtual PHI arguments.
23298 (rewrite_update_phi_arguments): Likewise.
23300 2013-01-10 Joel Sherrill <joel.sherrill@OARcorp.com>
23302 * config/v850/rtems.h (ASM_SPEC): Pass -m8byte-align and -mgcc-abi
23305 2013-01-10 Jakub Jelinek <jakub@redhat.com>
23307 PR tree-optimization/55921
23308 * tree-complex.c (expand_complex_asm): New function.
23309 (expand_complex_operations_1): Call it for GIMPLE_ASM.
23311 2013-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
23314 * config/s390/s390.c (s390_symref_operand_p)
23315 (s390_loadrelative_operand_p): Merge the two functions.
23316 (s390_check_qrst_address, print_operand_address): Add parameters
23317 to s390_loadrelative_operand_p invokation.
23318 (s390_check_symref_alignment): Use s390_loadrelative_operand_p.
23319 (s390_reload_larl_operand, s390_secondary_reload): Use
23320 s390_loadrelative_operand_p instead of s390_symref_operand_p.
23321 (legitimize_pic_address): Handle @GOTENT and @PLT + addend.
23323 2013-01-09 Mike Stump <mikestump@comcast.net>
23325 * dse.c (record_store): Remove unnecessary assert.
23327 2013-01-09 Jan Hubicka <jh@suse.cz>
23329 PR tree-optimization/55569
23330 * cfgloopmanip.c (scale_loop_profile): Make ITERATION_BOUND gcov_type.
23331 * cfgloop.h (scale_loop_profile): Likewise.
23333 2013-01-09 Jan Hubicka <jh@suse.cz>
23336 * ipa-inline.c (ipa_inline): Remove extern inlines and virtual
23338 * cgraphclones.c (cgraph_clone_node): Cpoy also LTO file data.
23340 2013-01-09 Richard Sandiford <rdsandiford@googlemail.com>
23342 PR middle-end/55114
23343 * expr.h (maybe_emit_group_store): Declare.
23344 * expr.c (maybe_emit_group_store): New function.
23345 * builtins.c (expand_builtin_int_roundingfn): Call it.
23346 (expand_builtin_int_roundingfn_2): Likewise.
23348 2013-01-09 Vladimir Makarov <vmakarov@redhat.com>
23350 PR rtl-optimization/55829
23351 * lra-constraints.c (match_reload): Add code for absent output.
23352 (curr_insn_transform): Add code for reloads of matched inputs
23355 2013-01-09 Uros Bizjak <ubizjak@gmail.com>
23357 * config/i386/sse.md (*vec_interleave_highv2df): Change mode
23358 attribute of movddup insn to DF.
23359 (*vec_interleave_lowv2df): Ditto.
23360 (vec_dupv2df): Ditto.
23362 2013-01-09 Jan Hubicka <jh@suse.cz>
23364 PR tree-optimiation/55875
23365 * tree-ssa-loop-niter.c (number_of_iterations_cond): Add
23366 EVERY_ITERATION parameter.
23367 (number_of_iterations_exit): Check if exit is executed every iteration.
23368 (idx_infer_loop_bounds): Similarly here.
23369 (n_of_executions_at_most): Simplify
23370 to only test for cases where statement is dominated by the
23371 particular bound; handle correctly the "postdominance" test.
23372 (scev_probably_wraps_p): Use max loop iterations info
23373 as a global bound first.
23375 2013-01-09 Nguyen Duy Dat <dat.nguyen.yn@rvc.renesas.com>
23376 Nick Clifton <nickc@redhat.com>
23378 * config/v850/v850.md (cbranchsf4): New pattern.
23379 (cstoresf4): New pattern.
23380 (cbranchdf4): New pattern.
23381 (cstoredf4): New pattern.
23382 (movsicc): Disallow floating point comparisons.
23383 (cmpsf_le_insn): Fix order of operators.
23384 (cmpsf_lt_insn): Likewise.
