platform/upstream/gcc.git
7 years agoDaily bump.
GCC Administrator [Wed, 4 Jan 2017 00:16:15 +0000 (00:16 +0000)]
Daily bump.

From-SVN: r244040

7 years agoPR tree-optimization/78696 - [7 Regression] -fprintf-return-value misoptimizes %...
Martin Sebor [Tue, 3 Jan 2017 23:14:44 +0000 (16:14 -0700)]
PR tree-optimization/78696 - [7 Regression] -fprintf-return-value misoptimizes %.Ng where N is greater than 10

gcc/ChangeLog:

PR tree-optimization/78696
* gimple-ssa-sprintf.c (format_floating): Correct handling of
precision.  Use MPFR for %f for greater fidelity.  Correct handling
of %g.
(pass_sprintf_length::compute_format_length): Set width and precision
specified by asrerisk to void_node for vararg functions.
(try_substitute_return_value): Adjust dump output.

gcc/testsuite/ChangeLog:

PR tree-optimization/78696
* gcc.dg/tree-ssa/builtin-sprintf-5.c: Remove incorrect test cases.
* gcc.dg/tree-ssa/builtin-sprintf-warn-7.c: Correct off-by-1 errors.
* gcc.dg/tree-ssa/builtin-sprintf-warn-9.c: New test.
* gcc.dg/tree-ssa/builtin-sprintf.c: Add test cases.

From-SVN: r244037

7 years agoruntime: remove __go_alloc and __go_free
Ian Lance Taylor [Tue, 3 Jan 2017 22:58:48 +0000 (22:58 +0000)]
runtime: remove __go_alloc and __go_free

    Move allocg and handling of allgs slice from C to Go.

    Reviewed-on: https://go-review.googlesource.com/34797

From-SVN: r244036

7 years ago* doc/invoke.texi (RS6000 options): LRA is enabled by default.
David Edelsohn [Tue, 3 Jan 2017 22:08:39 +0000 (22:08 +0000)]
* doc/invoke.texi (RS6000 options): LRA is enabled by default.

From-SVN: r244035

7 years agoinvoke.texi (SPARC options): Document -mlra as the default.
Eric Botcazou [Tue, 3 Jan 2017 21:13:49 +0000 (21:13 +0000)]
invoke.texi (SPARC options): Document -mlra as the default.

* doc/invoke.texi (SPARC options): Document -mlra as the default.
* config/sparc/sparc.c (sparc_option_override): Force LRA unless
-mlra/-mno-lra was passed to the compiler.

From-SVN: r244034

7 years agoUpdate .po files from TP.
Joseph Myers [Tue, 3 Jan 2017 20:52:48 +0000 (20:52 +0000)]
Update .po files from TP.

gcc/po:
* be.po, da.po, de.po, el.po, es.po, fi.po, fr.po, hr.po, id.po,
ja.po, nl.po, ru.po, sr.po, sv.po, tr.po, uk.po, vi.po, zh_CN.po,
zh_TW.po: Update.

libcpp/po:
* be.po, ca.po, da.po, de.po, el.po, eo.po, es.po, fi.po, fr.po,
id.po, ja.po, nl.po, pr_BR.po, ru.po, sr.po, sv.po, tr.po, uk.po,
vi.po, zh_CN.po, zh_TW.po: Update.

From-SVN: r244032

7 years agore PR go/78789 (Error: no such instruction: `aesenc %xmm0,%xmm2' when compiling libgo...
Ian Lance Taylor [Tue, 3 Jan 2017 20:41:54 +0000 (20:41 +0000)]
re PR go/78789 (Error: no such instruction: `aesenc %xmm0,%xmm2' when compiling libgo/runtime/aeshash.c)

PR go/78789
    runtime: don't build aeshash.c if the assembler doesn't support it

    This is for CentOS 5, whose assembler does not know the aesinc
    instruction.

    Fixes GCC PR 78789.

    Patch by Uros Bizjak.

    Reviewed-on: https://go-review.googlesource.com/34796

From-SVN: r244031

7 years agore PR rtl-optimization/65618 (gnat bootstrap comparison failure on mips{,el}-linux...
James Cowgill [Tue, 3 Jan 2017 18:41:05 +0000 (18:41 +0000)]
re PR rtl-optimization/65618 (gnat bootstrap comparison failure on mips{,el}-linux-gnu)

PR rtl-optimization/65618
* emit-rtl.c (try_split): Move initialization of "before" and
"after" to just before the call to emit_insn_after_setloc.

From-SVN: r244029

7 years agoPR 78534 Revert r244011
Janne Blomqvist [Tue, 3 Jan 2017 18:01:30 +0000 (20:01 +0200)]
PR 78534 Revert r244011

r244011 caused regressions on 32-bit hosts.

From-SVN: r244027

7 years agos-linux-mips.ads: Use correct signal and errno constants.
James Cowgill [Tue, 3 Jan 2017 17:52:28 +0000 (17:52 +0000)]
s-linux-mips.ads: Use correct signal and errno constants.

* s-linux-mips.ads: Use correct signal and errno constants.
(sa_handler_pos, sa_mask_pos): Fix offsets for 64-bit MIPS.

From-SVN: r244026

7 years agos-linux-mips.ads: Rename from s-linux-mipsel.ads.
James Cowgill [Tue, 3 Jan 2017 17:51:02 +0000 (17:51 +0000)]
s-linux-mips.ads: Rename from s-linux-mipsel.ads.

* s-linux-mips.ads: Rename from s-linux-mipsel.ads.
* gcc-interface/Makefile.in (MIPS/Linux): Merge mips and mipsel
sections.

From-SVN: r244025

7 years agogcc-dg.exp: support more than one digit in relative line number notation
David Malcolm [Tue, 3 Jan 2017 16:25:31 +0000 (16:25 +0000)]
gcc-dg.exp: support more than one digit in relative line number notation

gcc/testsuite/ChangeLog:
* gcc.dg/dg-test-1.c: Add tests of relative line specifications
with more than one digit.
* lib/gcc-dg.exp (process-message): Support more than one digit
in relative line specifications.

From-SVN: r244024

7 years ago* doc/md.texi (Standard Names): Remove reference to Java frontend.
Gerald Pfeifer [Tue, 3 Jan 2017 13:49:19 +0000 (13:49 +0000)]
* doc/md.texi (Standard Names): Remove reference to Java frontend.

From-SVN: r244023

7 years agoAdd deleted std::thread(const thread&&) constructor
Jonathan Wakely [Tue, 3 Jan 2017 13:31:26 +0000 (13:31 +0000)]
Add deleted std::thread(const thread&&) constructor

PR libstdc++/78956
* include/std/thread (thread(const thread&&)): Add deleted
constructor.
* testsuite/30_threads/thread/cons/lwg2097.cc: New test.

