1 2011-11-11 Jason Merrill <jason@redhat.com>
3 * varasm.c (default_elf_asm_output_limited_string): #ifdef
5 (default_elf_asm_output_ascii): Likewise.
7 2011-11-11 Uros Bizjak <ubizjak@gmail.com>
10 * config/i386/constraints.md (j): New address constraint.
11 * config/i386/predicates.md (lea_address_operand): Redefine as
13 * config/i386/i386.md (*lea_3_zext): Use "j" constraint for operand 1.
16 2011-11-11 David S. Miller <davem@davemloft.net>
19 2011-11-05 David S. Miller <davem@davemloft.net>
21 2011-11-11 Jakub Jelinek <jakub@redhat.com>
23 * opts-common.c (generate_canonical_option): Free opt_text
24 it it has been allocated here and not stored anywhere.
26 * tree-vect-slp.c (vect_free_slp_tree): Also free SLP_TREE_CHILDREN
28 (vect_create_new_slp_node): Don't allocate node before checking stmt
30 (vect_free_oprnd_info): Remove FREE_DEF_STMTS argument, always
31 free def_stmts vectors and additionally free oprnd_info.
32 (vect_build_slp_tree): Adjust callers. Call it even if
33 stop_recursion. If vect_create_new_slp_node or
34 vect_build_slp_tree fails, properly handle freeing memory.
35 If it succeeded, clear def_stmts in oprnd_info.
37 2011-08-12 Dimitrios Apostolou <jimis@gmx.net>
39 * final.c, output.h (fprint_whex, fprint_w, fprint_ul, sprint_ul):
40 New functions serving as fast replacements for fprintf() integer
41 to string conversions. They were used in the following changes.
42 * final.c (sprint_ul_rev): Internal helper for the above.
43 (output_addr_const): case CONST_INT: don't use fprintf().
44 * elfos.h (ASM_GENERATE_INTERNAL_LABEL): Don't use sprintf("%u"),
45 use sprint_ul() and stpcpy() which are much faster.
46 (TARGET_ASM_INTERNAL_LABEL): Define as default_elf_internal_label.
47 (ELF_ASCII_ESCAPES, ELF_STRING_LIMIT): Are the old ESCAPES and
49 (ASM_OUTPUT_LIMITED_STRING, ASM_OUTPUT_ASCII): Macros now just
50 call respective functions that provide the same
51 functionality. Those are default_elf_asm_output_limited_string()
52 and default_elf_asm_output_ascii() in varasm.c.
53 * varasm.c: Fixed some whitespace inconsistencies.
54 (default_elf_asm_output_limited_string)
55 (default_elf_asm_output_ascii): The above macros from elfos.h are
56 implemented here as these functions, avoiding superfluous calls to
58 (default_elf_internal_label): Hook for
59 targetm.asm_out.internal_label and ASM_OUTPUT_DEBUG_LABEL.
60 * i386.c: Don't call fprintf("%u") but fprint_ul() instead.
61 * defaults.h (ASM_OUTPUT_LABEL, ASM_OUTPUT_INTERNAL_LABEL):
62 Expanded the macros on multiple lines for readability.
63 (ASM_OUTPUT_LABELREF): Have two calls to fputs() instead of one to
65 * dwarf2asm.c (dw2_assemble_integer, dw2_asm_output_data)
66 (dw2_asm_data_uleb128, dw2_asm_delta_uleb128)
67 (dw2_asm_delta_sleb128): Convert fprintf() calls to the new
69 * dwarf2out.c (dwarf2out_source_line): Convert fprintf() calls to
70 the new faster functions.
72 2011-11-10 Andrew MacLeod <amacleod@redhat.com>
74 * doc/extend.texi: Document __atomic_test_and_set and __atomic_clear.
76 2011-11-10 Andrew MacLeod <amacleod@redhat.com>
79 * builtins.c (expand_builtin_atomic_clear): New. Expand atomic_clear.
80 (expand_builtin_atomic_test_and_set): New. Expand atomic test_and_set.
81 (expand_builtin): Add cases for test_and_set and clear.
82 * sync-builtins.def (BUILT_IN_ATOMIC_TEST_AND_SET): New.
83 (BUILT_IN_ATOMIC_CLEAR): New.
85 2011-11-10 Roberto Agostino Vitillo <ravitillo@lbl.gov>
88 * dwarf2out.c (set_cur_line_info_table): Restore the last is_stmt
89 value in the current line table.
91 2011-11-10 Nathan Sidwell <nathan@acm.org>
93 * gcov.c (struct function_info): Make src an index, not a pointer.
94 (struct source_info): Remove index and next source fields.
95 (fn_end): New static var.
96 (sources_index): Remove.
97 (sources): Now a pointer to an array, not a list.
98 (n_sources, a_sources): New.
99 (process_file): Adjust for changes to read_graph_file. Insert
100 functions into source lists and check line numbers here.
101 (generate_results): Only allocate lines for sources with
102 contents. Adjust for source array.
103 (release_structures): Likewise.
104 (find_source): Return source index, adjust for source array.
105 (read_graph_file): Return function list. Don't insert into source
107 (read_count_file): Take list of functions.
108 (solve_flow_graph): Reverse the arc lists here.
109 (add_line_counts): Adjust for source array.
111 2011-11-10 Jakub Jelinek <jakub@redhat.com>
114 * tree-object-size.c (addr_object_size): Check TREE_CODE of
115 MEM_REF's operand rather than code of the MEM_REF itself.
117 2011-11-10 Andrew MacLeod <amacleod@redhat.com>
119 PR rtl-optimization/51040
120 * optabs.c (expand_atomic_fetch_op): Patchup code for NAND should
121 be AND followed by NOT.
122 * builtins.c (expand_builtin_atomic_fetch_op): Patchup code for NAND
123 should be AND followed by NOT.
125 2011-11-10 Jakub Jelinek <jakub@redhat.com>
127 * vec.h (VEC_BASE): If base is at offset 0 in the structure,
128 use &(P)->base even if P is NULL.
130 PR rtl-optimization/51023
131 * combine.c (simplify_comparison) <case SIGN_EXTEND>: Don't use
132 val_signbit_known_clear_p for signed comparison narrowing
133 optimization. Don't check for non-VOIDmode, use
134 HWI_COMPUTABLE_MODE_P macro.
135 <case ZERO_EXTEND>: Don't check for non-VOIDmode.
136 Optimize even when const_op is equal to GET_MODE_MASK (mode),
137 don't optimize if const_op is negative.
139 2011-11-10 Richard Guenther <rguenther@suse.de>
141 PR tree-optimization/51042
142 * tree-ssa-pre.c (phi_translate_1): Avoid recursing on
143 self-referential expressions. Refactor code to avoid duplication.
145 2011-11-10 Richard Guenther <rguenther@suse.de>
147 PR tree-optimization/51070
148 * tree-loop-distribution.c (generate_builtin): Do not replace
149 the loop with a builtin if the partition contains statements which
150 results are used outside of the loop.
151 (pass_loop_distribution): Verify and collect.
153 2011-11-10 Richard Guenther <rguenther@suse.de>
155 PR tree-optimization/51030
156 * tree-ssa-phiopt.c (jump_function_from_stmt): New function.
157 (value_replacement): Use it to handle trivial non-empty
160 2011-11-10 Richard Guenther <rguenther@suse.de>
163 * gimple.c (gimple_has_side_effects): Remove checking code
164 that doesn't belong here.
166 2011-11-10 Ira Rosen <ira.rosen@linaro.org>
168 PR tree-optimization/51058
169 * tree-vect-slp.c (vect_get_constant_vectors): Handle CALL_EXPR.
171 2011-11-10 Jakub Jelinek <jakub@redhat.com>
173 PR tree-optimization/51000
174 * tree-vect-patterns.c (vect_recog_bool_pattern): If adding
175 a pattern stmt for a bool store, adjust DR_STMT too.
176 Don't handle bool conversions to single bit precision lhs.
177 * tree-vect-stmts.c (vect_remove_stores): If next is a pattern
178 stmt, remove its related stmt and free its stmt_vinfo.
179 (free_stmt_vec_info): Free also pattern stmt's vinfo and
180 pattern def stmt's vinfo.
181 * tree-vect-loop.c (destroy_loop_vec_info): Don't try to
182 free pattern stmt's vinfo here.
183 (vect_transform_loop): When calling vect_remove_stores,
184 do gsi_next first and don't call gsi_remove. If not strided
185 store, free stmt vinfo for gsi_stmt (si) rather than stmt.
187 2011-11-09 Jakub Jelinek <jakub@redhat.com>
190 * config/i386/i386.c (expand_vec_perm_interleave2): If d->vmode is
191 V4SImode, !TARGET_SSE2 and punpck[lh]* is needed, change dremap.vmode
194 * function.h (requires_stack_frame_p): New prototype.
195 * function.c (requires_stack_frame_p): No longer static.
196 * config/i386/i386.c (ix86_finalize_stack_realign_flags): If
197 stack_realign_fp was just a conservative guess for a function
198 which doesn't use sp/fp/argp at all, clear frame_pointer_needed
199 and stack realignment.
201 2011-11-09 Paolo Carlini <paolo.carlini@oracle.com>
203 PR preprocessor/51061
204 * incpath.c (merge_include_chains): Make sure to not pass null
205 pointers to free_path.
207 2011-11-09 Paolo Carlini <paolo.carlini@oracle.com>
209 * doc/invoke.texi ([Option Summary, C++ Language Options]):
210 Add -Wzero-as-null-pointer-constant.
212 2011-11-07 Matthew Gretton-Dann <matthew.gretton-dann@arm.com>
214 * config/arm/arm-cores.def: Add -mcpu=cortex-a7.
215 * config/arm/arm-tables.opt: Regenerate.
216 * config/arm/arm-tune.md: Likewise.
217 * config/arm/bpabi.h (BE8_LINK_SPEC): Add Cortex A-7.
218 * doc/invoke.texi: Document -mcpu=cortex-a7.
220 2011-11-09 Richard Guenther <rguenther@suse.de>
223 2011-11-09 Richard Guenther <rguenther@suse.de>
225 * tree-cfg.c (verify_gimple_call): Verify that
226 gimple_call_cannot_inline_p is returning a conservative
227 correct result according to gimple_check_call_matching_types.
229 2011-11-09 Richard Guenther <rguenther@suse.de>
231 PR tree-optimization/51039
232 * tree-cfg.c (verify_gimple_call): Verify that
233 gimple_call_cannot_inline_p is returning a conservative
234 correct result according to gimple_check_call_matching_types.
235 * ipa-inline-analysis.c (estimate_function_body_sizes): Remove
236 code dealing with un-inlinablility.
237 * gimple-streamer-in.c (input_gimple_stmt): Update the
240 2011-11-09 Richard Guenther <rguenther@suse.de>
242 PR tree-optimization/51039
243 * tree-inline.c (setup_one_parameter): Always perform a
244 valid gimple type change.
245 (declare_return_variable): Likewise.
247 2011-11-09 Jakub Jelinek <jakub@redhat.com>
249 * config/rs6000/vector.md (vcondv4sfv4si, vcondv4siv4sf,
250 vconduv4sfv4si): New patterns.
251 * config/rs6000/rs6000.c (rs6000_emit_vector_cond_expr): Handle
252 different dest_mode from comparison mode.
254 2011-11-09 Richard Guenther <rguenther@suse.de>
256 * gimple-fold.c (canonicalize_constructor_val): Mark
257 address-taken variables addressable.
259 2011-11-09 Ira Rosen <ira.rosen@linaro.org>
261 PR tree-optimization/51015
262 * tree-vect-loop.c (vect_determine_vectorization_factor): Expect
263 vectype to be set for pattern def stmts.
265 2011-11-09 Alan Modra <amodra@gmail.com>
267 * function.c (bb_active_p): Delete.
268 (dup_block_and_redirect, active_insn_between): New functions.
269 (convert_jumps_to_returns, emit_return_for_exit): New functions,
271 (thread_prologue_and_epilogue_insns): ..here. Delete
272 shadowing variables. Don't do prologue register clobber tests
273 when shrink wrapping already failed. Delete all last_bb_active
274 code. Instead compute tail block candidates for duplicating
275 exit path. Remove these from antic set. Duplicate tails when
276 reached from both blocks needing a prologue/epilogue and
277 blocks not needing such.
278 * ifcvt.c (dead_or_predicable): Test both flag_shrink_wrap and
280 * bb-reorder.c (get_uncond_jump_length): Make global.
281 * bb-reorder.h (get_uncond_jump_length): Declare.
282 * cfgrtl.c (rtl_create_basic_block): Comment typo fix.
283 (rtl_split_edge): Likewise. Warning fix.
284 (rtl_duplicate_bb): New function.
285 (rtl_cfg_hooks): Enable can_duplicate_block_p and duplicate_block.
286 * Makefile.in (function.o): Update dependencies.
288 2011-11-08 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
289 Georg-Johann Lay <avr@gjlay.de>
291 * doc/tm.texi.in (MODE_CODE_BASE_REG_CLASS): Add address space
293 (REGNO_MODE_CODE_OK_FOR_BASE_P): Likewise.
294 * doc/tm.texi: Regenerate.
296 * config/cris/cris.h (MODE_CODE_BASE_REG_CLASS): Add address
298 (REGNO_MODE_CODE_OK_FOR_BASE_P): Likewise.
299 * config/bfin/bfin.h (MODE_CODE_BASE_REG_CLASS): Likewise.
300 (REGNO_MODE_CODE_OK_FOR_BASE_P): Likewise.
301 * config/avr/avr.h (MODE_CODE_BASE_REG_CLASS): Add address space
303 (REGNO_MODE_CODE_OK_FOR_BASE_P): Ditto.
304 * config/avr/avr-protos.h (avr_mode_code_base_reg_class): Ditto.
305 (avr_regno_mode_code_ok_for_base_p): Ditto.
306 * config/avr/avr.c (avr_mode_code_base_reg_class): Ditto.
307 (avr_regno_mode_code_ok_for_base_p): Ditto.
308 (avr_reg_ok_for_addr_p): Pass AS down to
309 avr_regno_mode_code_ok_for_base_p.
311 * addresses.h (base_reg_class): Add address space argument.
312 Pass to MODE_CODE_BASE_REG_CLASS.
313 (ok_for_base_p_1): Add address space argument. Pass to
314 REGNO_MODE_CODE_OK_FOR_BASE_P.
315 (regno_ok_for_base_p): Add address space argument. Pass to
318 * regrename.c (scan_rtx_address): Add address space argument.
319 Pass address space to regno_ok_for_base_p and base_reg_class.
320 Update recursive calls.
321 (scan_rtx): Pass address space to scan_rtx_address.
322 (build_def_use): Likewise.
323 * regcprop.c (replace_oldest_value_addr): Add address space
324 argument. Pass to regno_ok_for_base_p and base_reg_class.
325 Update recursive calls.
326 (replace_oldest_value_mem): Pass address space to
327 replace_oldest_value_addr.
328 (copyprop_hardreg_forward_1): Likewise.
330 * reload.c (find_reloads_address_1): Add address space argument.
331 Pass address space to base_reg_class and regno_ok_for_base_p.
332 Update recursive calls.
333 (find_reloads_address): Pass address space to base_reg_class,
334 regno_ok_for_base_p, and find_reloads_address_1.
335 (find_reloads): Pass address space to base_reg_class.
336 (find_reloads_subreg_address): Likewise.
338 * ira-costs.c (record_reg_classes): Update calls to base_reg_class.
339 (ok_for_base_p_nonstrict): Add address space argument. Pass to
341 (record_address_regs): Add address space argument. Pass to
342 base_reg_class and ok_for_base_p_nonstrict. Update recursive calls.
343 (record_operand_costs): Pass address space to record_address_regs.
344 (scan_one_insn): Likewise.
346 * caller-save.c (init_caller_save): Update call to base_reg_class.
347 * ira-conflicts.c (ira_build_conflicts): Likewise.
348 * reload1.c (maybe_fix_stack_asms): Likewise.
350 2011-11-08 Michael Matz <matz@suse.de>
352 * gengtype.c (write_field_root): Avoid out-of-scope access of newv.
354 * tree-stdarg.c (execute_optimize_stdarg): Accept clobbers.
356 * tree.h (TREE_CLOBBER_P): New macro.
357 * gimple.h (gimple_clobber_p): New inline function.
358 * gimplify.c (gimplify_bind_expr): Add clobbers for all variables
359 that go out of scope and live in memory.
360 * tree-ssa-operands.c (get_expr_operands): Transfer volatility also
362 * cfgexpand.c (decl_to_stack_part): New static variable.
363 (add_stack_var): Allocate it, and remember mapping.
364 (fini_vars_expansion): Deallocate it.
365 (stack_var_conflict_p): Add early outs.
366 (visit_op, visit_conflict, add_scope_conflicts_1,
367 add_scope_conflicts): New static functions.
368 (expand_used_vars_for_block): Don't call add_stack_var_conflict, tidy.
369 (expand_used_vars): Add scope conflicts.
370 (expand_gimple_stmt_1): Expand clobbers to nothing.
371 (expand_debug_expr): Ditto.
373 * tree-pretty-print.c (dump_generic_node): Dump clobbers nicely.
374 * tree-ssa-live.c (remove_unused_locals): Remove clobbers that
375 refer to otherwise unused locals.
376 * tree-sra.c (build_accesses_from_assign): Ignore clobbers.
377 * tree-ssa-dce.c (mark_stmt_if_obviously_necessary): Clobbers of
378 SSA names aren't necessary.
379 (propagate_necessity): Accept and ignore constructors on the rhs, tidy.
380 * gimple.c (walk_gimple_op): Accept constructors like mem_rhs.
381 * tree-ssa-structalias.c (find_func_aliases): Clobbers don't store
383 * tree-ssa-sccvn.c (vn_reference_lookup_3): Ditto, in particular they
384 don't zero-initialize something.
385 * tree-ssa-phiopt.c (cond_if_else_store_replacement_1): Ignore
386 clobber RHS, we don't want PHI nodes with those.
388 2011-11-08 Jakub Jelinek <jakub@redhat.com>
390 * config/i386/i386.c (ix86_preferred_simd_mode): Even for TARGET_AVX
391 if not TARGET_PREFER_AVX128 return 32-byte integer vectors.
393 2011-11-08 Richard Guenther <rguenther@suse.de>
395 PR tree-optimization/51012
396 * ipa-prop.c (update_indirect_edges_after_inlining): Fixup
398 * cgraph.c (cgraph_make_edge_direct): Likewise.
400 2011-11-08 Eric Botcazou <ebotcazou@adacore.com>
402 PR rtl-optimization/47698
403 * ifcvt.c (noce_operand_ok): Move around comment.
405 2011-11-08 Richard Guenther <rguenther@suse.de>
408 * lto-opts.c (append_to_collect_gcc_options): Split out from...
409 (lto_write_options): ... here. Prepend frontend specific flags.
411 2011-11-08 Jakub Jelinek <jakub@redhat.com>
413 * config/i386/sse.md (*avx_vperm2f128<mode>3_nozero): Emit mask
414 0x12 and 0x20 as vinsert[fi]128 instead of vperm2[fi]128.
416 * tree-vect-stmts.c (vectorizable_call): Add SLP_NODE argument.
417 Handle vectorization of SLP calls.
418 (vect_analyze_stmt): Adjust caller, add call to it for SLP too.
419 (vect_transform_stmt): Adjust vectorizable_call caller, remove
421 * tree-vect-slp.c (vect_get_and_check_slp_defs): For calls start
423 (vect_build_slp_tree): Allow CALL_EXPR.
425 2011-11-08 Richard Guenther <rguenther@suse.de>
427 * gimple-fold.c (canonicalize_constructor_val): Make sure
428 we have referenced vars setup before adding to them.
430 2011-11-07 Richard Henderson <rth@redhat.com>
431 Aldy Hernandez <aldyh@redhat.com>
432 Andrew MacLeod <amacleod@redhat.com>
433 Torvald Riegel <triegel@redhat.com>
435 Merged from transactional-memory.
437 * gtm-builtins.def: New file.
438 * trans-mem.c: New file.
439 * trans-mem.h: New file.
441 * opts.c (finish_options): Error out when using -flto and -fgnu-tm.
443 * config/i386/i386.c: Define TARGET_VECTORIZE* transactional variants.
444 (ix86_handle_tm_regparm_attribute, struct bdesc_tm,
445 ix86_builtin_tm_load, ix86_builtin_tm_store,
446 ix86_init_tm_builtins): New.
447 (ix86_init_builtins): Initialize TM builtins.
448 (struct ix86_attribute_table): Add "*tm regparm".
449 * config/i386/i386-builtin-types.def (PV2SI): Define.
451 Define V2SI_FTYPE_PCV2SI.
452 Define V4SF_FTYPE_PCV4SF.
453 Define V8SF_FTYPE_PCV8SF.
454 Define VOID_PV2SI_V2SI.
456 * doc/invoke.texi (C Dialect Options): Document -fgnu-tm and
457 tm-max-aggregate-size.
458 * doc/tm.texi.in: Add TARGET_VECTORIZE_BUILTIN_TM_LOAD and
459 TARGET_VECTORIZE_BUILTIN_TM_STORE hooks.
460 * doc/tm.texi: Regenerate.
462 * attribs.c (apply_tm_attr): New.
463 (init_attributes): Allow '*' prefix for overrides.
464 (register_attribute): Likewise.
465 * builtin-attrs.def (ATTR_TM_TMPURE, ATTR_TM_REGPARM): New.
466 (ATTR_TM_NOTHROW_LIST, ATTR_TM_TMPURE_NOTHROW_LIST,
467 ATTR_TM_PURE_TMPURE_NOTHROW_LIST, ATTR_TM_NORETURN_NOTHROW_LIST,
468 ATTR_TM_CONST_NOTHROW_LIST, ATTR_TMPURE_MALLOC_NOTHROW_LIST,
469 ATTR_TMPURE_NOTHROW_LIST): New.
470 * builtin-types.def (BT_FN_I[1248]_VPTR, BT_FN_FLOAT_VPTR,
471 BT_FN_DOUBLE_VPTR, BT_FN_LDOUBLE_VPTR, BT_FN_VOID_VPTR_I[1248],
472 BT_FN_VOID_VPTR_FLOAT, BT_FN_VOID_VPTR_DOUBLE,
473 BT_FN_VOID_VPTR_LDOUBLE, BT_FN_VOID_VPTR_SIZE): New.
474 * builtins.def: Include gtm-builtins.def. Add comments regarding
475 transactional memory synchronization.
476 (DEF_TM_BUILTIN): New.
477 * c-parser.c (struct c_parser): Add in_transaction.
478 (c_parser_transaction, c_parser_transaction_expression,
479 c_parser_transaction_cancel, c_parser_transaction_attributes): New.
480 (c_parser_attribute_any_word): Split out from c_parser_attributes.
481 (c_parser_statement_after_labels): Handle RID_TRANSACTION*.
482 (c_parser_unary_expression): Same.
483 * c-tree.h (c_finish_transaction): Declare.
484 * c-typeck.c (c_finish_transaction): New.
485 (build_function_call_vec): Call tm_malloc_replacement.
486 * calls.c (is_tm_builtin): New.
487 (flags_from_decl_or_type): Add ECF_TM_BUILTIN and ECF_TM_PURE.
488 * cfgbuild.c (make_edges): Add edges for REG_TM notes.
489 * cfgexpand.c (expand_call_stmt): Call
490 mark_transaction_restart_calls.
491 (gimple_expand_cfg): Free the tm_restart map.
492 (mark_transaction_restart_calls): New.
493 * cfgrtl.c (purge_dead_edges): Look for REG_TM notes.
494 * cgraph.c (dump_cgraph_node): Handle tm_clone.
495 * cgraph.h (struct cgraph_node): Add tm_clone field.
496 (decl_is_tm_clone): New.
497 (struct cgraph_local_info): Add tm_may_enter_irr.
498 (cgraph_copy_node_for_versioning): Declare.
499 * cgraphunit.c (cgraph_copy_node_for_versioning): Export;
500 copy analyzed from old version.
501 * combine.c (distribute_notes): Handle REG_TM notes.
502 * common.opt: Add -fgnu-tm.
503 * crtstuff.c (__TMC_LIST__, __TMC_END__): New.
504 (__do_global_dtors_aux): Deregister clone table.
505 (frame_dummy): Register clone table.
506 * emit-rtl.c (try_split): Handle REG_TM. Early return if no function
508 * gimple-low.c (lower_stmt): Handle GIMPLE_EH_ELSE and
510 (gimple_stmt_may_fallthru): Handle GIMPLE_EH_ELSE.
511 * gimple-pretty-print.c: Include trans-mem.h.
512 (dump_gimple_fmt): Add %x.
513 (dump_gimple_call): Dump arguments for calls to _ITM_beginTransaction.
514 (dump_gimple_eh_else, dump_gimple_transaction): New.
515 (dump_gimple_stmt): Handle GIMPLE_EH_ELSE and GIMPLE_TRANSACTION.
516 * gimple.c (gimple_build_eh_else, gimple_build_transaction): New.
517 (walk_gimple_seq): Honor removed_stmt. Document usage of removed_stmt
519 (walk_gimple_op): Handle GIMPLE_TRANSACTION.
520 (walk_gimple_stmt): Initialize and honor removed_stmt.
521 Handle GIMPLE_EH_ELSE and GIMPLE_TRANSACTION.
522 (gimple_copy): Handle GIMPLE_EH_ELSE and GIMPLE_TRANSACTION.
523 * gimple.def (GIMPLE_TRANSACTION, GIMPLE_EH_ELSE): New.
524 * gimple.h (struct gimple_statement_eh_else,
525 gimple_statement_transaction, GTMA_*): New.
526 (gimple_statement_d): Add gimple_statement_eh_else and
528 (gimple_build_eh_else, gimple_build_transaction,
529 gimple_fold_call, diagnose_tm_safe_errors): Declare.
530 (get_call_expr_in): Remove prototype.
531 (gimple_has_substatements): Add GIMPLE_EH_ELSE and GIMPLE_TRANSACTION.
532 (gimple_eh_else_n_body, gimple_eh_else_e_body,
533 gimple_eh_else_set_n_body, gimple_eh_else_set_e_body,
534 gimple_transaction_body, gimple_transaction_label,
535 gimple_transaction_label_ptr, gimple_transaction_subcode,
536 gimple_transaction_set_body, gimple_transaction_set_label,
537 gimple_transaction_set_subcode): New.
538 (struct walk_stmt_info): Use BOOL_BITFIELD; add removed_stmt.
539 * gimplify.c (create_tmp_var_name): Use clean_symbol_name.
540 (voidify_wrapper_expr): Handle TRANSACTION_EXPR.
541 (gimplify_transaction): New.
542 (gimplify_expr): Handle TRANSACTION_EXPR.
543 * gsstruct.def (GSS_EH_ELSE, GSS_TRANSACTION): New.
544 * ipa-inline.c (can_inline_edge_p): Do not inline TM safe calling
546 * Makefile.in: Add trans-mem.o and dependencies.
547 (BUILTINS_DEF): Add gtm-builtins.def.
548 (gimple-pretty-print.o): Depend on TRANS_MEM_H.
549 (GTFILES): Add trans-mem.c.
550 * omp-low.c (WALK_SUBSTMTS): Add GIMPLE_TRANSACTION.
551 * output.h (record_tm_clone_pair, finish_tm_clone_pairs,
552 get_tm_clone_pair): Declare.
553 * params.def (PARAM_TM_MAX_AGGREGATE_SIZE): New.
554 * passes.c (init_optimization_passes): Place transactional memory
556 * print-tree.c (print_node): Dump tm-clone.
557 * recog.c (peep2_attempt): Handle REG_TM.
558 * reg-notes.def (TM): New.
559 * rtlanal.c (alloc_reg_note): Handle REG_TM.
560 * target.def (builtin_tm_load, builtin_tm_store): New.
561 * targhooks.c (default_builtin_tm_load_store): New.
562 * targhooks.h (default_builtin_tm_load_store): Declare.
563 * timevar.def (TV_TRANS_MEM): New.
564 * toplev.c (compile_file): Call finish_tm_clone_pairs.
565 * tree-cfg.c (make_edges): Handle GIMPLE_TRANSACTION.
566 (cleanup_dead_labels): Handle GIMPLE_TRANSACTION. Avoid unnecessary
567 writes into the statements to update labels.
568 (is_ctrl_altering_stmt): Add TM ending statements. Handle
570 (verify_gimple_transaction): New.
571 (verify_gimple_stmt): Handle GIMPLE_TRANSACTION.
572 (verify_gimple_in_seq_2): Handle GIMPLE_EH_ELSE and GIMPLE_TRANSACTION.