23385 (cmpsf_eq_insn): Likewise.
23386 (cmpdf_le_insn): Likewise.
23387 (cmpdf_lt_insn): Likewise.
23388 (cmpdf_eq_insn): Likewise.
23389 (cmpsf_ge_insn): Use LE comparison.
23390 (cmpdf_ge_insn): Likewise.
23391 (cmpsf_gt_insn): Use LT comparison.
23392 (cmpdf_gt_insn): Likewise.
23393 (cmpsf_ne_insn): Delete pattern.
23394 (cmpdf_ne_insn): Delete pattern.
23395 * config/v850/v850.c (v850_gen_float_compare): Use
23396 gen_cmpdf_eq_insn for NE comparison.
23397 (v850_float_z_comparison_operator)
23398 (v850_float_nz_comparison_operator): Move from here ...
23399 * config/v850/predicates.md: ... to here. Move GT and GE
23400 comparisons into v850_float_z_comparison_operator.
23401 * config/v850/v850-protos.h (v850_float_z_comparison_operator):
23403 (v850_float_nz_comparison_operator): Likewise.
23405 2013-01-09 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
23407 * config/pa/pa.c (pa_emit_move_sequence): Replace calls to gen_insv
23408 with calls to gen_insvsi/gen_insvdi.
23410 2013-01-09 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
23412 * config/i386/i386.c (initial_ix86_tune_features): Set up
23413 X86_TUNE_AVX128_OPTIMAL for m_BTVER2.
23415 2013-01-09 Steven Bosscher <steven@gcc.gnu.org>
23416 Jakub Jelinek <jakub@redhat.com>
23418 PR tree-optimization/48189
23419 * predict.c (predict_loops): If max is 0, don't call compare_tree_int.
23420 If nitercst is 0, don't predict the exit edge.
23422 2013-01-08 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
23424 * config/aarch64/aarch64.c (aarch64_print_operand): Replace %r
23425 in asm_fprintf with reg_names.
23426 (aarch64_print_operand_address): Likewise.
23427 (aarch64_return_addr): Likewise.
23428 * config/aarch64/aarch64.h (ASM_FPRINTF_EXTENSIONS): Remove.
23430 2013-01-08 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
23432 * config/pa/pa.h (VAL_U6_BITS_P): Define.
23433 (INT_U6_BITS): Likewise.
23434 * config/pa/predicates.md (uint6_operand): New predicate.
23435 (shift5_operand, shift6_operand): Likewise.
23436 * config/pa/pa.md (lshrsi3, rotrsi3): Use shift5_operand instead of
23438 (lshrdi3): Use shift6_operand.
23439 (shrpsi4, shrpdi4): New insn patterns.
23440 (extzv): Delete expander.
23441 (extzvsi, extzvdi): New expanders. Use uint5_operand and uint6_operand
23442 predicates in unamed zero extract patterns. Tighten common constraint.
23443 (extv): Delete expander.
23444 (extvsi, extvdi): New expanders. Use uint5_operand and uint6_operand
23445 predicates in unamed sign extract patterns. Tighten common constraint.
23446 (insv): Delete expander.
23447 (insvsi, insvdi): New expanders. Use uint5_operand and uint6_operand
23448 predicates in unamed insert patterns. Tighten common constraint.
23449 Change uint32_operand predicate to uint6_operand predicate in unamed
23450 DImode pattern to insert constant values of type 1...1xxxx.
23452 2013-01-04 Jan Hubicka <jh@suse.cz>
23454 PR tree-optimization/55823
23455 * ipa-prop.c (update_indirect_edges_after_inlining): Fix ordering
23458 2013-01-08 Jakub Jelinek <jakub@redhat.com>
23459 Uros Bizjak <ubizjak@gmail.com>
23461 PR rtl-optimization/55845
23462 * df-problems.c (can_move_insns_across): Stop scanning at
23463 volatile_insn_p source instruction or give up if
23464 across_from .. across_to range contains any volatile_insn_p
23467 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
23469 * config/aarch64/aarch64-simd.md (vec_init<mode>): New.
