Douglas Gregor [Wed, 26 Mar 2008 04:17:57 +0000 (04:17 +0000)]
pt.c (coerce_template_template_parm): Moved the body of the loop of coerce_template_template_parms here...
2008-03-26 Douglas Gregor <doug.gregor@gmail.com>
* pt.c (coerce_template_template_parm): Moved the body of the loop
of coerce_template_template_parms here, to make iteration over a
template argument pack simpler.
Also, allow matching of a template parameter pack in the template
template parameter to a template parameter in the template
template argument.
(coerce_template_template_parms): Deal with variadic template
template parameters. Use coerce_template_template_parm.
(unify): Make sure we coerce the template template argument's
template arguments to the template template parameter's template
parameters, not the other way around.
2008-03-26 Douglas Gregor <doug.gregor@gmail.com>
* g++.dg/cpp0x/variadic84.C: Update to reflect the change in
variadic template template parameter binding semantics.
* g++.dg/cpp0x/variadic85.C: Ditto.
* g++.dg/cpp0x/variadic88.C: New.
* g++.dg/cpp0x/variadic89.C: New.
* g++.dg/cpp0x/variadic90.C: New.
* g++.dg/cpp0x/variadic-ex14.C: Update to reflect the change in
variadic template template parameter binding semantics.
* g++.dg/cpp0x/variadic-lambda.C: New.
re PR rtl-optimization/35232 (ICE in fp-int-convert-double.c at -O2)
gcc/
PR rtl-optimization/35232
* reload1.c (reg_reloaded_call_part_clobbered): Clarify comment.
(forget_old_reloads_1, forget_marked_reloads): Don't clear
reg_reloaded_call_part_clobbered here.
(reload_regs_reach_end_p): New function.
(reload_reg_rtx_for_input): New variable.
(reload_reg_rtx_for_output): Likewise.
(emit_input_reload_insns): Use reloadreg rather than rl->reg_rtx
when reassigning a pseudo register. Load reloadreg from
reload_reg_rtx_for_input, moving the mode and register
calculation to...
(do_input_reload): ...here. Use the mode-adjusted reg_rtx
instead of the original when deciding whether an input reload
would be a no-op or whether an output reload can be deleted.
(emit_output_reload_insns): Use the mode-adjusted reg_rtx
when setting up new_spill_reg_store. Load it from
reload_reg_rtx_for_output, moving the mode and register
calculation to...
(do_output_reload): ...here. Use the mode-adjusted reg_rtx
instead of the original when deciding whether an output reload
would be a no-op. Do the same when modifying insn notes.
Use rtx_equal_p instead of == to compare the registers.
(inherit_piecemeal_p): Take a mode and two register numbers
as argument.
(emit_reload_insns): Clear new_spill_reg_store for every hard
register in the reload register. Remove spill registers
from reg_reloaded_valid before considering whether to record
inheritance information for them. Use reload_reg_rtx_for_output
instead of reg_rtx when recording output reloads. Use
reload_reg_rtx_for_input instead of reg_rtx when recording
input reloads. Set or clear reg_reloaded_call_part_clobbered
at the same time as setting reg_reloaded_valid.
(delete_output_reload): Add a new_reload_reg parameter and use it
instead of rld[j].reg_rtx.
(emit_input_reload_insns, do_input_reload, do_output_reload): Adjust
calls accordingly.
gcc/testsuite/
PR rtl-optimization/35232
* gcc.target/mips/pr35232.c: New test.
Eric Botcazou [Tue, 25 Mar 2008 17:58:54 +0000 (17:58 +0000)]
revert: re PR ada/35186 (implicit assumption about alignment of DImode)
Revert
2008-03-05 Eric Botcazou <ebotcazou@adacore.com>
PR ada/35186
* decl.c (maybe_pad_type): Avoid padding an integral type when
bumping its alignment is sufficient.
Bob Wilson [Tue, 25 Mar 2008 17:56:31 +0000 (17:56 +0000)]
xtensa.c (xtensa_va_start): Use build_int_cst instead of size_int for integer types.
* config/xtensa/xtensa.c (xtensa_va_start): Use build_int_cst
instead of size_int for integer types.
(xtensa_gimplify_va_arg_expr): Likewise. Convert index to sizetype
to match type of MINUS_EXPR.
