PR tree-optimization/19283
* fold-const.c (fold_widened_comparison): Return NULL if shorter_type
is not shorter than the original type.
* gcc.c-torture/execute/20050106-1.c: New test.
From-SVN: r93043
PR rtl-optimization/18861
* cfgbuild.c (BLOCK_USED_BY_TABLEJUMP): Define.
(FULL_STATE): Define.
(mark_tablejump_edge): New function.
(purge_dead_tablejump_edges): New function.
(find_bb_boundaries): Use it.
* gcc.dg/20050105-1.c: New test.
From-SVN: r93041
PR tree-optimization/18828
* builtins.c (expand_builtin_next_arg): Remove argument and all
the argument checking.
(expand_builtin): Adjust caller.
(expand_builtin_va_start): Likewise. Remove error for too many
arguments.
(fold_builtin_next_arg): Issue error for too many arguments.
After checking arguments, replace them with magic arguments that
prevent further checking of the args.
* gcc.dg/20050105-2.c: New test.
From-SVN: r93040
PR target/6077
* gcc.c (input_suffix_matches): Tweak the semantics of %{.s:...}
and %{.S:...} (and their negative variants) to test whether the
input file is assembler or pre-processed-assembler independent of
the actual filename extension.
From-SVN: r93023
2005-01-06 Daniel Berlin <dberlin@dberlin.org>
Fix PR tree-optimization/18792
* tree-data-ref.c (build_classic_dist_vector): Change first_loop
to first_loop_depth, and use loop depth instead of loop number.
(build_classic_dir_vector): Ditto.
(compute_data_dependences_for_loop): Use depth, not loop number.
* tree-loop-linear.c (try_interchange_loops): Use loop depth, not loop
number. Pass in loops, instead of loop numbers.
(gather_interchange_stats): Ditto.
(linear_transform_loops): Ditto.
From-SVN: r93008
* gcc.c (process_command): Change year in 'gcc --version' to 2005.
fortran/
* gfortranspec.c (lang_specific_driver): Change year to 2005 in
output of 'gfortran --version'.
From-SVN: r92995
2005-01-05 Daniel Berlin <dberlin@dberlin.org>
Fix PR middle-end/19286
Fix PR debug/19267
* dwarf2out.c (gen_subprogram_die): If we've already tried to
output this subprogram, simply ignore this attempt to do it again.
(add_abstract_origin_attribute): Don't abort trying to add the abstract
origin attribute if it's not possible.
(gen_block_die): Don't ignore subblocks of "unused" blocks.
(decls_for_scope): Ditto.
* gimple-low.c (mark_blocks_with_used_subblocks): Remove.
(mark_used_blocks): Don't call mark_blocks_with_used_subblocks.
From-SVN: r92991
PR target/11327
* config/i386/i386.c (BUILTIN_DESC_SWAP_OPERANDS): New.
(bdesc_2arg): Use it.
(ix86_expand_binop_builtin): Force operands into registers
when optimizing.
(ix86_expand_unop_builtin, ix86_expand_unop1_builtin,
ix86_expand_sse_compare, ix86_expand_sse_comi,
ix86_expand_builtin): Likewise.
From-SVN: r92988
* combine.c (expand_compound_operation) <ZERO_EXTRACT>: Add
comment that we fall through after case.
(mark_used_regs_combine): Don't expect a SIGN_EXTRACT in a SET_DEST.
(distribute_links): Likewise.
* cse.c (cse_insn): Likewise.
* cselib.c (cselib_invalidate_mem): Likewise.
* df.c: Update comments at the top of the file.
(read_modify_subreg_p): Update comments here too.
(df_def_record_1): Don't expect a SIGN_EXTRACT in a SET_DEST.
* flow.c (mark_set_1): Likewise.
(mark_used_regs): Likewise.
* gcse.c (mems_conflict_for_gcse_p): Likewise.
(canon_list_insert): Likewise.
(mark_set): Likewise.
(try_replace_reg): Likewise.
(store_killed_in_insn): Likewise.
* loop.c (count_one_set): Likewise.
(basic_induction_var): Likewise.
* postreload-gcse.c (find_mem_conflicts): Likewise.
* postreload.c (reload_combine_note_store): Likewise.
(move2add_note_store): Likewise.
* reload.c (find_equiv_reg): Likewise.