573 (gimple_redirect_edge_and_branch): Handle TM_TRANSACTION.
574 (dump_function_to_file): Display [tm-clone] if applicable.
575 * tree-eh.c (struct_ptr_eq): Make inline and move to tree.h.
576 (struct_ptr_hash): Same.
577 (collect_finally_tree): Handle GIMPLE_EH_ELSE.
578 (replace_goto_queue_1): Likewise.
580 (honor_protect_cleanup_actions): Handle GIMPLE_EH_ELSE.
581 (lower_try_finally_nofallthru): Likewise.
582 (lower_try_finally_onedest): Likewise.
583 (lower_try_finally_copy): Likewise.
584 (lower_try_finally_switch): Likewise.
585 (lower_try_finally): Likewise.
586 (decide_copy_try_finally): Likewise.
587 (lower_eh_constructs_2): Likewise.
588 (refactor_eh_r): Likewise.
589 * tree-flow.h (struct gimple_df): Add tm_restart field.
590 Define tm_restart_node.
591 * tree-inline.c (remap_gimple_stmt): Handle GIMPLE_TRANSACTION.
592 (estimate_num_insns): Likewise.
593 (init_inline_once): Init tm_cost.
594 * tree-inline.h (struct eni_weights_d): Add tm_cost.
595 * tree-pass.h (pass_diagnose_tm_blocks, pass_lower_tm, pass_tm_init,
596 pass_tm_mark, pass_tm_memopt, pass_tm_edges, pass_ipa_tm): Declare.
597 * tree-pretty-print.c (dump_generic_node): Handle TRANSACTION_EXPR.
598 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Handle
599 BUILT_IN_TM_MEMSET, BUILT_IN_TM_MEMCPY, BUILT_IN_TM_MEMMOVE.
600 Add support for TM vector loads. Add support for TM logging builtins.
601 (call_may_clobber_ref_p_1): Add support for vector stores.
602 * tree-ssa-structalias.c (find_func_aliases): Add support for TM
603 vector stores and loads. Handle BUILT_IN_TM_MEMSET,
604 BUILT_IN_TM_MEMCPY, BUILT_IN_TM_MEMMOVE.
605 * tree.c (strip_invariant_refs): Moved from gimple.c to here.
606 (local_define_builtin): Handle ECF_TM_PURE.
607 (build_common_builtin_nodes): Set __builtin_eh_pointer to ECF_TM_PURE.
608 * tree.def (TRANSACTION_EXPR): New.
609 * tree.h (strip_invariant_refs): Moved from gimple.h to here.
610 (TRANSACTION_EXPR_BODY, TRANSACTION_EXPR_CHECK,
611 TRANSACTION_EXPR_OUTER, TRANSACTION_EXPR_RELAXED,
612 BUILTIN_TM_LOAD_STORE_P, BUILTIN_TM_LOAD_P, BUILTIN_TM_STORE_P,
613 CASE_BUILT_IN_TM_LOAD, CASE_BUILT_IN_TM_STORE): New.
614 (ECF_TM_PURE, ECF_TM_BUILTIN): New.
615 (struct tree_function_decl): Add tm_clone_flag.
616 (struct_ptr_eq, struct_ptr_hash): New.
617 (apply_tm_attr): Declare.
618 (is_tm_safe_or_pure): New.
619 (build_tm_abort_call, is_tm_safe, is_tm_pure,
620 is_tm_may_cancel_outer, is_tm_ending_fndecl, record_tm_replacement,
621 tm_malloc_replacement): Declare.
622 * varasm.c (tm_clone_hash): New.
623 (record_tm_clone_pair, finish_tm_clone_pairs, get_tm_clone_pair,
624 dump_tm_clone_to_vec, dump_tm_clone_pairs, tm_alias_pair_cmp): New.
625 (struct tm_alias_pair): New. Declare VEC types for object.
627 2011-11-07 Richard Henderson <rth@redhat.com>
629 * optabs.h (OTI_sync_compare_and_swap, OTI_sync_lock_test_and_set,
630 OTI_sync_old_add, OTI_sync_old_sub, OTI_sync_old_ior,
631 OTI_sync_old_and, OTI_sync_old_xor, OTI_sync_old_nand,
632 OTI_sync_new_add, OTI_sync_new_sub, OTI_sync_new_ior,
633 OTI_sync_new_and, OTI_sync_new_xor, OTI_sync_new_nand): Move and
634 rename from the direct_optab_index enum.
635 (sync_compare_and_swap_optab, sync_lock_test_and_set_optab,
636 sync_old_add_optab, sync_old_sub_optab, sync_old_ior_optab,
637 sync_old_and_optab, sync_old_xor_optab, sync_old_nand_optab,
638 sync_new_add_optab, sync_new_sub_optab, sync_new_ior_optab,
639 sync_new_and_optab, sync_new_xor_optab, sync_new_nand_optab): Read
640 from the optab_table, not the direct_optab_table.
641 (init_sync_libfuncs): Declare.
642 (can_compare_and_swap_p): Update parameters.
643 * optabs.c (init_sync_libfuncs_1, init_sync_libfuncs): New.
644 (can_compare_and_swap_p): Add allow_libcall parameter; if true,
645 test for the legacy compare-and-swap libcall.
646 (expand_atomic_exchange): Use the legacy test-and-set libcall.
647 (expand_atomic_compare_and_swap): Use the legacy CAS libcall.
648 (struct atomic_op_functions): Update for optab type changes.
649 (maybe_emit_op): Likewise.
650 (expand_atomic_fetch_op): Use the legacy fetch-op libcalls.
651 * builtins.c (fold_builtin_atomic_always_lock_free): Update call
652 to can_compare_and_swap_p.
653 * omp-low.c (expand_omp_atomic_fetch_op): Likewise.
654 (expand_omp_atomic_pipeline): Likewise.
655 * genopinit.c (optabs): Make sync_old_*_optab, sync_new_*_optab,
656 sync_compare_and_swap_optab, sync_lock_test_and_set_optab regular
658 * doc/md.texi (sync_compare_and_swap): Update docs for libcalls.
660 2011-11-07 Jakub Jelinek <jakub@redhat.com>
662 * config/i386/i386-bultin-types.def (V8SI_FTYPE_V4DF_V4DF): Add.
663 * config/i386/i386.c (enum ix86_builtins): Add
664 IX86_BUILTIN_VEC_PACK_SFIX256.
665 (bdesc_args): Add __builtin_ia32_vec_pack_sfix256.
666 (ix86_expand_args_builtin): Handle V8SI_FTYPE_V4DF_V4DF.
667 (ix86_builtin_vectorized_function): Also vectorize lrint using
668 256-bit vectors for -mavx.
670 2011-11-07 Anatoly Sokolov <aesok@post.ru>
672 * config/cris/constraints.md: New file.
673 * config/cris/cris.h (REG_CLASS_FROM_LETTER, CONSTRAINT_LEN,
674 CRIS_CONST_OK_FOR_LETTER_P, CONST_OK_FOR_CONSTRAINT_P,
675 CONST_DOUBLE_OK_FOR_LETTER_P, EXTRA_MEMORY_CONSTRAINT,
676 EXTRA_CONSTRAINT, EXTRA_CONSTRAINT_Q, EXTRA_CONSTRAINT_R,
677 EXTRA_CONSTRAINT_T, EXTRA_CONSTRAINT_S, EXTRA_CONSTRAINT_U): Remove.
678 * config/cris/cris.c: Incule "tm-constrs.h".
679 (cris_print_operand): Use satisfies_constraint_O.
680 (cris_normal_notice_update_cc, cris_rtx_costs): Use
681 satisfies_constraint_I.
682 (cris_address_cost): Use satisfies_constraint_L.
683 * config/cris/cris.md: Include "constraints.md".
684 (*mov_side<mode>, *mov_sidesisf, *mov_side<mode>_mem,
685 *mov_sidesisf_mem, *clear_side<mode>, *ext_sideqihi,
686 *ext_side<mode>si, *op_side<mode>, *op_swap_side<mode>,
687 *extopqihi_side, *extop<mode>si_side, *extopqihi_swap_side,
688 *extop<mode>si_swap_side): Use satisfies_constraint_N and
689 satisfies_constraint_J.
690 (moversideqi movemsideqi mover2side peephole2): Use
691 satisfies_constraint_N and satisfies_constraint_J.
692 (andu peephole2): Use satisfies_constraint_I and
693 satisfies_constraint_O.
695 2011-11-07 Uros Bizjak <ubizjak@gmail.com>
697 * config/i386/i386.c (ix86_builtin_vectorized_function): Handle
698 BUILT_IN_IRINT, BUILT_IN_IRINTF, BUILT_IN_LLRINT and BUILT_IN_LLRINTF.
700 2011-11-07 Andrew MacLeod <amacleod@redhat.com>
702 * optabs.c (get_atomic_op_for_code): Fill in optab table at runtime so
703 SWITCHABLE_TARGET can change the values during compilation.
704 (expand_atomic_fetch_op): Handle parameter change ripples for
705 get_atomic_op_for_code call.
707 2011-11-07 Andrew MacLeod <amacleod@redhat.com>
709 * doc/extend.texi: Docuemnt behaviour change for __atomic_exchange and
711 * optabs.c (expand_atomic_exchange): Expand to __sync_lock_test_and_set
712 only when originated from that builtin.
713 (expand_atomic_store): Expand to __sync_lock_release when originated
715 * builtins.c (expand_builtin_sync_lock_test_and_set): Add flag that
716 expand_atomic_exchange call originated from here.
717 (expand_builtin_sync_lock_release): Add flag that expand_atomic_store
718 call originated from here.
719 (expand_builtin_atomic_exchange): Add origination flag.
720 (expand_builtin_atomic_store): Add origination flag.
721 * expr.h (expand_atomic_exchange, expand_atomic_store): Add boolean
722 parameters to indicate implementation fall back options.
724 2011-11-07 Georg-Johann Lay <avr@gjlay.de>
726 * config/avr/avr.c (output_reload_in_const): Can handle CONSTANT_P
727 now, not only CONST_INT and CONST_DOUBLE.
728 (output_movqi): Use output_reload_in_const.
729 (output_reload_inhi): Ditto.
730 (output_reload_insisf): Move assertion to output_reload_in_const.
731 (avr_out_reload_inpsi): Ditto.
733 2011-11-07 Nathan Sidwell <nathan@acm.org>
735 * gcov.c (object_summary): Replace with ...
736 (object_runs): ... this.
737 (process_file): Remove functions with no data.
738 (generate_results): Ignore files with no lines.
739 (release_function): New helper, broken out of ...
740 (release_structures): ... here. Use it.
741 (read_count_file): Adjust for new data file format.
742 (output_lines): Use object_runs.
743 * gcov-io.h (GCOV_TAG_OBJECT_SUMMARY): Obsolete.
744 (struct gcov_ctr_info): Move definition.
745 (struct gcov_fn_info): Add key field, use gcov_ctr_info for
747 (struct gcov_info): Add merge function array, remove mask and
748 counts. Trailing array is array of pointers to function info.
749 * coverage.c (struct function_list): Replace counter numbers with
750 counter arrays. Add fndecl field. GTYify.
751 (counts_entry): Remove chain workspace.
752 (functions_head): GTYify.
753 (prg_n_ctrs): Remove.
755 (tree_ctr_tables): Remove.
756 (read_counts_file): Cope with blank entries and expect program
757 summaries before functions. Don't warn on missing entries.
758 (coverage_counter_alloc): Allocate individual function arrays.
759 (tree_coverage_counter_ref, tree_coverage_counter_addr): Adjust
760 for individual function arrays.
761 (coverage_end_function): GTYify function list object. Finalize
762 function's counter arrays.
763 (build_var): New. Create a counter-related variable with
765 (build_fn_info_type): Adjust for new runtime structure.
766 (build_fn_info_value): Rename to ...
767 (build_fn_info): ... here. Build new format data.
768 (build_ctr_info_type, build_ctr_info_value): Remove.
769 (build_info_type): New. Build new format data structure.
770 (build_info): Adjust for new format data.
771 (create_coverage): Likewise.
772 * gcov-dump.c (tag_function): Recognize placeholders.
774 2011-11-07 Georg-Johann Lay <avr@gjlay.de>
776 * config/avr/constraints.md (Cm2): New constraint for int -2.
777 * config/avr/avr.md (addqi3): Use it. New alternatives for +/-2.
778 (*negqihi2): New insn.
780 2011-11-07 H.J. Lu <hongjiu.lu@intel.com>
782 * dwarf2cfi.c (dwarf2out_frame_debug_expr): Check
783 HARD_FRAME_POINTER_REGNUM instead of hard_frame_pointer_rtx in Rule 18.
785 2011-11-07 Jakub Jelinek <jakub@redhat.com>
787 * config/i386/i386.c (ix86_expand_builtin): If gather mask
788 argument is known to have all high bits set, pass pc_rtx as
789 second argument to the expander instead of op0.
790 * config/i386/sse.md (*avx2_gathersi<mode>_2,
791 *avx2_gatherdi<mode>_2): New patterns.
792 * config/i386/avx2intrin.h (_mm256_i32gather_pd,
793 _mm256_i64gather_pd, _mm256_i32gather_ps): Set mask using
794 _mm256_cmp_pd with zero vector arguments and _CMP_EQ_OQ instead of
797 PR tree-optimization/50789
798 * tree-vect-stmts.c (process_use): Add force argument, avoid
799 exist_non_indexing_operands_for_use_p check if true.
800 (vect_mark_stmts_to_be_vectorized): Adjust callers. Handle
802 (gen_perm_mask): New function.
803 (perm_mask_for_reverse): Use it.
804 (reverse_vec_element): Rename to...
805 (permute_vec_elements): ... this. Add Y and MASK_VEC arguments,
806 generalize for any permutations.
807 (vectorizable_load): Adjust caller. Handle STMT_VINFO_GATHER_P.
808 * target.def (TARGET_VECTORIZE_BUILTIN_GATHER): New hook.
809 * doc/tm.texi.in (TARGET_VECTORIZE_BUILTIN_GATHER): Document it.
810 * doc/tm.texi: Regenerate.
811 * tree-data-ref.c (initialize_data_dependence_relation,
812 compute_self_dependence): No longer static.
813 * tree-data-ref.h (initialize_data_dependence_relation,
814 compute_self_dependence): New prototypes.
815 * tree-vect-data-refs.c (vect_check_gather): New function.
816 (vect_analyze_data_refs): Detect possible gather load data refs.
817 * tree-vectorizer.h (struct _stmt_vec_info): Add gather_p field.
818 (STMT_VINFO_GATHER_P): Define.
819 (vect_check_gather): New prototype.
820 * config/i386/i386-builtin-types.def: Add types for alternate
822 * config/i386/sse.md (AVXMODE48P_DI): Remove.
823 (VEC_GATHER_MODE): Rename mode_attr to...
824 (VEC_GATHER_IDXSI): ... this.
825 (VEC_GATHER_IDXDI, VEC_GATHER_SRCDI): New mode_attrs.
826 (avx2_gathersi<mode>, *avx2_gathersi<mode>): Use <VEC_GATHER_IDXSI>
827 instead of <VEC_GATHER_MODE>.
828 (avx2_gatherdi<mode>): Use <VEC_GATHER_IDXDI> instead of
829 <AVXMODE48P_DI> and <VEC_GATHER_SRCDI> instead of VEC_GATHER_MODE
830 on src and mask operands.
831 (*avx2_gatherdi<mode>): Likewise. Use VEC_GATHER_MODE iterator
832 instead of AVXMODE48P_DI.
833 (avx2_gatherdi<mode>256, *avx2_gatherdi<mode>256): Removed.
834 * config/i386/i386.c (enum ix86_builtins): Add
835 IX86_BUILTIN_GATHERALTSIV4DF, IX86_BUILTIN_GATHERALTDIV8SF,
836 IX86_BUILTIN_GATHERALTSIV4DI and IX86_BUILTIN_GATHERALTDIV8SI.
837 (ix86_init_mmx_sse_builtins): Create those builtins.
838 (ix86_expand_builtin): Handle those builtins and adjust expansions
839 of other gather builtins.
840 (ix86_vectorize_builtin_gather): New function.
841 (TARGET_VECTORIZE_BUILTIN_GATHER): Define.
843 2011-11-07 Uros Bizjak <ubizjak@gmail.com>
845 * config/i386/f16cintrin.h: Remove extra _X86INTRIN_H_INCLUDED check.
847 2011-11-07 Tristan Gingold <gingold@adacore.com>
849 * config.gcc (*-*-*vms*): Set c_target_objs, cxx_target_objs.
850 * config/vms/t-vms: Add vms-c.o rule.
851 * config/vms/vms-c.c: New file.
852 * config/vms/vms-protos.h (vms_c_register_pragma): New prototype.
853 * config/vms/vms.h (REGISTER_TARGET_PRAGMAS): Define.
855 2011-11-07 Tristan Gingold <gingold@adacore.com>
857 * config/alpha/vms.h (TARGET_OBJECT_SUFFIX,
858 TARGET_EXECUTABLE_SUFFIX, TARGET_OS_CPP_BUILTINS,
859 TARGET_ABI_OPEN_VMS, LONG_TYPE_SIZE, ADA_LONG_TYPE_SIZE,
860 POINTER_SIZE, POINTERS_EXTEND_UNSIGNED): Move to config/vms/vms.h
861 (SUBTARGET_OS_CPP_BUILTINS): Define.
862 (TARGET_DEFAULT): Tune according to POINTER_SIZE.
863 (MASK_RETURN_ADDR): Define in 64 bit mode.
864 * config/ia64/vms.h: Likewise.
865 * config/vms/vms.h: New file.
866 * config/vms/vms64.h: New file.
867 * config/alpha/vms64.h: Removed.
868 * config/ia64/vms64.h: Removed.
869 * config.gcc (alpha64-dec-*vms*, alpha*-dec-*vms*, ia64-hp-*vms*):
870 Adjust for above change.
872 2011-11-07 Enkovich Ilya <ilya.enkovich@intel.com>
875 * config/i386/i386-protos.h (ix86_use_lea_for_mov): New.
876 * config/i386/i386.c (ix86_use_lea_for_mov): Likewise.
877 * config/i386/i386.md (movsi_internal): Emit lea if profitable.
878 (movdi_internal_rex64): Likewise.
880 2011-11-07 Sergey Ostanevich <sergos.gnu@gmail.com>
882 PR rtl-optimization/47698
883 * ifcvt.c (noce_operand_ok): Return false for mems with side effects.
885 2011-11-07 Tristan Gingold <gingold@adacore.com>
887 * common/config/alpha/alpha-common.c (alpha_option_init_struct):
889 (TARGET_OPTION_INIT_STRUCT): Define.
890 * config/alpha/alpha.c (TARGET_MIN_ANCHOR_OFFSET)
891 (TARGET_MAX_ANCHOR_OFFSET)
892 (TARGET_USE_BLOCKS_FOR_CONSTANT_P): Redefine.
894 2011-11-06 Quentin Neill <quentin.neill@amd.com>
896 Fix r180999, update ChangeLog
897 * config.gcc: Add f16cintrin.h.
898 * config/i386/f16cintrin.h: Add missing endif.
900 2011-11-06 Sebastian Huber <sebastian.huber@embedded-brains.de>
902 * config.gcc (arm*-*-rtemseabi*): New target.
903 * config/arm/rtems-eabi.h: New.
904 * config/arm/t-rtems-eabi: New.
906 2011-11-06 David S. Miller <davem@davemloft.net>
908 * config/sparc/sparc.md (UNSPEC_SHORT_LOAD): Delete.
909 (zero_extend_v8qi_vis, zero_extend_v4hi_vis,
910 *zero_extend_v8qi_<P:mode>_insn,
911 *zero_extend_v4hi_<P:mode>_insn): Express using vec_merge
912 and vec_duplicate instead of using an UNSPEC.
914 2011-11-07 Alan Modra <amodra@gmail.com>
917 * config/rs6000/rs6000.c (rs6000_emit_stack_reset): Always emit
920 2011-11-06 Dave Korn <dave.korn.cygwin@gmail.com>
922 * config/i386/cygwin.h (LIBGCJ_SONAME): Updated to match recent
924 * config/i386/mingw32.h (LIBGCJ_SONAME): Likewise.
926 2011-11-06 Eric Botcazou <ebotcazou@adacore.com>
928 * gcse.c: Adjust outdated comments throughout.
929 (struct mem_conflict_info): New structure.
930 (mems_conflict_for_gcse_p): Use it to communicate with caller.
931 (load_killed_in_block_p): Pass it to note_stores.
932 (hash_expr): Remove superfluous line break.
933 (hash_scan_set): Rename PAT parameter into SET.
934 (hash_scan_insn): Reorder cases.
935 (canon_list_insert): Fix long line.
937 (prune_expressions): Rename E local variable into EXPR.
938 (compute_pre_data): Return struct edge_list * object.
939 (pre_expr_reaches_here_p_work): Fix formatting.
940 (process_insert_insn): Move around comment.
941 (pre_edge_insert): Fix long line.
942 (pre_insert_copies): Likewise.
943 (gcse_emit_move_after): Swap SRC and DEST parameters.
944 (pre_delete): Adjust call to gcse_emit_move_after.
945 (pre_gcse): Take struct edge_list * parameter. Fix long line.
946 (one_pre_gcse_pass): Use flag_gcse_lm condition for all routines.
947 Use a local list of edges.
948 (hoist_code): Fix long line. Adjust call to gcse_emit_move_after.
949 (pre_ldst_expr_hash): Fix long line.
950 (free_ldst_mems): Rename into...
951 (free_ld_motion_mems): ...this.
952 (first_ls_expr): Delete.
953 (next_ls_expr): Likewise.
954 (print_ldst_list): Do not use above two functions.
955 (simple_mem): Adjust interface.
956 (compute_ld_motion_mems): Fix formatting.
957 (update_ld_motion_stores): Reuse local variable.
959 2011-11-06 Joseph Myers <joseph@codesourcery.com>
961 * c-decl.c (shadow_tag_warned, grokdeclarator): Handle _Alignas
963 (build_null_declspecs): Initialize align_log and alignas_p fields.
964 (declspecs_add_alignas): New.
965 * c-parser.c (c_token_starts_declspecs): Handle RID_ALIGNAS.
966 (c_parser_declspecs): Handle _Alignas specifiers.
967 (c_parser_alignas_specifier): New.
968 (c_parser_alignof_expression): Diagnose alignof use for non-C1X.
969 Diagnose _Alignof (expression).
970 * c-tree.h (struct c_declspecs): Add align_log and alignas_p fields.
971 (declspecs_add_alignas): Declare.
972 * ginclude/stddef.h (max_align_t): Define for C1X and C++11.
973 * ginclude/stdalign.h: New.
974 * Makefile.in (USER_H): Add stdalign.h.
976 2011-11-06 Joern Rennecke <joern.rennecke@embecosm.com>
978 * regset.h (fixed_reg_set_regset): Declare.
979 * dse.c: Include regset.h .
980 (struct insn_info): Add member fixed_regs_live.
981 (note_add_store_info): New typedef.
982 (note_add_store): New function.
983 (emit_inc_dec_insn_before): Expect arg to be of type insn_info_t.
984 Use gen_add3_insn / gen_move_insn.
985 Check new insn for unwanted clobbers before emitting it.
986 (check_for_inc_dec): Rename to...
987 (check_for_inc_dec_1:) ... this. Return bool. Take insn_info
988 parameter. Changed all callers in file.
989 (check_for_inc_dec, copy_fixed_regs): New functions.
990 (scan_insn): Set fixed_regs_live field of insn_info.
991 * rtl.h (check_for_inc_dec): Update prototype.
992 * postreload.c (reload_cse_simplify): Take new signature of
993 check_ind_dec into account.
994 * reginfo.c (fixed_reg_set_regset): New variable.
995 (init_reg_sets_1): Initialize it.
997 2011-11-06 Jakub Jelinek <jakub@redhat.com>
999 * tree-cfg.c (gimple_can_merge_blocks_p): For -O0 don't remove
1002 2011-11-06 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
1004 * config/pa/pa.c (pa_hpux_init_libfuncs): Rename to pa_init_libfuncs.
1005 Remove dependence of declaration and target define on definition of
1006 HPUX_LONG_DOUBLE_LIBRARY. Update implementation.
1008 2011-11-06 Andrew Macleod <amacleod@redhat.com>
1009 Richard Henderson <rth@redhat.com>
1010 Aldy Hernandez <aldyh@redhat.com>
1012 Merged from cxx-mem-model.
1014 * cppbuiltin.c (define__GNUC__): Define __ATOMIC memory models
1015 * coretypes.h (enum memmodel): New. enumerated memory model type.
1016 * Makefile.in (cppbuiltin.o) Add missing dependency on $(TREE_H)
1017 * genopinit,c (optabs): Add atomic direct optab handlers.
1018 * sync-builtins.def (BUILT_IN_ATOMIC_*): New builtins.
1019 * builtin-types.def (BT_CONST_VOLATILE_PTR,
1020 BT_FN_I{1,2,4,8,16}_CONST_VPTR_INT, BT_FN_VOID_VPTR_INT,
1021 BT_FN_BOOL_VPTR_INT, BT_FN_BOOL_SIZE_CONST_VPTR,
1022 BT_FN_I{1,2,4,8,16}_VPTR_I{1,2,4,8,16}_INT,
1023 BT_FN_VOID_VPTR_I{1,2,4,8,16}_INT, BT_FN_VOID_SIZE_VPTR_PTR_INT,
1024 BT_FN_VOID_SIZE_CONST_VPTR_PTR_INT, BT_FN_VOID_SIZE_VPTR_PTR_PTR_INT,
1025 BT_FN_BOOL_VPTR_PTR_I{1,2,4,8,16}_BOOL_INT_INT): New builtin types.
1026 * expr.h (expand_atomic_*): Add prototypes.
1027 (expand_{bool,val}_compare_and_swap): Remove prototypes.
1028 * c-typeck.c (build_function_call_vec): Don't reprocess __atomic
1030 * common.opt (Winvalid-memory-model): New warning flag.
1031 (finline-atomics): New. Flag to disable atomic inlining.
1032 * params.h (ALLOW_LOAD_DATA_RACES): New.
1033 (ALLOW_PACKED_LOAD_DATA_RACES): New.
1034 (ALLOW_PACKED_STORE_DATA_RACES): New.
1035 * params.def (PARAM_ALLOW_LOAD_DATA_RACES): New.
1036 (PARAM_ALLOW_PACKED_LOAD_DATA_RACES): New.
1037 (PARAM_ALLOW_PACKED_STORE_DATA_RACES): New.
1038 * builtins.c (is_builtin_name): Handle __atomic.
1039 (get_memmodel): New. Extract memory model.
1040 (expand_expr_force_mode): New. Factor out common code for ensuring an
1041 integer argument is in the proper mode.
1042 (expand_builtin_sync_operation): Remove ignore param. Always call
1043 expand_atomic_fetch_op instead of the old expanders.
1044 (expand_builtin_compare_and_swap,
1045 expand_builtin_sync_lock_test_and_set): Use expand_expr_force_mode,
1046 call atomic expanders instead of sync expanders.
1047 (expand_builtin_sync_lock_release): Call atomic_store expander.
1048 (expand_builtin_atomic_compare_exchange, expand_builtin_atomic_load,
1049 expand_builtin_atomic_store, expand_builtin_atomic_fetch_op): New.
1050 (expand_builtin_atomic_exchange): New.
1051 (fold_builtin_atomic_always_lock_free,
1052 expand_builtin_atomic_always_lock_free,
1053 fold_builtin_atomic_is_lock_free, expand_builtin_atomic_is_lock_free):
1055 (expand_builtin_mem_thread_fence, expand_builtin_atomic_thread_fence,
1056 expand_builtin_atomic_signal_fence): New.
1057 (expand_builtin_mem_signal_fence): New.
1058 (expand_builtin): Add cases for BUILT_IN_ATOMIC_*.
1059 (fold_builtin_2): Add cases for BUILT_IN_ATOMIC_{IS,ALWAYS}_LOCK_FREE.
1060 * optabs.h (DOI_atomic_*): Define new atomics.
1061 (atomic_*_optab): Define.
1062 (can_compare_and_swap_p, expand_atomic_compare_and_swap): New
1064 * optabs.c (expand_sync_operation, expand_sync_fetch_operation): Remove.
1065 (expand_sync_lock_test_and_set): Remove.
1066 (expand_atomic_load, expand_atomic_store): New.
1067 (expand_atomic_exchange): New.
1068 (expand_atomic_compare_and_swap): New. Implements
1069 atomic_compare_exchange via compare and swap.