Jayant R Sonar [Tue, 25 Mar 2008 13:36:11 +0000 (13:36 +0000)]
constraints.md (I28): New constraint.
* config/sh/constraints.md (I28): New constraint.
* config/sh/sh.c (broken_move): Add support for movi20s.
* config/sh/sh.md (movsi_ie): Add the alternative for
movi20s.
* gcc.target/sh/sh2a-resbank.c: New test.
* gcc.target/sh/sh2a-tbr-jump.c: New test.
* gcc.target/sh/sh2a-jsrn.c: New test.
* gcc.target/sh/sh2a-rtsn.c: New test.
Co-Authored-By: Jayant R Sonar <jayant.sonar@kpitcummins.com> Co-Authored-By: Naveen.H.S <naveen.hs@kpitcummins.com>
From-SVN: r133513
Uros Bizjak [Tue, 25 Mar 2008 08:02:14 +0000 (09:02 +0100)]
sse-17.c: Include sse2-check.h.
* gcc.target/i386/sse-17.c: Include sse2-check.h.
(main): Change to void and rename to sse2_test.
* gcc.dg/vect/costmodel/i386/i386-costmodel-vect.exp: Set
dg-do-what-default to "compile" if SSE2 hardware is not available.
* gcc.dg/vect/costmodel/i386/x86_64-costmodel-vect.exp: Ditto.
* gcc.dg/vect/costmodel/x86_64/costmodel-vect-reduc-1char.c: Remove
xfail vect_no_int_max from "vectorization not profitable" check.
Paul Thomas [Mon, 24 Mar 2008 19:11:24 +0000 (19:11 +0000)]
re PR fortran/34813 (ICE on incorrect nested type constructor (fold-const.c (fold_convert):2629))
2008-03-24 Paul Thomas <pault@gcc.gnu.org>
PR fortran/34813
* resolve.c (resolve_structure_cons): It is an error to assign
NULL to anything other than a pointer or allocatable component.
PR fortran/33295
* resolve.c (resolve_symbol): If the symbol is a derived type,
resolve the derived type. If the symbol is a derived type
function, ensure that the derived type is visible in the same
namespace as the function.
2008-03-24 Paul Thomas <pault@gcc.gnu.org>
PR fortran/34813
* gfortran.dg/null_3.f90 : New test
PR fortran/33295
* gfortran.dg/module_function_type_1.f90 : New test
Richard Guenther [Mon, 24 Mar 2008 15:08:52 +0000 (15:08 +0000)]
re PR c/22371 (C front-end produces mis-match types in MODIFY_EXPR)
2008-03-24 Richard Guenther <rguenther@suse.de>
PR c/22371
* gimplify.c (gimplify_modify_expr): For frontend type-correct
pointer assignments change conversions according to middle-end rules.
(gimplify_modify_expr_rhs): Deal with NULL TARGET_EXPR_INITIAL.
* configure.ac: Include type checking in yes.
* configure: Regenerate.
PR documentation/15479
* Make-lang.in (doc/gnat_ugn.texi) Renamed from ...
(doc/gnat_ugn_unw.texi): ... this, and adjusted.
(doc/gnat_ugn.info): Renamed from ...
(doc/gnat_ugn_unw.info): ... this.
(doc/gnat_ugn.dvi): Renamed from ...
(doc/gnat_ugn_unw.dvi): ... this.
(doc/gnat_ugn.pdf): Renamed from ...
(doc/gnat_ugn_unw.pdf): ... this.
(ADA_INFOFILES, ADA_PDFFILES, ada.install-info, ada.dvi):
Adjusted.
* gnat_ugn.texi (FILE): Hard-code gnat_ugn; set filename
unconditionally to gnat_ugn.info. Fix cross references to the
GNAT Reference Manual. Convert links to the GCC, GDB, Emacs,
and GNU make manuals to be proper texinfo links.
* gnat_rm.texi: Fix cross references to the GNAT User's Guide.