23470 * config/aarch64/aarch64-protos.h (aarch64_expand_vector_init):
23472 * config/aarch64/aarch64.c (aarch64_simd_dup_constant,
23473 aarch64_simd_make_constant, aarch64_expand_vector_init): New.
23475 2013-01-08 Jakub Jelinek <jakub@redhat.com>
23478 * asan.c (asan_clear_shadow): New function.
23479 (asan_emit_stack_protection): Use it.
23481 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
23483 * config/aarch64/aarch64-simd.md (aarch64_simd_vec_<su>mult_lo_<mode>,
23484 aarch64_simd_vec_<su>mult_hi_<mode>): Separate instruction and operand
23485 with tab instead of space.
23487 2013-01-08 Nick Clifton <nickc@redhat.com>
23489 * config/rl78/rl78.c (rl78_expand_prologue): Always select
23490 register bank 0 at the start of an interrupt handler.
23491 * config/rl78/rl78.md (mulsi3_g13): Correct values for MDBL and
23494 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
23496 * config/aarch64/aarch64-simd.md
23497 (aarch64_simd_bsl<mode>_internal): Add floating-point modes.
23498 (aarch64_simd_bsl): Likewise.
23499 (aarch64_vcond_internal<mode>): Likewise.
23500 (vcond<mode><mode>): Likewise.
23501 (aarch64_cm<cmp><mode>): Fix constraints, add new modes.
23502 * config/aarch64/iterators.md (V_cmp_result): Add V2DF.
23504 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
23506 * config/aarch64/aarch64-builtins.c
23507 (aarch64_builtin_vectorized_function): Handle sqrt, sqrtf.
23509 2013-01-08 Martin Jambor <mjambor@suse.cz>
23512 * tree-sra.c (analyze_access_subtree): Return true also after
23513 potentially creating a debug-only replacement.
23515 2013-01-08 Jakub Jelinek <jakub@redhat.com>
23517 PR middle-end/55890
23518 * tree-ssa-ccp.c (evaluate_stmt): Use gimple_call_builtin_p.
23520 PR tree-optimization/54120
23521 * tree-vrp.c (range_fits_type_p): Don't allow
23522 src_precision < precision from signed vr to unsigned_p
23523 if vr->min or vr->max is negative.
23524 (simplify_float_conversion_using_ranges): Test can_float_p
23525 against CODE_FOR_nothing.
23527 2013-01-08 Jakub Jelinek <jakub@redhat.com>
23528 Richard Biener <rguenther@suse.de>
23530 PR middle-end/55851
23531 * fold-const.c (int_binop_types_match_p): Allow all INTEGRAL_TYPE_P
23532 types instead of just INTEGER_TYPE types.
23534 2013-01-07 Mark Kettenis <kettenis@openbsd.org>
23536 * config/i386/openbsdelf.h (LIBGCC2_HAS_TF_MODE, LIBGCC2_TF_CEXT,
23539 2013-01-07 Steve Ellcey <sellcey@mips.com>
23542 * config/mips/mips.opt: Change mad to mmad to match documentation.
23544 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
23547 * doc/extend.texi (AVR Named Address Spaces): __memx goes into
23548 .progmemx.data now.
23550 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
23553 * config/avr/avr.h (ADDR_SPACE_COUNT): New enum.
23554 (avr_addrspace_t): Add .section_name field.
23555 * config/avr/avr.c (progmem_section): Use ADDR_SPACE_COUNT as
23557 (avr_addrspace): Same. Initialize .section_name. Remove last
23558 NULL entry. Put __memx into .progmemx.data.
23559 (progmem_section_prefix): Remove.
23560 (avr_asm_init_sections): No need to initialize progmem_section.
23561 (avr_asm_named_section): Use avr_addrspace[].section_name to get
23562 section name prefix.