1070 (struct atomic_op_functions): Opcode table struct for fetch ops.
1071 (get_atomic_op_for_code): New. Return an opcode table entry.
1072 (maybe_emit_op): New. Try to emit a fetch op.
1073 (expand_atomic_fetch_op): New.
1074 (expand_val_compare_and_swap_1): Remove.
1075 (expand_val_compare_and_swap, expand_bool_compare_and_swap): Remove.
1076 (expand_atomic_compare_and_swap): Rename from
1077 expand_atomic_compare_exchange. Rewrite to return both success and
1078 oldval return values; expand via both atomic and sync optabs.
1079 (can_compare_and_swap_p): New.
1080 (expand_compare_and_swap_loop): Use expand_atomic_compare_and_swap.
1081 (maybe_gen_insn): Handle 7 and 8 operands.
1082 * omp-low.c (expand_omp_atomic_fetch_op): Don't test individual
1083 fetch_op optabs, only test can_compare_and_swap_p. Use __atomic
1084 builtins instead of __sync builtins.
1085 (expand_omp_atomic_pipeline): Use can_compare_and_swap_p.
1086 * doc/extend.texi: Document __atomic built-in functions.
1087 * doc/invoke.texi: Document data race parameters.
1088 * doc/md.texi: Document atomic patterns.
1089 * config/i386/i386.md (UNSPEC_MOVA): New.
1090 (UNSPECV_CMPXCHG): Split into ...
1091 (UNSPECV_CMPXCHG_1, UNSPECV_CMPXCHG_2,
1092 UNSPECV_CMPXCHG_3, UNSPECV_CMPXCHG_4): New.
1093 * config/i386/sync.md (ATOMIC): New mode iterator.
1094 (atomic_load<ATOMIC>, atomic_store<ATOMIC>): New.
1095 (atomic_loaddi_fpu, atomic_storedi_fpu, movdi_via_fpu): New.
1096 (mem_thread_fence): Rename from memory_barrier.
1097 Handle the added memory model parameter.
1098 (mfence_nosse): Rename from memory_barrier_nosse.
1099 (sync_compare_and_swap<CASMODE>): Split into ...
1100 (atomic_compare_and_swap<SWI124>): this and ...
1101 (atomic_compare_and_swap<CASMODE>): this. Handle the new parameters.
1102 (atomic_compare_and_swap_single<SWI>): Rename from
1103 sync_compare_and_swap<SWI>; rewrite to use split unspecs.
1104 (atomic_compare_and_swap_double<DCASMODE>): Rename from
1105 sync_double_compare_and_swap<DCASMODE>; rewrite to use split unspecs.
1106 (*atomic_compare_and_swap_doubledi_pic): Rename from
1107 sync_double_compare_and_swapdi_pic; rewrite to use split unspecs.
1108 (atomic_fetch_add<SWI>): Rename from sync_old_add<SWI>; add memory
1110 (*atomic_fetch_add_cmp<SWI>): Similarly.
1111 (atomic_add<SWI>, atomic<any_logic><SWI>): Similarly.
1112 (atomic_sub<SWI>): Similarly. Use x86_maybe_negate_const_int.
1113 (sync_lock_test_and_set<SWI>): Merge with ...
1114 (atomic_exchange<SWI>): ... this.
1116 2011-11-6 Richard Guenther <rguenther@suse.de>
1118 * ipa-prop.c (ipa_modify_call_arguments): Re-compute inlinable flag.
1120 2011-11-06 Ira Rosen <ira.rosen@linaro.org>
1122 * tree-vectorizer.h (vectorizable_condition): Add argument.
1123 * tree-vect-loop.c (vectorizable_reduction): Fail for condition
1124 in SLP. Update calls to vectorizable_condition.
1125 * tree-vect-stmts.c (vect_is_simple_cond): Add basic block info to
1126 the arguments. Pass it to vect_is_simple_use_1.
1127 (vectorizable_condition): Add slp_node to the arguments. Support
1128 vectorization of basic blocks. Fail for reduction in SLP. Update
1129 calls to vect_is_simple_cond and vect_is_simple_use. Support SLP:
1130 call vect_get_slp_defs to get vector operands.
1131 (vect_analyze_stmt): Update calls to vectorizable_condition.
1132 (vect_transform_stmt): Likewise.
1133 * tree-vect-slp.c (vect_create_new_slp_node): Handle COND_EXPR.
1134 (vect_get_and_check_slp_defs): Handle COND_EXPR. Allow pattern
1136 (vect_build_slp_tree): Handle COND_EXPR.
1137 (vect_analyze_slp_instance): Push pattern statements to root node.
1138 (vect_get_constant_vectors): Fix comments. Handle COND_EXPR.
1140 2011-11-05 David S. Miller <davem@davemloft.net>
1142 * config/sparc/sparc.md (UNSPEC_SHORT_LOAD): New unspec.
1143 (zero-extend_v8qi_vis, zero_extend_v4hi_vis): New expanders.
1144 (*zero_extend_v8qi_<P:mode>_insn,
1145 *zero_extend_v4hi_<P:mode>_insn): New insns.
1146 * config/sparc/sparc.c (vector_init_move_words,
1147 vector_init_prepare_elts, sparc_expand_vector_init_vis2,
1148 sparc_expand_vector_init_vis1): New functions.
1149 (vector_init_bshuffle): Rewrite to handle more cases and make use
1150 of locs[] array prepared by vector_init_prepare_elts.
1151 (vector_init_fpmerge, vector_init_faligndata): Delete.
1152 (sparc_expand_vector_init): Rewrite using new infrastructure.
1154 2011-11-05 Joern Rennecke <joern.rennecke@embecosm.com>
1156 * config.gcc (epiphany-*-*): New architecture.
1157 (epiphany-*-elf): New configuration.
1158 * config/epiphany, common/config/epiphany : New directories.
1159 * doc/extend.texi (disinterrupt attribute): Add Epiphany.
1160 (interrupt attribute): Add Epiphany.
1161 (long_call, short_call attribute): Add Epiphany.
1162 * doc/invoke.texi (Options): Add Epiphany options.
1163 * doc/md.texi (Machine Constraints): Add Epiphany constraints.
1164 * doc/install.texi (Options specification):
1165 Add --with-stack-offset=@var{num} description.
1166 (host/target specific issues): Add epiphany-*-elf.
1167 * doc/contrib.texi (Contributors): Mention Epiphany port.
1169 2011-11-05 Jakub Jelinek <jakub@redhat.com>
1171 PR tree-optimization/50693
1172 * tree-cfg.c (gimple_can_merge_blocks_p): Allow merging with
1173 non-forced user labels.
1174 (gimple_merge_blocks): Turn non-forced user labels into
1175 debug bind stmt with the label as first operand and reset value.
1176 (gimple_duplicate_bb): Don't duplicate label debug stmts.
1177 * dwarf2out.c (gen_label_die): Handle NOTE_INSN_DELETED_DEBUG_LABEL.
1178 * final.c (final_scan_insn): Likewise.
1179 (rest_of_clean_state): Don't dump NOTE_INSN_DELETED_DEBUG_LABEL.
1180 * var-tracking.c (debug_label_num): New variable.
1181 (delete_debug_insns): Don't delete DEBUG_INSNs for LABEL_DECLs,
1182 instead turn them into NOTE_INSN_DELETED_DEBUG_LABEL notes.
1183 * cfglayout.c (skip_insns_after_block, duplicate_insn_chain): Handle
1184 NOTE_INSN_DELETED_DEBUG_LABEL.
1185 (duplicate_insn_chain): Don't duplicate LABEL_DECL DEBUG_INSNs.
1186 * insn-notes.def (DELETED_DEBUG_LABEL): New note kind.
1187 * print-rtl.c (print_rtx): Handle NOTE_INSN_DELETED_DEBUG_LABEL.
1188 * gengtype.c (adjust_field_rtx_def): Likewise.
1189 * config/i386/i386.c (ix86_output_function_epilogue): For MachO
1190 clear CODE_LABEL_NUMBER of NOTE_INSN_DELETED_DEBUG_LABEL
1191 if their are at the end of function and nop hasn't been emitted.
1192 * config/rs6000/rs6000.c (rs6000_output_function_epilogue): Likewise.
1194 2011-11-05 Georg-Johann Lay <avr@gjlay.de>
1196 PR rtl-optimization/50448
1197 * cprop.c (try_replace_reg): Also try to replace uses of FROM that
1200 2011-11-05 Peter Dufault <dufault@hda.com>,
1201 Sebastian Huber <sebastian.huber@embedded-brains.de>
1203 * config/rs6000/rtems.h (SUBSUBTARGET_OVERRIDE_OPTIONS):
1204 Extend rs6000_spe handling.
1206 2011-11-05 Ralf Corsépius <ralf.corsepius@rtems.org>
1208 * config/rs6000/t-rtems: Add -mcpu=8540/-mfloat-gprs=double multilib.
1209 Remove -mcpu=601 multilib.
1210 Remove -Dmpc8260 multilib.
1211 * config/rs6000/rtems.h: Allow --float-gprs=... to override grps
1214 2011-11-05 Quentin Neill <quentin.neill@amd.com>
1216 Piledriver f16cintrin.h fix.
1217 * config/i386/f16cintrin.h: Contents moved from immintrin.h.
1218 * config/i386/immintrin.h: Include f16cintrin.h.
1219 * config.gcc (i[34567]86-*-*, x86_64-*-*): Add f16cintrin.h.
1222 2011-11-04 Eric Botcazou <ebotcazou@adacore.com>
1225 * c-parser.c (c_parser_postfix_expression) <RID_OFFSETOF>: Adjust call
1227 * c-typeck.c (build_unary_op) <ADDR_EXPR>: Call fold_offsetof_1.
1229 2011-11-04 Alan Modra <amodra@gmail.com>
1231 * reload1.c (gen_reload): Don't use REGNO on SUBREGs.
1232 * print-rtl.c (print_rtx): Don't segfault on negative regno.
1234 2011-11-04 David S. Miller <davem@davemloft.net>
1237 * config/sparc/sparc.c (sparc_expand_conditional_move): Handle the
1238 fact that sparc_emit_float_lib_cmp modifies the comparison in
1241 2011-11-04 Ralf Corsépius <ralf.corsepius@rtems.org>
1243 * config/lm32/t-rtems: New.
1244 * config.gcc (lm32-*-rtems*): Add t-rtems.
1246 2011-11-04 Eric Botcazou <ebotcazou@adacore.com>
1249 * config/sparc/sparc.h (ASM_CPU_SPEC): Pass -Av8 if -mcpu=v8.
1251 2011-11-04 Jiangning Liu <jiangning.liu@arm.com>
1253 PR rtl-optimization/38644
1254 * config/arm/arm.c (thumb1_expand_epilogue): Add memory barrier
1255 for epilogue having stack adjustment.
1257 2011-11-04 Georg-Johann Lay <avr@gjlay.de>
1260 * config/avr/avr-modes.def: New file defining PSImode.
1261 * config/avr/avr-c.c (__INT24_MAX__, __INT24_MIN__,
1262 __UINT24_MAX__): New built-in defines.
1263 * config/avr/avr.md (adjust_len): Add tstpsi, mov24, reload_in24,
1264 ashlpsi, ashrpsi, lshrpsi.
1265 (QISI, QIDI, HISI, HIDI, MPUSH, rotx, rotsmode): Add PSI.
1266 (MOVMODE): New mode iterator.
1267 (movpsi): New expander.
1268 (movqi, movhi, movsi, movsf, movpsi): Write as one using MOVMODE.
1269 (*reload_inpsi, *movpsi): New insns.
1270 (*reload_inpsi): New RTL peephole.
1271 (addpsi3, *addpsi3_zero_extend.qi, *addpsi3_zero_extend.hi,
1272 *addpsi3_sign_extend.hi): New insns.
1273 (subpsi3, *subpsi3_zero_extend.qi, *subpsi3_zero_extend.hi,
1274 *subpsi3_sign_extend.hi): New insns.
1275 (divmodpsi4, udivmodpsi4): New define insn-and-split.
1276 (*divmodpsi4_call, *udivmodpsi4_call): New insns.
1277 (andpsi3, iorpsi3, xorpsi3): New insns.
1278 (*rotlpsi2.1, *rotlpsi2.23): New insns.
1279 (*rotw<mode>): Insn condition only allow even-sized modes.
1280 (*rotb<mode>): Insn condition allows odd-sized modes.
1281 (ashlpsi3, ashrpsi3, lshrpsi3, *addpsi3.lt0): New insns.
1282 (negpsi2, one_cmplpsi2): New insns.
1283 (extendqipsi2, extendhipsi2, extendpsisi2): New insns.
1284 (zero_extendqipsi2, zero_extendhipsi2, zero_extendpsisi2): New
1286 (*cmppsi, *negated_tstpsi, *reversed_tstpsi): New insns.
1287 (cbranchpsi4): New expander.
1288 * config/avr/constraints.md (Ca3, Co3, Cx3): New constraints.
1289 * config/avr/avr-protos.h (avr_out_tstpsi, avr_out_movpsi,
1290 avr_out_ashlpsi3, avr_out_ashrpsi3, avr_out_lshrpsi3,
1291 avr_out_reload_inpsi): New prototypes.
1293 * config/avr/avr.c (TARGET_SCALAR_MODE_SUPPORTED_P): Define to...
1294 (avr_scalar_mode_supported_p): ...this new static function.
1295 (avr_asm_len): Always return "".
1296 (avr_out_load_psi, avr_out_store_psi): New static functions.
1297 (avr_out_movpsi, avr_out_reload_inpsi): New functions.
1298 (avr_out_tstpsi): New function.
1299 (avr_out_ashlpsi3, avr_out_ashrpsi3, avr_out_lshrpsi3): New functions.
1300 (avr_out_plus_1, output_reload_in_const): Handle 3-byte types.
1301 (avr_simplify_comparison_p): Ditto.
1302 (adjust_insn_length): Handle ADJUST_LEN_RELOAD_IN24,
1303 ADJUST_LEN_MOV24, ADJUST_LEN_TSTPSI, ADJUST_LEN_ASHLPSI,
1304 ADJUST_LEN_ASHRPSI, ADJUST_LEN_LSHRPSI.
1305 (avr_rtx_costs_1): Report PSI costs.
1306 (avr_libcall_value): Handle odd-sized parameters.
1307 (avr_init_builtin_int24): New static function to define built-in
1308 24-bit types __int24 and __uint24.
1309 (avr_init_builtins): Use it.
1311 2011-11-04 Thomas Doerfler <thomas.doerfler@embedded-brains.de>
1314 * config/arm/rtems-elf.h, config/arm/t-rtems: Add optional
1315 support for VFP floating point model.
1317 2011-11-04 Tristan Gingold <gingold@adacore.com>
1319 * config/alpha/vms.h (ASM_OUTPUT_DEF): Do not switch section.
1321 2011-11-04 Ira Rosen <ira.rosen@linaro.org>
1324 2011-10-24 Ira Rosen <ira.rosen@linaro.org>
1326 PR tree-optimization/50730
1327 * tree-vect-data-refs.c (vect_analyze_data_refs): Stop basic block
1328 analysis if encountered unsupported data-ref.
1330 2011-11-04 Jakub Jelinek <jakub@redhat.com>
1332 * config/i386/i386.c (ix86_expand_vector_convert_uns_vsivsf): New
1334 * config/i386/i386-protos.h (ix86_expand_vector_convert_uns_vsivsf):
1336 * config/i386/sse.md (floatuns<sseintvecmodelower><mode>2): Use it.
1337 For floatunsv8siv8sf2 require TARGET_AVX2.
1339 * config/i386/i386.c (ix86_expand_adjust_ufix_to_sfix_si): Add
1340 XORP argument. Subtract 0x1p31 instead of 0x1p32. Use normal
1341 signalling comparison instead of non-signalling. Store into
1342 *XORP pseudo holding 0x80000000 integers if 0x1p31 has been
1343 subtracted and 0 otherwise.
1344 * config/i386/i386-protos.h (ix86_expand_adjust_ufix_to_sfix_si):
1346 * config/i386/sse.md (fixuns_trunc<mode><sseintvecmodelower>2): Enable
1347 already for TARGET_SSE2. Xor in vector initialized by
1348 ix86_expand_adjust_ufix_to_sfix_si at the end.
1349 (vec_pack_ufix_trunc_<mode>): Likewise.
1351 * tree-vect-stmts.c (vectorizable_conversion): Rewritten to handle
1352 not just FLOAT_EXPR and FIX_TRUNC_EXPR, but also CONVERT_EXPR_CODE_P,
1353 WIDEN_MULT_EXPR and WIDEN_LSHIFT_EXPR to handle what
1354 vectorizable_type_demotion and vectorizable_type_promotion did.
1355 Additionally handle FLOAT_EXPR and FIX_TRUNC_EXPR where the integer
1356 is {,un}signed {char,short}.
1357 (vect_create_vectorized_demotion_stmts): Fix comment typo. For
1358 recursive calls unconditionally use VEC_PACK_TRUNC_EXPR.
1359 Push vec_dest back to the vec_dsts vector at the end.
1360 (vect_create_vectorized_promotion_stmts): Don't recurse, do just
1361 one step. Removed multi_step_cvt, vec_dsts, slp_node and
1362 prev_stmt_info arguments, add vec_dest argument. Push always
1363 into vec_tmp, not just when multi_step_cvt != 0, replace *vec_oprdn0
1364 with vec_tmp at the end after freeing old *vec_oprnd0 vector.
1365 (vectorizable_type_demotion, vectorizable_type_promotion): Removed.
1366 (vect_analyze_stmt): Don't call vectorizable_type_demotion and
1367 vectorizable_type_promotion. Call vectorizable_conversion even
1368 for SLP bb vectorization.
1369 (vect_transform_stmt): Call vectorizable_conversion instead of
1370 vectorizable_type_demotion and vectorizable_type_promotion.
1371 (supportable_widening_operation): Clear *multi_step_cvt first,
1372 simplify c1/c2 computation, free *interm_types vector on failure.
1373 (supportable_narrowing_operation): Clear *multi_step_cvt first,
1374 free *interm_types vector on failure, handle multi-step
1377 2011-11-04 Tristan Gingold <gingold@adacore.com>
1379 * config/alpha/alpha.c (alpha_write_linkage): Remove fundecl
1380 argument. Conditionally generate crash debug info. Adjust
1381 for alpha_funcs_tree removal.
1382 (machine_function): Add links field.
1383 (alpha_start_function): Conditionally generate crash debug info.
1384 (alpha_end_function): Adjust call to alpha_write_linkage.
1385 (alpha_funcs): Remove.
1386 (links_kind): Remove.
1387 (alpha_links): Remove num, target and lkind field. Add func field.
1388 (alpha_links_tree): Remove.
1389 (alpha_funcs_tree): Remove.
1390 (alpha_need_linkage): Remove.
1391 (alpha_use_linkage): Change prototype. Adjust.
1392 (alpha_write_one_linkage): Use ASM_OUTPUT_INTERNAL_LABEL.
1393 Use SYMBOL_REF_EXTERNAL_P and SYMBOL_REF_LOCAL_P macro.
1394 * config/alpha/alpha-protos.h (alpha_use_linkage): Update.
1395 (alpha_need_linkage): Remove.
1396 * config/alpha/alpha.md: Update calls to alpha_use_linkage.
1397 Adjust calls to alpha_need_linkage.
1399 2011-11-03 Uros Bizjak <ubizjak@gmail.com>
1401 * sched-vis.c (print_value): Handle STRICT_LOW_PART.
1403 2011-11-03 Uros Bizjak <ubizjak@gmail.com>
1405 * config/i386/i386.md (lround<X87MODEF:mode><SWI248x:mode>2,
1406 rint<mode>2, floor<mode>2, lfloor<MODEF:mode><SWI48:mode>2,
1407 btrunc<mode>2, lwp_lwpval<mode>3): Use operands[N] instead of operandN.
1409 2011-11-03 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
1412 * config/arm/t-bpabi: New file.
1413 * config.gcc (arm*-*-linux*): Add arm/t-bpabi to tmake_file for
1415 (arm*-*-uclinux*): Add arm/t-bpabi to tmake_file for
1416 arm*-*-uclinux*eabi.
1417 (arm*-*-eabi*, arm*-*-symbianelf*): Add arm/t-bpabi to tmake_file
1420 2011-11-03 Michael Matz <matz@suse.de>
1423 * configure.ac: Check for -fno-exceptions -fno-rtti.
1424 * configure: Regenerate.
1425 * Makefile.in (NOEXCEPTION_FLAGS): New flag.
1426 (ALL_CXXFLAGS): Use it.
1428 2011-11-03 Uros Bizjak <ubizjak@gmail.com>
1430 * config/i386/i386.md: Use {} for multi-line preparation statements.
1432 2011-11-03 Eric Botcazou <ebotcazou@adacore.com>
1434 * config/sparc/sparc.md (movtf_insn_sp32_no_fpu): Consolidate into...
1435 (movtf_insn_sp32): ...this.
1436 (movtf_insn_sp64_no_fpu): Consolidate into...
1437 (movtf_insn_sp64): ...this.
1438 (movtf_insn_sp64_hq): Do not test TARGET_FPU.
1439 * config/sparc/sparc.c (sparc_legitimate_address_p): Likewise.
1441 2011-11-03 Tristan Gingold <gingold@adacore.com>
1443 * config/vms/vms.c (vms_patch_builtins): Fix typo.
1445 2011-11-03 Richard Guenther <rguenther@suse.de>
1448 * lto-opts.c: Re-implement.
1449 * lto-streamer.h (lto_register_user_option): Remove.
1450 (lto_read_file_options): Likewise.
1451 (lto_reissue_options): Likewise.
1452 (lto_clear_user_options): Likewise.
1453 (lto_clear_file_options): Likewise.
1454 * opts-global.c (post_handling_callback): Remove.
1455 (set_default_handlers): Do not set post_handling_callback.
1456 (decode_options): Remove LTO specific code.
1457 * lto-wrapper.c (merge_and_complain): New function.
1458 (run_gcc): Read all input file options and
1459 prepend a merged set before the linker driver options.
1460 * gcc.c (driver_post_handling_callback): Remove.
1461 (set_option_handlers): Do not set post_handling_callback.
1462 * opts-common.c (handle_option): Do not call post_handling_callback.
1463 * opts.h (struct cl_option_handlers): Remove post_handling_callback.
1465 2011-11-03 Richard Guenther <rguenther@suse.de>
1467 * collect2.c (main): Guard object_nbr variable with TARGET_AIX_VERSION.
1469 2011-11-03 Martin Jambor <mjambor@suse.cz>
1471 * ipa-prop.c (type_change_info): New fields offset, object,
1472 known_current_type and multiple_types_encountered.
1473 (extr_type_from_vtbl_ptr_store): New function.
1474 (check_stmt_for_type_change): Use it, set multiple_types_encountered if
1475 the result is different from the previous one.
1476 (detect_type_change): Renamed to detect_type_change_1. New parameter
1477 comp_type. Set up new fields in tci, build known type jump
1478 functions if the new type can be identified.
1479 (detect_type_change): New function.
1480 * tree.h (DECL_CONTEXT): Comment new use.
1482 2011-11-03 Richard Guenther <rguenther@suse.de>
1485 * lto-wrapper.c (get_options_from_collect_gcc_options): Properly
1486 decode an encoded literal '.
1488 2011-11-03 Tristan Gingold <gingold@adacore.com>
1490 * collect2.c (main): Add support of -f (response file) on AIX.
1492 2011-11-03 Ira Rosen <ira.rosen@linaro.org>
1494 PR tree-optimization/50912
1495 * tree-vectorizer.h (slp_void_p): New.
1496 (struct _slp_tree): Replace left and right with children. Update
1498 (struct _slp_oprnd_info): New.
1499 (vect_get_vec_defs): Declare.
1500 (vect_get_slp_defs): Update arguments.
1501 * tree-vect-loop.c (vect_create_epilog_for_reduction): Call
1502 vect_get_vec_defs instead of vect_get_slp_defs.
1503 (vectorizable_reduction): Likewise.
1504 * tree-vect-stmts.c (vect_get_vec_defs): Remove static, add argument.
1505 Update call to vect_get_slp_defs.
1506 (vectorizable_conversion): Update call to vect_get_vec_defs.
1507 (vectorizable_assignment, vectorizable_shift,
1508 vectorizable_operation): Likewise.
1509 (vectorizable_type_demotion): Call vect_get_vec_defs instead of
1511 (vectorizable_type_promotion, vectorizable_store): Likewise.
1512 (vect_analyze_stmt): Fix typo.
1513 * tree-vect-slp.c (vect_free_slp_tree): Update SLP tree traversal.
1514 (vect_print_slp_tree, vect_mark_slp_stmts,
1515 vect_mark_slp_stmts_relevant, vect_slp_rearrange_stmts,
1516 vect_detect_hybrid_slp_stmts, vect_slp_analyze_node_operations,
1517 vect_schedule_slp_instance): Likewise.
1518 (vect_create_new_slp_node): New.
1519 (vect_create_oprnd_info, vect_free_oprnd_info): Likewise.
1520 (vect_get_and_check_slp_defs): Pass information about defs using
1521 oprnds_info, allow any number of operands.
1522 (vect_build_slp_tree): Likewise. Update calls to
1523 vect_get_and_check_slp_defs. Fix comments.
1524 (vect_analyze_slp_instance): Move node creation to
1525 vect_create_new_slp_node.
1526 (vect_get_slp_defs): Allow any number of operands.
1528 2011-11-02 Peter Bergner <bergner@vnet.ibm.com>
1529 Iain Sandoe <iains@gcc.gnu.org>
1531 * config/rs6000/rs6000.c (USE_HIDDEN_LINKONCE): New define.
1532 (get_ppc476_thunk_name): Use it.
1533 (rs6000_code_end): Likewise.
1534 (macho_branch_islands): Fix typo.
1536 2011-11-02 Paolo Carlini <paolo.carlini@oracle.com>
1537 Jason Merrill <jason@redhat.com>
1540 * configure.ac: Add -Wno-narrowing to warning options.
1541 * doc/invoke.texi ([-Wnarrowing], [-Wc++0x-compat]): Update.
1543 2011-11-02 Eric Botcazou <ebotcazou@adacore.com>
1546 * config/sparc/sparc.md (movsf_insn): Reindent constraints.
1547 (movdf_insn_sp32): Likewise. Remove redundant G constraint.
1548 (movdf_insn_sp64): Likewise.
1549 (DFmode splitter): Do not test TARGET_FPU.
1550 (movtf_insn_sp32): Reindent constraints.
1551 (movtf_insn_sp32_no_fpu): Likewise.
1552 (movtf_insn_sp64): Likewise.
1553 (movtf_insn_sp64_hq): Likewise.
1554 (movtf_insn_sp64_no_fpu): Likewise.
1556 2011-11-02 Paolo Carlini <paolo.carlini@oracle.com>
1559 * builtins.c (fold_builtin_memchr): Fix cast.
1561 2011-11-02 Teresa Johnson <tejohnson@google.com>
1563 * config/i386/predicates.md (promotable_binary_operator): Add minus
1564 to the list of promotable operators.
1566 2011-11-02 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
1568 * gthr-single.h, gthr.h: Move to ../libgcc.
1569 * gthr-aix.h: Move to ../libgcc/config/rs6000.
1570 * gthr-dce.h: Move to ../libgcc/config/pa.
1571 * gthr-lynx.h: Move to ../libgcc/config.
1572 * gthr-mipssde.h: Move to ../libgcc/config/mips.
1573 * gthr-posix.h: Move to ../libgcc/config.
1574 * gthr-rtems.h: Likewise.
1575 * gthr-tpf.h: Move to ../libgcc/config/s390.
1576 * gthr-vxworks.h: Move to ../libgcc/config.
1577 * gthr-win32.h: Move to ../libgcc/config/i386.
1578 * configure.ac (gthread_flags): Remove
1579 (gthr-default.h): Don't create.
1580 (thread_file): Don't substitute.
1581 * configure: Regenerate.
1582 * Makefile.in (GCC_THREAD_FILE): Remove.
1583 (GTHREAD_FLAGS): Remove.
1584 (libgcc.mvars): Remove GTHREAD_FLAGS.
1585 * config/t-vxworks (EXTRA_HEADERS): Remove.
1587 2011-11-02 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
1588 Paolo Bonzini <bonzini@gnu.org>
1590 * configure.ac (libgcc_tm_file_list, libgcc_tm_include_list): Remove.
1591 * configure: Regenerate.
1592 * Makefile.in (libgcc_tm_file_list, libgcc_tm_include_list): Remove.