Thomas Koenig [Sun, 23 Mar 2008 22:19:19 +0000 (22:19 +0000)]
re PR libfortran/32972 (performance of pack/unpack)
2007-03-23 Thomas Koenig <tkoenig@gcc.gnu.org
PR libfortran/32972
* Makefile.am: Add new variable, i_unpack_c, containing
unpack_i1.c, unpack_i2.c, unpack_i4.c, unpack_i8.c,
unpack_i16.c, unpack_r4.c, unpack_r8.c, unpack_r10.c,
unpack_r16.c, unpack_c4.c, unpack_c8.c, unpack_c10.c
and unpack_c16.c
Add i_unpack_c to gfor_built_src.
Add rule to generate i_unpack_c from m4/unpack.m4.
* Makefile.in: Regenerated.
* libgfortran.h: Add prototypes for unpack0_i1, unpack0_i2,
unpack0_i4, unpack0_i8, unpack0_i16, unpack0_r4, unpack0_r8,
unpack0_r10, unpack0_r16, unpack0_c4, unpack0_c8, unpack0_c10,
unpack0_c16, unpack1_i1, unpack1_i2, unpack1_i4, unpack1_i8,
unpack1_i16, unpack1_r4, unpack1_r8, unpack1_r10, unpack1_r16,
unpack1_c4, unpack1_c8, unpack1_c10 and unpack1_c16.
* intrinsics/pack_generic.c (unpack1): Add calls to specific
unpack1 functions.
(unpack0): Add calls to specific unpack0 functions.
* m4/unpack.m4: New file.
* generated/unpack_i1.c: New file.
* generated/unpack_i2.c: New file.
* generated/unpack_i4.c: New file.
* generated/unpack_i8.c: New file.
* generated/unpack_i16.c: New file.
* generated/unpack_r4.c: New file.
* generated/unpack_r8.c: New file.
* generated/unpack_r10.c: New file.
* generated/unpack_r16.c: New file.
* generated/unpack_c4.c: New file.
* generated/unpack_c8.c: New file.
* generated/unpack_c10.c: New file.
* generated/unpack_c16.c: New file.
2007-03-23 Thomas Koenig <tkoenig@gcc.gnu.org
PR libfortran/32972
* gfortran.dg/intrinsic_unpack_1.f90: New test case.
* gfortran.dg/intrinsic_unpack_2.f90: New test case.
* gfortran.dg/intrinsic_unpack_3.f90: New test case.
Uros Bizjak [Sun, 23 Mar 2008 12:07:34 +0000 (13:07 +0100)]
Revert:
2008-03-05 H.J. Lu <hongjiu.lu@intel.com>
* config/i386/i386-modes.def: Use 4 byte alignment on DI for
32bit host.
2008-03-19 Uros Bizjak <ubizjak@gmail.com>
PR target/35496
* stor-layout.c (update_alignment_for_field): Set minimum alignment
of the underlying type of a MS bitfield layout to the natural
alignment of the type.
2008-03-22 Uros Bizjak <ubizjak@gmail.com>
* config/i386/i386.c (assign_386_stack_local): Align DImode slots
to their natural alignment to avoid store forwarding stalls.
Tobias Schlüter [Sun, 23 Mar 2008 09:28:03 +0000 (10:28 +0100)]
trans.h: Use fold_build in build1_v, build2_v and build3_v macros.
2008-03-23 Tobias Schlueter <tobi@gcc.gnu.org>
* trans.h: Use fold_build in build1_v, build2_v and build3_v
macros.
* trans-openmp.c (gfc_trans_omp_critical, gfc_trans_omp_single):
Don't use build2_v macro.
Jerry DeLisle [Sat, 22 Mar 2008 22:03:13 +0000 (22:03 +0000)]
re PR fortran/35632 (stream io broken on FreeBSD due to ftruncate changes.)
2008-03-22 Jerry DeLisle <jvdelisle@gcc.gnu.org>
PR libfortran/35632
* io/transfer.c (data_transfer_init): Fix whitespace.
(next_record_w): Truncate the file only if the stream
position is short of the file end.
Richard Guenther [Sat, 22 Mar 2008 19:50:48 +0000 (19:50 +0000)]
tree-cfg.c (verify_expr): Recurse again for invariant addresses.
2008-03-22 Richard Guenther <rguenther@suse.de>
* tree-cfg.c (verify_expr): Recurse again for invariant addresses.
For PHI nodes verify the address is invariant.
