2007-01-25 Razya Ladelsky <razya@il.ibm.com>
* ipa-cp.c (ipcp_insert_stage, ipcp_driver): Support for SSA.
(ipcp_driver): Change to static definition.
Add dumping of the ifunctions.
(constant_val_insert): Remove unused parameter. Support for SSA.
(ipcp_propagate_const): Support for SSA.
(ipcp_profile_bb_print): Print only analyzed nodes.
(ipcp_replace_map_create): Remove support for Fortran constant
for now.
* ipa-prop.c (ipa_method_modify_stmt,
ipa_callsite_compute_param): Support for SSA.
* ipa-prop.h (ipcp_driver): Remove declaration.
(IS_VALID_TREE_MAP_INDEX): Add define.
From-SVN: r121167
PR middle-end/30447
* builtins.c (fold_builtin_cabs): Use MPFR to evaluate a
constant argument to cabs and do it without checking for
-funsafe-math-optimizations.
From-SVN: r121163
* ipa-inline.c (initial_insns, max_insns): Delete.
(compute_max_insns): New function.
(cgraph_decide_inlining_of_small_function): Use it; take minimal amount
of insns as base for code growth.
(cgraph_decide_inlining): Make initial_insns local; do not compute
max_insns.
* params.def (PARAM_INLINE_UNIT_GROWTH): Set to 60.
* doc/invoke.texi (inline-unit-growth): Update docs.
From-SVN: r121144
PR middle-end/27416
* gimplify.c (omp_check_private): New function.
(gimplify_scan_omp_clauses): Use it for
firstprivate/lastprivate/reduction.
* gcc.dg/gomp/pr27416.c: New test.
From-SVN: r121133
PR middle-end/30421
* omp-low.c (lower_omp_for_lastprivate): Add dlist argument.
If lower_lastprivate_clauses emits some statements, append them
to dlist rather than body_p and to body_p append an initializer.
(lower_omp_for): Adjust caller.
* gcc.dg/gomp/pr30421.c: New test.
From-SVN: r121131
* gcc.dg/tree-ssa/pr23382.c: Avoid DCE from eliminating dead variable.
* tree-ssa-dce.c (eliminate_unnecesary_stmts): Remove dead LHS of calls.
From-SVN: r121108
2007-01-24 Andreas Krebbel <krebbel1@de.ibm.com>
* c-cppbuiltin.c (builtin_define_type_sizeof): New function.
(c_cpp_builtins): New builtin macros: __SIZEOF_INT__, __SIZEOF_LONG__,
__SIZEOF_LONG_LONG__, __SIZEOF_SHORT__, __SIZEOF_POINTER__,
__SIZEOF_FLOAT__, __SIZEOF_DOUBLE__, __SIZEOF_LONG_DOUBLE__,
__SIZEOF_SIZE_T__, __SIZEOF_WCHAR_T__, __SIZEOF_WINT_T__ and
__SIZEOF_PTRDIFF_T__.
* doc/cpp.texi: Documentation for the new builtin macros added.
2007-01-24 Andreas Krebbel <krebbel1@de.ibm.com>
* gcc.c-torture/compile/sizeof-macros-1.c: New testcase.
From-SVN: r121107
* config/i386/i386.md (tanxf2, tan<mode>2, atan<mode>2, log<mode>2,
log10<mode>2, log2<mode>2, expxf2, exp10xf2, exp2xf2): Use op2
instead of operands[2] to avoid access past the end of array.
From-SVN: r121106
gcc:
* config/rs6000/rs6000-c.c (rs6000_cpu_cpp_builtins): Define
_SOFT_DOUBLE if doubles use software floating-point.
* config/rs6000/libgcc-ppc-glibc.ver: Export additional long
double functions if _SOFT_DOUBLE, not _SOFT_FLOAT.
* config/rs6000/darwin-ldouble.c: Also compile functions for
hard-float without FPRs. Use fmsub function for all __NO_FPRS__
cases. Compile extra functions if _SOFT_DOUBLE, not _SOFT_FLOAT.
* config/rs6000/linuxspe.h (SUBSUBTARGET_OVERRIDE_OPTIONS): Remove
commented-out long double override.
(CPP_LONGDOUBLE_DEFAULT_SPEC): Likewise.
* config/rs6000/eabispe.h: Likewise.
* config/rs6000/rs6000.c (rs6000_override_options): Don't override
long double for non-SPE.
(rs6000_handle_option): Likewise.
(invalid_e500_subreg): Disallow more subregs involding DImode,
DFmode, TImode or TFmode.
(rs6000_legitimate_offset_address_p): Check TFmode offsets for
E500 double.