23563 (avr_asm_select_section): Ditto. And use get_unnamed_section to
23564 retrieve the progmem section.
23565 * avr-c.c (avr_cpu_cpp_builtins): Use ADDR_SPACE_COUNT as loop
23566 boundary to run over avr_addrspace[].
23567 (avr_register_target_pragmas): Ditto.
23569 2013-01-06 Jakub Jelinek <jakub@redhat.com>
23571 * varasm.c (output_constant_def_contents): For asan_protect_global
23572 protected strings, adjust DECL_ALIGN if needed, before testing for
23574 (place_block_symbol): Adjust size for asan protected STRING_CSTs if
23575 TREE_CONSTANT_POOL_ADDRESS_P. Increase alignment for asan protected
23577 (output_object_block): For asan protected decls, emit asan padding
23578 after their contents.
23579 * asan.c (asan_protect_global): Don't check TREE_ASM_WRITTEN here.
23580 (asan_finish_file): Test it here instead.
23582 2013-01-07 Nick Clifton <nickc@redhat.com>
23583 Matthias Klose <doko@debian.org>
23584 Doug Kwan <dougkwan@google.com>
23585 H.J. Lu <hongjiu.lu@intel.com>
23588 * collect2.c (main): Support -fuse-ld=bfd and -fuse-ld=gold.
23590 * common.opt: Add fuse-ld=bfd and fuse-ld=gold.
23592 * gcc.c (LINK_COMMAND_SPEC): Pass -fuse-ld=* to collect2.
23594 * opts.c (comman_handle_option): Ignore -fuse-ld=bfd and -fuse-ld=gold.
23596 * doc/invoke.texi: Document -fuse-ld=bfd and -fuse-ld=gold.
23598 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
23601 * doc/install.texi (Cross-Compiler-Specific Options): Document
23604 2013-01-07 Tejas Belagod <tejas.belagod@arm.com>
23606 * config/aarch64/arm_neon.h (vmovn_high_is16, vmovn_high_s32,
23607 vmovn_high_s64, vmovn_high_u16, vmovn_high_u32, vmovn_high_u64,
23608 vqmovn_high_s16, vqmovn_high_s32, vqmovn_high_s64, vqmovn_high_u16,
23609 vqmovn_high_u32, vqmovn_high_u64, vqmovun_high_s16, vqmovun_high_s32,
23610 vqmovun_high_s64): Fix source operand number and update copyright.
23612 2013-01-07 Richard Biener <rguenther@suse.de>
23614 PR middle-end/55890
23615 * gimple.h (gimple_call_builtin_p): New overload.
23616 * gimple.c (validate_call): New function.
23617 (gimple_call_builtin_p): Likewise.
23618 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
23619 Use gimple_call_builtin_p.
23620 (find_func_clobbers): Likewise.
23621 * tree-ssa-strlen.c (adjust_last_stmt): Likewise.
23622 (strlen_optimize_stmt): Likewise.
23624 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
23626 * config/aarch64/arm_neon.h (vld1_dup_*): Make argument const.
23627 (vld1q_dup_*): Likewise.
23628 (vld1_*): Likewise.
23629 (vld1q_*): Likewise.
23630 (vld1_lane_*): Likewise.
23631 (vld1q_lane_*): Likewise.
23633 2013-01-07 Richard Biener <rguenther@suse.de>
23635 * lto-streamer.h (LTO_minor_version): Bump to 2.
23637 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
23639 * config/aarch64/aarch64-protos.h
23640 (aarch64_const_double_zero_rtx_p): Rename to...
23641 (aarch64_float_const_zero_rtx_p): ...this.
23642 (aarch64_float_const_representable_p): New.
23643 (aarch64_output_simd_mov_immediate): Likewise.
23644 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): Refactor
23645 move immediate case.
23646 * config/aarch64/aarch64.c
23647 (aarch64_const_double_zero_rtx_p): Rename to...