* tree-ssa-ccp.c (ccp_decl_initial_min_invariant): Remove.
(get_symbol_constant_value): Use is_gimple_min_invariant.
(maybe_fold_stmt_indirect): Likewise.
re PR rtl-optimization/33927 (replace_read in dse.c could handle cases where GET_MODE_CLASS (read_mode) != GET_MODE_CLASS (store_mode) (and the size is the same))
gcc/
PR rtl-optimization/33927
* Makefile.in (dse.o): Depend on $(TM_P_H).
* expr.h (extract_low_bits): Declare.
* expmed.c (extract_low_bits): New function.
* rtlhooks.c (gen_lowpart_general): Generalize SUBREG handling.
* dse.c: Include tm_p.h.
(find_shift_sequence): Remove the read_reg argument and return the
read value. Emit the instructions instead of returning them.
Iterate on new_mode rather than calculating it each time.
Check MODES_TIEABLE_P. Use simplify_gen_subreg to convert the
source to NEW_MODE and extract_low_bits to convert the shifted
value to READ_MODE.
(replace_read): Allow the load and store to have different mode
classes. Use extract_low_bits when SHIFT == 0. Create the shift
or extraction instructions before trying the replacement. Update
dump-file code accordingly, avoiding use of REGNO (store_info->rhs).
gcc/testsuite/
* gcc.target/mips/dse-1.c: Add checks for zeros.
Andrew Pinski [Fri, 21 Mar 2008 22:18:23 +0000 (15:18 -0700)]
re PR target/27946 (double to long long and back to double stores to the stack)
2008-03-21 Andrew Pinski <andrew_pinski@playstation.sony.com>
PR target/27946
* config/rs6000/rs6000.md (floatdidf2): Discouraging fprs and
encouraging but not allowing gprs for input;
change the input constraint to !f#r.
(fix_truncdfdi2): Discouraging fprs and encouraging but not allowing
gprs for output;
change the output constraint to !f#r.
2008-03-21 Andrew Pinski <andrew_pinski@playstation.sony.com>
PR target/27946
* gcc.target/powerpc/ppc64-double-1.c: New testcase.
Andrew Pinski [Fri, 21 Mar 2008 22:16:49 +0000 (22:16 +0000)]
re PR target/27946 (double to long long and back to double stores to the stack)
2008-03-21 Andrew Pinski <andrew_pinski@playstation.sony.com>
PR target/27947
* config/rs6000/rs6000.md (floatdidf2): Discouraging fprs and
encouraging but not allowing gprs for input;
change the input constraint to !f#r.
(fix_truncdfdi2): Discouraging fprs and encouraging but not allowing
gprs for output;
change the output constraint to !f#r.
2008-03-21 Andrew Pinski <andrew_pinski@playstation.sony.com>
PR target/27947
* gcc.target/powerpc/ppc64-double-1.c: New testcase.
Uros Bizjak [Fri, 21 Mar 2008 20:43:12 +0000 (21:43 +0100)]
re PR target/13958 (Conversion from unsigned to double is painfully slow on P4)
PR target/13958
* config/i386/i386.md ("*floatunssi<mode2>_1"): New pattern with
corresponding post-reload splitters.
("floatunssi<mode>2"): Expand to unsigned_float x87 insn pattern
when x87 FP math is selected.
* config/i386/i386-protos.h (ix86_expand_convert_uns_sixf_sse):
New function prototype.
* config/i386/i386.c (ix86_expand_convert_uns_sixf_sse): New
unreachable function to ease macroization of insn patterns.
Olivier Hainque [Fri, 21 Mar 2008 11:43:54 +0000 (11:43 +0000)]
trans.c (addressable_p): Accept COND_EXPR when both arms are addressable.
2008-03-21 Olivier Hainque <hainque@adacore.com>
Ed Schonberg <schonberg@adacore.com>
* trans.c (addressable_p): Accept COND_EXPR when both arms
are addressable.
(gnat_gimplify_expr): Let the gimplifier handle &COND_EXPR.
(call_to_gnu): Do not use name reference in the error message
for a misaligned by_reference_parameter. The actual may be a
general expression.
Kaz Kojima [Thu, 20 Mar 2008 23:25:41 +0000 (23:25 +0000)]
linux-atomic.asm (ATOMIC_TEST_AND_SET): Take unsigned extension into account.