1593 (TM_H): Remove libgcc_tm.h, $(libgcc_tm_file_list).
1594 (libgcc_tm.h, cs-libgcc_tm.h): Remove.
1595 (clean): Remove libgcc_tm.h
1596 * mkconfig.sh: Don't include libgcc_tm.h in tm.h.
1597 * config.gcc (libgcc_tm_file): Remove.
1598 (arm*-*-linux*): Remove libgcc_tm_file for arm*-*-linux-*eabi.
1599 (arm*-*-uclinux*): Remove libgcc_tm_file for arm*-*-uclinux*eabi.
1600 (arm*-*-eabi*, arm*-*-symbianelf*): Remove libgcc_tm_file.
1601 (avr-*-rtems*): Likewise.
1602 (avr-*-*): Likewise.
1603 (frv-*-elf): Likewise.
1604 (frv-*-*linux*): Likewise.
1605 (h8300-*-rtems*): Likewise.
1606 (h8300-*-elf*): Likewise.
1607 (i[34567]86-*-darwin*): Likewise.
1608 (x86_64-*-darwin*): Likewise.
1609 (rx-*-elf*): Likewise.
1610 (tic6x-*-elf): Likewise.
1611 (tic6x-*-uclinux): Likewise.
1612 (i[34567]86-*-linux*, x86_64-*-linux*): Likewise.
1614 2011-11-02 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
1616 * Makefile.in (LIBGCC2_DEBUG_CFLAGS LIBGCC2_CFLAGS)
1617 (LIBGCC2_INCLUDES, TARGET_LIBGCC2_CFLAGS, LIB2FUNCS_EXTRA)
1618 (LIB2FUNCS_STATIC_EXTRA, LIB2FUNCS_EXCLUDE, T, T_TARGET)
1619 (INCLUDES_FOR_TARGET): Remove.
1620 (LIBGCC2_CFLAGS): Don't export.
1621 (LIB2FUNCS_ST, LIB2_DIVMOD_FUNCS, LIB2ADD, LIB2ADD_ST, srcdirify):
1623 (libgcc-support): Remove $(LIB2ADD), $(LIB2ADD_ST) dependencies.
1624 (libgcc.mvars): Likewise.
1625 Don't emit LIB2FUNCS_ST, LIB2FUNCS_EXCLUDE, LIB2ADD, LIB2ADD_ST,
1626 LIB2_SIDITI_CONV_FUNCS, LIB2_DIVMOD_FUNCS, LIBGCC2_CFLAGS,
1627 TARGET_LIBGCC2_CFLAGS.
1629 * libgcc2.c, libgcc2.h, gbl-ctors.h, longlong.h: Move to ../libgcc.
1630 * config/darwin-64.c: Move to ../libgcc/config.
1631 * config/divmod.c, config/floatunsidf.c, config/floatunsisf.c,
1632 config/floatunsitf.c, config/floatunsixf.c, config/udivmod.c,
1633 config/udivmodsi4.c: Move to ../libgcc/config.
1634 * config/gthr-posix.c: Move to ../libgcc/config/alpha.
1635 * config/memcmp.c, config/memcpy.c, config/memmove.c,
1636 config/memset.c: Move to ../libgcc/config.
1637 * config/t-darwin (TARGET_LIBGCC2_CFLAGS): Remove.
1638 * config/t-freebsd: Remove.
1639 * config/t-freebsd-thread: Move to ../libgcc/config.
1640 * config/t-libgcc-pic: Move to ../libgcc/config.
1641 * config/t-libunwind (TARGET_LIBGCC2_CFLAGS): Remove.
1642 * config/t-linux: Remove.
1643 * config/t-lynx (TARGET_LIBGCC2_CFLAGS, LIBGCC, INSTALL_LIBGCC):
1645 * config/t-openbsd-thread: Move to ../libgcc/config.
1646 * config/t-rtems (LIBGCC2_INCLUDES): Remove.
1647 * config/t-sol2 (TARGET_LIBGCC2_CFLAGS): Remove.
1648 * config/t-svr4: Remove.
1649 * config/t-vxworks (LIBGCC, INSTALL_LIBGCC, TARGET_LIBGCC2_CFLAGS)
1650 (LIBGCC2_DEBUG_CFLAGS, LIB2FUNCS_EXTRA, LIBGCC2_INCLUDES): Remove.
1651 * config/vxlib.c, config/vxlib-tls.c: Move to ../libgcc/config.
1652 * config/alpha/qrnnd.asm: Move to ../libgcc/config/alpha/qrnnd.S.
1653 * config/alpha/t-alpha, config/alpha/t-ieee: Remove.
1654 * config/alpha/t-vms (LIB2FUNCS_EXTRA, LIBGCC, INSTALL_LIBGCC): Remove.
1655 * config/alpha/vms-gcc_shell_handler.c: Move to ../libgcc/config/alpha.
1656 * config/arm/bpabi.c, config/arm/unaligned-funcs.c,
1657 config/arm/fp16.c, config/arm/linux-atomic.c,
1658 config/arm/linux-atomic-64bit.c: Move to ../libgcc/config/arm.
1659 * config/arm/t-arm-elf (LIBGCC, INSTALL_LIBGCC)
1660 (TARGET_LIBGCC2_CFLAGS): Remove.
1661 * config/arm/t-bpabi, config/arm/t-linux: Remove.
1662 * config/arm/t-linux-eabi (TARGET_LIBGCC2_CFLAGS)
1663 (LIB2FUNCS_STATIC_EXTRA): Remove.
1664 * config/arm/t-netbsd: Remove.
1665 * config/arm/t-strongarm-elf (LIBGCC, INSTALL_LIBGCC)
1666 (TARGET_LIBGCC2_CFLAGS): Remove.
1667 * config/arm/t-symbian (LIB2FUNCS_STATIC_EXTRA): Remove.
1668 * config/arm/t-wince-pe (LIBGCC, INSTALL_LIBGCC)
1669 (TARGET_LIBGCC2_CFLAGS): Remove.
1670 * config/avr/t-avr (LIB2FUNCS_EXCLUDE, TARGET_LIBGCC2_CFLAGS)
1671 (LIBGCC, INSTALL_LIBGCC): Remove.
1672 * config/bfin/t-bfin-elf (TARGET_LIBGCC2_CFLAGS): Remove.
1673 * config/bfin/t-bfin-linux: Likewise.
1674 * config/bfin/t-bfin-uclinux: Likewise.
1675 * config/c6x/eqd.c, config/c6x/eqf.c, config/c6x/ged.c,
1676 config/c6x/gef.c, config/c6x/gtd.c, config/c6x/gtf.c,
1677 config/c6x/led.c, config/c6x/lef.c, config/c6x/ltd.c,
1678 config/c6x/ltf.c: Move to ../libgcc/config/c6x.
1679 * config/c6x/t-c6x-elf (LIB2FUNCS_EXCLUDE, LIB2FUNCS_EXTRA): Remove.
1680 * config/c6x/t-c6x-uclinux (TARGET_LIBGCC2_CFLAGS): Remove.
1681 * config/cris/arit.c: Move to ../libgcc/config/cris.
1682 * config/cris/cris_abi_symbol.c: Remove.
1683 * config/cris/cris.h: Remove obsolete comment.
1684 * config/cris/mulsi3.asm: Move to ../libgcc/config/cris/mulsi3.S.
1685 * config/cris/t-cris (LIB2FUNCS_EXTRA, CRIS_LIB1CSRC)
1686 ($(LIB2FUNCS_EXTRA)): Remove.
1687 * config/cris/t-elfmulti (LIB2FUNCS_STATIC_EXTRA, INSTALL_LIBGCC)
1689 * config/cris/t-linux (TARGET_LIBGCC2_CFLAGS): Remove.
1690 * config/fr30/t-fr30: Remove.
1691 * config/frv/cmovd.c, config/frv/cmovh.c, config/frv/cmovw.c,
1692 config/frv/modi.c, config/frv/uitod.c, config/frv/uitof.c,
1693 config/frv/ulltod.c, config/frv/ulltof.c, config/frv/umodi.c: Move
1694 to ../libgcc/config/frv.
1695 * config/frv/t-frv (LIB2FUNCS_EXTRA, TARGET_LIBGCC2_CFLAGS)
1696 (cmovh.c, cmovw.c, cmovd.c, modi.c, umodi.c, uitof.c, uitod.c)
1697 (ulltof.c, LIBGCC, INSTALL_LIBGCC): Remove.
1698 * config/frv/t-linux (TARGET_LIBGCC2_CFLAGS): Remove.
1699 * config/h8300/clzhi2.c, config/h8300/ctzhi2.c,
1700 config/h8300/fixunssfsi.c, config/h8300/parityhi2.c,
1701 config/h8300/popcounthi2.c: Move to ../libgcc/config/h8300.
1702 * config/h8300/t-h8300 (LIB2FUNCS_EXTRA, TARGET_LIBGCC2_CFLAGS)
1703 (LIBGCC, INSTALL_LIBGCC): Remove.
1704 * config/i386/gthr-win32.c: Move to ../libgcc/config/i386.
1705 * config/i386/t-cygming (LIBGCC2_INCLUDES): Remove.
1706 * config/i386/t-cygwin: Remove.
1707 * config/i386/t-darwin (LIB2_SIDITI_CONV_FUNCS, LIB2FUNCS_EXTRA)
1708 (LIB2FUNCS_EXCLUDE): Remove.
1709 * config/i386/t-darwin64 (LIB2_SIDITI_CONV_FUNCS, LIB2FUNCS_EXTRA)
1710 (LIBGCC, INSTALL_LIBGCC): Remove.
1711 * config/i386/t-gthr-win32: Move to ../libgcc/config/i386.
1712 * config/i386/t-linux64 (LIBGCC, INSTALL_LIBGCC): Remove.
1713 * config/i386/t-mingw-w32: Likewise.
1714 * config/i386/t-mingw-w64: Likewise.
1715 * config/i386/t-openbsd: Likewise.
1716 * config/i386/t-nto: Remove.
1717 * config/ia64/quadlib.c: Move to ../libgcc/config/ia64.
1718 * config/ia64/t-hpux (LIBGCC, INSTALL_LIBGCC, LIB2FUNCS_EXTRA)
1719 (quadlib.c): Remove.
1720 * config/ia64/t-ia64: Remove comment.
1721 * config/iq2000/lib2extra-funcs.c: Move to
1722 ../libgcc/config/iq2000/lib2funcs.c.
1723 * config/iq2000/t-iq2000: Remove.
1724 * config/m32c/m32c-lib2.c: Move to ../libgcc/config/m32c/lib2funcs.c.
1725 * config/m32c/m32c-lib2-trapv.c: Move to ../libgcc/config/m32c/trapv.c.
1726 * config/m32r/t-linux (TARGET_LIBGCC2_CFLAGS): Remove.
1727 * config/m32c/t-m32c (LIB2FUNCS_EXTRA): Remove.
1728 * config/m32r/t-m32r (TARGET_LIBGCC2_CFLAGS, LIBGCC)
1729 (INSTALL_LIBGCC): Remove.
1730 * config/m68k/fpgnulib.c: Move to ../libgcc/config/m68k.
1731 * config/m68k/t-floatlib: Remove.
1732 * config/m68k/t-mlibs (LIBGCC, INSTALL_LIBGCC): Remove.
1733 * config/mcore/t-mcore (TARGET_LIBGCC2_CFLAGS): Remove.
1735 (LIBGCC, INSTALL_LIBGCC): Remove.
1736 * config/mep/mep-lib2.c: Move to ../libgcc/config/mep/lib2funcs.c.
1737 * config/mep/mep-tramp.c: Move to ../libgcc/config/mep/tramp.c.
1738 * config/mep/t-mep (LIB2FUNCS_EXTRA): Remove.
1739 * config/mips/t-elf (TARGET_LIBGCC2_CFLAGS, LIBGCC)
1740 (INSTALL_LIBGCC): Remove.
1741 * config/mips/t-isa3264: Likewise.
1742 * config/mips/t-mips (LIB2_SIDITI_CONV_FUNCS): Remove.
1743 * config/mips/t-r3900 (TARGET_LIBGCC2_CFLAGS, LIBGCC)
1744 (INSTALL_LIBGCC): Remove.
1745 * config/mips/t-sde (LIBGCC, INSTALL_LIBGCC): Remove.
1746 * config/mips/t-sr71k (TARGET_LIBGCC2_CFLAGS, LIBGCC)
1747 (INSTALL_LIBGCC): Remove.
1748 * config/mips/t-vr (TARGET_LIBGCC2_CFLAGS)
1749 (LIB2FUNCS_STATIC_EXTRA): Remove.
1750 * config/mips/vr4120-div.S: Move to ../libgcc/config/mips.
1751 * config/mmix/t-mmix (TARGET_LIBGCC2_CFLAGS): Remove.
1752 * config/mn10300/t-mn10300 (LIBGCC, INSTALL_LIBGCC): Remove.
1753 * config/pa/fptr.c, config/pa/linux-atomic.c: Move to
1754 ../libgcc/config/pa.
1755 * config/pa/lib2funcs.asm: Move to ../libgcc/config/pa/lib2funcs.S.
1756 * config/pa/quadlib.c: Move to ../libgcc/config/pa.
1757 * config/pa/t-dce-thr (LIBGCC, INSTALL_LIBGCC): Remove.
1758 * config/pa/t-linux, config/pa/t-linux64: Remove.
1759 * config/pa/t-pa-hpux, config/pa/t-pa-hpux10,
1760 config/pa/t-pa-hpux11, config/pa/t-pa64: Remove.
1761 * config/pdp11/t-pdp11 (TARGET_LIBGCC2_CFLAGS, LIB2FUNCS_EXTRA):
1763 * config/picochip/libgccExtras: Move to ../libgcc/config/picochip.
1764 * config/picochip/t-picochip (LIB2FUNCS_EXTRA, RANLIB_FOR_TARGET)
1765 (TARGET_LIBGCC2_CFLAGS, LIBGCC2_DEBUG_CFLAGS): Remove.
1766 * config/rs6000/crtresfpr.asm: Move to
1767 ../libgcc/config/rs6000/crtresfpr.S.
1768 * config/rs6000/crtresgpr.asm: Move to
1769 ../libgcc/config/rs6000/crtresgpr.S.
1770 * config/rs6000/crtresxfpr.asm: Move to
1771 ../libgcc/config/rs6000/crtresxfpr.S.
1772 * config/rs6000/crtresxgpr.asm: Move to
1773 ../libgcc/config/rs6000/crtresxgpr.S.
1774 * config/rs6000/crtsavfpr.asm: Move to
1775 ../libgcc/config/rs6000/crtsavfpr.S.
1776 * config/rs6000/crtsavgpr.asm: Move to
1777 ../libgcc/config/rs6000/crtsavgpr.S.
1778 * config/rs6000/darwin-asm.h: Move to ../libgcc/config/rs6000.
1779 * config/rs6000/darwin-fpsave.asm: Move to
1780 ../libgcc/config/rs6000/darwin-fpsave.S.
1781 * config/rs6000/darwin-gpsave.asm: Move to
1782 ../libgcc/config/rs6000/darwin-gpsave.S.
1783 * config/rs6000/darwin-tramp.asm: Move to
1784 ../libgcc/config/rs6000/darwin-tramp.S.
1785 * config/rs6000/darwin-vecsave.asm: Move to
1786 ../libgcc/config/rs6000/darwin-vecsave.S.
1787 * config/rs6000/darwin-world.asm: Move to
1788 ../libgcc/config/rs6000/darwin-world.S.
1789 * config/rs6000/e500crtres32gpr.asm: Move to
1790 ../libgcc/config/rs6000/e500crtres32gpr.S.
1791 * config/rs6000/e500crtres64gpr.asm: Move to
1792 ../libgcc/config/rs6000/e500crtres64gpr.S.
1793 * config/rs6000/e500crtres64gprctr.asm: Move to
1794 ../libgcc/config/rs6000/e500crtres64gprctr.S.
1795 * config/rs6000/e500crtrest32gpr.asm: Move to
1796 ../libgcc/config/rs6000/e500crtrest32gpr.S.
1797 * config/rs6000/e500crtrest64gpr.asm: Move to
1798 ../libgcc/config/rs6000/e500crtrest64gpr.S.
1799 * config/rs6000/e500crtresx32gpr.asm: Move to
1800 ../libgcc/config/rs6000/e500crtresx32gpr.S.
1801 * config/rs6000/e500crtresx64gpr.asm: Move to
1802 ../libgcc/config/rs6000/e500crtresx64gpr.S.
1803 * config/rs6000/e500crtsav32gpr.asm: Move to
1804 ../libgcc/config/rs6000/e500crtsav32gpr.S.
1805 * config/rs6000/e500crtsav64gpr.asm: Move to
1806 ../libgcc/config/rs6000/e500crtsav64gpr.S.
1807 * config/rs6000/e500crtsav64gprctr.asm: Move to
1808 ../libgcc/config/rs6000/e500crtsav64gprctr.S.
1809 * config/rs6000/e500crtsavg32gpr.asm: Move to
1810 ../libgcc/config/rs6000/e500crtsavg32gpr.S.
1811 * config/rs6000/e500crtsavg64gpr.asm: Move to
1812 ../libgcc/config/rs6000/e500crtsavg64gpr.S.
1813 * config/rs6000/e500crtsavg64gprctr.asm: Move to
1814 ../libgcc/config/rs6000/e500crtsavg64gprctr.S.
1815 * config/rs6000/eabi.asm: Move to ../libgcc/config/rs6000/eabi.S.
1816 * config/rs6000/t-aix43 (LIBGCC, INSTALL_LIBGCC, LIB2FUNCS_EXTRA)
1817 (TARGET_LIBGCC2_CFLAGS): Remove.
1818 * config/rs6000/t-aix52: Likewise.
1819 * config/rs6000/t-darwin: Remove.
1820 * config/rs6000/t-darwin64 (LIB2_SIDITI_CONV_FUNCS)
1821 (LIB2FUNCS_EXTRA): Remove.
1822 * config/rs6000/t-fprules (LIBGCC, INSTALL_LIBGCC): Remove.
1823 * config/rs6000/t-linux64 (TARGET_LIBGCC2_CFLAGS): Remove.
1824 * config/rs6000/t-lynx (LIB2FUNCS_EXTRA, tramp.S, LIBGCC)
1825 (INSTALL_LIBGCC): Remove.
1826 * config/rs6000/t-netbsd (LIB2FUNCS_EXTRA)
1827 (LIB2FUNCS_STATIC_EXTRA, tramp.S, crtsavfpr.S, crtresfpr.S)
1828 (crtsavgpr.S, crtresgpr.S, crtresxfpr.S, crtresxgpr.S, LIBGCC)
1829 (INSTALL_LIBGCC, $(T)crtsavfpr$(objext), $(T)crtresfpr$(objext))
1830 ($(T)crtsavgpr$(objext), $(T)crtresgpr$(objext))
1831 ($(T)crtresxfpr$(objext), $(T)crtresxgpr$(objext)): Remove.
1832 * config/rs6000/t-ppccomm (LIB2FUNCS_EXTRA)
1833 (LIB2FUNCS_STATIC_EXTRA, eabi.S, tramp.S): Remove.
1834 * config/rs6000/t-spe (LIBGCC, INSTALL_LIBGCC): Remove.
1835 * config/rs6000/t-vxworks: Remove comment.
1836 * config/rs6000/tramp.asm: Move to ../libgcc/config/rs6000/tramp.S.
1837 * config/rx/t-rx (LIBGCC, INSTALL_LIBGCC): Remove.
1838 * config/sh/linux-atomic.asm: Move to
1839 ../libgcc/config/sh/linux-atomic.S.
1840 * config/sh/t-linux (LIB2FUNCS_EXTRA): Remove.
1841 * config/sh/t-netbsd: Remove.
1842 * config/sh/t-sh (TARGET_LIBGCC2_CFLAGS, LIBGCC, INSTALL_LIBGCC):
1844 * config/sparc/t-elf (LIBGCC, INSTALL_LIBGCC): Remove.
1845 * config/sparc/t-leon: Likewise.
1846 * config/sparc/t-leon3: Likewise.
1847 * config/sparc/t-linux64: Likewise.
1848 * config/sparc/t-netbsd64: Fix typo.
1850 * config/spu/divmodti4.c, config/spu/divv2df3.c,
1851 config/spu/float_disf.c, config/spu/float_unsdidf.c,
1852 config/spu/float_unsdisf.c, config/spu/float_unssidf.c,
1853 config/spu/mfc_multi_tag_release.c,
1854 config/spu/mfc_multi_tag_reserve.c, config/spu/mfc_tag_release.c,
1855 config/spu/mfc_tag_reserve.c, config/spu/mfc_tag_table.c,
1856 config/spu/multi3.c: Move to ../libgcc/config/spu.
1857 * config/spu/t-spu-elf (TARGET_LIBGCC2_CFLAGS, LIB2FUNCS_EXCLUDE)
1858 (LIB2FUNCS_STATIC_EXTRA, LIB2_SIDITI_CONV_FUNCS, LIBGCC)
1859 (INSTALL_LIBGCC): Remove.
1860 * config/stormy16/stormy16-lib2.c: Move to
1861 ../libgcc/config/stormy16/lib2.c.
1862 * config/stormy16/stormy16-lib2-ashlsi3.c: Move to
1863 ../libgcc/config/stormy16/ashlsi3.c.
1864 * config/stormy16/stormy16-lib2-ashrsi3.c: Move to
1865 ../libgcc/config/stormy16/ashrsi3.c.
1866 * config/stormy16/stormy16-lib2-clzhi2.c: Move to
1867 ../libgcc/config/stormy16/clzhi2.c.
1868 * config/stormy16/stormy16-lib2-cmpsi2.c: Move to
1869 ../libgcc/config/stormy16/cmpsi2.c.
1870 * config/stormy16/stormy16-lib2-ctzhi2.c: Move to
1871 ../libgcc/config/stormy16/ctzhi2.c.
1872 * config/stormy16/stormy16-lib2-divsi3.c: Move to
1873 ../libgcc/config/stormy16/divsi3.c.
1874 * config/stormy16/stormy16-lib2-ffshi2.c: Move to
1875 ../libgcc/config/stormy16/ffshi2.c.
1876 * config/stormy16/stormy16-lib2-lshrsi3.c: Move to
1877 ../libgcc/config/stormy16/lshrsi3.c.
1878 * config/stormy16/stormy16-lib2-modsi3.c: Move to
1879 ../libgcc/config/stormy16/modsi3.c.
1880 * config/stormy16/stormy16-lib2-parityhi2.c: Move to
1881 ../libgcc/config/stormy16/parityhi2.c.
1882 * config/stormy16/stormy16-lib2-popcounthi2.c: Move to
1883 ../libgcc/config/stormy16/popcounthi2.c.
1884 * config/stormy16/stormy16-lib2-ucmpsi2.c: Move to
1885 ../libgcc/config/stormy16/ucmpsi2.c.
1886 * config/stormy16/stormy16-lib2-udivmodsi4.c: Move to
1887 ../libgcc/config/stormy16/udivmodsi4.c.
1888 * config/stormy16/stormy16-lib2-udivsi3.c: Move to
1889 ../libgcc/config/stormy16/udivsi3.c.
1890 * config/stormy16/stormy16-lib2-umodsi3.c: Move to
1891 ../libgcc/config/stormy16/umodsi3.c.
1892 * config/stormy16/t-stormy16: Move to ../libgcc/config/t-stormy16.
1893 * config/v850/t-v850 (INSTALL_LIBGCC): Remove.
1894 * config/xtensa/lib2funcs.S: Move to ../libgcc/config/xtensa.
1895 * config/xtensa/t-elf: Remove.
1896 * config/xtensa/t-xtensa (LIB2FUNCS_EXTRA): Remove.
1897 * config.gcc (*-*-freebsd*): Remove t-freebsd, t-freebsd-thread
1899 (*-*-linux*, frv-*-*linux*, *-*-kfreebsd*-gnu, *-*-knetbsd*-gnu,
1900 *-*-gnu*, *-*-kopensolaris*-gnu): Remove t-linux from tmake_file.
1901 (*-*-netbsd*): Remove t-libgcc-pic from tmake_file.
1902 (*-*-openbsd*): Likewise.
1903 Remove t-openbsd-thread for posix threads.
1904 (alpha*-*-linux*): Remove alpha/t-alpha, alpha/t-ieee from tmake_file.
1905 (alpha*-*-freebsd*): Likewise.
1906 (alpha*-*-netbsd*): Likewise.
1907 (alpha*-*-openbsd*): Likewise.
1908 (alpha64-dec-*vms*): Likewise.
1909 (alpha*-dec-*vms*): Likewise.
1910 (arm*-*-netbsdelf*): Remove arm/t-netbsd from tmake_file.
1911 (arm*-*-linux*): Remove t-linux from tmake_file.
1912 Remove arm/t-bpabi from tmake_file for arm*-*-linux-*eabi.
1913 (arm*-*-uclinux*): Remove arm/t-bpabi from tmake_file for
1914 arm*-*-uclinux*eabi.
1915 (arm*-*-eabi*, arm*-*-symbianelf* ): Remove arm/t-bpabi from
1916 tmake_file for arm*-*-eabi*.
1917 (fr30-*-elf): Remove tmake_file.
1918 (hppa*64*-*-linux*): Remove tmake_file.
1919 (hppa*-*-linux*): Likewise.
1920 (hppa[12]*-*-hpux10*): Remove pa/t-pa-hpux10, pa/t-pa-hpux from
1922 (hppa*64*-*-hpux11*): Remove pa/t-pa64, pa/t-pa-hpux from tmake_file.
1923 (hppa[12]*-*-hpux11*): Remove pa/t-pa-hpux11, pa/t-pa-hpux from
1925 (i[34567]86-*-elf*): Remove tmake_file.
1926 (x86_64-*-elf*): Likewise.
1927 (i[34567]86-*-nto-qnx*): Likewise.
1928 (i[34567]86-*-cygwin*): Remove i386/t-cygwin from tmake_file.
1929 (i[34567]86-*-mingw*, x86_64-*-mingw*): Remove i386/t-gthr-win32
1930 from tmake_file if using win32 threads.
1931 (iq2000*-*-elf*): Remove tmake-file.
1932 (microblaze*-linux*): Likewise.
1933 (sh-*-elf*, sh[12346l]*-*-elf*, sh-*-linux*)
1934 (sh[2346lbe]*-*-linux*, sh-*-netbsdelf*, shl*-*-netbsdelf*)
1935 (sh5-*-netbsd*, sh5l*-*-netbsd*, sh64-*-netbsd*)
1936 (sh64l*-*-netbsd*): Remove sh/t-netbsd from tmake_file for
1937 sh5*-*-netbsd*, sh64*-netbsd*, *-*-netbsd.
1938 (xtensa*-*-elf*): Remove tmake_file.
1940 2011-11-02 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
1942 * Makefile.in (LIB1ASMSRC): Don't export.
1943 (libgcc.mvars): Don't emit LIB1ASMFUNCS, LIB1ASMSRC.
1944 * config/arm/arm.c: Update lib1funcs.asm filename.
1945 * config/arm/linux-eabi.h: Likewise.
1946 * config/arm/bpabi-v6m.S, config/arm/bpabi.S,
1947 config/arm/ieee754-df.S, config/arm/ieee754-sf.S: Move to
1948 ../libgcc/config/arm.
1949 * config/arm/lib1funcs.asm: Move to ../libgcc/config/arm/lib1funcs.S.
1950 * config/arm/t-arm (LIB1ASMSRC, LIB1ASMFUNCS): Remove.
1951 * config/arm/t-arm-elf (LIB1ASMFUNCS): Remove.
1952 * config/arm/t-bpabi: Likewise.
1953 * config/arm/t-linux (LIB1ASMSRC, LIB1ASMFUNCS): Remove.
1954 * config/arm/t-linux-eabi (LIB1ASMFUNCS): Remove.
1955 * config/arm/t-strongarm-elf: Likewise.
1956 * config/arm/t-symbian: Likewise.
1957 * config/arm/t-vxworks: Likewise.
1958 * config/arm/t-wince-pe: Likewise.
1959 * config/avr/libgcc.S: Move to ../libgcc/config/avr.
1960 * config/avr/t-avr (LIB1ASMSRC, LIB1ASMFUNCS): Remove.
1961 * config/bfin/lib1funcs.asm: Move to ../libgcc/config/bfin/lib1funcs.S.
1962 * config/bfin/t-bfin: Remove.
1963 * config/bfin/t-bfin-elf (LIB1ASMSRC, LIB1ASMFUNCS): Remove.
1964 * config/bfin/t-bfin-linux: Likewise.
