Jan Hubicka [Tue, 28 Nov 2006 14:20:04 +0000 (15:20 +0100)]
builtins.c: Include tree-flow.h.
* builtins.c: Include tree-flow.h.
(fold_builtin_memory_op): Be more aggressive on converting memcpy to
assignment; convert memmove to memcpy for sizes greater than 1 where
alignment of operands prohibit the partial overlap.
Roger Sayle [Mon, 27 Nov 2006 18:47:41 +0000 (18:47 +0000)]
c-common.c (constant_fits_type_p): Delete.
2006-11-27 Roger Sayle <roger@eyesopen.com>
Manuel Lopez-Ibanez <manu@gcc.gnu.org>
* c-common.c (constant_fits_type_p): Delete.
(unsigned_conversion_warning): Delete.
(conversion_warning): Integrate Wconversion warning from
unsigned_conversion_warning.
(convert_and_check): Integrate Woverflow warning from
unsigned_conversion_warning. Reorganize and simplify to avoid
dependence upon the middle-end setting TREE_OVERFLOW on integral
conversions, by using int_fits_type_p directly.
Co-Authored-By: Manuel López-Ibáñez <manu@gcc.gnu.org>
From-SVN: r119258
Michael Matz [Mon, 27 Nov 2006 16:34:19 +0000 (16:34 +0000)]
re PR target/29319 (ICE unrecognizable insn: offset too large for larl (breaks glibc))
2006-11-27 Michael Matz <matz@suse.de>
Andreas Krebbel <krebbel1@de.ibm.com>
PR target/29319
* config/s390/predicates.md (larl_operand): Check addend of larl
operand to be in range of -/+2GB.
* config/s390/s390.c (legitimize_pic_address): Likewise.
Changed type of variable even to HOST_WIDE_INT.
2006-11-27 Michael Matz <matz@suse.de>
Andreas Krebbel <krebbel1@de.ibm.com>
PR target/29319
* gcc.dg/20061127-1.c: New testcase.
Co-Authored-By: Andreas Krebbel <krebbel1@de.ibm.com>
From-SVN: r119256
Richard Guenther [Mon, 27 Nov 2006 11:38:42 +0000 (11:38 +0000)]
re PR fortran/25620 (Missed optimization with power)
2006-11-27 Richard Guenther <rguenther@suse.de>
PR middle-end/25620
* builtins.c (expand_builtin_pow): Optimize non integer valued
constant exponents using sqrt or cbrt if possible. Always fall back
to expanding via optabs.
* gcc.target/i386/pow-1.c: New testcase.
* gcc.dg/builtins-58.c: Likewise.
Ira Rosen [Mon, 27 Nov 2006 08:10:23 +0000 (08:10 +0000)]
re PR tree-optimization/22372 (Vectorizer produces mis-match types)
PR tree-optimization/22372
* tree-vect-transform.c (vect_create_addr_base_for_vector_ref): Convert
the type of the base address.
(vect_create_data_ref_ptr): Add an argument, type of the data-ref. Use
it in the data-ref pointer creation.
(vect_init_vector): Add an argument, type of the init vector. Use it in
the stmt creation.
(vect_get_vec_def_for_operand): Fix calls to vect_init_vector.
(get_initial_def_for_reduction): Likewise.
(vect_create_epilog_for_reduction): Change type of MODIFY_EXPR to
void_type_node.
(vectorizable_assignment, vect_permute_store_chain): Likewise.
(vectorizable_store): Fix call to vect_create_data_ref_ptr. Use the
correct type in stmt creation.
(vect_setup_realignment): Use the correct type in stmt creation.
(vect_permute_load_chain): Change type of MODIFY_EXPR to
void_type_node.
(vectorizable_load): Fix calls to vect_create_data_ref_ptr. Change type
of MODIFY_EXPR to void_type_node.
(vectorizable_condition): Change type of MODIFY_EXPR to
void_type_node.
(vect_update_ivs_after_vectorizer): Use the correct type in stmt
creation.
Uros Bizjak [Mon, 27 Nov 2006 08:03:45 +0000 (09:03 +0100)]
revert: i386.c (ix86_rtx_costs): For standard 80387 constants...