* config/sh/linux-atomic.asm (ATOMIC_TEST_AND_SET): Take
unsigned extension into account.
(ATOMIC_COMPARE_AND_SWAP): Likewise.
(ATOMIC_FETCH_AND_OP, ATOMIC_FETCH_AND_COMBOP): Likewise.
Do computations on a scratch register.
Richard Guenther [Thu, 20 Mar 2008 22:06:40 +0000 (22:06 +0000)]
tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Remove special casing of constant qualifiers.
2008-03-20 Richard Guenther <rguenther@suse.de>
* tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Remove
special casing of constant qualifiers.
* tree-ssa.c (useless_type_conversion_p_1): Instead do not
care about them in general.
* tree-ssa-ccp.c (ccp_fold): Addresses are constant or not
regardless of their type.
(fold_stmt_r): Forcefully fold *& if we end up with that.
Victor Kaplansky [Thu, 20 Mar 2008 19:43:56 +0000 (19:43 +0000)]
re PR testsuite/34168 (runtime tests in gfortran.dg/vect fail for unsupported [non-SSE2] targets)
PR testsuite/34168
* lib/target-supports.exp (check_sse2_hw_available): New procedure.
* gcc.dg/vect/vect.exp: Set dg-do-what-default to "compile"
if SSE2 hardware is not available.
* g++.dg/vect/vect.exp: Update target-dependent overrides to match
gcc.dg/vect/vect.exp.
* gfortran.dg/vect/vect.exp: Ditto.
(check_effective_target_lp64_or_vect_no_align): Remove procedure.
Andreas Schwab [Thu, 20 Mar 2008 13:31:25 +0000 (13:31 +0000)]
re PR testsuite/35647 (FAIL: gcc.dg/cpp/cmdlne-d(I|M)-M.c scan-file (^|\\n)cmdlne-d(I|M)-M[^\\n]*:[^\\n]*cmdlne-d(I|M)-M.c)
PR testsuite/35647
* gcc.dg/cpp/cmdlne-dI-M.c: Match optional continuation.
* gcc.dg/cpp/cmdlne-dM-M.c: Likewise.
* gcc.dg/cpp/cmdlne-dN-M.c: Match correct file name and optional
continuation and remove xfail.
Kai Tietz [Thu, 20 Mar 2008 08:24:42 +0000 (08:24 +0000)]
ms format support for mingw.
for gcc:
* c-format.c (replace_format_name_to_system_name): New.
(cmp_attribs): New.
(convert_format_name_to_system_name): New.
(decode_format_attr): Add use of convert_format_name_to_system_name.
(format_types_orig): Add gnu_ prefix to names.
(check_format_info_main): Special treating of \0 escaped names for
supporting multi-character format specifiers as I32, I64.
(TARGET_OVERRIDES_FORMAT_ATTRIBUTES): Use of user defined attributes.
(gnu_target_overrides_format_attributes): New.
* c-format.h: Add structure target_ovr_attr to hold
system specific formatter names.
* config.gcc: Add for x86&x86_64 cygwin and mingw32 targets the
msformat-c.o file to c_target_objs and cxx_target_objs.
* config/i386/mingw32.h (TARGET_OVERRIDES_FORMAT_ATTRIBUTES): New.
(TARGET_OVERRIDES_FORMAT_ATTRIBUTES_COUNT): New.
(TARGET_N_FORMAT_TYPES): New.
* config/i386/msformat-c.c: New.
* config/i386/t-cygming: Add build rule for msformat-c.o.
* doc/extend.texi: Add new format names gnu_* and ms_* and
further details.
* doc/tm.texi: (TARGET_OVERRIDES_FORMAT_ATTRIBUTES): New.
Ira Rosen [Thu, 20 Mar 2008 07:05:53 +0000 (07:05 +0000)]
invoke.texi (-O3): Add -ftree-vectorize to the list of optimizations turned on under -O3.
* doc/invoke.texi (-O3): Add -ftree-vectorize to the list of
optimizations turned on under -O3.
(ftree-vectorize): Add that the flag is turned on with -O3.