(legitimate_lo_sum_address_p): Also check for TFmode for E500
double.
(rs6000_legitimize_address): Also handle TFmode for E500 double.
(rs6000_legitimize_reload_address): Also handle TFmode for E500
double.
(rs6000_legitimate_address): Also check for TFmode for E500
double.
(rs6000_emit_move): Use DFmode subregs of TFmode for E500 double.
(spe_build_register_parallel): Handle TFmode and TCmode.
(rs6000_spe_function_arg): Handle TFmode and TCmode for E500
double.
(function_arg): Handle TFmode and TCmode for E500 double.
(rs6000_init_libfuncs): Initialize extra libfuncs for soft double
in general.
(print_operand): Handle TFmode and TImode for %y.
(rs6000_generate_compare): Handle TFmode comparisons for E500
double.
(spe_func_has_64bit_regs_p): Check for TFmode for E500 double.
(rs6000_function_value): Handle TFmode and TCmode for E500 double.
(rs6000_libcall_value): Handle TFmode and TCmode for E500 double.
* config/rs6000/rs6000.h (CANNOT_CHANGE_MODE_CLASS): Check for
TFmode for E500 double.
* config/rs6000/rs6000.md (FP): Allow TF for E500 double.
(floatsidf2): Enable for E500 double.
(movtf_softfloat): Use rs6000_nonimmediate_operand.
(extenddftf2): Change to extenddftf2_fprs.
(extenddftf2): Call gen_spe_extenddftf2 or gen_extenddftf2_fprs
depending on TARGET_E500_DOUBLE.
(extendsftf2): Enable for E500 double.
(trunctfdf2): Enable for E500 double.
(trunctfsf2): Change to trunctfsf2_fprs.
(trunctfsf2): Call gen_spe_trunctfsf2 or gen_trunctfsf2_fprs
depending on TARGET_E500_DOUBLE.
(floatsitf2): Enable for E500 double.
(fix_trunctfsi2): Change to fix_trunctfsi2_fprs.
(fix_trunctfsi2): Call gen_spe_fix_trunctfsi2 or
gen_fix_trunctfsi2_fprs depending on TARGET_E500_DOUBLE.
(negtf2): Change to negtf2_internal.
(negtf2): New expander.
(abstf2): Enable for E500 double. Call gen_spe_abstf2_tst,
gen_spe_abstf2_cmp or gen_abstf2_internal depending on
TARGET_E500_DOUBLE and flag_unsafe_math_optimizations.
(movdi_internal32): Use rs6000_nonimmediate_operand.
(unnamed splitter): Likewise.
* config/rs6000/spe.md (CMPTFEQ_GPR, TSTTFEQ_GPR, CMPTFGT_GPR,
TSTTFGT_GPR, CMPTFLT_GPR, TSTTFLT_GPR): New unspecs.
(SPE64TF, DITI): New mode macros.
(frob_df_di): Change to frob_<SPE64:mode>_<DITI:mode>; allow more
modes.
(frob_tf_ti): New.
(frob_<mode>_di_2): New.
(frob_tf_di_8_2): New.
(frob_di_df): Change to frob_di_<mode>; allow more modes.
(frob_ti_tf): New.
(frob_di_df_2): Change to frob_<DITI:mode>_<SPE64:mode>_2; allow
more modes.
(frob_ti_<mode>_8_2): New.
(frob_ti_tf_2): New.
(mov_si<mode>_e500_subreg0, mov_si<mode>_e500_subreg0_2,
mov_si<mode>_e500_subreg4, mov_si<mode>_e500_subreg4_2): Allow
TFmode.
(mov_sitf_e500_subreg8, mov_sitf_e500_subreg8_2,
mov_sitf_e500_subreg12, mov_sitf_e500_subreg12_2): New.
(spe_trunctfdf2_internal1, spe_trunctfsf2, spe_extenddftf2,
spe_fix_trunctfsi2, spe_fix_trunctfsi2_internal,
spe_negtf2_internal, spe_abstf2_cmp, spe_abstf2_tst): New.
(cmptfeq_gpr, tsttfeq_gpr, cmptfgt_gpr, tsttfgt_gpr, cmptflt_gpr,
tsttflt_gp): New.
libgcc:
* config/rs6000/t-ldbl128: Always use -mlong-double-128.
From-SVN: r121085
PR rtl-optimization/29329
* combine.c (replaced_rhs_insn): Rename to i2mod.
(replaced_rhs_value): Rename to i2mod_new_rhs.
(i2mod_old_rhs): New global variable.