Revert:
2006-05-08 Uros Bizjak <uros@kss-loka.si>
* config/i386/i386.c (ix86_rtx_costs) [FLOAT_EXTEND]: For
standard 80387 constants, raise the cost to prevent
compress_float_constant() to generate load from memory.
* config/i386/i386.md: Add new splitter pattern to split
float_extended load of constant from constant pool into
pure constant load.
gcc:
* doc/install.texi: Move GMP/MPFR configure options from Fortran
section to general section. Remove documentation for
--with-mpfr-dir and --with-gmp-dir, add documentation for
--with-mpfr-include, --with-mpfr-lib, --with-gmp-include and
--with-gmp-lib.
gcc:
* doc/install.texi: Move GMP/MPFR configure options from Fortran
section to general section. Remove documentation for
--with-mpfr-dir and --with-gmp-dir, add documentation for
--with-mpfr-include, --with-mpfr-lib, --with-gmp-include and
--with-gmp-lib.
re PR fortran/29892 (substring out of bounds: Missing variable name for variables with parameter attribute)
PR fortran/29892
* trans-intrinsic.c (gfc_conv_intrinsic_bound): Use a locus in
the call to gfc_trans_runtime_check.
* trans-array.c (gfc_trans_array_bound_check): Try harder to find
the variable or function name for the runtime error message.
(gfc_trans_dummy_array_bias): Use a locus in the call to
gfc_trans_runtime_check
Paolo Carlini [Sun, 26 Nov 2006 10:04:25 +0000 (10:04 +0000)]
PR libstdc++/29385 (2nd part, based on an idea by Ion Gaztanaga)
2006-11-26 Paolo Carlini <pcarlini@suse.de>
PR libstdc++/29385 (2nd part, based on an idea by Ion Gaztanaga)
* include/bits/stl_tree.h (_Rb_tree<>::_M_equal_range): Add.
(equal_range(const key_type&)): Use it.
Razya Ladelsky [Sun, 26 Nov 2006 10:00:18 +0000 (10:00 +0000)]
re PR tree-optimization/29122 (ICE with -ipa-cp and -m64 (tail calls))
2006-11-26 Razya Ladklesky <razya@il.ibm.com>
PR tree-optimization/29122
* tree-inline.c (tree_function_versioning): Update DECL_ASSEMBLER_NAME
for the new version. Assign NULL to DECL_RTL of the new version.
Steven Bosscher [Sat, 25 Nov 2006 21:23:27 +0000 (21:23 +0000)]
cse.c (cse_altered): Remove.
* cse.c (cse_altered): Remove.
(record_jump_equiv): Make true/false argument a bool instead
of an int.
(cse_insn): Don't set cse_altered. Use delete_insn_and_edges
where appropriate. Emit a new jump before the existing one
instead of after so that delete_insn_and_edges removes the
dead edges properly. Check any_condjump_p before calling
record_jump_equiv.
(cse_basic_block): Check any_condjump_p before calling
record_jump_equiv.
(cse_main): Don't set/check cse_altered. Remove USE_C_ALLOCA.
Uros Bizjak [Sat, 25 Nov 2006 15:05:53 +0000 (16:05 +0100)]
config/i386/i386.md (fpremxf4): Rename to fpremxf4_i387.
(fprem1xf4): Rename to fprem1xf4_i387.
(fmodsf3, fmoddf3): Macroize patterns using X87MODEF12 mode macro.
Rename patterns to fmod<mode>3. Use general_operand operand
constraint for operands 1 and 2. Use SSE_FLOAT_MODE_P to disable
patterns for SSE math.
(remaindersf3, remainderdf3): Ditto.
Steven Bosscher [Sat, 25 Nov 2006 10:53:06 +0000 (10:53 +0000)]
jump.c (mark_all_labels): Work in cfglayout mode.
* jump.c (mark_all_labels): Work in cfglayout mode.
* cfgcleanup.c (cleanup_cfg): Do not call delete_dead_jumptables
when in cfglayout mode, because there are no dead jumptables
visible.
* cfgrtl.c (commit_one_edge_insertion): Don't set bb->aux when
in cfglayout mode.