23648 (aarch64_float_const_zero_rtx_p): ...this.
23649 (aarch64_print_operand): Allow printing of new constants.
23650 (aarch64_valid_floating_const): New.
23651 (aarch64_legitimate_constant_p): Check for valid floating-point
23653 (aarch64_simd_valid_immediate): Likewise.
23654 (aarch64_vect_float_const_representable_p): New.
23655 (aarch64_float_const_representable_p): Likewise.
23656 (aarch64_simd_imm_zero_p): Also allow for floating-point 0.0.
23657 (aarch64_output_simd_mov_immediate): New.
23658 * config/aarch64/aarch64.md (*movsf_aarch64): Add new alternative.
23659 (*movdf_aarch64): Likewise.
23660 * config/aarch64/constraints.md (Ufc): New.
23661 (Y): call aarch64_float_const_zero_rtx.
23662 * config/aarch64/predicates.md (aarch64_fp_compare_operand): New.
23664 2013-01-07 Richard Biener <rguenther@suse.de>
23666 PR tree-optimization/55888
23667 PR tree-optimization/55862
23668 * tree-ssa-pre.c (phi_translate_1): Revert previous change.
23669 (valid_in_sets): Check if a NAME has a leader in AVAIL_OUT,
23670 not if it is contained therein.
23672 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
23674 * config/avr/t-avr: Typo.
23676 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
23679 * config/avr/t-avr: Don't automatically rebuild
23680 $(srcdir)/config/avr/t-multilib
23681 $(srcdir)/config/avr/avr-tables.opt
23682 $(srcdir)/doc/avr-mmcu.texi
23683 (avr-mcus): New phony target to build them on request.
23684 (s-avr-mlib, s-avr-mmcu-texi): Remove.
23685 * avr/avr-mcus.def: Adjust comments.
23687 2013-01-07 Uros Bizjak <ubizjak@gmail.com>
23689 * config/i386/i386.c (DEFAULT_PCC_STRUCT_RETURN): Remove.
23691 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
23693 * file-find.c, file-find.h, realmpfr.c: Add FSF as copyright holder.
23695 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
23697 * config/tilepro/gen-mul-tables.cc: Put copyright on one line.
23699 2013-01-05 David Edelsohn <dje.gcc@gmail.com>
23701 * config/rs6000/aix53.h (LIB_SPEC): Add -lpthreads when compiling
23702 to generate profiling.
23703 * config/rs6000/aix64.h (LIB_SPEC): Same.
23705 2013-01-04 Andrew Pinski <apinski@cavium.com>
23707 * config/aarch64/aarch64.c (aarch64_fixed_condition_code_regs):
23709 (TARGET_FIXED_CONDITION_CODE_REGS): Define.
23711 2013-01-04 Uros Bizjak <ubizjak@gmail.com>
23713 * config/i386/i386.c (ix86_legitimize_address): Call convert_to_mode
23715 (ix86_expand_move): Ditto.
23716 (ix86_zero_extend_to_Pmode): Ditto.
23717 (ix86_expand_call): Ditto.
23718 (ix86_expand_special_args_builtin): Ditto.
23719 (ix86_expand_builtin): Ditto.
23721 2013-01-04 Richard Biener <rguenther@suse.de>
23723 PR tree-optimization/55862
23724 * tree-ssa-pre.c (phi_translate_1): Valueize SSA names after
23725 translating them through PHI nodes.
23727 2013-01-04 Martin Jambor <mjambor@suse.cz>
23729 PR tree-optimization/55755
23730 * tree-sra.c (sra_modify_assign): Do not check that an access has no
23731 children when trying to avoid producing a VIEW_CONVERT_EXPR.
23733 2013-01-04 Marek Polacek <polacek@redhat.com>
23735 PR middle-end/55859
23736 * opts.c (default_options_optimization): Clarify error message.