From-SVN: r244022

7 years agoFix typos in libstdc++ docs and update copyright years
Jonathan Wakely [Tue, 3 Jan 2017 11:16:31 +0000 (11:16 +0000)]
Fix typos in libstdc++ docs and update copyright years

* doc/xml/manual/spine.xml: Update copyright years.
* doc/xml/manual/build_hacking.xml: Fix spelling of libbuilddir.
* doc/xml/manual/test.xml: Likewise.
* doc/html/*: Regenerate.

From-SVN: r244017

7 years agoDWARF: add DW_AT_encoding attributes for DW_TAG_enumeration_type DIEs
Pierre-Marie de Rodat [Tue, 3 Jan 2017 08:54:57 +0000 (08:54 +0000)]
DWARF: add DW_AT_encoding attributes for DW_TAG_enumeration_type DIEs

Currently, the DWARF description does not specify the signedness of the
representation of enumeration types.  This is a problem in some
contexts where DWARF consumers need to determine if value X is greater
than value Y.

For instance in Ada:

    type Enum_Type is ( A, B, C, D);
    for Enum_Type use (-1, 0, 1, 2);

    type Rec_Type (E : Enum_Type) is record
       when A .. B => null;
       when others => B : Booleann;
    end record;

The above can be described in DWARF the following way:

    DW_TAG_enumeration_type(Enum_Type)
    | DW_AT_byte_size: 1
      DW_TAG_enumerator(A)
      | DW_AT_const_value: -1
      DW_TAG_enumerator(B)
      | DW_AT_const_value: 0
      DW_TAG_enumerator(C)
      | DW_AT_const_value: 1
      DW_TAG_enumerator(D)
      | DW_AT_const_value: 2

    DW_TAG_structure_type(Rec_Type)
      DW_TAG_member(E)
      | DW_AT_type: <Enum_Type>
      DW_TAG_variant_part
      | DW_AT_discr: <E>
        DW_TAG_variant
        | DW_AT_discr_list: DW_DSC_range 0x7f 0
        DW_TAG_variant
        | DW_TAG_member(b)

DWARF consumers need to know that enumerators (A, B, C and D) are signed
in order to determine the set of E values for which Rec_Type has a B
field.  In practice, they need to know how to interpret the 0x7f LEB128
number above (-1, not 127).

When in non-strict DWARF mode, this patch adds a DW_AT_encoding
attribute to generated DW_TAG_enumeration_type DIEs to make this
signedness explicit.

gcc/

* dwarf2out.c (gen_enumeration_type_die): When
-gno-strict-dwarf, add a DW_AT_encoding attribute.

From-SVN: r244015

7 years agore PR tree-optimization/78965 (Invalid -fprintf-return-value optimization)
Jakub Jelinek [Tue, 3 Jan 2017 07:23:11 +0000 (08:23 +0100)]
re PR tree-optimization/78965 (Invalid -fprintf-return-value optimization)

PR tree-optimization/78965
* gimple-ssa-sprintf.c (pass_sprintf_length::compute_format_length):
Change first argument from const call_info & to call_info &.  For %n
set info.nowrite to false.

* gcc.dg/pr78965.c: New test.

From-SVN: r244014

7 years agore PR middle-end/78901 (ICE: verify_gimple failed (error: statement marked for throw...
Jakub Jelinek [Tue, 3 Jan 2017 07:20:04 +0000 (08:20 +0100)]
re PR middle-end/78901 (ICE: verify_gimple failed (error: statement marked for throw in middle of block))

PR middle-end/78901
* gimple-ssa-sprintf.c (try_substitute_return_value): Don't change
possibly throwing calls.

* g++.dg/opt/pr78901.C: New test.

From-SVN: r244013

7 years agogenmatch.c (dt_node::gen_kids_1): If generic_exprs include SSA_NAME and exprs_len...
Jakub Jelinek [Tue, 3 Jan 2017 07:17:38 +0000 (08:17 +0100)]
genmatch.c (dt_node::gen_kids_1): If generic_exprs include SSA_NAME and exprs_len || fns_len...

* genmatch.c (dt_node::gen_kids_1): If generic_exprs include SSA_NAME
and exprs_len || fns_len, emit the code for SSA_NAME next to the exprs
and fns handling, rather than in a separate case SSA_NAME.

From-SVN: r244012

7 years agoPR 78534 Change character length from int to size_t
Janne Blomqvist [Tue, 3 Jan 2017 07:04:01 +0000 (09:04 +0200)]
PR 78534 Change character length from int to size_t

In order to handle large character lengths on (L)LP64 targets, switch
the GFortran character length from an int to a size_t.

This is an ABI change, as procedures with character arguments take
hidden arguments with the character length.

I also changed the _size member in vtables from int to size_t, as
there were some cases where character lengths and sizes were
apparently mixed up and caused regressions otherwise. Although I
haven't tested, this might enable very large derived types as well.

Also, as there are some places in the frontend were negative character
lengths are used as special flag values, in the frontend the character
length is handled as a signed variable of the same size as a size_t,
although in the runtime library it really is size_t.

I haven't changed the character length variables for the co-array
intrinsics, as this is something that may need to be synchronized with
OpenCoarrays.

This is v3 of the patch. All the issues pointed out by FX's review of
v2 have been fixed. In particular, there are now new functions
gfc_mpz_get_hwi and gfc_mpz_set_hwi, similar to the GMP functions
mpz_get_si and mpz_set_si, except that they get/set a HOST_WIDE_INT
instead of a long value. Similarly, gfc_get_int_expr now takes a
HOST_WIDE_INT instead of a long, gfc_extract_long is replaced by
gfc_extract_hwi. Also, the preliminary work to handle
gfc_charlen_type_node being unsigned has been removed.