1965 * config/bfin/t-bfin-uclinux: Likewise.
1966 * config/c6x/lib1funcs.asm: Move to ../libgcc/config/c6x/lib1funcs.S.
1967 * config/c6x/t-c6x-elf (LIB1ASMSRC, LIB1ASMFUNCS): Remove.
1968 * config/fr30/lib1funcs.asm: Move to ../libgcc/config/fr30/lib1funcs.S.
1969 * config/fr30/t-fr30 (LIB1ASMSRC, LIB1ASMFUNCS): Remove.
1970 * config/frv/lib1funcs.asm: Move to ../libgcc/config/frv/lib1funcs.S.
1971 * config/frv/t-frv (CROSS_LIBGCC1, LIB1ASMSRC, LIB1ASMFUNCS): Remove.
1972 * config/h8300/fixunssfsi.c: Update lib1funcs.asm filename.
1973 * config/h8300/lib1funcs.asm: Move to
1974 ../libgcc/config/h8300/lib1funcs.S.
1975 * config/h8300/t-h8300 (LIB1ASMSRC, LIB1ASMFUNCS): Remove.
1976 * config/i386/cygwin.asm: Move to ../libgcc/config/i386/cygwin.S.
1977 * config/i386/t-cygming (LIB1ASMSRC, LIB1ASMFUNCS): Remove.
1978 * config/i386/t-interix: Likewise.
1979 * config/ia64/lib1funcs.asm: Move to ../libgcc/config/ia64/lib1funcs.S.
1980 * config/ia64/t-hpux (LIB1ASMFUNCS, LIBGCC1_TEST): Remove.
1981 * config/ia64/t-ia64 (LIB1ASMSRC, LIB1ASMFUNCS): Remove.
1982 * config/iq2000/t-iq2000 (LIBGCC1, CROSS_LIBGCC1): Remove.
1983 * config/m32c/m32c.c: Update m32c-lib1.S filename.
1984 * config/m32c/m32c-lib1.S: Move to ../libgcc/config/m32c/lib1funcs.S.
1985 * config/m32c/t-m32c (LIB1ASMSRC, LIB1ASMFUNCS): Remove.
1986 * config/m32r/t-linux (CROSS_LIBGCC1, LIBGCC1, LIBGCC1_TEST): Remove.
1987 * config/m68k/lb1sf68.asm: Move to ../libgcc/config/m68k/lb1sf68.S.
1988 * config/m68k/t-floatlib (LIB1ASMSRC, LIB1ASMFUNCS): New file.
1989 * config/mcore/lib1.asm: Move to ../libgcc/config/mcore/lib1funcs.S.
1990 * config/mcore/t-mcore (LIB1ASMSRC, LIB1ASMFUNCS): Remove.
1991 * config/mep/mep-lib1.asm: Move to ../libgcc/config/mep/lib1funcs.S.
1992 * config/mep/t-mep (LIB1ASMSRC, LIB1ASMFUNCS): Remove.
1993 * config/mips/mips16.S: Move to ../libgcc/config/mips.
1994 * config/mips/t-libgcc-mips16: Remove.
1995 * config/mips/t-sr71k (LIBGCC1, CROSS_LIBGCC1): Remove.
1996 * config/pa/milli64.S: Move to ../libgcc/config/pa.
1997 * config/pa/t-linux (LIB1ASMFUNCS, LIB1ASMSRC): Remove.
1998 * config/pa/t-linux64: Likewise.
1999 * config/picochip/libgccExtras/fake_libgcc.asm: Move to
2000 ../libgcc/config/picochip/lib1funcs.S.
2001 * config/picochip/t-picochip (LIB1ASMFUNCS, LIB1ASMSRC): Remove.
2002 * config/sh/lib1funcs.asm: Move to ../libgcc/config/sh/lib1funcs.S.
2003 * config/sh/lib1funcs.h: Move to ../libgcc/config/sh.
2004 * config/sh/sh.h: Update lib1funcs.asm filename.
2005 * config/sh/t-linux (LIB1ASMFUNCS_CACHE): Remove.
2006 * config/sh/t-netbsd: Likewise.
2007 * config/sh/t-sh (LIB1ASMSRC, LIB1ASMFUNCS, LIB1ASMFUNCS_CACHE):
2009 * config/sh/t-sh64 (LIB1ASMFUNCS): Remove.
2010 * config/sparc/lb1spc.asm: Move to ../libgcc/config/sparc/lb1spc.S.
2011 * config/sparc/lb1spl.asm: Remove.
2012 * config/sparc/t-elf (LIB1ASMSRC, LIB1ASMFUNCS): Remove.
2013 * config/sparc/t-leon: Likewise.
2014 * config/spu/t-spu-elf (LIBGCC1, CROSS_LIBGCC1): Remove.
2015 * config/v850/lib1funcs.asm: Move to ../libgcc/config/v850/lib1funcs.S.
2016 * config/v850/t-v850 (LIB1ASMSRC, LIB1ASMFUNCS): Remove
2017 * config/vax/lib1funcs.asm: Move to ../libgcc/config/vax/lib1funcs.S.
2018 * config/vax/t-linux: Remove.
2019 * config/xtensa/ieee754-df.S, config/xtensa/ieee754-sf.S: Move to
2020 ../libgcc/config/xtensa.
2021 * config/xtensa/lib1funcs.asm: Move to
2022 ../libgcc/config/xtensa/lib1funcs.S.
2023 * config/xtensa/t-xtensa (LIB1ASMSRC, LIB1ASMFUNCS): Remove.
2024 * config.gcc (bfin*-rtems*): Remove bfin/t-bfin from tmake_file.
2025 (bfin*-*): Likewise.
2026 (mips64*-*-linux*, mipsisa64*-*-linux*): Remove
2027 mips/t-libgcc-mips16 from tmake_file.
2028 (mips*-*-linux*): Likewise.
2029 (mips*-sde-elf*): Likewise.
2030 (mipsisa32-*-elf*, mipsisa32el-*-elf*, mipsisa32r2-*-elf*)
2031 (mipsisa32r2el-*-elf*, mipsisa64-*-elf*, mipsisa64el-*-elf*)
2032 (mipsisa64r2-*-elf*, mipsisa64r2el-*-elf*): Likewise.
2033 (mipsisa64sb1-*-elf*, mipsisa64sb1el-*-elf*): Likewise.
2034 (mips-*-elf*, mipsel-*-elf*): Likewise.
2035 (mips64-*-elf*, mips64el-*-elf*): Likewise.
2036 (mips64orion-*-elf*, mips64orionel-*-elf*): Likewise.
2037 (mips*-*-rtems*): Likewise.
2038 (mipstx39-*-elf*, mipstx39el-*-elf*): Likewise.
2039 (vax-*-linux*): Remove vax/t-linux from tmake_file.
2041 2011-11-02 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
2043 * config.gcc (extra_parts): Remove.
2044 (*-*-freebsd*): Remove extra_parts.
2045 (*-*-linux*, frv-*-*linux*, *-*-kfreebsd*-gnu, *-*-knetbsd*-gnu,
2046 *-*-gnu*, *-*-kopensolaris*-gnu): Likewise.
2047 (*-*-netbsd*): Remove t-libc-ok, t-netbsd from tmake_file.
2048 Remove extra_parts for *-*-netbsd*1.[7-9]*, *-*-netbsd[2-9]*,
2049 *-*-netbsdelf[2-9]*.
2050 (*-*-openbsd*): Remove t-libc-ok from tmake_file.
2051 (alpha*-*-linux*): Remove extra_parts.
2052 (alpha*-*-freebsd*): Likewise.
2053 (bfin*-linux-uclibc*): Likewise.
2054 (fr30-*-elf): Likewise.
2055 (moxie-*-elf): Likewise.
2056 (moxie-*-uclinux*): Likewise.
2057 (h8300-*-rtems*): Remove h8300/t-elf from tmake_file.
2058 (h8300-*-elf*): Likewise.
2059 (hppa*64*-*-hpux11*): Remove extra_parts.
2060 (i[34567]86-*-elf*): Remove i386/t-i386elf, i386/t-crtstuff from
2062 (x86_64-*-elf*): Likewise.
2063 (i[34567]86-*-freebsd*): Remove tmake_file.
2064 (x86_64-*-freebsd*): Likewise.
2065 (x86_64-*-netbsd*): Likewise.
2066 (i[34567]86-*-openbsd2.*, i[34567]86-*openbsd3.[0123]): Remove
2067 t-libc-ok from tmake_file.
2068 (i[34567]86-*-linux*, i[34567]86-*-kfreebsd*-gnu,
2069 i[34567]86-*-knetbsd*-gnu, i[34567]86-*-gnu*,
2070 i[34567]86-*-kopensolaris*-gnu): Remove i386/t-crtstuff from
2073 (x86_64-*-linux*, x86_64-*-kfreebsd*-gnu, x86_64-*-knetbsd*-gnu):
2074 Remove i386/t-crtstuff from tmake_file.
2075 (i[34567]86-*-lynxos*): Likewise.
2077 (ia64*-*-elf*): Remove extra_parts.
2078 (ia64*-*-freebsd*): Likewise.
2079 (ia64*-*-linux*): Likewise.
2080 (ia64-hp-*vms*): Remove ia64/t-vms from tmake_file.
2081 (m32r-*-elf*): Remove extra_parts.
2082 (m32rle-*-elf*): Likewise.
2083 (m32r-*-rtems*): Likewise.
2084 (m68k-*-elf*, fido-*-elf*): Likewise.
2085 (m68k*-*-openbsd*): Remove t-libc-ok from tmake_file.
2086 (m68k-*-rtems*): Remove extra_parts.
2087 (mep-*-*): Likewise.
2088 (microblaze*-linux*): Likewise.
2089 (mips64*-*-linux*, mipsisa64*-*-linux*): Likewise.
2090 (mips*-*-linux*): Likewise.
2091 (powerpc-*-lynxos*): Likewise.
2092 (s390x-ibm-tpf*): Likewise.
2093 (score-*-elf): Likewise.
2095 (sh-*-elf*, sh[12346l]*-*-elf*, sh-*-linux*, sh[2346lbe]*-*-linux*,
2096 sh-*-netbsdelf*, shl*-*-netbsdelf*, sh5-*-netbsd*, sh5l*-*-netbsd*,
2097 sh64-*-netbsd*, sh64l*-*-netbsd*): Remove sh/t-elf from tmake_file.
2098 Remove sh/t-superh from tmake_file for sh*-superh-elf.
2099 Remove sh/t-linux64 from tmake_file for sh64*-*-linux*.
2100 (sh-*-rtems*): Remove sh/t-elf from tmake_file.
2101 (sh-wrs-vxworks): Likewise.
2102 (sparc-*-linux*): Remove extra_parts.
2103 (sparc64-*-linux*): Likewise.
2104 (sparc64-*-freebsd*, ultrasparc-*-freebsd*): Likewise.
2105 (xstormy16-*-elf): Likewise.
2106 (xtensa*-*-linux*): Remove xtensa/t-linux from tmake_file.
2107 (am33_2.0-*-linux*): Remove extra_parts.
2108 * configure.ac (extra_parts): Don't substitute.
2109 * configure: Regenerate.
2110 * crtstuff.c: Move to ../libgcc.
2111 * Makefile.in (CRTSTUFF_CFLAGS): Remove.
2112 (EXTRA_PARTS): Remove.
2113 (CRTSTUFF_T_CFLAGS): Remove.
2114 (MOSTLYCLEANFILES): Remove $(EXTRA_PARTS).
2115 (GCC_EXTRA_PARTS): Remove.
2116 (libgcc.mvars): Remove GCC_EXTRA_PARTS, CRTSTUFF_CFLAGS,
2117 CRTSTUFF_T_CFLAGS, CRTSTUFF_T_CFLAGS_S.
2118 Emit GCC_CFLAGS, INHIBIT_LIBC_CFLAGS.
2119 ($(T)crtbegin.o, $(T)crtend.o, $(T)crtbeginS.o, $(T)crtendS.o)
2120 ($(T)crtbeginT.o): Remove.
2121 * config/alpha/t-vms (EXTRA_PARTS): Remove.
2122 ($(T)vms-dwarf2.o, $(T)vms-dwarf2eh.o): Remove.
2123 * config/alpha/vms-dwarf2.asm: Move to
2124 ../libgcc/config/alpha/vms-dwarf2.S.
2125 * config/alpha/vms-dwarf2eh.asm: Move to
2126 ../libgcc/config/alpha/vms-dwarf2eh.S.
2127 * config/arm/crti.asm: Move to ../libgcc/config/arm/crti.S.
2128 * config/arm/crtn.asm: Move to ../libgcc/config/arm/crtn.S.
2129 * config/arm/t-arm-elf (EXTRA_MULTILIB_PARTS): Remove.
2130 ($(T)crti.o, $(T)crtn.o): Remove.
2131 * config/arm/t-linux: Remove comment.
2132 * config/arm/t-linux-eabi (EXTRA_MULTILIB_PARTS): Remove.
2133 * config/arm/t-strongarm-elf (EXTRA_MULTILIB_PARTS): Remove.
2134 ($(T)crti.o, $(T)crtn.o): Remove.
2135 * config/arm/t-symbian (EXTRA_MULTILIB_PARTS): Remove.
2136 * config/bfin/crti.s: Move to ../libgcc/config/bfin/crti.S.
2137 * config/bfin/crtn.s: Move to ../libgcc/config/bfin/crtn.S.
2138 * config/bfin/crtlibid.s: Move to ../libgcc/config/bfin/crtlibid.S.
2139 * config/bfin/t-bfin (EXTRA_PARTS): Remove.
2140 ($(T)crti.o, $(T)crtn.o): Remove.
2141 * config/bfin/t-bfin-elf (CRTSTUFF_T_CFLAGS): Remove.
2142 ($(T)crti.o, $(T)crtn.o, $(T)crtlibid.o): Remove
2143 (EXTRA_MULTILIB_PARTS): Remove.
2144 * config/bfin/t-bfin-linux (CRTSTUFF_T_CFLAGS,
2145 EXTRA_MULTILIB_PARTS): Remove.
2146 * config/bfin/t-bfin-uclinux (CRTSTUFF_T_CFLAGS): Remove.
2147 ($(T)crtlibid.o): Remove.
2148 (EXTRA_MULTILIB_PARTS): Remove.
2149 * config/c6x/crti.s: Move to ../libgcc/config/c6x/crti.S.
2150 * config/c6x/crtn.s: Move to ../libgcc/config/c6x/crtn.S.
2151 * config/c6x/t-c6x-elf ($(T)crti.o, $(T)crtn.o): Remove.
2152 (EXTRA_MULTILIB_PARTS): Remove.
2153 (CRTSTUFF_T_CFLAGS, CRTSTUFF_T_CFLAGS_S): Remove.
2154 * config/c6x/t-c6x-uclinux (CRTSTUFF_T_CFLAGS,
2155 CRTSTUFF_T_CFLAGS_S): Remove.
2156 * config/cris/t-elfmulti (CRTSTUFF_T_CFLAGS): Remove.
2157 * config/cris/t-linux (CRTSTUFF_T_CFLAGS_S): Remove.
2158 * config/fr30/crti.asm: Move to ../libgcc/config/fr30/crti.S.
2159 * config/fr30/crtn.asm: Move to ../libgcc/config/fr30/crtn.S.
2160 * config/fr30/t-fr30 ($(T)crti.o, $(T)crtn.o): Remove.
2161 * config/frv/frvbegin.c, config/frv/frvend.c: Move to
2162 ../libgcc/config/frv.
2163 * config/frv/t-frv (EXTRA_MULTILIB_PARTS): Remove.
2164 (FRVSTUFF_CFLAGS, $(T)frvbegin$(objext), $(T)frvend$(objext)): Remove.
2165 * config/frv/t-linux (EXTRA_MULTILIB_PARTS): Remove.
2166 (CRTSTUFF_T_CFLAGS): Remove.
2167 * config/h8300/crti.asm: Move to ../libgcc/config/h8300/crti.S.
2168 * config/h8300/crtn.asm: Move to ../libgcc/config/h8300/crtn.S.
2169 * config/h8300/t-elf: Remove.
2170 * config/i386/cygming-crtbegin.c, config/i386/cygming-crtend.c:
2171 Move to ../libgcc/config/i386.
2172 * config/i386/t-crtstuff: Remove.
2173 * config/i386/t-i386elf: Remove.
2174 * config/i386/t-linux64 (EXTRA_MULTILIB_PARTS): Remove.
2175 * config/i386/t-nto (CRTSTUFF_T_CFLAGS, EXTRA_PARTS): Remove.
2176 * config/ia64/crtbegin.asm: Move to ../libgcc/config/ia64/crtbegin.S.
2177 * config/ia64/crtend.asm: Move to ../libgcc/config/ia64/crtend.S.
2178 * config/ia64/crti.asm: Move to ../libgcc/config/ia64/crti.S.
2179 * config/ia64/crtn.asm: Move to ../libgcc/config/ia64/crtn.S.
2180 * config/ia64/t-vms: Remove.
2181 * config/ia64/vms-crtinit.asm: Move to
2182 ../libgcc/config/ia64/vms-crtinit.S.
2183 * config/m32c/t-m32c (EXTRA_MULTILIB_PARTS): Remove.
2184 * config/m32r/initfini.c: Move to ../libgcc/config/m32r.
2185 * config/m32r/t-linux (CRTSTUFF_T_CFLAGS_S): Remove.
2186 * config/m32r/t-m32r (CRTSTUFF_T_CFLAGS): Remove.
2187 ($(T)crtinit.o, $(T)crtfini.o): Remove.
2188 (m32rx, m32r2): Remove.
2189 (EXTRA_MULTILIB_PARTS): Remove.
2190 * config/m68k/crti.s: Move to ../libgcc/config/m68k/crti.S.
2191 * config/m68k/crtn.s: Move to ../libgcc/config/m68k/crtn.S.
2192 * config/m68k/t-crtstuff: Remove.
2193 * config/m68k/t-linux (EXTRA_MULTILIB_PARTS): Remove.
2194 * config/m68k/t-m68kelf: Remove.
2195 * config/m68k/t-uclinux (EXTRA_MULTILIB_PARTS): Remove.
2196 * config/mcore/crti.asm: Move to ../libgcc/config/mcore/crti.S.
2197 * config/mcore/crtn.asm: Move to ../libgcc/config/mcore/crtn.S.
2198 * config/mcore/t-mcore ($(T)crti.o, $(T)crtn.o): Remove.
2199 (EXTRA_PARTS, EXTRA_MULTILIB_PARTS): Remove.
2200 * config/mep/t-mep (CRTSTUFF_CFLAGS): Remove.
2201 (EXTRA_MULTILIB_PARTS): Remove.
2202 * config/microblaze/crti.s: Move to ../libgcc/config/microblaze/crti.S.
2203 * config/microblaze/crtn.s: Move to ../libgcc/config/microblaze/crtn.S.
2204 * config/microblaze/t-microblaze (EXTRA_MULTILIB_PARTS,
2205 EXTRA_PARTS): Remove.
2206 ($(T)crti$(objext), $(T)crtn$(objext)): Remove.
2207 * config/mips/crti.asm: Move to ../libgcc/config/mips/crti.S.
2208 * config/mips/crtn.asm: Move to ../libgcc/config/mips/crtn.S.
2209 * config/mips/t-elf (CRTSTUFF_T_CFLAGS): Remove.
2210 ($(T)crti.o, $(T)crtn.o): Remove.
2211 (EXTRA_MULTILIB_PARTS): Remove.
2212 * config/mips/t-isa3264: Likewise.
2213 * config/mips/t-linux64 (EXTRA_MULTILIB_PARTS): Remove.
2214 * config/mips/t-r3900 (EXTRA_MULTILIB_PARTS): Remove.
2215 (CRTSTUFF_T_CFLAGS): Remove.
2216 * config/mips/t-sde (CRTSTUFF_T_CFLAGS): Remove.
2217 ($(T)crti.o, $(T)crtn.o): Remove.
2218 (EXTRA_MULTILIB_PARTS): Remove.
2219 * config/mips/t-sr71k (EXTRA_MULTILIB_PARTS, CRTSTUFF_T_CFLAGS):
2221 ($(T)crti.o, $(T)crtn.o): Remove.
2222 * config/mips/t-st (EXTRA_MULTILIB_PARTS): Remove.
2223 * config/mips/t-vr (CRTSTUFF_T_CFLAGS): Remove.
2224 (EXTRA_MULTILIB_PARTS): Remove.
2225 ($(T)crti.o, $(T)crtn.o): Remove.
2226 * config/mmix/crti.asm: Move to ../libgcc/config/crti.S.
2227 * config/mmix/crtn.asm: Move to ../libgcc/config/crtn.S.
2228 * config/mmix/t-mmix (CRTSTUFF_T_CFLAGS): Remove.
2229 * config/moxie/crti.asm, config/moxie/crtn.asm: Remove.
2230 * config/pa/stublib.c: Move to libgcc/config/pa.
2231 * config/pa/t-linux (CRTSTUFF_T_CFLAGS_S): Remove.
2232 * config/pa/t-linux64 (CRTSTUFF_T_CFLAGS_S): Remove.
2233 * config/pa/t-pa-hpux11 (LIBGCCSTUB_OBJS, stublib.c): Remove.
2234 (pthread_default_stacksize_np-stub.o, pthread_mutex_lock-stub.o)
2235 (pthread_mutex_unlock-stub.o, pthread_once-stub.o)
2236 ($(T)libgcc_stub.a): Remove.
2237 * config/pa/t-pa64 (LIBGCCSTUB_OBJS, stublib.c): Remove.
2238 (rfi-stub.o, dfi-stub.o, cxaf-stub.o, jvrc-stub.o)
2239 (pthread_default_stacksize_np-stub.o, pthread_mutex_lock-stub.o)
2240 (pthread_mutex_unlock-stub.o, pthread_once-stub.o)
2241 ($(T)libgcc_stub.a): Remove.
2242 * config/rs6000/eabi-cn.asm: Move to
2243 ../../../libgcc/config/rs6000/eabi-cn.S.
2244 * config/rs6000/eabi-ci.asm: Move to
2245 ../../../libgcc/config/rs6000/eabi-ci.S.
2246 * config/rs6000/sol-ci.asm: Move to
2247 ../../../libgcc/config/rs6000/sol-ci.S.
2248 * config/rs6000/sol-cn.asm: Move to
2249 ../../../libgcc/config/rs6000/sol-cn.S.
2250 * config/rs6000/t-lynx (EXTRA_MULTILIB_PARTS): Remove.
2251 (CRTSTUFF_T_CFLAGS, CRTSTUFF_T_CFLAGS_S): Remove.
2252 * config/rs6000/t-netbsd (CRTSTUFF_T_CFLAGS, CRTSTUFF_T_CFLAGS_S):
2254 (EXTRA_MULTILIB_PARTS): Remove.
2255 * config/rs6000/t-ppccomm (EXTRA_MULTILIB_PARTS): Remove.
2256 (ecrti.S, ecrtn.S, ncrti.S, ncrtn.S): Remove.
2257 ($(T)ecrti$(objext), $(T)ecrtn$(objext), $(T)ncrti$(objext),
2258 ($(T)ncrtn$(objext)): Remove.
2259 (CRTSTUFF_T_CFLAGS, CRTSTUFF_T_CFLAGS_S): Remove.
2260 * config/rs6000/t-vxworks (EXTRA_MULTILIB_PARTS): Remove.
2261 * config/rx/t-rx (EXTRA_MULTILIB_PARTS): Remove.
2262 * config/score/crti.asm: Move to ../libgcc/config/score/crti.S.
2263 * config/score/crtn.asm: Move to ../libgcc/config/score/crtn.S.
2264 * config/score/t-score-elf: Remove.
2265 * config/sh/crt1.asm: Move to ../libgcc/config/sh/crt1.S.
2266 * config/sh/crti.asm: Move to ../libgcc/config/sh/crti.S.
2267 * config/sh/crtn.asm: Move to ../libgcc/config/sh/crtn.S.
2268 * config/sh/lib1funcs-4-300.asm: Move to
2269 ../../../libgcc/config/sh/lib1funcs-4-300.S.
2270 * config/sh/lib1funcs-Os-4-200.asm: Move to
2271 ../libgcc/config/sh/lib1funcs-Os-4-200.S.
2272 * config/sh/t-elf: Remove.
2273 * config/sh/t-linux (EXTRA_MULTILIB_PARTS): Remove.
2274 * config/sh/t-linux64: Remove.
2275 * config/sh/t-netbsd (EXTRA_MULTILIB_PARTS): Remove.
2276 * config/sh/t-sh ($(T)crt1.o, $(T)crti.o, $(T)crtn.o): Remove.
2277 (IC_EXTRA_PARTS, OPT_EXTRA_PARTS, EXTRA_MULTILIB_PARTS): Remove.
2278 ($(T)ic_invalidate_array_4-100.o)
2279 ($(T)libic_invalidate_array_4-100.a)
2280 ($(T)ic_invalidate_array_4-200.o)
2281 ($(T)libic_invalidate_array_4-200.a, $(T)ic_invalidate_array_4a.o)
2282 ($(T)libic_invalidate_array_4a.a, $(T)sdivsi3_i4i-Os-4-200.o)
2283 ($(T)udivsi3_i4i-Os-4-200.o, $(T)unwind-dw2-Os-4-200.o)
2284 ($(T)libgcc-Os-4-200.a, $(T)div_table-4-300.o)
2285 ($(T)libgcc-4-300.a): Remove.
2286 * config/sh/t-superh: Remove.
2287 * config/sh/t-vxworks (EXTRA_MULTILIB_PARTS): Remove.
2288 * config/sparc/t-linux64 (CRTSTUFF_T_CFLAGS): Remove.
2289 * config/spu/cache.S: Move to ../libgcc/config/spu.
2290 * config/spu/cachemgr.c: Move to ../libgcc/config/spu.
2291 * config/spu/t-spu-elf (CRTSTUFF_T_CFLAGS): Remove.
2292 (EXTRA_MULTILIB_PARTS): Remove.
2293 ($(T)cachemgr.o, $(T)cachemgr_nonatomic.o, $(T)libgcc_%.a): Remove.
2294 ($(T)cache8k.o, $(T)cache16k.o, $(T)cache32k.o, $(T)cache32k.o)
2295 ($(T)cache64k.o, $(T)cache128k.o): Remove.
2296 * config/t-freebsd (CRTSTUFF_T_CFLAGS_S): Remove.
2297 * config/t-libc-ok: Remove.
2298 * config/t-linux (CRTSTUFF_T_CFLAGS_S): Remove.
2299 * config/t-lynx (CRTSTUFF_T_CFLAGS_S): Remove.
2300 * config/t-netbsd: Remove.
2301 * config/t-svr4 (CRTSTUFF_T_CFLAGS_S): Remove.
2302 * config/t-vxworks (EXTRA_MULTILIB_PARTS): Remove.
2303 * config/vms/t-vms (VMS_EXTRA_PARTS): Remove.
2304 ($(T)vcrt0.o, $(T)pcrt0.o): Remove.
2305 * config/vms/vms-ucrt0.c: Move to ../libgcc/config/vms.
2306 * config/xtensa/crti.asm: Move to ../libgcc/config/xtensa/crti.S.
2307 * config/xtensa/crtn.asm: Move to ../libgcc/config/xtensa/crtn.S.
2308 * config/xtensa/t-elf (CRTSTUFF_T_CFLAGS, CRTSTUFF_T_CFLAGS_S): Remove.
2309 (EXTRA_MULTILIB_PARTS): Remove.
2310 * config/xtensa/t-linux: Remove.
2311 * config/xtensa/t-xtensa ($(T)crti.o, $(T)crtn.o): Remove.
2313 2011-11-02 Uros Bizjak <ubizjak@gmail.com>
2315 * config/i386/i386.c (bdesc_args) [IX86_BUILTIN_CVTTPD2DQ256]: Use
2316 CODE_FOR_fix_truncv4dfv4si2, not CODE_FOR_fix_truncv4sfv4si2.
2318 2011-11-02 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
2320 PR translation/45116
2321 * Makefile.in (slibdir): Remove, don't export.
2322 (SHLIB_NM_FLAGS): Remove.
2323 (libgcc.mvars): Don't emit SHLIB_LINK, SHLIB_INSTALL,
2324 SHLIB_DLLDIR, SHLIB_EXT, SHLIB_MKMAP, SHLIB_MKMAP_OPTS,
2325 SHLIB_MAPFILES, SHLIB_NM_FLAGS.
2326 (DRIVER_DEFINES): Test SHLIB instead of SHLIB_LINK.
2327 (gcc.o): Pass SHLIB instead of SHLIB_LINK.