Jerry DeLisle [Thu, 20 Mar 2008 02:05:05 +0000 (02:05 +0000)]
re PR fortran/35627 (namelist read error)
2008-03-19 Jerry DeLisle <jvdelisle@gcc.gnu.org>
PR libfortran/35627
* io/list_read.c (free_line): Clear the line buffer enable flag and
reset the index into line_buffer, aka item_count.
(next_char): Cleanup whitespace.
(read_logical): Use unget_char to assure that the first character of the
bad logical is saved in case it is part of an object name. Remove the
clearing of index and flag that is now in free_line.
(read_real): Likewise.
Thomas Koenig [Wed, 19 Mar 2008 15:42:55 +0000 (15:42 +0000)]
re PR libfortran/32972 (performance of pack/unpack)
2008-03-19 Thomas Koenig <tkoenig@gcc.gnu.org>
PR libfortran/32972
* Makefile.am (in_pack_c): Add in_pack_i1.c, in_pack_i2.c,
in_pack_r4.c, in_pack_r8.c, in_pack_r10.c and in_pack_r16.c.
(in_unpack_c): Add in_unpack_i1.c, in_unpack_i2.c,
in_unpack_r4.c, in_unpack_r8.c, in_unpack_r10.c and
in_unpack_r16.c.
* Makefile.in: Regenerate.
* libgfortran.h: Add prototypes for internal_pack_1,
internal_pack_2, internal_pack_16, internal_pack_r4,
internal_pack_r8, internal_pack_r10, internal_pack_r16,
internal_pack_c10 and internal_pack_c16. Add prototypes for
internal_unpack_1, internal_unpack_2, internal_unpack_16,
internal_unpack_r4, internal_unpack_r8, internal_unpack_r10,
internal_unpack_r16, internal_unpack_c10 and
internal_unpack_c16.
* runtime/in_pack_generic.c (internal_pack): Use sizeof instead
of hardwired sizes.
Add calls to internal_pack_1, internal_pack_2,
internal_pack_16, internal_pack_r4, internal_pack_r8,
internal_pack_r10, internal_pack_r16, internal_pack_c10 and
internal_pack_c16.
* runtime/in_unpack_generic.c (internal_unpack): Use sizeof
instead of hardwired sizes.
Add calls to internal_unpack_1, internal_unpack_2,
internal_unpack_16, internal_unpack_r4, internal_unpack_r8,
internal_unpack_r10, internal_unpack_r16, internal_unpack_c10
and internal_unpack_c16.
* generated/in_pack_r4.c: New file.
* generated/in_pack_i2.c: New file.
* generated/in_unpack_i1.c: New file.
* generated/in_pack_r10.c: New file.
* generated/in_unpack_r4.c: New file.
* generated/in_unpack_i2.c: New file.
* generated/in_unpack_r16.c: New file.
* generated/in_pack_r8.c: New file.
* generated/in_unpack_r10.c: New file.
* generated/in_unpack_r8.c: New file.
* generated/in_pack_r16.c: New file.
* generated/in_pack_i1.c: New file.
2008-03-19 Thomas Koenig <tkoenig@gcc.gnu.org>
PR libfortran/32972
* gfortran.dg/internal_pack_1.f90: New test case.
* gfortran.dg/internal_pack_2.f90: New test case.
* gfortran.dg/internal_pack_3.f90: New test case.
Uros Bizjak [Wed, 19 Mar 2008 14:50:15 +0000 (15:50 +0100)]
re PR target/35496 (test failures between revs. 132950 and 132974)
PR target/35496
* stor-layout.c (update_alignment_for_field): Set minimum alignment
of the underlying type of a MS bitfield layout to the natural
alignment of the type.
Richard Guenther [Wed, 19 Mar 2008 10:44:52 +0000 (10:44 +0000)]
re PR tree-optimization/35609 ("is used uninitialized in this function" should be may warning)
2008-03-19 Richard Guenther <rguenther@suse.de>
PR middle-end/35609
* tree-ssa.c (always_executed): New global flag.
(warn_uninitialized_var): If !always_executed warn with "maybe"
instead of "is".
(execute_early_warn_uninitialized): Compute post-dominators.
Initialize always_executed before processing each basic block.
* gcc.dg/testsuite/uninit-15.c: New testcase.
* gcc.dg/testsuite/uninit-16.c: Likewise.