Regtested on x86_64-pc-linux-gnu.

frontend:

2017-01-03  Janne Blomqvist  <jb@gcc.gnu.org>

PR fortran/78534
PR fortran/66310
* class.c (gfc_find_derived_vtab): Use gfc_size_kind instead of
hardcoded kind.
(find_intrinsic_vtab): Likewise.
* expr.c (gfc_get_character_expr): Length parameter of type
gfc_charlen_t.
(gfc_get_int_expr): Value argument of type HOST_WIDE_INT.
(gfc_extract_hwi): New function.
(simplify_const_ref): Make string_len of type gfc_charlen_t.
(gfc_simplify_expr): Use HOST_WIDE_INT for substring refs.
* gfortran.h (gfc_mpz_get_hwi): New prototype.
(gfc_mpz_set_hwi): Likewise.
(gfc_charlen_t): New typedef.
(gfc_expr): Use gfc_charlen_t for character lengths.
(gfc_size_kind): New extern variable.
(gfc_extract_hwi): New prototype.
(gfc_get_character_expr): Use gfc_charlen_t for character length.
(gfc_get_int_expr): Use HOST_WIDE_INT type for value argument.
* iresolve.c (gfc_resolve_repeat): Pass string length directly without
temporary, use gfc_charlen_int_kind.
* match.c (select_intrinsic_set_tmp): Use HOST_WIDE_INT for charlen.
* misc.c (gfc_mpz_get_hwi): New function.
(gfc_mpz_set_hwi): New function.
* module.c (atom_int): Change type from int to HOST_WIDE_INT.
(parse_integer): Don't complain about large integers.
(write_atom): Use HOST_WIDE_INT for integers.
(mio_integer): Handle integer type mismatch.
(mio_hwi): New function.
(mio_intrinsic_op): Use HOST_WIDE_INT.
(mio_array_ref): Likewise.
(mio_expr): Likewise.
* resolve.c (resolve_select_type): Use HOST_WIDE_INT for charlen,
use snprintf.
(resolve_charlen): Use mpz_sgn to determine sign.
* simplify.c (gfc_simplify_repeat): Use HOST_WIDE_INT/gfc_charlen_t
instead of long.
* target-memory.c (size_character): Length argument of type
gfc_charlen_t.
(gfc_encode_character): Likewise.
(gfc_interpret_character): Use gfc_charlen_t.
* target-memory.h (gfc_encode_character): Modify prototype.
* trans-array.c (get_array_ctor_var_strlen): Use
gfc_conv_mpz_to_tree_type.
* trans-const.c (gfc_conv_mpz_to_tree_type): New function.
* trans-const.h (gfc_conv_mpz_to_tree_type): New prototype.
* trans-expr.c (gfc_class_len_or_zero_get): Build const of type
gfc_charlen_type_node.
(gfc_conv_intrinsic_to_class): Use gfc_charlen_int_kind instead of
4, fold_convert to correct type.
(gfc_conv_class_to_class): Build const of type size_type_node for
size.
(gfc_copy_class_to_class): Likewise.
(gfc_conv_string_length): Use same type in expression.
(gfc_conv_substring): Likewise, use HOST_WIDE_INT for charlen.
(gfc_conv_string_tmp): Make sure len is of the right type.
(gfc_conv_concat_op): Use same type in expression.
(gfc_conv_procedure_call): Likewise.
(alloc_scalar_allocatable_for_subcomponent_assignment):
fold_convert to right type.
(gfc_trans_subcomponent_assign): Likewise.
(trans_class_vptr_len_assignment): Build const of correct type.
(gfc_trans_pointer_assignment): Likewise.
(alloc_scalar_allocatable_for_assignment): fold_convert to right
type in expr.
(trans_class_assignment): Build const of correct type.
* trans-intrinsic.c (gfc_conv_associated): Likewise.
(gfc_conv_intrinsic_repeat): Do calculation in sizetype.
* trans-io.c (gfc_build_io_library_fndecls): Use
gfc_charlen_type_node for character lengths.
* trans-stmt.c (gfc_trans_label_assign): Build const of
gfc_charlen_type_node.
(gfc_trans_character_select): Likewise.
(gfc_trans_allocate): Likewise, don't typecast strlen result.
(gfc_trans_deallocate): Don't typecast strlen result.
* trans-types.c (gfc_size_kind): New variable.
(gfc_init_types): Determine gfc_charlen_int_kind and gfc_size_kind
from size_type_node.

testsuite:

2017-01-03  Janne Blomqvist  <jb@gcc.gnu.org>

PR fortran/78534
PR fortran/66310
* gfortran.dg/dependency_49.f90: Change scan-tree-dump-times
          due to gfc_trans_string_copy change to avoid
          -Wstringop-overflow.
* gfortran.dg/repeat_4.f90: Use integers of kind C_SIZE_T.
* gfortran.dg/repeat_7.f90: New test for PR 66310.
* gfortran.dg/scan_2.f90: Handle potential cast in assignment.
* gfortran.dg/string_1.f90: Limit to ilp32 targets.
* gfortran.dg/string_1_lp64.f90: New test.
* gfortran.dg/string_3.f90: Limit to ilp32 targets.
* gfortran.dg/string_3_lp64.f90: New test.
* gfortran.dg/transfer_intrinsic_1.f90: Change
          scan-tree-dump-times due to gfc_trans_string_copy change to
          avoid -Wstringop-overflow.

libgfortran:

2017-01-03  Janne Blomqvist  <jb@gcc.gnu.org>

PR fortran/78534
* intrinsics/args.c (getarg_i4): Use gfc_charlen_type.
(get_command_argument_i4): Likewise.
(get_command_i4): Likewise.
* intrinsics/chmod.c (chmod_internal): Likewise.
* intrinsics/env.c (get_environment_variable_i4): Likewise.
* intrinsics/extends_type_of.c (struct vtype): Use size_t for size
member.
* intrinsics/gerror.c (gerror): Use gfc_charlen_type.
* intrinsics/getlog.c (getlog): Likewise.
* intrinsics/hostnm.c (hostnm_0): Likewise.
* intrinsics/string_intrinsics_inc.c (string_len_trim): Rework to
work if gfc_charlen_type is unsigned.
(string_scan): Likewise.
* io/transfer.c (transfer_character): Modify prototype.
(transfer_character_write): Likewise.
(transfer_character_wide): Likewise.
(transfer_character_wide_write): Likewise.
(transfer_array): Typecast to avoid signed-unsigned comparison.
* io/unit.c (is_trim_ok): Use gfc_charlen_type.
* io/write.c (namelist_write): Likewise.
* libgfortran.h (gfc_charlen_type): Change typedef to size_t.

From-SVN: r244011

7 years agodarwin-driver.c (darwin_driver_init): Const-correctness fixes for first_period and...
Jeff Law [Tue, 3 Jan 2017 05:36:40 +0000 (22:36 -0700)]
darwin-driver.c (darwin_driver_init): Const-correctness fixes for first_period and second_period variables.

* config/darwin-driver.c (darwin_driver_init): Const-correctness
fixes for first_period and second_period variables.

From-SVN: r244010

7 years agoDaily bump.
GCC Administrator [Tue, 3 Jan 2017 00:16:13 +0000 (00:16 +0000)]
Daily bump.