23738 2013-01-04 Richard Biener <rguenther@suse.de>
23740 PR middle-end/55863
23741 * fold-const.c (split_tree): Undo -X - 1 to ~X folding for
23744 2013-01-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
23747 * config/pa/pa.md (movsi): Revert previous change.
23748 * config/pa/pa.c (pa_legitimate_constant_p): Reject all TLS symbol
23751 2013-01-03 Richard Henderson <rth@redhat.com>
23753 * config/i386/i386.c (ix86_expand_move): Always assign to op1
23754 after eliminating TLS symbols.
23756 2013-01-03 Marc Glisse <marc.glisse@inria.fr>
23759 * graphite-interchange.c (pdr_stride_in_loop): Use gmp_fprintf.
23760 * graphite-poly.c (debug_gmp_value): Likewise.
23762 2013-01-03 Uros Bizjak <ubizjak@gmail.com>
23765 * config/i386/i386-c.c (ix86_target_macros_internal): Depending on
23766 selected code model, define __code_mode_small__, __code_model_medium__,
23767 __code_model_large__, __code_model_32__ or __code_model_kernel__.
23768 * config/i386/cpuid.h (__cpuid, __cpuid_count) [__i386__]: Prefix
23769 xchg temporary register with %k. Declare temporary register as
23771 [__x86_64__]: For medium and large code models, preserve %rbx register.
23773 2013-01-03 Richard Biener <rguenther@suse.de>
23775 * tree-data-ref.c (dump_conflict_function): Use less vertical spacing.
23776 (dump_subscript): Adjust.
23777 (finalize_ddr_dependent): Do not dump redundant info.
23778 (analyze_siv_subscript): Adjust.
23779 (subscript_dependence_tester): Likewise.
23780 (compute_affine_dependence): Likewise.
23782 2013-01-03 Richard Biener <rguenther@suse.de>
23785 2013-01-03 Richard Biener <rguenther@suse.de>
23787 PR tree-optimization/55857
23788 * tree-vect-stmts.c (vectorizable_load): Do not setup
23789 re-alignment for invariant loads.
23791 2013-01-02 Richard Biener <rguenther@suse.de>
23793 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
23794 invariant load do not generate a vector load from the scalar location.
23796 2013-01-03 Richard Biener <rguenther@suse.de>
23798 * tree-vect-loop.c (vect_analyze_loop_form): Clarify reason
23799 for not vectorizing.
23800 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref): Do
23801 not build INDIRECT_REFs, call get_name once only.
23802 (vect_create_data_ref_ptr): Likewise. Dump base object kind
23803 based on DR_BASE_OBJECT, not DR_BASE_ADDRESS.
23805 2013-01-03 Richard Biener <rguenther@suse.de>
23807 PR tree-optimization/55857
23808 * tree-vect-stmts.c (vectorizable_load): Do not setup
23809 re-alignment for invariant loads.
23811 2013-01-03 Richard Biener <rguenther@suse.de>
23814 * lto-symtab.c (lto_symtab_merge_decls_1): As last resort, always
23815 prefer a built-in decl.
23817 2013-01-03 Jakub Jelinek <jakub@redhat.com>
23819 * gcc.c (process_command): Update copyright notice dates.
23820 * gcov.c (print_version): Likewise.
23821 * gcov-dump.c (print_version): Likewise.
23823 PR rtl-optimization/55838
23824 * loop-iv.c (iv_number_of_iterations): Call lowpart_subreg on
23825 iv0.step, iv1.step and step.
23827 2013-01-03 Jakub Jelinek <jakub@redhat.com>
23828 Marc Glisse <marc.glisse@inria.fr>
23830 PR tree-optimization/55832
23831 * fold-const.c (fold_binary_loc): For ABS_EXPR<x> >= 0 and
23832 ABS_EXPR<x> < 0 folding use constant_boolean_node instead of
23833 integer_{one,zero}_node.
23835 2013-01-03 Jakub Jelinek <jakub@redhat.com>
23838 * params.def (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE): New param.