(commit_edge_insertions): Do not allow insertion of instructions
with control flow insns when in cfglayout mode.
Uros Bizjak [Sat, 25 Nov 2006 09:17:57 +0000 (10:17 +0100)]
config/i386/i386.md (UNSPEC_TRUNC_NOOP): New unspec definition.
(X87MODEF): New mode macro.
(ssemodefsuffix): New mode attribute.
(truncxf<mode>2_i387_noop_unspec): New insn pattern.
(sqrt_extend<mode>xf2_i387): New insn pattern.
(sqrt<mode>2): For non-SSE sqrt, emit sqrt_extend<mode>xf2_i387
insn and truncate result back to original mode using
UNSPEC_TRUNC_NOOP truncation.
(*sqrt<mode>2_sse): Implement using SSEMODEF mode macro and
ssemodefsuffix mode attribute.
(*sqrtsf2_mixed, *sqrtsf2_i387, *sqrtdf2_mixed, *sqrtdf2_i387)
(*sqrtextendsfdf2_i387, *sqrtextendsfxf2_i387)
(*sqrtextenddfxf2_i387): Remove insn patterns.
(fmodsf3, fmoddf3, remaindersf3, remainderdf3): Use noop
truncation patterns.
Paul Thomas [Fri, 24 Nov 2006 22:22:40 +0000 (22:22 +0000)]
re PR fortran/20880 (USE association of procedure's own interface)
2006-11-24 Paul Thomas <pault@gcc.gnu.org>
PR fortran/20880
* parse.c (parse_interface): Error if procedure name is that of
encompassing scope.
* resolve.c (resolve_fl_procedure): Error if procedure is
ambiguous.
PR fortran/29387
* interface.c (compare_actual_formal): Add missing condition
that 'where' be present for error that asserts that actual
arguments be definable.
2006-11-24 Paul Thomas <pault@gcc.gnu.org>
PR fortran/20880
* gfortran.dg/interface_3.f90: New test.
PR fortran/29387
* gfortran.dg/generic_8.f90: New test.
Andrew Pinski [Fri, 24 Nov 2006 18:32:14 +0000 (18:32 +0000)]
re PR tree-optimization/29964 (function with volatile operators still found to be pure)
2006-11-24 Andrew Pinski <andrew_pinski@playstation.sony.com>
PR tree-opt/29964
* ipa-pure-const.c (check_tree): If the original tree
is volatile return early and say the function is not pure
nor const. Remove the volatile check for writes.
(analyze_function): Print out the result of the local
analysis pass.
2006-11-24 Andrew Pinski <andrew_pinski@playstation.sony.com>
Joseph Myers [Fri, 24 Nov 2006 16:25:25 +0000 (16:25 +0000)]
eabispe.h (TARGET_DEFAULT): Include MASK_STRICT_ALIGN.
* config/rs6000/eabispe.h (TARGET_DEFAULT): Include
MASK_STRICT_ALIGN.
* config/rs6000/linuxspe.h (TARGET_DEFAULT): Likewise.
* config/rs6000/rs6000.c (rs6000_override_options): Use
MASK_STRICT_ALIGN for 8540 and 8548. Add MASK_STRICT_ALIGN to
POWERPC_MASKS.
re PR c/9072 (-Wconversion should be split into two distinct flags)
2006-11-23 Manuel Lopez-Ibanez <manu@gcc.gnu.org>
PR c/9072
* c.opt (Wtraditional-conversion): New.
(Wconversion): Update description.
* c-typeck.c (convert_arguments): Warnings for prototypes causing
type conversions different from what would happen in the absence
of prototype are now handled by Wtraditional-conversion.
* doc/invoke.texi (Wtraditional-conversion): New.
(Wconversion): Update description.
* doc/trouble.texi (Wconversion): Replace Wconversion by
Wtraditional-conversion.
testsuite/
PR c/9072
* gcc.dg/builtin-protos-1.c: Replace Wconversion by
Wtraditional-conversion.
* gcc.dg/overflow-warn-2.c: Likewise.
* gcc.dg/Wconversion.c: Likewise. Renamed as
Wtraditional-conversion.c .