From-SVN: r244009

7 years agore PR target/78967 (inserts are not effective)
Uros Bizjak [Mon, 2 Jan 2017 22:08:18 +0000 (23:08 +0100)]
re PR target/78967 (inserts are not effective)

 target/78967
* config/i386/i386.md (UNSPEC_NOREX_MEM): New unspec.
(*insvqi_1): New insn pattern.
(*insvqi_1_mem_rex64): Ditto.
(*insvqi_2): Ditto.
(*insvqi_3): Rename from *insvqi.

(*extzvqi_mem_rex64): Add UNSPEC_NOREX_MEM tag.

testsuite/ChangeLog:

PR target/78967
* gcc.target/i386/pr78967-1.c: New test.
* gcc.target/i386/pr78967-2.c: Ditto.
* gcc.target/i386/pr78967-3.c: Ditto.

* gcc.target/i386/pr78904-2.c: Tighten scan-asm patterns.
* gcc.target/i386/pr78904-4.c: Ditto.
* gcc.target/i386/pr78904-6.c: Ditto.

From-SVN: r244006

7 years ago* es.po: Update.
Joseph Myers [Mon, 2 Jan 2017 20:20:34 +0000 (20:20 +0000)]
* es.po: Update.

From-SVN: r244004

7 years agoPR 78534 Modify string copy to avoid -Wstringop-overflow warning
Janne Blomqvist [Mon, 2 Jan 2017 20:00:18 +0000 (22:00 +0200)]
PR 78534 Modify string copy to avoid -Wstringop-overflow warning

When the character length is changed from int to size_t the existing
algorithm causes a -Wstringop-overflow warning with -O1 on the
gfortran.dg/allocate_deferred_char_scalar_1.f03 testcase. This change
is committed separately from the character length size change in order
to make bisecting potential performance issues easier.

2017-01-02  Janne Blomqvist  <jb@gcc.gnu.org>

        PR fortran/78534
* trans-expr.c (gfc_trans_string_copy): Rework string copy
algorithm to avoid -Wstringop-overflow warning.

From-SVN: r244003

7 years agocfg.texi (Edges): Remove reference to Java.
Gerald Pfeifer [Mon, 2 Jan 2017 14:23:06 +0000 (14:23 +0000)]
cfg.texi (Edges): Remove reference to Java.

* doc/cfg.texi (Edges): Remove reference to Java.
(Maintaining the CFG): Ditto.

From-SVN: r244002

7 years agofaq.xml: Update address of C++ ABI link.
Gerald Pfeifer [Mon, 2 Jan 2017 00:33:17 +0000 (00:33 +0000)]
faq.xml: Update address of C++ ABI link.

* doc/xml/faq.xml: Update address of C++ ABI link.
* doc/xml/manual/abi.xml: Ditto.

From-SVN: r244001

7 years agoDaily bump.
GCC Administrator [Mon, 2 Jan 2017 00:16:15 +0000 (00:16 +0000)]
Daily bump.

From-SVN: r244000

7 years agore PR ipa/77674 (ICE in binds_to_current_def_p with -fkeep-inline-functions)
Jan Hubicka [Sun, 1 Jan 2017 23:31:53 +0000 (00:31 +0100)]
re PR ipa/77674 (ICE in binds_to_current_def_p with -fkeep-inline-functions)

PR middle-end/77674
* symtab.c (symtab_node::binds_to_current_def_p): Fix handling of
transparent aliases.

PR middle-end/77674
* g++.dg/torture/pr77674.C: New testcase.

From-SVN: r243997

7 years ago* gnatvsn.ads: Bump copyright year.
Eric Botcazou [Sun, 1 Jan 2017 18:08:36 +0000 (18:08 +0000)]
* gnatvsn.ads: Bump copyright year.

From-SVN: r243996

7 years agore PR middle-end/77484 (Static branch predictor causes ~6-8% regression of SPEC2000...
Jan Hubicka [Sun, 1 Jan 2017 15:40:29 +0000 (16:40 +0100)]
re PR middle-end/77484 (Static branch predictor causes ~6-8% regression of SPEC2000 GAP)

PR middle-end/77484
* predict.def (PRED_CALL): Update hitrate.
(PRED_INDIR_CALL, PRED_POLYMORPHIC_CALL): New predictors.
* predict.c (tree_estimate_probability_bb): Split CALL predictor
into direct/indirect/polymorphic variants.

From-SVN: r243995

7 years agoUpdate copyright years.
Jakub Jelinek [Sun, 1 Jan 2017 12:07:43 +0000 (13:07 +0100)]
Update copyright years.

From-SVN: r243994

7 years agoupdate-copyright.py (TestsuiteFilter): Skip params/README.
Jakub Jelinek [Sun, 1 Jan 2017 11:50:10 +0000 (12:50 +0100)]
update-copyright.py (TestsuiteFilter): Skip params/README.

* update-copyright.py (TestsuiteFilter): Skip params/README.
(GCCCmdLine): Remove libjava and libjava/testsuite add_dir.

From-SVN: r243993

7 years agogcc.c (process_command): Update copyright notice dates.
Jakub Jelinek [Sun, 1 Jan 2017 11:31:28 +0000 (12:31 +0100)]
gcc.c (process_command): Update copyright notice dates.

gcc/
* gcc.c (process_command): Update copyright notice dates.
* gcov-dump.c (print_version): Ditto.
* gcov.c (print_version): Ditto.
* gcov-tool.c (print_version): Ditto.
* gengtype.c (create_file): Ditto.
* doc/cpp.texi: Bump @copying's copyright year.
* doc/cppinternals.texi: Ditto.
* doc/gcc.texi: Ditto.
* doc/gccint.texi: Ditto.
* doc/gcov.texi: Ditto.
* doc/install.texi: Ditto.
* doc/invoke.texi: Ditto.
gcc/fortran/
* gfortranspec.c (lang_specific_driver): Update copyright notice
dates.
* gfc-internals.texi: Bump @copying's copyright year.
* gfortran.texi: Ditto.
* intrinsic.texi: Ditto.
* invoke.texi: Ditto.
gcc/ada/
* gnat_ugn.texi: Bump @copying's copyright year.
* gnat_rm.texi: Likewise.
gcc/go/
* gccgo.texi: Bump @copyrights-go year.
libitm/
* libitm.texi: Bump @copying's copyright year.
libgomp/
* libgomp.texi: Bump @copying's copyright year.
libquadmath/
* libquadmath.texi: Bump @copying's copyright year.

From-SVN: r243992

7 years agoRegenerate .pot files.
Joseph Myers [Sun, 1 Jan 2017 00:55:25 +0000 (00:55 +0000)]
Regenerate .pot files.

gcc/po:
* gcc.pot: Regenerate.

libcpp/po:
* cpplib.pot: Regenerate.