2328 (gccspec.o): Likewise.
2329 (installdirs): Don't create $(DESTDIR)$(slibdir).
2330 * configure.ac (slibdir): Remove.
2331 * configure: Regenerate.
2332 * libgcc-libsystem.ver: Move to ../libgcc/config.
2333 * mkmap-flat.awk, mkmap-symver.awk: Move to ../libgcc.
2334 * config/libgcc-glibc.ver: Move to ../libgcc/config.
2335 * config/t-libunwind (SHLIB_LC): Remove.
2336 * config/t-linux (SHLIB_MAPFILES): Remove.
2337 * config/t-slibgcc-dummy: Rename to config/t-slibgcc.
2338 * config/t-slibgcc-elf-ver: Remove.
2339 * config/t-slibgcc-libgcc, config/t-slibgcc-nolc-override: Move to
2341 * config/alpha/libgcc-alpha-ldbl.ver, config/alpha/t-linux: Move
2342 to ../libgcc/config/alpha.
2343 * config/alpha/t-vms (shlib_version, SHLIB_EXT, SHLIB_OBJS,
2344 SHLIB_NAME, SHLIB_MULTILIB, SHLIB_INSTALL, SHLIB_SYMVEC,
2345 SHLIB_SYMVECX2, SHLIB_LINK): Remove.
2346 * config/arm/libgcc-bpabi.ver: Move to ../libgcc/config/arm.
2347 * config/arm/t-bpabi (SHLIB_MAPFILES): Remove.
2348 * config/arm/t-netbsd (SHLIB_EXT, SHLIB_NAME, SHLIB_SONAME,
2349 SHLIB_OBJS, SHLIB_LINK, SHLIB_INSTALL): Remove.
2350 * config/arm/t-symbian (SHLIB_LC): Remove.
2351 * config/bfin/libgcc-bfin.ver: Move to
2352 ../libgcc/config/bfin/libgcc-glibc.ver.
2353 * config/bfin/t-bfin-linux (SHLIB_MAPFILES): Remove.
2354 * config/c6x/libgcc-c6xeabi.ver: Move to
2355 ../libgcc/config/c6x/libgcc-eabi.ver.
2356 * config/c6x/t-c6x-elf (SHLIB_MAPFILES): Remove.
2357 * config/cris/libgcc.ver: Move to
2358 ../libgcc/config/cris/libgcc-glibc.ver.
2359 * config/cris/t-linux (SHLIB_MAPFILES): Remove.
2360 * config/frv/libgcc-frv.ver: Move to ../libgcc/config/frv.
2361 * config/frv/t-linux (SHLIB_MAPFILES): Remove.
2362 * config/i386/darwin-libgcc.10.4.ver: Move to
2363 ../libgcc/config/i386/libgcc-darwin.10.4.ver.
2364 * config/i386/darwin-libgcc.10.5.ver: Move to
2365 ../libgcc/config/i386/libgcc-darwin.10.5.ver.
2366 * config/i386/libgcc-glibc.ver: Move to ../libgcc/config/i386.
2367 * config/i386/t-cygming (SHLIB_EXT, SHLIB_IMPLIB, SHLIB_SOVERSION,
2368 SHLIB_SONAME, SHLIB_MAP, SHLIB_OBJS, SHLIB_DIR, SHLIB_SLIBDIR_QUAL)
2369 SHLIB_PTHREAD_CFLAG, SHLIB_PTHREAD_LDFLAG, SHLIB_LINK,
2370 SHLIB_INSTALL, SHLIB_MKMAP, SHLIB_MKMAP_OPTS, SHLIB_MAPFILES): Remove.
2371 * config/i386/t-cygwin (SHLIB_LC, SHLIB_EH_EXTENSION,
2372 SHLIB_IMPLIB, SHLIB_SONAME, SHLIB_MKMAP_OPTS): Remove.
2373 * config/i386/t-dlldir, config/i386/t-dlldir-x: Move to
2374 ../libgcc/config/i386.
2375 * config/i386/t-dw2-eh, config/i386/t-sjlj-eh: Move to
2376 ../libgcc/config/i386.
2377 * config/i386/t-linux: Move to ../libgcc/config/i386.
2378 * config/i386/t-mingw-pthread: Move to ../libgcc/config/i386.
2379 * config/i386/t-mingw-w32 (SHLIB_LC): Remove.
2380 * config/i386/t-mingw-w64: Likewise.
2381 * config/i386/t-mingw32: Remove.
2382 * config/ia64/libgcc-glibc.ver, config/ia64/libgcc-ia64.ver: Move
2383 to ../libgcc/config/ia64.
2384 * config/ia64/t-glibc: Remove.
2385 * config/ia64/t-hpux (SHLIB_EXT, SHLIB_LINK, SHLIB_INSTALL): Remove.
2386 * config/ia64/t-ia64 (SHLIB_MAPFILES): Remove.
2387 * config/ia64/t-vms (shlib_version, SHLIB_EXT, SHLIB_OBJS,
2388 SHLIB_NAME, SHLIB_MULTILIB, SHLIB_INSTALL, SHLIB_LINK): Remove.
2389 * config/ia64/vms_symvec_libgcc_s.opt: Remove.
2390 * config/m32r/libgcc-glibc.ver: Move to ../libgcc/config/m32r.
2391 * config/m32r/t-linux (SHLIB_MAPFILES): Remove.
2392 * config/m68k/t-slibgcc-elf-ver: Move to ../libgcc/config/m68k.
2393 * config/mips/t-libgcc-mips16 (SHLIB_MAPFILES): Remove.
2394 * config/pa/t-hpux-shlib: Move to ../libgcc/config/pa/t-slibgcc-hpux.
2395 * config/pa/t-slibgcc-dwarf-ver, config/pa/t-slibgcc-sjsj-ver:
2396 Move to ../libgcc/config/pa.
2397 * config/rs6000/darwin-libgcc.10.4.ver: Move to
2398 ../libgcc/config/rs6000/libgcc-darwin.10.4.ver.
2399 * config/rs6000/darwin-libgcc.10.5.ver: Move to
2400 ../libgcc/config/rs6000/libgcc-darwin.10.5.ver.
2401 * config/rs6000/t-aix43 (SHLIB_EXT, SHLIB_LINK, SHLIB_INSTALL,
2402 SHLIB_LIBS, SHLIB_MKMAP, SHLIB_NM_FLAGS, AR_FLAGS_FOR_TARGET): Remove.
2403 * config/rs6000/t-aix52: Likewise.
2404 * config/sh/libgcc-excl.ver, config/sh/libgcc-glibc.ver: Move to
2405 ../libgcc/config/sh.
2406 * config/sparc/libgcc-sparc-glibc.ver: Move to
2407 ../libgcc/config/sparc/libgcc-glibc.ver.
2408 * config/sparc/t-linux: Move to ../libgcc/config/sparc.
2409 * config/xtensa/t-linux (SHLIB_MAPFILES): Remove.
2410 * config/xtensa/libgcc-xtensa.ver: Move to
2411 ../libgcc/config/xtensa/libgcc-glibc.ver.
2412 * config.gcc (*-*-freebsd*): Replace t-slibgcc-elf-ver with
2413 t-slibgcc in tmake_file.
2414 Remove t-slibgcc-nolc-override for *-*-freebsd[34],
2415 *-*-freebsd[34].* with pthreads.
2416 (*-*-linux*, frv-*-*linux*, *-*-kfreebsd*-gnu,
2417 *-*-knetbsd*-gnu, *-*-gnu*, *-*-kopensolaris*-gnu): Replace
2418 t-slibgcc-elf-ver with t-slibgcc in tmake_file.
2419 (*-*-netbsd*): Likewise.
2420 (*-*-solaris2*): Replace t-slibgcc-dummy with t-slibgcc in tmake_file.
2421 (*-*-*vms*): Add t-slibgcc to tmake_file.
2422 (alpha*-*-linux*): Remove alpha/t-linux from tmake_file.
2423 (alpha*-dec-osf5.1*): Replace t-slibgcc-dummy with t-slibgcc in
2425 (arm*-*-linux*): Remove t-slibgcc-libgcc from tmake_file for
2427 (bfin*-linux-uclibc*): Replace t-slibgcc-dummy with t-slibgcc in
2429 (crisv32-*-linux*, cris-*-linux*): Likewise.
2430 (hppa*-*-linux*): Remove t-slibgcc-libgcc, pa/t-slibgcc-sjlj-ver,
2431 pa/t-slibgcc-dwarf-ver from tmake_file.
2432 (hppa[12]*-*-hpux10*): Replace pa/t-hpux-shlib with t-slibgcc in
2434 Remove pa/t-slibgcc-sjlj-ver, pa/t-slibgcc-dwarf-ver from tmake_file.
2435 (hppa*64*-*-hpux11*): Likewise.
2436 (hppa[12]*-*-hpux11*): Likewise.
2437 (i[34567]86-*-darwin*): Replace t-slibgcc-dummy in t-slibgcc in
2439 (x86_64-*-darwin*): Likewise.
2440 (i[34567]86-*-cygwin*): Remove tmake_eh_file, tmake_dlldir_file.
2441 Add t-slibgcc to tmake_file.
2442 (i[34567]86-*-mingw*, x86_64-*-mingw*): Likewise.
2443 Remove i386/t-mingw32 from tmake_file unless x86_64-w64-*,
2445 Remove i386/t-mingw-pthread from tmake_file.
2446 (ia64*-*-linux*): Remove ia64/t-glibc from tmake_file.
2447 (ia64*-*-hpux*): Add t-slibgcc to tmake_file.
2448 (ia64-hp-*vms*): Likewise.
2449 (m32r-*-linux*): Replace t-slibgcc-elf-ver with t-slibgcc in
2451 (m32rle-*-linux*): Likewise.
2452 (m68k-*-linux*): Remove m68k/t-slibgcc-elf-ver from tmake_file.
2453 (microblaze*-linux*): Remove t-slibgcc-elf-ver,
2454 t-slibgcc-nolc-override from tmake_file.
2455 (mips-sgi-irix6.5*): Replace t-slibgcc-dummy with t-slibgcc in
2457 (powerpc-*-darwin*): Likewise.
2458 (powerpc64-*-darwin*): Likewise.
2459 (powerpc-*-freebsd*): Remove t-slibgcc-libgcc from tmake_file.
2460 (powerpc-*-linux*, powerpc64-*-linux*): Likewise.
2461 (rs6000-ibm-aix4.[3456789]*, powerpc-ibm-aix4.[3456789]*): Add
2462 t-slibgcc to tmake_file.
2463 (rs6000-ibm-aix5.1.*, powerpc-ibm-aix5.1.*): Likewise.
2464 (rs6000-ibm-aix5.2.*, powerpc-ibm-aix5.2.*): Likewise.
2465 (rs6000-ibm-aix5.3.*, powerpc-ibm-aix5.3.*): Likewise.
2466 (rs6000-ibm-aix[6789].*, powerpc-ibm-aix[6789].*): Likewise.
2467 (sparc-*-linux*): Remove sparc/t-linux from tmake_file.
2468 (sparc64-*-linux*): Likewise.
2469 (tic6x-*-uclinux): Replace t-slibgcc-elf-ver with t-slibgcc in
2471 (i[34567]86-*-linux*, x86_64-*-linux*, i[34567]86-*-kfreebsd*-gnu,
2472 x86_64-*-kfreebsd*-gnu, i[34567]86-*-gnu*): Remove i386/t-linux
2475 2011-11-02 Richard Guenther <rguenther@suse.de>
2477 PR tree-optimization/50902
2478 * tree-vect-stmts.c (vectorizable_load): Properly convert
2479 an invariant initializer element.
2481 2010-11-02 Richard Guenther <rguenther@suse.de>
2483 PR tree-optimization/50890
2484 * gimple.h (gimple_fold_call): Remove.
2485 * gimple-fold.c (fold_stmt_1): Move all call related code to ...
2486 (gimple_fold_call): ... here. Make static. Update the
2487 cannot-inline flag on direct calls.
2488 * ipa-inline.c (early_inliner): Copy the cannot-inline flag
2489 from the statements to the edges.
2491 2011-11-01 Ian Lance Taylor <iant@google.com>
2493 * godump.c (struct macro_hash_value): Define.
2494 (macro_hash_hashval): New static function.
2495 (macro_hash_eq, macro_hash_del): New static functions.
2496 (go_define): Use macro_hash_value to store values in macro_hash.
2497 Replace an old value on a redefinition. Don't print anything to
2499 (go_undef): Delete the entry from the hash table.
2500 (go_output_typedef): For an enum, use macro_hash_value, and don't
2501 print anything to go_dump_file.
2502 (go_print_macro): New static function.
2503 (go_finish): Traverse macro_hash with go_print_macro.
2504 (dump_go_spec_init): Update macro_hash creation for macro_hash_value.
2506 2011-11-02 Alan Modra <amodra@gmail.com>
2508 * config/rs6000/rs6000.c (rs6000_code_end): Declare ATTRIBUTE_UNUSED.
2510 2011-11-01 Paolo Carlini <paolo.carlini@oracle.com>
2513 * doc/invoke.texi: Document -Wzero-as-null-pointer-constant.
2515 2011-11-01 Andrew Stubbs <ams@codesourcery.com>
2517 * config/arm/bpabi.h (BE8_LINK_SPEC): Recognize generic-armv7 tuning.
2519 2011-11-01 Uros Bizjak <ubizjak@gmail.com>
2521 * config/i386/i386.md (splitters for int-float conversion): Use
2522 SUBREG_REG on SUBREGs in splitter constraints.
2524 2011-11-01 Jakub Jelinek <jakub@redhat.com>
2526 * config/i386/i386-protos.h (ix86_expand_adjust_ufix_to_sfix_si): New
2528 * config/i386/i386.c (ix86_expand_adjust_ufix_to_sfix_si): New
2530 * config/i386/sse.md (fixuns_trunc<mode><sseintvecmodelower>2): Use it.
2531 (ssepackfltmode): New mode attr.
2532 (vec_pack_ufix_trunc_<mode>): New expander.
2534 2011-11-01 Uros Bizjak <ubizjak@gmail.com>
2537 * config/i386/i386.md (floatsi<mode>2_vector_sse_with_temp splitter):
2538 Compare <ssevecmode>mode to V4SFmode, not V4SImode.
2540 2011-11-01 Peter Bergner <bergner@vnet.ibm.com>
2542 * config.gcc (powerpc*-*-linux*): Add powerpc*-*-linux*ppc476* variant.
2543 * config/rs6000/476.h: New file.
2544 * config/rs6000/476.opt: Likewise.
2545 * config/rs6000/rs6000.h (TARGET_LINK_STACK): New define.
2546 (SET_TARGET_LINK_STACK): Likewise.
2547 (TARGET_ASM_CODE_END): Define.
2548 * config/rs6000/rs6000.c (rs6000_option_override_internal): Enable
2549 TARGET_LINK_STACK for -mtune=476 and -mtune=476fp.
2550 (rs6000_legitimize_tls_address): Emit the link stack preserving GOT
2551 code if TARGET_LINK_STACK.
2552 (rs6000_emit_load_toc_table): Likewise.
2553 (output_function_profiler): Likewise
2554 (macho_branch_islands): Likewise
2555 (machopic_output_stub): Likewise
2556 (get_ppc476_thunk_name): New function.
2557 (rs6000_code_end): Likewise.
2558 * config/rs6000/rs6000.md (load_toc_v4_PIC_1, load_toc_v4_PIC_1b):
2559 Convert to a define_expand.
2560 (load_toc_v4_PIC_1_normal): New define_insn.
2561 (load_toc_v4_PIC_1_476): Likewise.
2562 (load_toc_v4_PIC_1b_normal): Likewise.
2563 (load_toc_v4_PIC_1b_476): Likewise.
2565 2011-11-01 Georg-Johann Lay <avr@gjlay.de>
2568 * config/avr/avr.opt (-mbranch-cost=): New option.
2569 * config/avr/avr.h (BRANCH_COST): Define to avr_branch_cost.
2570 * config/avr/avr.c (avr_rtx_costs_1): Adjust [U]DIV/[U]MOD costs.
2571 * config/avr/avr.md (*addqi3.lt0, *addhi3.lt0, *addsi3.lt0): New insns.
2572 (*addhi3_zero_extend1): Remove % in constraint of operand 1.
2573 (*addhi3.sign_extend1, *subhi3.sign_extend2): New insns.
2575 2011-11-01 Tom de Vries <tom@codesourcery.com>
2577 PR tree-optimization/50908
2578 * tree-ssa-tail-merge.c (update_vuses): Now that edges are removed
2579 before update_vuses, test for 1 predecessor rather than two.
2580 (delete_block_update_dominator_info): New function, part of it factored
2582 (replace_block_by): Use delete_block_update_dominator_info. Call
2583 update_vuses after deleting bb1 and updating dominator info, instead of
2586 2011-11-01 David S. Miller <davem@davemloft.net>
2588 * config/sparc/sparc.c (vector_init_faligndata): New function.
2589 (sparc_expand_vector_init): Use it for V4HImode on VIS1.
2591 * config/sparc/sparc.c (sparc_expand_vcond): New function.
2592 * config/sparc/sparc-protos.h (sparc_expand_vcond): Declare it.
2593 * config/sparc/sparc.md (vcond<mode><mode>): New VIS3 expander.
2594 (vconduv8qiv8qi): Likewise.
2596 2011-11-01 Alexandre Oliva <aoliva@redhat.com>
2599 * cselib.c (cfa_base_preserved_regno): Initialize.
2600 (cselib_expand_value_rtx_1): Don't expand it.
2601 * var-tracking.c (vt_expand_var_loc_chain): Initialize depth.
2602 Check it's only zero if result is NULL.
2604 2011-11-01 Jakub Jelinek <jakub@redhat.com>
2606 * config/i386/sse.md (fixuns_trunc<mode><sseintvecmodelower>2): New
2609 * config/i386/sse.md (sseintvecmode): Remove duplicate modes.
2610 (sseintvecmodelower): New mode iterator.
2611 (floatv8siv8sf2, floatunsv4siv4sf2): Macroize into...
2612 (float<sseintvecmodelower><mode>2): ... this using VF1 iterator.
2613 (floatunsv4siv4sf2): Macroize into...
2614 (floatuns<sseintvecmodelower><mode>2): ... this using VF1 iterator.
2616 2011-10-31 David S. Miller <davem@davemloft.net>
2618 * config/sparc/sparc.md (cmask patterns): Allow zero operand.
2620 * dwarf2out.c (cached_next_real_insn): New.
2621 (dwarf2out_end_epilogue): Set it to NULL_RTX.
2622 (dwarf2out_var_location): Remove cached_next_real_insn local static.
2624 2011-10-31 Richard Henderson <rth@redhat.com>
2626 * config/i386/sse.md (floatv8siv8sf2): Rename from avx_cvtdq2ps256.
2627 (floatv4siv4sf2): Rename from sse2_cvtdq2ps.
2628 (floatunsv4siv4sf2): Rename from sse2_cvtudq2ps.
2629 (fix_truncv8sfv8si2): Rename from avx_cvttps2dq256.
2630 (fix_truncv4sfv4si2): Rename from sse2_cvttps2dq.
2631 (floatv4siv4df2): Rename from avx_cvtdq2pd256.
2632 (fix_truncv4dfv4si2): Rename from avx_cvttpd2dq256.
2633 (vec_unpacku_float_hi_v8si): Update for insn pattern name changes.
2634 * config/i386/i386.md (splitters for int-float conversion): Likewise.
2635 * config/i386/i386.c (ix86_split_convert_uns_si_sse): Likewise.
2636 (bdesc_args): Likewise.
2637 (enum ix86_builtins) [IX86_BUILTIN_CVTUDQ2PS]: Remove.
2638 (ix86_vectorize_builtin_conversion): Remove.
2639 (TARGET_VECTORIZE_BUILTIN_CONVERSION): Remove.
2641 2011-10-31 Martin Jambor <mjambor@suse.cz>
2643 * ipa-prop.c (mark_modified): Moved up in the file.
2644 (is_parm_modified_before_call): Renamed to
2645 is_parm_modified_before_stmt, moved up in the file.
2646 (load_from_unmodified_param): New function.
2647 (compute_complex_assign_jump_func): Also attempt to create pass
2648 through jump functions for values loaded from (addressable)
2651 2011-10-31 Jakub Jelinek <jakub@redhat.com>
2653 * tree-vect-stmts.c (vectorizable_shift): If op1 is vect_external_def
2654 in a loop and has different type from op0, cast it to op0's type
2655 before the loop first. For slp give up. Don't crash if op1_vectype
2658 2011-10-31 Paul Brook <paul@codesourcery.com>
2660 * cgraphunit.c: Don't mark clones as static constructors.
2662 2011-10-31 David Edelsohn <dje.gcc@gmail.com>
2664 * gcc-ar: Do not include stdio.h.
2666 2011-10-31 Diego Novillo <dnovillo@google.com>
2668 * tree-streamer-out.c (pack_ts_base_value_fields): Emit
2670 * tree-streamer-in.c (unpack_ts_base_value_fields): Read
2673 2011-10-30 David S. Miller <davem@davemloft.net>
2675 * config/sparc/sparc.c (vector_init_bshuffle): New function.
2676 (vector_init_fpmerge): New function.
2677 (sparc_expand_vector_init): Use them to improve non-const cases.
2679 * dwarf2out.c (dwarf2out_var_location): When processing several
2680 consecutive location notes, cache the result of next_real_insn().
2682 2011-10-30 Uros Bizjak <ubizjak@gmail.com>
2684 * config/i386/i386.md (avx2_vec_dup<mode>): Macroize insn from
2685 avx2_vec_dup{v8sf,v4sf} using VF1 mode iterator.
2686 (vec_dupv4sf): Remove expander.
2687 (vec_dupv4sf): Merge from *vec_dupv4sf and *vec_dupv4sf_avx.
2688 (vec_dupv2df): Remove expander.
2689 (vec_dupv2df): Merge from *vec_dupv2df and *vec_dupv2df_sse3.
2690 (*vec_concatv2df): Merge *vec_concatv2df_sse3.
2691 (*vec_dupv4si): Merge *vec_dupv4si_avx.
2692 (*vec_dupv2di): Merge *vec_dupv2di_sse3.
2694 2011-10-30 Dmitry Plotnikov <dplotnikov@ispras.ru>
2696 * tree-cfg.c (verify_gimple_assign_unary): Allow vector conversions.
2697 * optabs.c (supportable_convert_operation): New function.
2698 * optabs.h (supportable_convert_operation): New prototype.
2699 * tree-vect-stmts.c (vectorizable_conversion): Change condition and
2700 behavior for NONE modifier case.
2701 * tree.h (VECTOR_INTEGER_TYPE_P): New macro.
2703 2011-10-30 Jakub Jelinek <jakub@redhat.com>
2705 * tree-vectorizer.h (NUM_PATTERNS): Bump to 9.
2706 * tree-vect-patterns.c (vect_recog_vector_vector_shift_pattern): New
2708 (vect_vect_recog_func_ptrs): Add it.
2710 2011-10-30 David S. Miller <davem@davemloft.net>
2712 * reorg.c (label_before_next_insn): New function.
2713 (relax_delay_slots): Use it instead of prev_label.
2714 * rtl.h (prev_label): Delete declaration.
2715 * emit-rtl.c (prev_label): Remove.
2717 2011-10-30 Revital Eres <revital.eres@linaro.org>
2719 * modulo-sched.c (generate_prolog_epilog): Mark prolog and epilog
2720 as BB_DISABLE_SCHEDULE.
2721 (mark_loop_unsched): New function.
2722 (sms_schedule): Call it.
2724 2011-10-29 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
2727 * config/pa/protos.h (attr_length_save_restore_dltp): Delete.
2728 (cmpib_comparison_operator): Likewise.
2729 (following_cal, output_and, output_ior, output_move_double,
2730 output_fp_move_double, output_block_move, output_block_clear,
2731 output_cbranch, output_lbranch, output_bb, output_bvb, output_dbra,
2732 output_movb, output_parallel_movb, output_parallel_addb, output_call,
2733 output_indirect_call, output_millicode_call, output_mul_insn,
2734 output_div_insn, output_mod_insn, singlemove_string,
2735 output_arg_descriptor, output_global_address, print_operand,
2736 legitimize_pic_address, hppa_encode_label, symbolic_expression_p,
2737 fmpyaddoperands, fmpysuboperands, emit_bcond_fp, emit_move_sequence,
2738 emit_hpdiv_const, is_function_label_plus_const, jump_in_call_delay,
2739 hppa_fpstore_bypass_p, attr_length_millicode_call, attr_length_call,
2740 attr_length_indirect_call, return_addr_rtx, function_arg_padding,
2741 insn_refs_are_delayed, get_deferred_plabel, ldil_cint_p, zdepi_cint_p,
2742 output_ascii, compute_frame_size, and_mask_p, cint_ok_for_move,
2743 hppa_expand_prologue, hppa_expand_epilogue, ior_mask_p,
2744 compute_zdepdi_operands, output_64bit_and, output_64bit_ior,
2745 reloc_needed, magic_milli, shadd_constant_p): Consistently prefix
2746 exported functions and variables with "pa_".
2747 * config/pa/predicates.md: Likewise.
2748 * config/pa/pa64-hpux.h: likewise.
2749 * config/pa/som.h: Likewise.
2750 * config/pa/elf.h: Likewise.
2751 * config/pa/pa64-linux.h: Likewise.
2752 * config/pa/pa.md: Likewise.
2753 * config/pa/pa.c: Likewise.
2754 * config/pa/pa-linux.h: Likewise.
2755 * config/pa/pa.h: Likewise.
2756 * config/pa/constraints.md: Likewise.
2758 2011-10-29 Uros Bizjak <ubizjak@gmail.com>
2760 * config/i386/i386.md (xop_sha<mode>3): Rename from xop_ashl<mode>3.
2762 (xop_shl<mode>3): Rename from xop_lshl<mode>3. Update all uses.
2763 * config/i386/i386.c: Update all uses.
2765 2011-10-29 Uros Bizjak <ubizjak@gmail.com>
2767 * config/i386/i386.md (lshlv16qi3): Remove expander.
2768 (lshrv16qi3): New expander.
2769 (<shift_insn>v16qi3): Macroize expander from ashrv16qi3 and lshrv16qi3
2770 using any_shiftrt code iterator. Cleanup.
2771 (ashlv16qi3): Cleanup.
2774 2011-10-29 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
2777 * config/pa/pa.c (emit_move_sequence): Legitimize TLS symbol
2779 (pa_legitimate_constant_p): Return false for TLS_MODEL_GLOBAL_DYNAMIC
2780 and TLS_MODEL_LOCAL_DYNAMIC symbol references.
2782 2011-10-29 Georg-Johann Lay <avr@gjlay.de>
2785 * config/avr/avr.opt (-maccumulate-args): New option.
2786 * config/avr/avr.h (STARTING_FRAME_OFFSET): Redefine to
2787 avr_starting_frame_offset.
2788 (ACCUMULATE_OUTGOING_ARGS): Define to avr_accumulate_outgoing_args.
2789 * config/avr/avr.md (UNSPECV_WRITE_SP_IRQ_ON): Remove.
2790 (UNSPECV_WRITE_SP_IRQ_OFF): Remove.
2791 (UNSPECV_WRITE_SP): New constant.
2792 (*addhi3_sp_R): Rewrite to...
2793 (*addhi3_sp): ...this new insn.
2794 (movhi_sp_r_irq_off, movhi_sp_r_irq_on): Combine to...
2795 (movhi_sp_r): ...this new insn.
2796 * config/avr/avr-protos.h (avr_accumulate_outgoing_args): New.
2797 (avr_starting_frame_offset): New.
2798 * config/avr/avr.c (avr_accumulate_outgoing_args): New function.
2799 (avr_starting_frame_offset): New function.
2800 (avr_outgoing_args_size): New static function.
2801 (avr_initial_elimination_offset): Use it.
2802 (avr_simple_epilogue): Use it.
2803 (avr_asm_function_end_prologue): Use it.
2804 (expand_epilogue): Use it.
2805 (expand_prologue): Use it. Break out code to...
2806 (avr_prologue_setup_frame): ...this new static function.
2807 (avr_can_eliminate): Allow eliminating to frame pointer if there
2809 (avr_frame_pointer_required_p): Use frame pointer if target has a
2811 * config/avr/constraints.md (R): Remove.
2812 (Csp): New constraint.
2813 * config/avr/predicates.md (avr_sp_immediate_operand): Use it.
2815 2011-10-29 Andi Kleen <ak@linux.intel.com>
2817 * gcc-ar.c (target_machine): Add.