(combine_instructions): Adjust for above change. Save a copy of
the old RHS into i2mod_old_rhs when the contents of a REG_EQUAL
note are substituted in the second instruction.
(distribute_notes) <REG_DEAD>: Adjust for above change. Do not
ditch the note if it pertains to the second eliminated register
and this register is mentioned in i2mod_old_rhs.
Revert:
2006-09-12 Eric Botcazou <ebotcazou@libertysurf.fr>
* combine.c (distribute_notes) <REG_DEAD>: Do not consider SETs past
the insn to which the note was originally attached.
From-SVN: r121037
2007-01-21 Dirk Mueller <dmueller@suse.de>
PR bootstrap/30511
* tree-vrp.c (check_array_bounds): do not warn
about ADDR_EXPR's of ARRAY_REF's which are immediately
used in binary expressions.
From-SVN: r121032
* tree-vectorizer.h (struct _stmt_vec_info): Add new field
read_write_dep and macros for its access.
* tree-vectorizer.c (new_stmt_vec_info): Initialize the new field.
* tree-vect-analyze.c (vect_analyze_data_ref_dependence): Remove
argument, call vect_check_interleaving for every independent pair of
data-refs. Mark loads that access the same memory location as a store
in the loop.
(vect_check_dependences): Remove.
(vect_analyze_data_ref_dependences): Remove vect_check_dependences
call, fix the call to vect_analyze_data_ref_dependence.
(vect_analyze_data_ref_access): For statements that access the same
data-ref, check that they are not stores; for loads, check that there
is no store that access the same location.
From-SVN: r121026
* doc/invoke.texi (-fdump-rtl-combine): Fix under/overfull hbox.
(-fdump-rtl-stack): Likewise.
(-fno-signed-zeros): Use @minus{} for a minus sign. Correct typo.
(-mcheck-zero-division, -mcpu): Fix under/overfull hbox.
(-mpt-fixed): Use @minus{} for minus sign.
(Using Precompiled Headers): Fix under/overfull hbox.
Co-Authored-By: Joseph Myers <joseph@codesourcery.com>
From-SVN: r121023
* tree-ssa-operands.h (struct vuse_vec_d): Change num_vuse field
to unsigned.
(VUSE_VECT_ELEMENT) [ENABLE_CHECKING]: Use unsigned comparison.
(VUSE_ELEMENT_PTR) [ENABLE_CHECKING]: Likewise.
(SET_VUSE_VECT_ELEMENT) [ENABLE_CHECKING]: Likewise.
(SET_VUSE_ELEMENT_VAR) [ENABLE_CHECKING]: Likewise.
(SET_VUSE_ELEMENT_PTR) [ENABLE_CHECKING]: Likewise.
(realloc_vdef, realloc_vuse): Change second parameter to
unsigned.
(ssa_operand_iterator_d): Change vuse_index and mayuse_index
fields to unsigned.
* tree-ssa-operands.c (realloc_vop): Change num_elem parameter to
unsigned. Change x and lim locals to unsigned.
(realloc_vdef, realloc_vuse): Change num_elem parameter to
unsigned.
(finalize_ssa_vuse_ops): Change old_i local to unsigned.
(copy_virtual_operands): Change i and n locals to unsigned.
From-SVN: r120994
* tree-ssa.c (init_tree_ssa): Do not call init_alias_heapvars.
* tree-ssa-structalias.c (compute_points_to_sets): Do call
init_alias_heapvars.
(init_alias_heapvars): Initialize only when not already initialized.
(delete_alias_heapvars): Set heapvar_for_stmt to NULL.
From-SVN: r120983
* common.opt (fsigned-zeros): New command line option.
* flags.h (HONOR_SIGNED_ZEROS): Control via flag_signed_zeros instead
of flag_unsafe_math_optimizations.
* opts.c (set_fast_math_flags): The -ffast-math command line option
implies -fno-signed-zeros.
(fast_math_flags_set_p): Likewise.
* doc/invoke.texi: Document new -fno-signed-zeros option, and update
the documentation of -ffast-math appropriately. Wrap long lines.
* gcc.dg/pr30172-1.c: Specify the -fno-signed-zeros command line
option instead of -funsafe-math-optimizations.