From-SVN: r243991

7 years agoDaily bump.
GCC Administrator [Sun, 1 Jan 2017 00:16:16 +0000 (00:16 +0000)]
Daily bump.

From-SVN: r243990

7 years agoDaily bump.
GCC Administrator [Sat, 31 Dec 2016 00:16:22 +0000 (00:16 +0000)]
Daily bump.

From-SVN: r243987

7 years agocppopts.texi: Reorder table entries to put the most commonly-used options first and...
Sandra Loosemore [Fri, 30 Dec 2016 21:22:33 +0000 (16:22 -0500)]
cppopts.texi: Reorder table entries to put the most commonly-used options first and debug...

2016-12-30  Sandra Loosemore  <sandra@codesourcery.com>

gcc/
* doc/cppopts.texi: Reorder table entries to put the most
commonly-used options first and debug options last.

From-SVN: r243984

7 years agore PR translation/78745 (Truncated messages in po file)
Jakub Jelinek [Fri, 30 Dec 2016 19:15:42 +0000 (20:15 +0100)]
re PR translation/78745 (Truncated messages in po file)

PR translation/78745
* exgettext: Handle multi-line help texts in *.opt files.
* gcc.pot: Regenerate.

From-SVN: r243981

7 years ago* es.po, fr.po: Update.
Joseph Myers [Fri, 30 Dec 2016 18:46:27 +0000 (18:46 +0000)]
* es.po, fr.po: Update.

From-SVN: r243979

7 years agocompile: add type conversion in Type::gc_symbol_pointer
Ian Lance Taylor [Fri, 30 Dec 2016 18:03:39 +0000 (18:03 +0000)]
compile: add type conversion in Type::gc_symbol_pointer

    Wrap the return from Type::gc_symbol_pointer with a type conversion to
    uintptr, since the values returned are stored into structure fields
    with that type.

    Reviewed-on: https://go-review.googlesource.com/34625

From-SVN: r243978

7 years agoi386.md (*testqi_ext_3): Merge insn pattern and corresponding splitter to define_insn...
Uros Bizjak [Fri, 30 Dec 2016 17:14:16 +0000 (18:14 +0100)]
i386.md (*testqi_ext_3): Merge insn pattern and corresponding splitter to define_insn_and_split.

* config/i386/i386.md (*testqi_ext_3): Merge insn pattern and
corresponding splitter to define_insn_and_split.  Use wi::shifted_mask
helper function to calculate mask.

From-SVN: r243977

7 years agopredicates.md (ext_register_operand): Do not reject registers without upper parts...
Uros Bizjak [Fri, 30 Dec 2016 15:10:45 +0000 (16:10 +0100)]
predicates.md (ext_register_operand): Do not reject registers without upper parts here.

* config/i386/predicates.md (ext_register_operand): Do not reject
registers without upper parts here.
* config/i386/i386.md (extv<mode>): Copy registers without
upper parts in operand 1 to a pseudo.
(extzv<mode>): Ditto.
(insv<mode>): Ditto.

From-SVN: r243976

7 years agostandards.texi (Standards): Remove broken reference to objc.toodarkpark.net and avoid...
Gerald Pfeifer [Fri, 30 Dec 2016 15:07:15 +0000 (15:07 +0000)]
standards.texi (Standards): Remove broken reference to objc.toodarkpark.net and avoid list with now just...

* doc/standards.texi (Standards): Remove broken reference to
objc.toodarkpark.net and avoid list with now just one item.

From-SVN: r243975

7 years agocompiler: use more accurate array type in Gogo::register_gc_vars
Ian Lance Taylor [Fri, 30 Dec 2016 07:08:00 +0000 (07:08 +0000)]
compiler: use more accurate array type in Gogo::register_gc_vars

    The length of the roots array in Gogo::register_gc_vars was being
    computed incorrectly, potentially leading to type clashes in the back
    end. Compute an accurate length for this array ahead of time.

    Reviewed-on: https://go-review.googlesource.com/34624

From-SVN: r243974

7 years agoc11-atomic-exec-5.c: Lower ITER_COUNT to 100 for AARCH64.
Andrew Pinski [Fri, 30 Dec 2016 04:24:40 +0000 (04:24 +0000)]
c11-atomic-exec-5.c: Lower ITER_COUNT to 100 for AARCH64.

2016-12-29  Andrew Pinski  <apinski@cavium.com>

        * gcc.dg/atomic/c11-atomic-exec-5.c: Lower ITER_COUNT to 100
        for AARCH64.

From-SVN: r243973

7 years agoDaily bump.
GCC Administrator [Fri, 30 Dec 2016 00:16:17 +0000 (00:16 +0000)]
Daily bump.

From-SVN: r243972

7 years ago* config/i386/i386.md (divmodqi4): Use lowpart_subreg.
Uros Bizjak [Thu, 29 Dec 2016 23:19:32 +0000 (00:19 +0100)]
* config/i386/i386.md (divmodqi4): Use lowpart_subreg.

From-SVN: r243969

7 years agore PR target/78904 (zero-extracts are not effective)
Uros Bizjak [Thu, 29 Dec 2016 21:41:07 +0000 (22:41 +0100)]
re PR target/78904 (zero-extracts are not effective)

PR target/78904
* config/i386/i386.md (*extvqi): Remove insn pattern.
(divmodqi4): Update expander to generate QImode zero-extract from AH.

testsuite/ChangeLog:

PR target/78904
* gcc.target/i386/pr78904-6.c: New test.

From-SVN: r243968

7 years agoImplement LWG 2842, in_place_t check for optional::optional(U&&) should decay U.
Ville Voutilainen [Thu, 29 Dec 2016 19:19:59 +0000 (21:19 +0200)]
Implement LWG 2842, in_place_t check for optional::optional(U&&) should decay U.

Implement LWG 2842, in_place_t check for optional::optional(U&&)
should decay U.
* include/std/optional (_Optional_base(in_place_t, _Args&&...)):
Constrain.
(_Optional_base(in_place_t, initializer_list<_Up>, _Args&&...)):
Turn the int-pack constraint hack into a saner bool.
(_Optional_base<_Tp, false>::_Optional_base(in_place_t, _Args&&...)):
Constrain.
(_Optional_base<_Tp, false>::_Optional_base(in_place_t,
initializer_list<_Up>, _Args&&...)):
Turn the int-pack constraint hack into a saner bool.
(optional(_Up&&)): Constrain against in_place_t.
(optional(in_place_t, _Args&&...)): Constrain.
(constexpr optional(in_place_t, initializer_list<_Up>, _Args&&...)):
Turn the int-pack constraint hack into a saner bool.
* testsuite/20_util/optional/cons/value_neg.cc: Add a test for
a type that is constructible from in_place.