23839 * var-tracking.c (reverse_op): Don't add reverse ops to
23840 VALUEs that have already
23841 PARAM_VALUE (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE) or longer locs list.
23843 2013-01-02 Gerald Pfeifer <gerald@pfeifer.com>
23845 * doc/contrib.texi: Note years as release manager for Mark Mitchell.
23847 2013-01-02 Teresa Johnson <tejohnson@google.com>
23849 * dumpfile.c (dump_loc): Print filename with location.
23850 * tree-ssa-loop-ivcanon.c (try_unroll_loop_completely): Use
23851 new location_t parameter to emit complete unroll message with
23852 new dump framework.
23853 (canonicalize_loop_induction_variables): Compute loops location
23854 and pass to try_unroll_loop_completely.
23855 * loop-unroll.c (report_unroll_peel): New function.
23856 (peel_loops_completely): Use new dump format with location
23857 for main dumpfile message, and invoke report_unroll_peel on success.
23858 (decide_unrolling_and_peeling): Ditto.
23859 (decide_peel_once_rolling): Remove old dumpfile message subsumed
23860 by report_unroll_peel.
23861 (decide_peel_completely): Ditto.
23862 (decide_unroll_constant_iterations): Ditto.
23863 (decide_unroll_runtime_iterations): Ditto.
23864 (decide_peel_simple): Ditto.
23865 (decide_unroll_stupid): Ditto.
23866 * cfgloop.c (get_loop_location): New function.
23867 * cfgloop.h (get_loop_location): Declare.
23869 2013-01-02 Sriraman Tallam <tmsriram@google.com>
23871 * config/i386/i386.c (fold_builtin_cpu): Remove unnecessary checks for
23874 2013-01-02 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
23876 PR middle-end/55198
23877 * expr.c (expand_expr_real_1): Don't use bitfield extraction for non
23878 BLKmode objects when EXPAND_MEMORY is specified.
23880 2013-01-02 Sriraman Tallam <tmsriram@google.com>
23882 * config/i386/i386.c (ix86_get_function_versions_dispatcher): Fix bug
23884 (fold_builtin_cpu): Do not share cpu model decls across statements.
23886 2013-01-02 Jason Merrill <jason@redhat.com>
23889 * tree.c (build_array_type_1): Revert earlier change.
23891 2013-01-02 Yufeng Zhang <yufeng.zhang@arm.com>
23893 * config/aarch64/aarch64-cores.def: Add entries for "cortex-a53" and
23895 * config/aarch64/aarch64-tune.md: Re-generate.
23897 2013-01-02 Richard Biener <rguenther@suse.de>
23899 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
23900 invariant load do not generate a vector load from the scalar location.
23902 2013-01-02 Richard Biener <rguenther@suse.de>
23905 * configure.ac: Add $GMPINC to CFLAGS/CXXFLAGS.
23906 * configure: Regenerate.
23908 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
23910 * builtins.c (expand_builtin_mathfn, expand_builtin_mathfn_2)
23911 (expand_builtin_mathfn_ternary, expand_builtin_mathfn_3)
23912 (expand_builtin_int_roundingfn_2): Keep the original target around
23913 for the fallback case.
23915 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
23917 * tree-vrp.c (range_fits_type_p): Require the MSB of the double_int
23918 to be clear for sign changes.
23920 2013-01-01 Jan Hubicka <jh@suse.cz>
23922 * ipa-inline-analysis.c: Fix formatting.
23924 2013-01-01 Jakub Jelinek <jakub@redhat.com>
23926 PR tree-optimization/55831
23927 * tree-vect-loop.c (get_initial_def_for_induction): Use
23928 gsi_after_labels instead of gsi_start_bb.
23930 Copyright (C) 2013 Free Software Foundation, Inc.
23932 Copying and distribution of this file, with or without modification,
23933 are permitted in any medium without royalty provided the copyright
23934 notice and this notice are preserved.