From-SVN: r120978
Patch by Tomas Bily <tbily@suse.cz>
* cgraphunit.c (cgraph_finalize_function): Updating of pid
* tree-profile.c:
(tree_init_ic_make_global_vars): New function
(tree_init_edge_profiler): call of tree_init_ic_make_global_vars
(tree_gen_ic_profiler): New function
(tree_gen_ic_func_profiler): New function
(tree_profiling): Added calling of tree_gen_ic_func_profiler
(tree_profile_hooks): Added hook for indirec/virtual calls
* value-prof.c (tree_find_values_to_profile): New case for
indirect calls
(tree_values_to_profile): Call for determining indirect/virtual
counters
(tree_indirect_call_to_profile): New function
(tree_ic_transform): New function
(tree_ic): New function
(find_func_by_pid): New function
(init_pid_map): New function
(tree_value_profile_transformations): Added check for
indirect/virtual call transformation
* value-prof.h (enum hist_type): New counter type for
indirect/virtual calls
(profile_hooks): Added new hook for profiling indirect/virtual
calls
* profile.c (instrument_values): New case for indirect/virtual
call added
* gcov-io.h (GCOV_LAST_VALUE_COUNTER): Changed to 6
(GCOV_COUNTER_V_INDIR): New counter type
(GCOV_COUNTER_NAMES): New name of counter "indirect" added
(GCOV_MERGE_FUNCTIONS): New merge function for indirect/virtual
call added
* cgraph.c: Definition of cgraph_max_pid
(cgraph_create_node): Default init of pid attribute
* cgraph.h: Declaration of cgraph_max_pid
(struct cgraph_node): Added pid attribute
* libgcov.c (__gcov_indirect_call_profiler): New function
(__gcov_one_value_profiler_body): New function
(__gcov_one_value_profiler): Body was moved to
__gcov_one_value_profiler_body and calls it
gcc.dg/tree-prof/indir-call-prof.c: New.
g++.dg/dg.exp: Add tree-prof subdirectory.
g++.dg/tree-prof/indir-call-prof.C: New.
g++.dg/tree-prof/tree-prof.exp: New.
From-SVN: r120975
* ipa-inline.c (cgraph_decide_inlining_incrementally): Instead of 'early' argument
take inlining mode argument specifying whether to inline for size/speeed or all
functions; add support for flattening; improve dumpting.
(cgraph_early_inlining): Update call of decide_inlining_incrementally.
From-SVN: r120970
2007-01-19 Manuel Lopez-Ibanez <manu@gcc.gnu.org>
PR c++/17947
* toplev.c (warn_deprecated_use): Use %qD instead of %qs to
print the name of the declared identifier.
testsuite/
* g++.dg/warn/deprecated.C: Update warning output.
* g++.dg/warn/deprecated-2.C: Likewise.
* g++.dg/warn/deprecated-3.C: New.
From-SVN: r120969
2007-01-19 Dirk Mueller <dmueller@suse.de>
* config/i386.h (CONDITIONAL_REGISTER_USAGE): Store
result of PIC_OFFSET_TABLE_REGNUM in temporary variable to avoid
duplicate evaluation.
From-SVN: r120967
* config/i386/i386.md (acos<mode>2): Rename from acossf2 and acosdf2.
Macroize expander using X87MODEF12 mode macro. Extend operand 1
to XFMode, use acosxf2 and truncate result to requested mode.
Use SSE_FLOAT_MODE_P to disable patterns for SSE math.
(asin<mode>2): Similarly, with asin expanders.
(*fscalexf4_i387): Rename from *fscalexf4.
(expNcorexf3): New expander.
(expxf2, exp10xf2, exp2xf2): Use expNcorexf3 expander.
(exp<mode>2): Rename from expsf2 and expdf2. Macroize expander using
X87MODEF12 mode macro. Extend operand 1 to XFMode, use expxf2 and
truncate result to requested mode. Use SSE_FLOAT_MODE_P to disable
patterns for SSE math.
(exp10<mode>2): Similarly, with exp10 expanders.
(exp2<mode>2): Similarly, with exp2 expanders.
(expm1<mode>2): Similarly, with expm1 expanders.
(ldexp<mode>3): Similarly, with ldexp expanders.
(log<mode>2, log10<mode>2, log2<mode>2, log1p<mode>2, logb<mode>2):
Use gen_truncxf<mode>2_i387_noop to truncate result.
From-SVN: r120966
DJ Delorie
Geoffrey Keating
Steve Ellcey
Richard Guenther
Razya Ladelsky
Geoffrey Keating
Kaveh R. Ghazi
Douglas Gregor
Jan Hubicka
Jakub Jelinek
Steve Ellcey
Andreas Krebbel
Uros Bizjak
Richard Sandiford
Joseph Myers
Ian Lance Taylor
Uros Bizjak
Andreas Schwab
Eric Botcazou
Dirk Mueller
Ira Rosen
Roger Sayle
Basile Starynkevitch
John David Anglin
Tomas Bily
Manuel López-Ibáñez