From-SVN: r243966

7 years agomemcpy-1.c: Include mpx-check.h.
Jakub Jelinek [Thu, 29 Dec 2016 18:51:29 +0000 (19:51 +0100)]
memcpy-1.c: Include mpx-check.h.

* gcc.target/i386/mpx/memcpy-1.c: Include mpx-check.h.
(main): Renamed to ...
(mpx_test): ... this.  Add argc and argv arguments.

From-SVN: r243965

7 years agors6000.c (altivec_expand_builtin): Fix typos in error messages.
Michael Meissner [Thu, 29 Dec 2016 17:19:28 +0000 (17:19 +0000)]
rs6000.c (altivec_expand_builtin): Fix typos in error messages.

2016-12-29  Michael Meissner  <meissner@linux.vnet.ibm.com>

* config/rs6000/rs6000.c (altivec_expand_builtin): Fix typos in
error messages.

From-SVN: r243963

7 years agoextend.texi (Cilk Plus Builtins): cilkplus.org now uses https by default.
Gerald Pfeifer [Thu, 29 Dec 2016 14:52:23 +0000 (14:52 +0000)]
extend.texi (Cilk Plus Builtins): cilkplus.org now uses https by default.

        * doc/extend.texi (Cilk Plus Builtins): cilkplus.org now uses
        https by default.
        * doc/passes.texi (Cilk Plus Transformation): Ditto.
        * doc/generic.texi (Statements for C++): Ditto, and use @uref.

From-SVN: r243962

7 years ago* config.sub: Import latest version.
Ben Elliston [Thu, 29 Dec 2016 11:04:53 +0000 (22:04 +1100)]
* config.sub: Import latest version.

From-SVN: r243961

7 years agoAdd RejectNegative for a c option.
Martin Liska [Thu, 29 Dec 2016 10:35:33 +0000 (11:35 +0100)]
Add RejectNegative for a c option.

PR c/78933
* g++.dg/pr78933.C: New test.
PR c/78933
* c.opt (strong-eval-order): Add RejectNegative keyword.

From-SVN: r243960

7 years agoDaily bump.
GCC Administrator [Thu, 29 Dec 2016 00:16:16 +0000 (00:16 +0000)]
Daily bump.

From-SVN: r243959

7 years ago* gcc.target/i386/pr78904-5.c: Remove duplicate part.
Uros Bizjak [Wed, 28 Dec 2016 15:41:29 +0000 (16:41 +0100)]
* gcc.target/i386/pr78904-5.c: Remove duplicate part.

From-SVN: r243956

7 years agore PR target/78904 (zero-extracts are not effective)
Uros Bizjak [Wed, 28 Dec 2016 15:16:26 +0000 (16:16 +0100)]
re PR target/78904 (zero-extracts are not effective)

PR target/78904
* config/i386/constraints.md (Bn): New special memory constraint.
* config/i386/predicates.md (norex_memory_operand): New predicate.
* config/i386/i386.md (*extzvqi_mem_rex64): New insn pattern and
corresponding peephole2 pattern.

testsuite/ChangeLog:

PR target/78904
* gcc.target/i386/pr78904-4.c: New test.
* gcc.target/i386/pr78904-5.c: Ditto.

From-SVN: r243955

7 years agocppdiropts.texi, [...]: New files, split from...
Sandra Loosemore [Wed, 28 Dec 2016 02:34:29 +0000 (21:34 -0500)]
cppdiropts.texi, [...]: New files, split from...

2016-12-27  Sandra Loosemore  <sandra@codesourcery.com>

gcc/
* doc/cppdiropts.texi, doc/cppwarnopts.texi:  New files, split from...
* doc/cppopts.texi: .... here.
* doc/cpp.texi (Invocation): Adjust includes.
* doc/invoke.texi (Option Summary): Add missing preprocesor-related
options.  Adjust sorting and formatting.
(Warning Options): Include cppwarnopts.texi.
(Preprocessor Options): Add pointers and list the specific
preprocessor options from cppopts.texi first instead of last.
(Directory Options): Move/merge documentation of -I, -iquote, and
-I- to cppdiropts.texi.  Include that file here.

From-SVN: r243954

7 years agoDaily bump.
GCC Administrator [Wed, 28 Dec 2016 00:16:14 +0000 (00:16 +0000)]
Daily bump.

From-SVN: r243952

7 years agopredicates.md (const_0_to_12_operand): Rename predicate and change test from 0..11...
Michael Meissner [Tue, 27 Dec 2016 23:19:15 +0000 (23:19 +0000)]
predicates.md (const_0_to_12_operand): Rename predicate and change test from 0..11 to 0..12 to match the semantics of...

[gcc]
2016-12-27  Michael Meissner  <meissner@linux.vnet.ibm.com>

* config/rs6000/predicates.md (const_0_to_12_operand): Rename
predicate and change test from 0..11 to 0..12 to match the
semantics of the word extract/insert instructions.  Change all
callers.
(const_0_to_11_operand): Likewise.
* config/rs6000/rs6000.c (altivec_expand_builtin): Likewise.
* config/rs6000/vsx.md (vextract4b): Likewise.
(vextract4b_internal): Likewise.
(vinsert4b): Likewise.
(vinsert4b_internal): Likewise.
(vinsert4b_di): Likewise.
(vinsert4b_di_internal): Likewise.
* config/rs6000/rs6000.md (zero_extendsi<mode>2): Fix offset used
in xxextractuw to zero extend the word in the vector registers.
(lfiwzx): Likewise.

[gcc/testsuite]
2016-12-27  Michael Meissner  <meissner@linux.vnet.ibm.com>

* gcc.target/powerpc/p9-vinsert4b-2.c: Update test to test for 13
being out of bounds instead of 12.

From-SVN: r243948

7 years ago* ChangeLog: Update my last entry.
Uros Bizjak [Tue, 27 Dec 2016 21:50:32 +0000 (22:50 +0100)]
* ChangeLog: Update my last entry.

From-SVN: r243946

7 years agoi386.c (ix86_secondary_reload): Require QImode intermediate for QImode mask register...
Uros Bizjak [Tue, 27 Dec 2016 21:25:09 +0000 (22:25 +0100)]
i386.c (ix86_secondary_reload): Require QImode intermediate for QImode mask register spill only for...

* config/i386/i386.c (ix86_secondary_reload): Require QImode
intermediate for QImode mask register spill only for !TARGET_AVX512DQ.
Always use true_regnum to determine operand regno.