* gcc.dg/Wconversion-2.c: Likewise. Renamed as
Wtraditional-conversion-2.c .
* gcc.dg/dfp/Wconversion-2.c: Likewise. Renamed as
Wtraditional-conversion-2.c
* gcc.dg/Wconversion-negative-constants.c: New.
Andrew Pinski [Thu, 23 Nov 2006 17:59:53 +0000 (09:59 -0800)]
predict.c (tree_estimate_probability): Check to make sure current_loops is non null before calling flow_loops_dump.
2006-11-23 Andrew Pinski <pinskia@gmail.com>
* predict.c (tree_estimate_probability): Check to make
sure current_loops is non null before calling flow_loops_dump.
2006-11-23 Andrew Pinski <pinskia@gmail.com>
gcc:
* config/mips/mips.h (ISA_HAS_FP4): Add MIPS32R2 + 64bit fpu
combination.
(ISA_HAS_MXHC1): True if ISA supports mfhc1 and mthc1 opcodes.
(ASM_SPEC): Pass along -mfp32 and -mfp64.
* config/mips/mips.c (mips_split_64bit_move): Use gen_mthc1 to set
high part of FP register when in 64-bit FP register mode. Similarly
use gen_mfhc1 to load high part of FP register.
(override_options): Allow -mgp32 and -mfp64 combination if
ISA_HAS_MXHC1 (currently for O32 only).
(mips_cannot_change_mode_class): If floating-point registers are
bigger than word size. disallow conversion of float register from a
large integer mode to a float mode smaller than the float register
size.
(mips_class_max_nregs): Handle float registers case seperately.
* config/mips/mips.md (define_constants): Add UNSPEC_MFHC1,
UNSPEC_MTHC1.
(movdi_32bit): Use !TARGET_FLOAT64 in condition pattern.
(movdf_hardfloat_32bit): Similarly.
(movdi_gp32_fp64): New DImode pattern for MIPS32R2 which optionally
support a full 64-bit fpu.
(mthc1): New pattern to generate MTHC1 instruction.
(mfhc1): New pattern to generate MFHC1 instruction.
* doc/invoke.texi (MIPS Options): Document the -mgp32 -mfp64
option for the MIPS32R2 and mention its use under O32 ABI.
gcc/testsuite:
* gcc.target/mips/mips.exp (dg-mips-options): Handle parsing of
-mfp64, allowable when ISA >= 33 and float is enabled.
* gcc.target/mips/mips32r2-mxhc1.c: New test for checking the use
of mthc1 and mfhc1 patterns.
Joseph Myers [Wed, 22 Nov 2006 16:41:29 +0000 (16:41 +0000)]
spe.md (SPE64): New mode macro.
* config/rs6000/spe.md (SPE64): New mode macro.
(mov_sidf_e500_subreg0): Change to mov_si<mode>_e500_subreg0. Add
memory load.
(mov_si<mode>_e500_subreg0_2): New.
(mov_sidf_e500_subreg4): Change to mov_si<mode>_e500_subreg4. Add
memory load.
(mov_si<mode>_e500_subreg4_2): New.
* config/rs6000/predicates.md (input_operand): Do not allow
invalid E500 subregs.
(rs6000_nonimmediate_operand): Check for invalid E500 subregs also
if TARGET_SPE.
* config/rs6000/rs6000.c (invalid_e500_subreg): Check for subregs
involving DFmode if TARGET_E500_DOUBLE. Check for subregs
involving vector modes if TARGET_SPE.
Bernd Schmidt [Wed, 22 Nov 2006 14:12:46 +0000 (14:12 +0000)]
predicates.md (d_register_operand, [...]): New predicates.
* config/bfin/predicates.md (d_register_operand, mem_p_address_operand,
mem_i_address_operand): New predicates.
* config/bfin/bfin.c (bfin_issue_rate): New function.
(TARGET_SCHED_ISSUE_RATE): New macro.
* config/bfin/bfin.md (addrtype): New attribute.
(slot0, slot1, slot2, store, pregs): New cpu_units.
(core): Now a define_reservation.
(alu): Remove some insn types from this reservation.
(dsp32, load32, loadp, loadi, store32, storep, storei, multi): New
insn reservations.