(mark_referenced_resources): Likewise.
* rtlanal.c (set_noop_p): Likewise.
(note_stores): Likewise.
(note_uses): Likewise.
* sched-deps.c (sched_analyze_1): Likewise.
* sched-rgn.c (check_live_1): Likewise.
(update_live_1): Likewise.
* config/i860/i860.c: Likewise.
* rtl.dec (SIGN_EXTRACT): Document that this cannot appear as
an lvalue.
(ZERO_EXTRACT): Mention that this one can be an lvalue.
* doc/rtl.texi: Update documentation for bit-fields and SET.
From-SVN: r92974
* unwind-dw2-fde-darwin.c (examine_objects): Use 64-bit
Mach-O getters if ppc64.
* config/darwin.c (darwin_asm_output_dwarf_delta): Obey
size argument.
* config/darwin.h (ASM_PREFERRED_EH_DATA_FORMAT): Use
signed four-byte field for global code case.
(STARTFILE_SPEC): Avoid crt2.o for 64-bit compilation.
From-SVN: r92962
PR target/16482
* lcm.c (create_pre_exit): New.
(optimize_mode_switching): In MODE_ENTRY / MODE_EXIT case, set
ENTRY_EXIT_EXTRA to 3. Use create_pre_exit.
Co-Authored-By: Kaz Kojima <kkojima@gcc.gnu.org>
From-SVN: r92950
* arm.h (TARGET_OPTIONS): Correctly record -mhard-float and
-msoft-float in target_float_switch.
* arm.c (arm_override_options): Fix processing of target_float_switch.
From-SVN: r92949
* ifcvt.c (find_if_case_1): Avoid creating an empty forwarder block,
if deleting the then-block allows the test-block to fallthru to the
else-block.
From-SVN: r92919
2005-01-04 Andrew Pinski <pinskia@physics.uc.edu>
PR c/19152
* c-decl.c (diagnose_mismatched_decls): Accept "extern inline" declared
after the full declaration if the are in two different TUs.
From-SVN: r92918
2005-01-04 Richard Henderson <rth@redhat.com>
PR tree-opt/19158
* tree-sra.c (generate_one_element_init): Just
call gimplify_and_add.
(generate_element_init): Record the
new referenced variables and mark them for renaming
and split out to ...
(generate_element_init_1): This.
(scalarize_init): Don't call push_gimplify_context/
pop_gimplify_context.
From-SVN: r92911
* fold-const.c (fold_single_bit_test): Delete unreachable handling
of TRUTH_NOT_EXPR.
(fold): Don't call fold_single_bit_test with a TRUTH_NOT_EXPR, as
all the cases handled by it are inverted by invert_truthvalue.
From-SVN: r92909
gcc:
2005-01-04 Paolo Bonzini <bonzini@gnu.org>
Devang Patel <dpatel@apple.com>
PR tree-optimization/18308
* tree-if-conv.c (add_to_dst_predicate_list): Gimplify
the operands before creating a new expression.
* dojump.c (do_jump): Make drop_through_label available
for all cases. Add expansion of COND_EXPR.
gcc/testsuite:
2005-01-04 Paolo Bonzini <bonzini@gnu.org>
* gcc.dg/vect/pr18308.c: New testcase.
Co-Authored-By: Devang Patel <dpatel@apple.com>
From-SVN: r92895
2005-01-04 Ira Rosen <irar@il.ibm.com>
* tree-vectorizer.c (vect_analyze_offset_expr): Test for
INTEGER_CST instead of TREE_CONSTANT.
(vect_gen_niters_for_prolog_loop): Test for INTEGER_CST
instead of TREE_CONSTANT.
(vect_analyze_pointer_ref_access): Test for INTEGER_CST
instead of TREE_CONSTANT.
From-SVN: r92894
Roger Sayle
Richard Sandiford
Eric Botcazou
Jakub Jelinek
David Edelsohn
Richard Henderson
John David Anglin
Geoffrey Keating
Daniel Berlin
Tobias Schlüter
Steven Bosscher
Jan Hubicka
Richard Henderson
Julian Brown
Stan Shebs
Joseph Myers
J"orn Rennecke
Richard Earnshaw
Uros Bizjak
Andrew Pinski
Paolo Bonzini
Ira Rosen