From-SVN: r243945

7 years agocppopts.texi: Delete redundant documentation for -x.
Sandra Loosemore [Tue, 27 Dec 2016 20:37:05 +0000 (15:37 -0500)]
cppopts.texi: Delete redundant documentation for -x.

2016-12-27  Sandra Loosemore  <sandra@codesourcery.com>

gcc/
* doc/cppopts.texi: Delete redundant documentation for -x.  Move
-fno-show-column documentation to...
* doc/invoke.texi (Diagnostic Message Formatting Options):  ...here.
Update the option summary.

From-SVN: r243944

7 years agoi386.md (VI_512): Remove.
Uros Bizjak [Tue, 27 Dec 2016 17:20:15 +0000 (18:20 +0100)]
i386.md (VI_512): Remove.

* config/i386/i386.md (VI_512): Remove.
(vcond<V_512:mode><VI_AVX512BW:mode>): Use VI_AVX512BW
mode iterator instead of VI_512.
(vcondu<V_512:mode><VI_AVX512BW:mode>): Ditto.

From-SVN: r243943

7 years agolibtool-version: New version.
Alexander Ivchenko [Tue, 27 Dec 2016 15:35:19 +0000 (15:35 +0000)]
libtool-version: New version.

libmpx/

2016-12-27  H.J. Lu  <hongjiu.lu@intel.com>
    Alexander Ivchenko  <alexander.ivchenko@intel.com>

* mpxwrap/libtool-version: New version.
* mpxwrap/mpx_wrappers.c (__mpx_wrapper_realloc): Make it
static with external alias.
(__mpx_wrapper_bzero): Ditto.
(mpx_wrapper_memcpy): Ditto.
(__mpx_wrapper_mempcpy): Ditto.

From-SVN: r243942

7 years agogfortran.h (gfc_error): Rename overload with OPT argument to...
Jakub Jelinek [Tue, 27 Dec 2016 15:17:19 +0000 (16:17 +0100)]
gfortran.h (gfc_error): Rename overload with OPT argument to...

* gfortran.h (gfc_error): Rename overload with OPT argument to...
(gfc_error_opt): ... this.
* error.c (gfc_error): Rename overloads with OPT argument to...
(gfc_error_opt): ... this.  Adjust callers.
(gfc_notify_std, gfc_error): Adjust callers.
* resolve.c (resolve_structure_cons, resolve_global_procedure): Use
gfc_error_opt instead of gfc_error.
* interface.c (argument_rank_mismatch, compare_parameter,
gfc_check_typebound_override): Likewise.  Fix up formatting.

From-SVN: r243941

7 years agore PR translation/78922 (Comment submitted for translation in stringop.opt)
Jakub Jelinek [Tue, 27 Dec 2016 14:41:04 +0000 (15:41 +0100)]
re PR translation/78922 (Comment submitted for translation in stringop.opt)

PR translation/78922
* config/i386/stringop.opt: Remove.

From-SVN: r243938

7 years agore PR target/78904 (zero-extracts are not effective)
Uros Bizjak [Tue, 27 Dec 2016 14:20:19 +0000 (15:20 +0100)]
re PR target/78904 (zero-extracts are not effective)

PR target/78904
* config/i386/constraints.md (Bc): New special memory constraint.
* config/i386/i386.md (*cmpqi_ext_1, *extvqi, *extzvqi): Use Bc
constraint with nonimmediate_operand to allow constant memory operands.
(*cmpqi_ext_3, insv<mode>_1, addqi_ext_1, *testqi_ext_1, andqi_ext_1)
(*<any_or:code>qi_ext_1, *xorqi_ext_1_cc): Use Bc constraint
with general_operand to allow constant memory operands.

testsuite/ChangeLog:

PR target/78904
* gcc.target/i386/pr78904-3.c: New test.

From-SVN: r243937

7 years agoc.opt (flag_chkp_flexible_struct_trailing_arrays): Add new option.
Alexander Ivchenko [Tue, 27 Dec 2016 13:31:43 +0000 (13:31 +0000)]
c.opt (flag_chkp_flexible_struct_trailing_arrays): Add new option.

2016-12-27  Alexander Ivchenko  <alexander.ivchenko@intel.com>

* c-family/c.opt (flag_chkp_flexible_struct_trailing_arrays):
  Add new option.
  (fchkp-narrow-to-innermost-array): Fix typo.
* doc/cpp.texi (flag_chkp_flexible_struct_trailing_arrays): Ditto.
* tree-chkp.c (chkp_may_narrow_to_field ): Forbid
  narrowing when flag_chkp_flexible_struct_trailing_arrays is used
  and the field is the last array field in the structure.

2016-12-27  Alexander Ivchenko  <alexander.ivchenko@intel.com>

* gcc.target/i386/mpx/vla-trailing-1-lbv.c: New test.
* gcc.target/i386/mpx/vla-trailing-1-nov.c: Ditto.
* gcc.target/i386/mpx/vla-trailing-1-ubv.c: Ditto.

From-SVN: r243936

7 years ago* gcc.target/i386/pr78904-2.c: Fix testcase.
Uros Bizjak [Tue, 27 Dec 2016 11:12:32 +0000 (12:12 +0100)]
* gcc.target/i386/pr78904-2.c: Fix testcase.

From-SVN: r243935

7 years agoi386.md (andqi_ext_1): Use general_operand predicate for operand 2.
Uros Bizjak [Tue, 27 Dec 2016 11:00:15 +0000 (12:00 +0100)]
i386.md (andqi_ext_1): Use general_operand predicate for operand 2.

* config/i386/i386.md (andqi_ext_1): Use general_operand
predicate for operand 2.

From-SVN: r243934

7 years agore PR target/78904 (zero-extracts are not effective)
Uros Bizjak [Tue, 27 Dec 2016 08:59:04 +0000 (09:59 +0100)]
re PR target/78904 (zero-extracts are not effective)

PR target/78904
* config/i386/i386.md (*cmpqi_ext_1, *extvqi, *extzvqi): Use
nonimmediate_operand instead of nonimmediate_x64nomem_operand.
(*cmpqi_ext_3, insv<mode>_1, addqi_ext_1, *testqi_ext_1, andqi_ext_1)
(*<any_or:code>qi_ext_1, *xorqi_ext_1_cc): Use general_operand
instead of general_x64nomem_operand.
* config/i386/predicates.md (nonimmediate_x64nomem_operand): Remove.
(general_x64nomem_operand): Ditto.

testsuite/ChangeLog:

PR target/78904
* gcc.target/i386/pr78904-2.c: New test.

From-SVN: r243933

7 years agoDaily bump.
GCC Administrator [Tue, 27 Dec 2016 00:16:15 +0000 (00:16 +0000)]
Daily bump.