(dummy reservation): Don't trigger for mcld insns.
(absence_sets): Two new absence sets to enforce slot ordering.
(popsi_insn): Set addrtype.
Ira Rosen [Wed, 22 Nov 2006 08:46:03 +0000 (08:46 +0000)]
c-tree.texi: Document new tree codes.
* doc/c-tree.texi: Document new tree codes.
* doc/md.texi: Document new optabs.
* tree-pretty-print.c (dump_generic_node): Handle print of new tree
codes.
* optabs.c (optab_for_tree_code, init_optabs): Handle new optabs.
* optabs.h (optab_index): Add new.
(vec_extract_even_optab, vec_extract_odd_optab,
vec_interleave_high_optab, vec_interleave_low_optab): New optabs.
* genopinit.c (vec_extract_even_optab, vec_extract_odd_optab,
vec_interleave_high_optab, vec_interleave_low_optab): Initialize
new optabs.
* expr.c (expand_expr_real_1): Add implementation for new tree codes.
* tree-vectorizer.c (new_stmt_vec_info): Initialize new fields.
* tree-vectorizer.h (stmt_vec_info): Add new fields for interleaving
along with macros for their access.
* tree-data-ref.h (first_location_in_loop, data_reference): Update
comment.
* tree-vect-analyze.c (toplev.h): Include.
(vect_determine_vectorization_factor): Fix indentation.
(vect_insert_into_interleaving_chain,
vect_update_interleaving_chain, vect_equal_offsets): New functions.
(vect_analyze_data_ref_dependence): Add argument for interleaving
check. Check for interleaving if it's true.
(vect_check_dependences): New function.
(vect_analyze_data_ref_dependences): Call vect_check_dependences for
every ddr. Call vect_analyze_data_ref_dependence with new argument.
(vect_update_misalignment_for_peel): Update for interleaving.
(vect_verify_datarefs_alignment): Check only first data-ref for
interleaving.
(vect_enhance_data_refs_alignment): Update for interleaving. Check
only first data-ref for interleaving.
(vect_analyze_data_ref_access): Check interleaving, update
interleaving data.
(vect_analyze_data_refs): Call compute_data_dependences_for_loop
with different parameters.
* tree.def (VEC_EXTRACT_EVEN_EXPR, VEC_EXTRACT_ODD_EXPR,
VEC_INTERLEAVE_HIGH_EXPR, VEC_INTERLEAVE_LOW_EXPR): New tree codes.
* tree-inline.c (estimate_num_insns_1): Add cases for new codes.
* tree-vect-transform.c (vect_create_addr_base_for_vector_ref):
Update step in case of interleaving.
(vect_strided_store_supported, vect_permute_store_chain): New
functions.
(vectorizable_store): Handle strided stores.
(vect_strided_load_supported, vect_permute_load_chain,
vect_transform_strided_load): New functions.
(vectorizable_load): Handle strided loads.
(vect_transform_stmt): Add argument. Handle strided stores. Check
that vectorized stmt exists for patterns.
(vect_gen_niters_for_prolog_loop): Update calculation for
interleaving.
(vect_transform_loop): Remove stmt_vec_info for strided stores after
whole chain vectorization.
* config/rs6000/altivec.md (UNSPEC_EXTEVEN, UNSPEC_EXTODD,
UNSPEC_INTERHI, UNSPEC_INTERLO): New constants.
(vpkuhum_nomode, vpkuwum_nomode, vec_extract_even<mode>,
vec_extract_odd<mode>, altivec_vmrghsf, altivec_vmrglsf,
vec_interleave_high<mode>, vec_interleave_low<mode>): Implement.
Jerry DeLisle [Wed, 22 Nov 2006 07:32:09 +0000 (07:32 +0000)]
io.h (unit_flags): Add new flag has_recl.
2006-11-22 Jerry DeLisle <jvdelisle@gcc.gnu.org>
* io/io.h (unit_flags): Add new flag has_recl.
* io.open.c (new_unit): Set flag if RECL= was specified.
* io/transfer.c (us_write): If flag set, leave recl as initialized by
new_unit.
projects / thirdparty / gcc.git / log