2819 2011-10-29 Anatoly Sokolov <aesok@post.ru>
2821 * config/cris/cris.c (reg_ok_for_base_p, reg_ok_for_index_p,
2822 cris_constant_index_p, cris_base_p, cris_index_p,
2823 cris_base_or_autoincr_p, cris_bdap_index_p, cris_biap_index_p,
2824 cris_legitimate_address_p): New functions.
2825 (TARGET_LEGITIMATE_ADDRESS_P): Define.
2826 (cris_pic_symbol_type, cris_valid_pic_const): Change arguments type
2827 from rtx to const_rtx.
2828 (cris_print_operand_address, cris_address_cost,
2829 cris_side_effect_mode_ok): Use
2830 cris_constant_index_p, cris_base_p, cris_base_or_autoincr_p,
2831 cris_biap_index_p and cris_bdap_index_p.
2832 * config/cris/cris.h (CONSTANT_INDEX_P, BASE_P, BASE_OR_AUTOINCR_P,
2833 BDAP_INDEX_P, BIAP_INDEX_P, GO_IF_LEGITIMATE_ADDRESS,
2834 REG_OK_FOR_BASE_P, REG_OK_FOR_INDEX_P): Remove.
2835 (EXTRA_CONSTRAINT_Q, EXTRA_CONSTRAINT_R, EXTRA_CONSTRAINT_T): Use
2836 cris_constant_index_p, cris_base_p, cris_base_or_autoincr_p,
2837 cris_biap_index_p and cris_bdap_index_p.
2838 * config/cris/cris.md (moversideqi movemsideqi peephole2): Use
2840 * config/cris/cris-protos.h (cris_constant_index_p, cris_base_p,
2841 cris_base_or_autoincr_p, cris_bdap_index_p, cris_biap_index_p): New
2843 (cris_pic_symbol_type, cris_valid_pic_const): Update prototype.
2845 2011-10-21 Andi Kleen <ak@linux.intel.com>
2847 * ggc-page (PAGE_ALIGN): Add.
2848 (alloc_page, ggc_pch_total_size, ggc_pch_this_base, ggc_pch_read):
2849 Replace ROUND_UP with PAGE_ALIGN.
2851 2011-10-20 Andi Kleen <ak@linux.intel.com>
2853 * ggc-page (alloc_anon): Add check argument.
2854 (alloc_page): Add fallback to 1 page allocation.
2855 Adjust alloc_anon calls to new argument.
2857 2011-10-18 Andi Kleen <ak@linux.intel.com>
2859 * ggc-page (release_pages): First free large continuous
2860 chunks in the madvise path.
2862 2011-10-18 Andi Kleen <ak@linux.intel.com>
2864 * ggc-page.c (alloc_pages): Always round up entry_size.
2866 2011-10-19 Andi Kleen <ak@linux.intel.com>
2868 * Makefile.in (MOSTLYCLEANFILES): Add gcc-ar/nm/ranlib.
2869 (native): Add gcc-ar, gcc-nm, gcc-ranlib.
2870 (AR_LIBS, gcc-ar, gcc-ar.o, gcc-ranlib, gcc-ranlib.o,
2871 gcc-nm, gcc-nm.o, gcc-ranlib.c, gcc-nm.c): Add.
2872 (install): Depend on install-gcc-ar.
2873 (install-gcc-ar): Add.
2874 (uninstall): Uninstall gcc-ar, gcc-nm, gcc-ranlib.
2875 * gcc-ar.c: Add new file.
2877 2011-10-28 Pat Haugen <pthaugen@us.ibm.com>
2879 * config/rs6000/rs6000.md (define_attr "type"): Add vecdouble.
2880 * config/rs6000/vsx.md (VStype_simple, VStype_mul): Use vecdouble
2882 (VStype_div): Use vector types for V2DF/V4SF.
2883 (VStype_sqrt): Use *sqrt types.
2884 (VS_spdp_type): Change type to vecdouble.
2885 (*vsx_fmav2df4, *vsx_nfmsv2df4, vsx_xvcvdpsxws, vsx_xvcvdpuxws,
2886 vsx_xvcvuxdsp, vsx_xvcvsxwdp, vsx_xvcvuxwdp, vsx_xvcvspsxds,
2887 vsx_xvcvspuxds): Likewise.
2888 (*vsx_fms<mode>4): Set type via <VStype_mul>.
2889 (*vsx_eq_<mode>_p, *vsx_gt_<mode>_p, *vsx_ge_<mode>_p): Set type via
2891 * config/rs6000/power7.md (power7-vecstore): Correct VSU pipe.
2892 (power7-fpcompare, power7-sdiv, power7-ddiv, power7-sqrt,
2893 power7-dsqrt): Correct insn latency.
2894 (power7-vecsimple): Add veccmp type and correct dispatch/VSU values.
2895 (power7-veccmp): Delete.
2896 (power7-vecfloat): Correct latency/dispatch/VSU values.
2897 (define_bypass "power7-vecfloat"): Correct latency and types.
2898 (power7-veccomplex, power7-vecperm): Correct dispatch/VSU values.
2899 (power7-vecdouble, power7-vecfdiv, power7-vecdiv): New.
2901 2011-10-28 Uros Bizjak <ubizjak@gmail.com>
2903 * config/i386/i386.md (shift_insn): Rename code attribute from
2904 shiftrt_insn. Also handle ashift RTX.
2905 (shift): Rename code attribute from shiftrt. Also handle ashift RTX.
2906 (vshift): New code attribute.
2907 (<shift_insn>*): Rename from <shiftrt_insn>*. Update asm templates.
2908 (any_lshift): Move and rename code iterator from ...
2909 * config/i386/sse.md (lshift): ... here.
2910 (lshift_insn): Remove code attribute.
2911 (lshift): Remove code attribute.
2912 (vlshr<mode>3): Use lshiftrt RTX.
2913 (vashr<mode>3, ashrv16qi3, ashrv2di3): Use ashiftrt RTX.
2914 (vashl<mode>3, ashlv16qi3): Use ashift RTX.
2915 (avx2_<lshift>v<mode>): Rename from avx2_<shift_insn>v<mode>. Use
2916 any_lshift code iterator. Update asm template.
2917 (<shift_insn><mode>3): Macroize insn from lshr<mode>3 and ashl<mode>3
2918 usign any_lshift code iterator.
2919 * config/i386/mmx.md (mmx_<shift_insn><mode>3): Macroize insn from
2920 mmx_lshr<mode>3 and mmx_ashl<mode>3 usign any_lshift code iterator.
2921 * config/i386/i386.c (bdesc_args) <__builtin_ia32_psll>: Update.
2923 2011-10-28 Georg-Johann Lay <avr@gjlay.de>
2926 * config/avr/avr.md (parityhi2): Expand allowing pseudos.
2927 (*parityhi2): New pre-reload insn-and-split to map 16-bit parity
2929 (*parityqihi2): Same for 8-bit parity.
2931 2011-10-28 Julian Brown <julian@codesourcery.com>
2933 PR rtl-optimization/47918
2934 * reload1.c (set_initial_label_offsets): Use initial offsets
2935 for labels on the nonlocal_goto_handler_labels chain.
2937 2011-10-28 Iain Sandoe <iains@gcc.gnu.org>
2939 * config/rs6000/t-darwin (LIB2FUNCS_STATIC_EXTRA):
2940 Move darwin-fpsave.asm from here to ... LIB2FUNCS_EXTRA.
2941 (LIB2FUNCS_EXTRA): Add darwin-gpsave.asm.
2942 (TARGET_LIBGCC2_CFLAGS): Ensure that fPIC and -pipe are inherited from
2944 * config/rs6000/darwin.h (FP_SAVE_INLINE): Adjust to enable.
2945 (GP_SAVE_INLINE): Likewise.
2946 (SAVE_FP_PREFIX, SAVE_FP_SUFFIX, RESTORE_FP_PREFIX,
2947 RESTORE_FP_SUFFIX): Set to empty strings.
2948 * config/rs6000/rs6000.c (rs6000_savres_strategy): Implement for Darwin.
2949 (debug_stack_info): Print savres_strategy.
2950 (rs6000_savres_routine_name): Implement for Darwin.
2951 (rs6000_make_savres_rtx): Adjust used register for Darwin.
2952 (rs6000_emit_prologue): Implement out-of-line saves for Darwin.
2953 (rs6000_output_function_prologue): Don't emit .extern for Mach-O.
2954 (rs6000_emit_epilogue): Implement out-of-line saves for Darwin.
2955 * config/rs6000/darwin-gpsave.asm: New file.
2957 2011-10-28 Jakub Jelinek <jakub@redhat.com>
2959 * config/i386/sse.md (VI4SD_AVX2): Removed.
2960 (VI48_AVX2, VI128_128, VI48_128, VI48_256): New mode iterators.
2961 (vashl<mode>3): Use VI12_128 iterator instead of VI124_128.
2962 Add another expander using VI48_128 iterator for
2963 TARGET_AVX2 || TARGET_XOP and another using VI48_256 iterator
2965 (vlshr<mode>3): Likewise. Change register_operand predicate to
2966 nonimmediate_operand on last operand in the VI12_128 expander.
2967 (vashr<mode>3): Use VI128_128 iterator instead of VI124_128.
2968 (vashrv4si3, vashrv8si3): New expanders.
2969 (avx2_ashrvv8si, avx2_ashrvv4si, avx2_<lshift>vv8si,
2970 avx2_<lshift>vv2di): Removed.
2971 (avx2_ashrv<mode>): New insn with VI4_AVX2 iterator.
2972 (avx2_<lshift>v<mode>): Macroize using VI48_AVX2
2973 iterator. Simplify pattern.
2975 2010-10-28 Richard Guenther <rguenther@suse.de>
2978 * lto-wrapper.c (get_options_from_collect_gcc_options):
2979 Properly count arguments.
2980 (run_gcc): Use an obstack to collect argv, properly separate
2981 switches and their arguments.
2983 2011-10-28 Jakub Jelinek <jakub@redhat.com>
2985 * tree-vect-stmts.c (vectorizable_shift): Give up if op1 has different
2986 vector mode from vectype's mode.
2988 2011-10-28 Chung-Lin Tang <cltang@codesourcery.com>
2990 PR rtl-optimization/49720
2991 * simplify-rtx.c (simplify_relational_operation_1): Detect
2992 infinite recursion condition in "(eq/ne (plus x cst1) cst2)
2993 simplifies to (eq/ne x (cst2 - cst1))" case.
2995 2011-10-27 David S. Miller <davem@davemloft.net>
2997 * config/sparc/sparc.md (snedi_special): Only match when not VIS3.
2998 (*snedi_zero): Likewise.
2999 (*snedi_zero_trunc): Likewise.
3000 (snedi_special_vis3): New expander.
3001 (*snedi_zero_vis3): New insn.
3002 (*snedi_zero_trunc_vis3): Likewise.
3003 (*sltu_insn_vis3): Likewise.
3004 (*sltu_insn_vis3_trunc): Likewise.
3006 (*addxc_trunc_sp64_vis3): Likewise.
3007 * config/sparc/sparc.c (emit_scc_insn): When VIS3 use the
3008 gen_snedi_special_vis3 expander, and try GTU/LTU addx based
3009 sequences on DImode values.
3011 * config/sparc/sparc.md (64-bit vector moves): Use 'e' not 'f'
3014 * regcprop.c (copyprop_hardreg_forward_1): Reject the
3015 transformation when we narrow the mode on big endian.
3017 2011-10-27 Jakub Jelinek <jakub@redhat.com>
3019 * config/i386/sse.md (avx_cvtpd2dq256_2, avx_cvttpd2dq256_2,
3020 vec_pack_sfix_trunc_v4df, vec_pack_sfix_v4df): New expanders.
3021 (*avx_cvtpd2dq256_2, *avx_cvttpd2dq256_2): New insns.
3023 * config/i386/i386.c (ix86_print_operand): Handle 'q' and 'x'
3024 overrides for -masm=intel memory.
3025 * config/i386/sse.md (sse2_cvtdq2pd, sse2_cvtps2pd,
3026 sse4_1_<code>v8qiv8hi2, avx2_<code>v8qiv8si2,
3027 sse4_1_<code>v4hiv4si2, avx2_<code>v4hiv4di2,
3028 sse4_1_<code>v2siv2di2): Use %q1 instead of %1 for -masm=intel.
3029 (sse4_1_<code>v4qiv4si2, avx2_<code>v4qiv4di2,
3030 sse4_1_<code>v2hiv2di2): Use %k1 instead of %1 for -masm=intel.
3031 (sse4_1_<code>v2qiv2di2): Use %w1 instead of %1 for -masm=intel.
3033 2011-10-27 Martin Jambor <mjambor@suse.cz>
3035 * ipa-prop.c (compute_pass_through_member_ptrs): Rename parm_info
3037 (ipa_compute_jump_functions_for_edge): Likewise.
3038 (ipa_compute_jump_functions): Likewise.
3039 (ipa_analyze_indirect_call_uses): Likewise.
3040 (ipa_analyze_call_uses): Likewise.
3041 (ipa_analyze_params_uses): Likewise.
3042 (ipa_analyze_node): Likewise.
3044 2011-10-27 Uros Bizjak <ubizjak@gmail.com>
3047 * config/i386/sse.md (*avx_unpcklpd256): Remove extra insn
3048 constraints. Change alternative 1 to "x,m,1".
3050 2011-10-27 Jakub Jelinek <jakub@redhat.com>
3052 * Makefile.in (build/gencheck.o): Depend on tree.def and
3053 c-family/c-common.def.
3055 * tree-ssa-strlen.c: Include expr.h.
3056 (get_stridx): Don't use c_strlen, instead use string_constant
3057 and compute string length from it.
3058 * Makefile.in (tree-ssa-strlen.o): Depend on $(EXPR_H).
3060 2011-10-27 Eric Botcazou <ebotcazou@adacore.com>
3062 PR rtl-optimization/46603
3064 * reload.c (push_reload): In the out case, restore previous behavior
3065 for subregs that don't have word mode.
3067 2011-10-27 Ian Lance Taylor <iant@google.com>
3069 * cppdefault.c: Undef NATIVE_SYSTEM_HEADER_DIR if
3070 CROSS_DIRECTORY_STRUCTURE is defined and TARGET_SYSTEM_ROOT is not.
3071 (cpp_include_defaults): Only use NATIVE_SYSTEM_HEADER_DIR if it is
3074 2011-10-27 Richard Henderson <rth@redhat.com>
3076 * optabs.c (expand_vec_perm): Use the correct mode for scaling the
3077 selector. Save the qimode constant selector for later use by the
3078 qimode vec_perm pattern.
3080 2011-10-27 Bernd Schmidt <bernds@codesourcery.com>
3082 * config/c6x/c6x.c (unit_req_imbalance, res_mii): Cast the first arg
3083 to unit_req_factor to the right enum type.
3084 (get_unit_operand_masks, reshuffle_units, try_rename_operands,
3085 hwloop_optimize): Remove unused variables.
3087 2010-10-27 Richard Guenther <rguenther@suse.de>
3090 * tree-vect-generic.c (do_binop): Handle scalar operands.
3092 2011-08-27 Uros Bizjak <ubizjak@gmail.com>
3095 * config/i386/sse.md (*vec_extract_v4sf_mem): Avoid combining registers
3096 from different units in a single alternative.
3098 2011-10-26 David S. Miller <davem@davemloft.net>
3100 * config/sparc/sparc.c (emit_scc_insn): Force attempt of v9 sequences
3101 if we're comparing DImode and comparison is other than EQ or NE.
3103 * config/sparc/sparc.c (emit_scc_insn): Do not try v9 sequences until
3104 LEU/LTU/GEU/GTU is attempted.
3105 * config/sparc/sparc.md (*neg_snesi_sign_extend): New 64-bit insn
3107 (*neg_seqsi_sign_extend): Likewise.
3108 (*sltu_extend_sp64, *neg_sltu_extend_sp64, *sgeu_extend_sp64,
3109 *neg_sgeu_extend_sp64): New insns.
3111 * config/sparc/sparc-protos.h (sparc_expand_conditional_move): Declare.
3112 * config/sparc/sparc.md (mov<I:mode>cc, mov<F:mode>cc): Call it.
3113 (*mov<I:mode>_cc_v9): Normalize to expect operand 0 always in operand 4.
3114 (*mov<I:mode>_cc_reg_sp64): Likewise.
3115 (*movsf_cc_v9): Likewise.
3116 (*movsf_cc_reg_sp64): Likewise.
3117 (*movdf_cc_v9): Likewise.
3118 (*movdf_cc_reg_sp64): Likewise.
3119 (*movtf_cc_hq_v9): Likewise.
3120 (*movtf_cc_reg_hq_sp64): Likewise.
3121 (*movtf_cc_v9): Likewise.
3122 (*movtf_cc_reg_sp64): Likewise.
3123 * config/sparc/sparc.c (sparc_expand_conditional_move): New function.
3124 (sparc_print_operand): Delete 'c' and 'd' handling, no longer used.
3126 2011-10-26 Eric Botcazou <ebotcazou@adacore.com>
3128 * reload.c (reload_inner_reg_of_subreg): Change type of return value
3129 and type of OUTPUT parameter to bool and adjust. Document MODE and
3130 OUTPUT parameters. Use HARD_REGISTER_P. Reorder final condition
3131 and improve associated comment.
3132 (push_reload): Clarify and update comments about reloading of subregs.
3133 Adjust calls to reload_inner_reg_of_subreg. Compute the class upfront
3134 for the reloading of subregs in the out case as well.
3136 2011-10-26 Alexandre Oliva <aoliva@redhat.com>
3139 * var-tracking.c (rtx_debug_expr_p): New.
3140 (use_type): Don't use debug exprs to track non-VTA variables.
3142 2011-10-26 Jeff Law <law@redhat.com>
3144 * doc/invoke.texi (sink-frequency-threshold): Document.
3145 * tree-ssa-sink.c: Include params.h.
3146 (select_best_block): New function.
3147 (statement_sink_location): Use it.
3148 * params.def (SINK_FREQUENCY_THRESHOLD): New PARAM.
3150 2011-10-26 Iain Sandoe <iains@gcc.gnu.org>
3153 * config/darwin.c (top level): Amend comments concerning LTO output.
3154 (lto_section_num): New variable. (darwin_lto_section_e): New GTY.
3155 (LTO_SECTS_SECTION, LTO_INDEX_SECTION): New.
3156 (LTO_NAMES_SECTION): Rename.
3157 (darwin_asm_named_section): Record LTO section counts and switches
3158 in a vec of darwin_lto_section_e.
3159 (darwin_file_start): Remove unused code.
3160 (darwin_file_end): Put an LTO section termination label. Handle
3161 output of the wrapped LTO sections, index and names table.
3163 2011-10-26 Alan Modra <amodra@gmail.com>
3165 * config/rs6000/rs6000.c (rs6000_make_savres_rtx): Delete unneeded
3167 (rs6000_emit_stack_reset): Only return insn emitted when it adjusts sp.
3168 (rs6000_make_savres_rtx): Rename to rs6000_emit_savres_rtx. Use
3169 simple_return in pattern, emit instruction, and set jump_label.
3170 (rs6000_emit_prologue): Update for rs6000_emit_savres_rtx. Use
3171 simple_return rather than return.
3172 (emit_cfa_restores): New function.
3173 (rs6000_emit_epilogue): Emit cfa_restores when flag_shrink_wrap.
3174 Add missing cfa_restores for SAVE_WORLD. Add missing LR cfa_restore
3175 when using out-of-line gpr restore. Add missing LR and FP regs
3176 cfa_restores for out-of-line fpr restore. Consolidate code setting
3177 up cfa_restores. Formatting. Use LR_REGNO define.
3178 (rs6000_output_mi_thunk): Use simple_return rather than return.
3179 * config/rs6000/rs6000.md (sibcall*, sibcall_value*): Likewise.
3180 (return_internal*): Likewise.
3181 (any_return, return_pred, return_str): New iterators.
3182 (return, conditional return insns): Provide both return and
3183 simple_return variants.
3184 * config/rs6000/rs6000.h (EARLY_R12, LATE_R12): Define.
3185 (REG_ALLOC_ORDER): Move r12 before call-saved regs when FIXED_R13.
3186 Move r11 and r0 later to suit shrink-wrapping.
3188 2011-10-26 Richard Guenther <rguenther@suse.de>
3190 * lto-wrapper.c (run_gcc): Properly init/free obstack.
3192 2011-10-26 Jakub Jelinek <jakub@redhat.com>
3194 * config/i386/i386.md (UNSPEC_VSIBADDR): New.
3195 * config/i386/predicates.md (vsib_address_operand,
3196 vsib_mem_operator): New predicates.
3197 * config/i386/i386.c (ix86_print_operand_address): Handle
3198 UNSPEC_VSIBADDR addresses.
3199 * config/i386/sse.md (avx2_gathersi<mode>, avx2_gatherdi<mode>,
3200 avx2_gatherdi<mode>256): Adjust expanders to use MEM with
3201 UNSPEC_VSIBADDR address.
3202 (*avx2_gathersi<mode>, *avx2_gatherdi<mode>, *avx2_gatherdi<mode>256):
3203 Adjust insns to use MEM with UNSPEC_VSIBADDR address.
3205 2011-10-26 Tom de Vries <tom@codesourcery.com>
3207 PR tree-optimization/50763
3208 * tree-ssa-tail-merge.c (replace_block_by): Update vops if phi_vuse1 or
3209 phi_vuse2 is NULL_TREE only if bb1 dominates or is dominated by bb2.
3211 2011-10-26 Richard Guenther <rguenther@suse.de>
3214 * Makefile.in (lto-wrapper): Depend on and link against opts-common.o.
3215 (lto-wrapper.o): Depend on $(OPTS_H) and $(OPTIONS_H).
3216 * lto-wrapper.c (get_options_from_collect_gcc_options): New function.
3217 (run_gcc): Use it. Filter out language specific options.
3219 2011-10-26 Andreas Tobler <andreast@fgznet.ch>
3221 * config/i386/freebsd64.h (LINK_SPEC): Emit the same warning as the
3224 2011-10-25 Paolo Carlini <paolo.carlini@oracle.com>
3227 * gcc.c (main): Fix fatal_error string for translation.
3229 2011-10-25 Ian Lance Taylor <iant@google.com>
3231 * tree-eh.c (do_return_redirection): Remove return_value_p
3232 parameter. Change all callers.
3233 (lower_try_finally_nofallthru): Remove local return_val.
3234 (lower_try_finally_onedest): Likewise.
3235 (lower_try_finally_copy): Likewise.
3236 (lower_try_finally_switch): Likewise.
3238 2011-10-25 H.J. Lu <hongjiu.lu@intel.com>
3240 * config/i386/mmx.md (*mmx_maskmovq): Replace :SI with :P and
3241 remove "&& !TARGET_64BIT"
3242 (*mmx_maskmovq_rex): Removed.
3244 2011-10-25 Eric Botcazou <ebotcazou@adacore.com>
3246 PR rtl-optimization/46603
3247 * reload.c (push_reload): In the out case, reload the subreg as well
3248 as the reg if it has word mode.
3250 2011-10-25 Eric Botcazou <ebotcazou@adacore.com>
3252 * dwarf2out.c (add_gnat_descriptive_type_attribute): Temporarily
3253 suppress debug info for the parent type.
3255 2011-10-25 Eric Botcazou <ebotcazou@adacore.com>
3257 * config/ia64/ia64.c (ia64_profile_hook): Fix thinko.
3259 2011-10-25 Richard Henderson <rth@redhat.com>
3261 * config/i386/sse.md (VEC_EXTRACT_EVENODD_MODE): Remove.
3262 (vec_extract_even<mode>, vec_extract_odd<mode>): Remove.
3264 * config/rs6000/altivec.md (vec_extract_evenv8hi,
3265 vec_extract_evenv16qi, vec_extract_oddv4si,
3266 vec_extract_oddv4sf): Remove.
3268 * config/spu/spu.md (vec_extract_evenv4si, vec_extract_evenv4sf,
3269 vec_extract_evenv8hi, vec_extract_evenv16qi, vec_extract_oddv4si,
3270 vec_extract_oddv4sf, vec_extract_oddv8hi, vec_extract_oddv16qi,
3271 vec_interleave_highv4sf, vec_interleave_lowv4sf,
3272 vec_interleave_highv4si, vec_interleave_lowv4si,
3273 vec_interleave_highv8hi, vec_interleave_lowv8hi,
3274 vec_interleave_highv16qi, vec_interleave_lowv16qi): Remove.
3276 * expr.c (expand_expr_real_2) [VEC_EXTRACT_EVEN_EXPR]: Use binop.
3277 [VEC_EXTRACT_ODD_EXPR, VEC_INTERLEAVE_HIGH_EXPR]: Likewise.
3278 [VEC_INTERLEAVE_LOW_EXPR]: Likewise.
3279 * optabs.c (expand_binop): Implement vec_interleave_high_optab,
3280 vec_interleave_low_optab, vec_extract_even_optab,
3281 vec_extract_odd_optab with expand_vec_perm.
3282 (can_vec_perm_for_code_p): New.
3284 * tree-vect-data-refs.c (vect_strided_store_supported): Allow for
3285 fallback via can_vec_perm_for_code_p.
3286 (vect_strided_load_supported): Likewise.
3287 * tree-vect-generic.c (expand_vector_operations_1): Never lower
3288 VEC_INTERLEAVE_HIGH_EXPR, VEC_INTERLEAVE_LOW_EXPR,
3289 VEC_EXTRACT_EVEN_EXPR, VEC_EXTRACT_ODD_EXPR.
3291 * target.def (vec_perm_const_ok): Change parameters to mode and
3293 * doc/tm.texi: Rebuild.
3294 * config/i386/i386.c (ix86_vectorize_vec_perm_const_ok): Change
3295 parameters to mode and array of indicies.
3296 * expr.c (expand_expr_real_2) [VEC_PERM_EXPR]: Expand operands here.
3297 * optabs.c (can_vec_perm_p): Rename from can_vec_perm_expr_p.
3298 Change parameters to mode and array of indicies.
3299 (expand_vec_perm_1): Rename from expand_vec_perm_expr_1.
3300 (expand_vec_perm): Rename from expand_vec_perm_expr. Change
3301 parameters to mode and rtx inputs. Try lowering to QImode
3302 vec_perm_const before trying fully variable permutation.
3303 * optabs.h: Update decls.
3304 * tree-vect-generic.c (lower_vec_perm): Extract array of indices from
3305 VECTOR_CST to pass to can_vec_perm_p.
3306 * tree-vect-slp.c (vect_get_mask_element): Change mask parameter type
3307 from int pointer to unsigned char pointer.
3308 (vect_transform_slp_perm_load): Update for change to can_vec_perm_p.
3309 * tree-vect-stmts.c (perm_mask_for_reverse): Likewise.
3311 * tree.def (VEC_EXTRACT_EVEN_EXPR): Fix typo in text name.
3312 (VEC_EXTRACT_ODD_EXPR, VEC_INTERLEAVE_HIGH_EXPR,
3313 VEC_INTERLEAVE_LOW_EXPR): Likewise.
3315 2011-10-25 Mike Stump <mikestump@comcast.net>
3317 * reload.c (regno_clobbered_p): Fix typo.
3319 2011-10-25 Dodji Seketeli <dodji@redhat.com>
3321 * input.c (expand_location): Rewrite using linemap_resolve_location
3322 and linemap_expand_location. Add a comment.
3324 2011-10-25 Jakub Jelinek <jakub@redhat.com>
3326 PR tree-optimization/50596
3327 * tree-vect-stmts.c (vect_mark_relevant): Only use
3328 FOR_EACH_IMM_USE_FAST if lhs is SSA_NAME.
3329 (vectorizable_store): If is_pattern_stmt_p look through
3330 VIEW_CONVERT_EXPR on lhs.
3331 * tree-vect-patterns.c (check_bool_pattern, adjust_bool_pattern):
3332 Use unsigned type instead of signed.
3333 (vect_recog_bool_pattern): Optimize also stores into bool memory in
3334 addition to casts from bool to integral types.
3335 (vect_mark_pattern_stmts): If pattern_stmt already has vinfo
3336 created, don't create it again.
3338 2011-10-25 Kai Tietz <ktietz@redhat.com>
3340 * config/i386/i386.c (ix86_frame_pointer_required): Require
3341 frame-pointer, if setjmp is used for 32-bit ms-abi.
3343 2011-10-24 Michael Meissner <meissner@linux.vnet.ibm.com>
3345 * builtins.c (set_builtin_user_assembler_name): Remove extra
3346 newline added in October 11th, 2011 change.
3348 2011-10-24 David S. Miller <davem@davemloft.net>
3350 * config/sparc/little-endian.opt: Delete.
3351 * config.gcc: Remove references to config/sparc/little-endian.opt
3352 * doc/invoke.texi: Remove documentation of -mlittl-endian on sparc.
3353 * config/sparc/linux64.h: Delete references to -mlittle-endian.
3354 * config/sparc/netbsd-elf.h: Likewise.
3355 * config/sparc/openbsd64.h: Likewise.
3356 * config/sparc/sparc.h: Likewise.
3357 * config/sparc/sp64-elf.h: Likewise and delete overrides for
3358 BYTES_BIG_ENDIAN and WORDS_BIG_ENDIAN.
3359 * config/sparc/sparc.c (dump_target_flag_bits): Remove reference
3360 to MASK_LITTLE_ENDIAN.
3361 * config/sparc/sparc.opt (Mask(LITTLE_ENDIAN)): Delete.
3363 * config/sparc/sparc.md: Only use F, G, and C constraints in FP insns.
3364 Only use D, Y, and Z constraints in vector insns.
3366 * config/sparc/sparc.md (cpu_feature, enabled): New attributes.
3367 (*movsi_insn_novis3, *movsi_insn_vis3): Consolidate into one pattern
3369 (*movdi_insn_sp32_v9_novis3, *movdi_insn_sp32_v9_vis3): Consolidate
3370 into *movdi_insn_sp32.
3371 (*movdi_insn_sp64_novis3, *movdi_insn_sp64_vis3): Consolidate into
3372 one pattern called *movdi_insn_sp64.
3373 (*movsf_insn_novis3, *movsf_insn_vis3, *movsf_insn_no_fpu):
3374 Consolidate into one pattern called *movsf_insn.
3375 (*movdf_insn_sp32_no_fpu, *movdf_insn_sp32_v9_novis3,
3376 *movdf_insn_sp32_v9_vis3, *movdf_insn_sp32_v9_no_fpu): Consolidate
3377 into *movdf_insn_sp32.
3378 (*movdf_insn_sp64_novis3, *movdf_insn_sp64_vis3,
3379 *movdf_insn_sp64_no_fpu): Consolidate into one pattern called
3381 (*zero_extendsidi2_insn_sp64_novis3,
3382 *zero_extendsidi2_insn_sp64_vis3): Consolidate into one pattern
3383 called *zero_extendsidi2_insn_sp64.
3384 (*sign_extendsidi2_insn_novis3, *sign_extendsidi2_insn_vis3):
3385 Consolidate into one pattern named *sign_extendsidi2_insn.
3386 (*mov<VM32:mode>_insn_novis3, *mov<VM32:mode>_insn_vis3):
3387 Consolidate into one pattern named *mov<VM32:mode>_insn.
3388 (*mov<VM64:mode>_insn_sp64_novis3,
3389 *mov<VM64:mode>_insn_sp64_novis3): Consolidate into one pattern
3390 named *mov<VM64:mode>_insn_sp64.
3391 (*mov<VM64:mode>_insn_sp32_novis3,
3392 *mov<VM64:mode>_insn_sp32_vis3): Consolidate into one pattern
3393 named *mov<VM64:mode>_insn_sp32.
3395 2011-10-24 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
3397 * tree-ssa-strlen.c (get_string_length): Change assertion to STPCPY.
3398 (zero_length_string): Change assertion to accept strinfo without
3399 length but with stmt instead.
3400 Set the endptr pointer also if starting a new chain.
3401 (adjust_related_strinfos): Ignore strinfos marked for delayed
3403 (handle_builtin_strcpy): Mark earlier strinfo elements also for
3404 delayed length computation.
3406 2011-10-24 Georg-Johann Lay <avr@gjlay.de>
3409 Port from 4.6 branch r180379
3410 * doc/invoke.texi (AVR Options): New subsubsection to explain EIND
3411 handling and indirect jump/calls on devices > 128k.
3413 2011-10-24 Anatoly Sokolov <aesok@post.ru>
3414 Georg-Johann Lay <avr@gjlay.de>
3417 * doc/extend.texi (Declaring Attributes of Functions):
3418 Document OS_main and OS_task attributes.
3419 (Specifying Attributes of Variables): Move up
3420 subsection "AVR Variable Attributes" as of alphabetical order.
3422 2011-10-24 Richard Guenther <rguenther@suse.de>
3424 * tree-vect-stmts.c (vect_get_vec_def_for_operand): Convert constants
3425 to vector element type.
3426 (vectorizable_assignment): Bail out for non-mode-precision operations.
3427 (vectorizable_shift): Likewise.
3428 (vectorizable_operation): Likewise.
3429 (vectorizable_type_demotion): Likewise.
3430 (vectorizable_type_promotion): Likewise.
3431 (vectorizable_store): Handle non-mode-precision stores.
3432 (vectorizable_load): Handle non-mode-precision loads.
3433 (get_vectype_for_scalar_type_and_size): Return a vector type
3434 for non-mode-precision integers.
3435 * tree-vect-loop.c (vectorizable_reduction): Bail out for
3436 non-mode-precision reductions.
3438 2011-10-24 Julian Brown <julian@codesourcery.com>
3440 * config/m68k/m68k.c (notice_update_cc): Tighten condition for
3441 setting CC_REVERSED for FP comparisons.
3443 2011-10-24 Richard Guenther <rguenther@suse.de>
3445 PR tree-optimization/50838
3446 * tree-data-ref.c (dr_analyze_indices): Properly canonicalize
3447 a MEM_REF base if we change it.
3449 2011-10-24 Bernd Schmidt <bernds@codesourcery.com>
3452 * rtlanal.c: Swap includes of "hard-reg-set.h" and "rtl.h".
3454 PR rtl-optimization/50833
3455 * function.c (thread_prologue_and_epilogue_insns): Expect the
3456 return insn optimization only if optimize.
3458 2011-10-24 Georg-Johann Lay <avr@gjlay.de>
3460 * config/avr/avr.c: Break long lines.
3461 Define target hooks on the fly if applicable.
3462 (TARGET_ASM_FUNCTION_RODATA_SECTION): Remove first definition
3464 (targetm): Move definition to end of file.
3465 (avr_can_eliminate): Make static on the fly.
3466 (avr_frame_pointer_required_p): Ditto.
3467 (avr_hard_regno_scratch_ok): Ditto.
3468 (avr_builtin_setjmp_frame_value): Make static on the fly.
3469 Indent according to coding rules.
3470 (avr_case_values_threshold): Ditto.
3471 (avr_attribute_table): Move down.
3473 2011-10-24 Ira Rosen <ira.rosen@linaro.org>
3475 PR tree-optimization/50730
3476 * tree-vect-data-refs.c (vect_analyze_data_refs): Stop basic block
3477 analysis if encountered unsupported data-ref.
3479 2011-10-23 David S. Miller <davem@davemloft.net>
3481 * config/sparc/sparc.c (sparc_option_override): Remove -mv8plus
3483 * config/sparc/linux64.h (CC1_SPEC): When defaulting to 64-bit,
3484 append -mcpu=v9 when -mv8plus is given.
3486 * config/sparc/sparc.h (SECONDARY_MEMORY_NEEDED): We can move
3487 between float and non-float regs when VIS3.
3488 * config/sparc/sparc.c (eligible_for_restore_insn): We can't
3489 use a restore when the source is a float register.
3490 (sparc_split_regreg_legitimate): When VIS3 allow moves between
3491 float and integer regs.
3492 (sparc_register_move_cost): Adjust to account for VIS3 moves.
3493 (sparc_preferred_reload_class): On 32-bit with VIS3 when moving an
3494 integer reg to a class containing EXTRA_FP_REGS, constrain to FP_REGS.
3495 (sparc_secondary_reload): On 32-bit with VIS3 when moving between
3496 float and integer regs we sometimes need a FP_REGS class
3497 intermediate move to satisfy the reload. When this happens
3498 specify an extra cost of 2.
3499 (*movsi_insn): Rename to have "_novis3" suffix and add !VIS3 guard.
3500 (*movdi_insn_sp32_v9): Likewise.
3501 (*movdi_insn_sp64): Likewise.
3502 (*movsf_insn): Likewise.
3503 (*movdf_insn_sp32_v9): Likewise.
3504 (*movdf_insn_sp64): Likewise.
3505 (*zero_extendsidi2_insn_sp64): Likewise.
3506 (*sign_extendsidi2_insn): Likewise.
3507 (*movsi_insn_vis3): New insn.
3508 (*movdi_insn_sp32_v9_vis3): New insn.
3509 (*movdi_insn_sp64_vis3): New insn.
3510 (*movsf_insn_vis3): New insn.
3511 (*movdf_insn_sp32_v9_vis3): New insn.
3512 (*movdf_insn_sp64_vis3): New insn.
3513 (*zero_extendsidi2_insn_sp64_vis3): New insn.
3514 (*sign_extendsidi2_insn_vis3): New insn.
3515 (TFmode reg/reg split): Make sure both REG operands are float.
3516 (*mov<VM32:mode>_insn): Add "_novis3" suffix and !VIS3 guard. Remove
3517 easy constant to integer reg alternatives.
3518 (*mov<VM64:mode>_insn_sp64): Likewise.
3519 (*mov<VM64:mode>_insn_sp32_novis3): Likewise.
3520 (*mov<VM32:mode>_insn_vis3): New insn.
3521 (*mov<VM64:mode>_insn_sp64_vis3): New insn.
3522 (*mov<VM64:mode>_insn_sp32_vis3): New insn.
3523 (VM64 reg<-->reg split): New splitter for 32-bit.
3525 * config/sparc/sparc.c (sparc_split_regreg_legitimate): New function.
3526 * config/sparc/sparc-protos.h (sparc_split_regreg_legitimate):
3528 * config/sparc/sparc.md (DImode reg/reg split): Use it.
3529 (DFmode reg/reg split): Likewise.
3531 * config/sparc/sparc.md (*movdi_insn_sp32_v9): Add alternatives for
3532 generating fzero and fone instructions.
3533 (DImode const_int --> reg splitter): Only trigger for integer regs.
3535 * config/sparc/predicates.md (input_operand): Disallow vector
3536 constants other than 0 and -1.
3537 * config/sparc/sparc.c (sparc_preferred_reload_class): Return
3538 NO_REGS for vector constants other than 0 and -1.
3540 * config/sparc/sparc.h (SPARC_FIRST_INT_REG, SPARC_LAST_INT_REG,
3541 SPARC_INT_REG_P): Define.
3542 (HARD_REGNO_NREGS): Use SPARC_INT_REG_P.
3543 (REGNO_OK_FOR_INDEX_P): Likewise.
3544 * config/sparc/sparc.c (gen_df_reg): Likewise.
3545 (eligible_for_return_delay): Likewise.
3546 (eligible_for_sibcall_delay): Likewise.
3547 (sparc_legitimate_address_p): Likewise.
3548 (emit_save_or_restore_regs): Likewise.
3549 (registers_ok_for_ldd_peep): Likewise.
3550 * config/spac/sparc.md (DI mode splitters): Likewise.
3551 (SF mode const splitters): Likewise.
3552 (DF mode splitters): Likewise.
3553 (32-bit DI mode logical op splitters): Likewise.
3555 2011-10-23 Paolo Carlini <paolo.carlini@oracle.com>
3559 2011-10-23 Paolo Carlini <paolo.carlini@oracle.com>
3562 * doc/invoke.texi ([-Wnarrowing], [-Wc++0x-compat]): Update.
3564 2011-10-23 Uros Bizjak <ubizjak@gmail.com>
3566 * config/i386/sse.md (avx2_lshl<mode>3): Remove insn pattern.
3567 (VI248_256): Remove mode iterator.
3568 * config/i386/i386.c (ix86_expand_vec_perm): Use gen_ashlv4di3
3569 instead of gen_avx2_lshlv4di3.
3570 (bdesc_args): Use CODE_FOR_ashl{v16hi,v8si,v4di}3 instead of
3571 CODE_FOR_avx2_lshl{v16hi,v8si,v4di}3.
3573 2011-10-23 Uros Bizjak <ubizjak@gmail.com>
3575 * config/i386/sse.md (sseintprefix): Rename from gthrfirstp.
3576 (<avx_avx2>_maskload<ssemodesuffix><avxsizesuffix>): Delete expander.
3577 (<avx_avx2>_maskload<ssemodesuffix><avxsizesuffix>) Merge insn
3578 pattern from *avx2_maskload<ssemodesuffix><avxsizesuffix> and
3579 *avx_maskload<ssemodesuffix><avxsizesuffix> using V48_AVX mode
3580 iterator. Use sseintprefix mode attribute.
3581 (<avx_avx2>_maskstore<ssemodesuffix><avxsizesuffix>): Delete expander.
3582 (<avx_avx2>_maskstore<ssemodesuffix><avxsizesuffix>) Merge insn
3583 pattern from *avx2_maskstore<ssemodesuffix><avxsizesuffix> and
3584 *avx_maskstore<ssemodesuffix><avxsizesuffix> using V48_AVX mode
3585 iterator. Use sseintprefix mode attribute.
3586 (*avx2_gathersi<mode>) Use sseintprefix and ssemodesuffix mode
3588 (*avx2_gatherdi<mode>): Ditto.
3589 (*avx2_gatherdi<mode>256): Ditto.
3590 (VI48_AVX2): Remove mode iterator.
3591 (gthrlastfp): Remove mode attribute.
3593 2011-10-23 Paolo Carlini <paolo.carlini@oracle.com>
3596 * doc/invoke.texi ([-Wnarrowing], [-Wc++0x-compat]): Update.
3598 2011-10-23 Tom de Vries <tom@codesourcery.com>
3600 PR tree-optimization/50763
3601 * tree-ssa-tail-merge.c (same_succ_flush_bb): New function, factored
3603 (same_succ_flush_bbs): Use same_succ_flush_bb.
3604 (purge_bbs): Remove argument. Remove calls to same_succ_flush_bbs,
3605 release_last_vdef and delete_basic_block.
3606 (unlink_virtual_phi): New function.
3607 (update_vuses): Add and use vuse1_phi_args argument. Set var to
3608 SSA_NAME_VAR of vuse1 or vuse2, and use var. Handle case that
3609 def_stmt2 is NULL. Use phi result as phi arg in case vuse1 or vuse2
3610 is NULL_TREE. Replace uses of vuse1 if vuse2 is NULL_TREE. Fix code
3611 to limit replacement of uses. Propagate phi argument for phis with a
3613 (replace_block_by): Update vops if phi_vuse1 or phi_vuse2 is NULL_TREE.
3614 Set vuse1_phi_args if vuse1 is a phi defined in bb1. Add
3615 vuse1_phi_args as argument to call to update_vuses. Call
3616 release_last_vdef, same_succ_flush_bb, delete_basic_block. Update
3617 CDI_DOMINATORS info.
3618 (tail_merge_optimize): Remove argument in call to purge_bbs. Remove
3619 call to free_dominance_info. Only call calculate_dominance_info once.
3621 2011-10-23 Eric Botcazou <ebotcazou@adacore.com>
3623 * fold-const.c (invert_tree_comparison): Always invert EQ_EXPR/NE_EXPR.
3625 PR tree-optimization/44683
3626 * tree-ssa-dom.c (record_edge_info): Record simple equivalences only if
3627 we can be sure that there are no signed zeros involved.
3629 2011-10-23 Jan Hubicka <jh@suse.cz>
3631 * ipa-inline.c (estimate_badness): Scale up and handle overflows.
3633 2011-10-23 Uros Bizjak <ubizjak@gmail.com>
3636 * config/i386/sse.md (<avx_avx2>_maskload<ssemodesuffix><avxsizesuffix>):
3637 Remove (match_dup 0).
3638 (*avx2_maskload<ssemodesuffix><avxsizesuffix>): New insn pattern.
3639 (*avx_maskload<ssemodesuffix><avxsizesuffix>): Ditto.
3640 (*avx2_maskstore<ssemodesuffix><avxsizesuffix>): Ditto.
3641 (*avx_maskstore<ssemodesuffix><avxsizesuffix>): Ditto.
3642 (*avx2_maskmov<ssemodesuffix><avxsizesuffix>): Remove insn pattern.
3643 (*avx_maskmov<ssemodesuffix><avxsizesuffix>): Ditto.
3645 2011-10-23 Ira Rosen <ira.rosen@linaro.org>
3647 PR tree-optimization/50819
3648 * tree-vectorizer.h (vect_analyze_data_ref_dependences): Remove
3650 * tree-vect-loop.c (vect_analyze_loop_2): Update call to
3651 vect_analyze_data_ref_dependences.
3652 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Remove
3653 the last argument. Check load-after-store dependence for unknown
3654 dependencies in basic blocks.
3655 (vect_analyze_data_ref_dependences): Update call to
3656 vect_analyze_data_ref_dependences.
3657 * tree-vect-patterns.c (vect_recog_widen_shift_pattern): Fix typo.
3658 * tree-vect-slp.c (vect_bb_vectorizable_with_dependencies): Remove.
3659 (vect_slp_analyze_bb_1): Update call to
3660 vect_analyze_data_ref_dependences. Don't call
3661 vect_bb_vectorizable_with_dependencies.
3663 2011-10-22 David S. Miller <davem@davemloft.net>
3665 * config/sparc/sparc.h (SECONDARY_INPUT_RELOAD_CLASS,
3666 SECONDARY_OUTPUT_RELOAD_CLASS): Delete.
3667 * config/sparc/sparc.c (TARGET_SECONDARY_RELOAD): Redefine.
3668 (sparc_secondary_reload): New function.
3670 * config/sparc/sparc.h (sparc_costs): Remove extern decl.
3671 (struct processor_costs): Move from here..
3672 * config/sparc/sparc.c (struct processor_costs): To here.
3673 (sparc_costs): Mark static.
3675 * config/sparc/sparc.c (short_branch, reg_unused_after): Delete.
3676 * config/sparc/sparc-protos.h (short_branch, reg_unused_after):
3677 Get rid of declarations.
3679 2011-10-21 Paul Brook <paul@codesourcery.com>
3681 * config/c6x/c6x.c (c6x_asm_emit_except_personality,
3682 c6x_asm_init_sections): New functions.
3683 (TARGET_ASM_EMIT_EXCEPT_PERSONALITY, TARGET_ASM_INIT_SECTIONS): Define.
3685 2011-10-21 Jakub Jelinek <jakub@redhat.com>
3688 * config/i386/i386.c (expand_vec_perm_even_odd_1): Handle
3689 V4DImode and V8SImode for !TARGET_AVX2.
3691 2011-10-21 Bernd Schmidt <bernds@codesourcery.com>
3694 * sched-deps.c (add_dependence): If not doing predication, promote
3695 REG_DEP_CONTROL to REG_DEP_ANTI.
3697 2011-10-21 Georg-Johann Lay <avr@gjlay.de>
3699 * config/avr/avr.h (LEGITIMIZE_RELOAD_ADDRESS): Pass address of X
3700 instead of X to avr_legitimize_reload_address.
3701 * config/avr/avr-protos.h (avr_legitimize_reload_address): Change
3702 first argument's type from rtx to rtx*.
3703 * config/avr/avr.c (avr_legitimize_reload_address): Ditto.
3704 Pass PX to push_reload instead of &X. Change log messages for
3707 2011-10-21 Roland Stigge <stigge@antcom.de>
3709 PR translation/47064
3710 * params.def: Fix typo "compilatoin" -> "compilation".
3712 2011-10-21 H.J. Lu <hongjiu.lu@intel.com>
3713 Kirill Yukhin <kirill.yukhin@intel.com>
3716 * config/i386/driver-i386.c (host_detect_local_cpu): Do cpuid 7 only
3717 if max_level allows that.
3719 2011-10-21 Bernd Schmidt <bernds@codesourcery.com>
3721 * reg-notes.def (DEP_CONTROL): New.
3722 * sched-ebb.c (add_deps_for_risky_insns): Add a REG_DEP_CONTROL when
3723 not doing speculation.
3724 * rtlanal.c (record_hard_reg_sets, find_all_hard_reg_sets,
3725 record_hard_reg_uses_1, record_hard_reg_uses): New functions.
3726 * function.c (record_hard_reg_sets, record_hard_reg_uses,
3727 record_hard_reg_uses_1): Remove; move to rtlanal.c.
3728 * lists.c (copy_INSN_LIST, concat_INSN_LIST): New functions.
3729 * haifa-sched.c: Swap includes of "rtl.h" and "hard-reg-set.h".
3730 (MUST_RECOMPUTE_SPEC_P): New macro.
3731 (real_insn_for_shadow): New function.
3732 (cond_clobbered_p, recompute_todo_spec, check_clobbered_conditions,
3733 toggle_cancelled_flags): New static functions.
3734 (schedule_insn): Relax an assert to only check for empty hard back
3735 dependencies. Skip cancelled dependencies. Call
3736 check_clobbered_conditions.
3737 (copy_insn_list): Remove function, renamed moved to lists.c.
3738 (save_backtrack_point): Use new spelling copy_INSN_LIST.
3739 (unschedule_insns_until): Ensure TODO_SPEC is reset properly.
3740 (restore_last_backtrack_point): Likewise. Call toggle_cancelled_flags.
3741 (estimate_insn_tick): Ignore cancelled dependencies.
3742 (haifa_speculate_insn): Move declaration.
3743 (try_ready): Move code into recompute_todo_spec and call it. Tweak
3744 some asserts. Ensure predicated patterns are restored if necessary.
3745 Dump DEP_CONTROL flag.
3746 (haifa_change_pattern): Merge with sched_change_pattern.
3747 (sched_change_pattern): Remove function.
3748 * sched-deps.c (NON_FLUSH_JUMP_KIND, NON_FLUSH_JUMP): Remove. All
3749 uses changed to simply not test NON_FLUSH_JUMP_P.
3750 (ds_to_dk, dk_to_ds, dump_dep, ds_to_dt, dump_ds, check_dep): Handle
3752 (dep_spec_p): If DO_PREDICATION, REG_DEP_CONTROL is speculative.
3753 (reg_pending_control_uses, control_dependency_cache): New static
3755 (sched_get_reverse_condition_uncached): New function.
3756 (sd_find_dep_between): Remove pointless assert. Look in
3757 control_dependency_cache.
3758 (ask_dependency_caches, set_dependency_caches, sd_delete_dep,
3759 extend_dependency_caches, sched_deps_finish): Handle REG_DEP_CONTROL
3760 and control_dependency_cache.
3761 (sd_unresolve_dep): Use dep_spec_p.
3762 (add_dependence): Now a wrapper around add_dependence_1, handling
3763 REG_DEP_CONTROL specially.
3764 (flush_pending_lists): Clear pending_jump_insns.
3765 (sched_analyze_1): Handle pending_jump_insns like a memory flush.
3766 (sched_analyze_2): Unconditionally add to pending memory flushes,
3767 keep previous behaviour but apply it to pending_jump_insns instead.
3768 (sched_analyze_insn): Defer adding jump reg dependencies using
3769 reg_pending_control_uses; add them to the control_uses list. Handle
3770 pending_jump_insns and control_uses when adding dependence lists.
3771 (deps_analyze_insn): Update INSN_COND_DEPS.
3772 (deps_analyze_insn): Add jumps to pending_jump_insns rather than
3773 last_pending_memory_flush.
3774 (init_deps): Initialize pending_jump_insns.
3775 (free_deps): Free control_uses.
3776 (remove_from_deps): Remove from pending_jump_insns.
3777 (init_deps_global): Allocate reg_pending_control_uses).
3778 (finish_deps_global): Free it.
3779 (add_dependence_1): Renamed from add_dependence. Handle
3781 * rtl.h (record_hard_reg_uses, find_all_hard_reg_sets): Declare.
3782 (copy_INSN_LIST, concat_INSN_LIST): Declare.
3783 * sched-int.h (struct deps_reg): Add control_uses.
3784 (struct deps_desc): Add pending_jump_insns.
3785 (struct _haifa_deps_insn_data): Add cond_deps.
3786 (struct _haifa_insn_data): Add must_recompute_spec and predicated_pat.
3787 (INSN_COND_DEPS, PREDICATED_PAT): New macros.
3788 (BITS_PER_DEP_WEAK): Adjust for two extra bits in the word.
3789 (DEP_CONTROL): New macro.
3790 (DEP_TYPES): Include it.
3791 (HARD_DEP): Adjust definition.
3792 (DEP_CANCELLED): New macro.
3793 (enum SCHED_FLAGS): Add DO_PREDICATION.
3794 (sched_get_reverse_condition_uncached, real_insn_for_shadow): Declare.
3795 * sched-rgn.c (concat_INSN_LIST): Remove function.
3796 (deps_join): Handle pending_jump_insns.
3797 (free_pending_lists): Likewise.
3798 * config/c6x/c6x.c (c6x_set_sched_flags): Set DO_PREDICATION for final
3801 2011-10-21 Georg-Johann Lay <avr@gjlay.de>
3804 * config/avr/libgcc.S (__EIND__): New define to 0x3C.
3805 (__tablejump__): Consistently use EIND for indirect jump/call.
3806 (__tablejump_elpm__): Ditto.
3808 2011-10-21 Bernd Schmidt <bernds@codesourcery.com>
3810 * config/c6x/c6x.md (attr "op_pattern"): New.
3811 (load_sdata_pic, mov<mode>_insn for QIHIM and SISFVM): Set it.
3812 * config/c6x/c6x-mult.md.in (mulhi3_VARIANT_, mulhisi3_insn_VARIANT_):
3814 * config/c6x/c6x-mult.md: Regenerate.
3815 * config/c6x/c6x.c: Include "regrename.h".
3816 (unit_req_table): New typedef.
3817 (unit_reqs): Use it for the declaration.
3818 (unit_req_factor, get_unit_reqs, merge_unit_reqs, unit_req_imbalance,
3819 get_unit_operand_masks, try_rename_operands, reshuffle_units): New
3821 (count_unit_reqs): New arg reqs. All callers changed. Use
3822 get_unit_reqs, and don't merge here.
3823 (res_mii): New arg reqs. All callers changed. Rewrite to use a loop
3824 using unit_req_factor.
3825 (hwloop_optimize): Call reshuffle_units. Call merge_unit_reqs after
3827 (c6x_reorg): Add reg notes problem, and call df_analyze.
3828 * Makefile.in ($(out_object_file)): Depend on regrename.h.
3830 2011-10-21 Kai Tietz <ktietz@redhat.com>
3832 * fold-const.c (simple_operand_p_2): Handle integral
3833 casts from boolean-operands.
3835 2011-10-21 Jan Hubicka <jh@suse.cz>
3837 * cgraph.c (dump_cgraph_node): Dump alias flag.
3838 * cgraphunit.c (handle_alias_pairs): Handle weakrefs with
3840 (get_alias_symbol): New function.
3841 (output_weakrefs): Output also weakrefs with no destinatoin.
3842 (lto_output_node): Output weakref alias flag when at function boundary.
3844 2011-10-21 Andrew Stubbs <ams@codesourcery.com>
3847 * config/arm/driver-arm.c (vendors): Make static.
3849 2011-10-21 Uros Bizjak <ubizjak@gmail.com>
3851 * config/i386/i386.c (ix86_emit_swdivsf): Force b into register.
3852 (ix86_emit_swsqrtsf): Force a into register.
3854 2011-10-20 Mike Stump <mikestump@comcast.net>
3856 * regcprop.c (copyprop_hardreg_forward_1): Update recog_data
3857 after validate_change wipes it out.
3859 2011-10-20 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
3861 * config/spu/spu.md ("vec_permv16qi"): Reduce selector modulo 32
3862 before using the shufb instruction.
3864 2011-10-20 Kirill Yukhin <kirill.yukhin@intel.com>
3867 * config/i386/i386.md (bmi_bextr_<mode>): Update register/
3868 memory operand order.
3869 (bmi2_bzhi_<mode>3): Ditto.
3870 (bmi2_pdep_<mode>3): Ditto.
3871 (bmi2_pext_<mode>3): Ditto.
3873 2011-10-20 Richard Henderson <rth@redhat.com>
3875 * target.def (vec_perm_const_ok): Rename from builtin_vec_perm_ok.
3876 * optabs.c (can_vec_perm_expr_p): Update to match.
3877 (expand_vec_perm_expr): Likewise.
3878 * config/i386/i386.c (TARGET_VECTORIZE_VEC_PERM_CONST_OK): Rename
3879 from TARGET_VECTORIZE_BUILTIN_VEC_PERM_OK.
3880 * doc/tm.texi.in: Likewise.
3882 2011-10-20 Sergey Ostanevich <sergos.gnu@gmail.com>
3885 * config/i386/i386.c (processor_target_table): Ch