From-SVN: r243932

7 years agore PR target/78904 (zero-extracts are not effective)
Uros Bizjak [Mon, 26 Dec 2016 19:00:47 +0000 (20:00 +0100)]
re PR target/78904 (zero-extracts are not effective)

PR target/78904
* config/i386/i386.md (addqi_ext_1): Canonicalize insn pattern w.r.t.
zero_extract RTXes.
(*addqi_ext_2): Ditto.
(testqi_ext_ccno_0): Canonicalize expander w.r.t. zero_extract RTXes.
(testqi_ext_1_ccno): Rename from testqi_ext_ccno_0.
(*testqi_ext_0): Merge with *testqi_ext_1.
(*testqi_ext_1): Canonicalize insn pattern w.r.t. zero_extract RTXes.
Update corresponding splitter.
(*testqi_ext_2): Canonicalize insn pattern w.r.t. zero_extract RTXes.
(*andqi_ext_0): Merge with *andqi_ext_1.
(andqi_ext_1): Canonicalize insn pattern w.r.t. zero_extract RTXes.
Rename from *andqi_ext_1.  Update corresponding splitter and
peephole2 patterns.
(*andqi_ext_1_cc): Rename from *andqi_ext_0_cc.
(*andqi_ext_2): Canonicalize insn pattern w.r.t. zero_extract RTXes.
(*<any_or:code>qi_ext_0): Merge with *andqi_ext_1.
(*<any_or:code>qi_ext_1): Canonicalize insn pattern w.r.t.
zero_extract RTXes.  Update corresponding splitter.
(*<any_or:code>qi_ext_2): Canonicalize insn pattern w.r.t.
zero_extract RTXes.
(xorqi_cc_ext_1): Canonicalize expander w.r.t. zero_extract RTXes.
(xorqi_ext_1_cc): Rename from xorqi_cc_ext_1.
(*xorqi_cc_ext_1): Canonicalize insn pattern w.r.t. zero_extract RTXes.
Update corresponding splitter.
(*xorqi_ext_1_cc): Rename from *xorqi_cc_ext_1.
(isinfxf2): Update calls to renamed expanders.
(isinf<mode>2): Ditto.
* config/i386/i386.c (ix86_expand_fp_compare): Ditto.
(ix86_emit_fp_unordered_jump): Ditto.
(ix86_emit_i387_round): Ditto.

testsuite/ChangeLog:

PR target/78904
* gcc.target/i386/pr78904.c: New test.

From-SVN: r243929

7 years agolibtool-version: New version.
Alexander Ivchenko [Mon, 26 Dec 2016 15:14:06 +0000 (15:14 +0000)]
libtool-version: New version.

2016-12-26  Alexander Ivchenko  <alexander.ivchenko@intel.com>

    * mpxrt/libtool-version: New version.
    * mpxrt/mpxrt-utils.c (set_mpx_rt_stop_handler): New function.
    (print_help): Add help for CHKP_RT_STOP_HANDLER environment
    variable.
    (__mpxrt_init_env_vars): Add initialization of stop_handler.
    (__mpxrt_stop_handler): New function.
    (__mpxrt_stop): Ditto.
    * mpxrt/mpxrt-utils.h (mpx_rt_stop_mode_handler_t): New enum.
    * mpxrt/mpxrt.c (handler): Replace exit(255) with __mpxrt_stop
    function call.

From-SVN: r243928

7 years agoMAINTAINERS (Write After Approval): Add myself.
Alexander Ivchenko [Mon, 26 Dec 2016 10:28:02 +0000 (10:28 +0000)]
MAINTAINERS (Write After Approval): Add myself.

2016-12-26  Alexander Ivchenko  <alexander.ivchenko@intel.com>

            * MAINTAINERS (Write After Approval): Add myself.

From-SVN: r243927

7 years agoSummary: Whitespace cleanup
Andreas Schwab [Mon, 26 Dec 2016 10:16:15 +0000 (10:16 +0000)]
Summary: Whitespace cleanup

From-SVN: r243926

7 years ago* doc/cpp.texi (Invocation): Remove space in command.
Eric Botcazou [Mon, 26 Dec 2016 09:36:33 +0000 (09:36 +0000)]
* doc/cpp.texi (Invocation): Remove space in command.

From-SVN: r243925

7 years ago* doc/invoke.texi (SPARC options): Add missing documentation for -mlra.
Eric Botcazou [Mon, 26 Dec 2016 09:22:56 +0000 (09:22 +0000)]
* doc/invoke.texi (SPARC options): Add missing documentation for -mlra.

From-SVN: r243924

7 years agocpp.texi (Invocation): Revise to indicate that GCC driver options are only documented...
Sandra Loosemore [Mon, 26 Dec 2016 00:31:02 +0000 (19:31 -0500)]
cpp.texi (Invocation): Revise to indicate that GCC driver options are only documented in the GCC manual.

2016-12-25  Sandra Loosemore  <sandra@codesourcery.com>

gcc/
* doc/cpp.texi (Invocation): Revise to indicate that GCC driver
options are only documented in the GCC manual.
* doc/cppopts.texi: Delete documentation of GCC driver options
-o, -Wall, -Wtraditional, -Werror, -Wsystem-headers, -w,
-pedantic, -pedantic-errors, -std=, -ansi, --help, --target-help,
-v, -version.  Update -Wcomment, -Wtrigraphs, -Wundef,
-Wexpansion-to-defined, -Wno-endif-labels, -traditional,
-traditional-cpp, -trigraphs to merge text previously in GCC manual.
* doc/invoke.texi (Option Summary): Move -trigraphs, -traditional,
and -traditional-cpp from C dialect options to preprocessor options.
(C Dialect Options): Likewise.
(Warning Options): Delete documentation of -Wcomment, -Wtrigraphs,
-Wexpansion-to-defined, -Wundef, and -Wno-endif-labels.

From-SVN: r243923

7 years agoDaily bump.
GCC Administrator [Mon, 26 Dec 2016 00:16:17 +0000 (00:16 +0000)]
Daily bump.

From-SVN: r243922

7 years agoDaily bump.
GCC Administrator [Sun, 25 Dec 2016 00:16:21 +0000 (00:16 +0000)]
Daily bump.

From-SVN: r243919

7 years agoMake it cheaper to test whether an SSA name is a virtual operand
Richard Sandiford [Sat, 24 Dec 2016 14:30:18 +0000 (14:30 +0000)]
Make it cheaper to test whether an SSA name is a virtual operand

virtual_operand_p handled SSA names by looking at the flags of the
underlying variable.  This seems to be a relatively common source
of cache misses, mainly because virtual_operand_p is the first thing
tested by is_gimple_reg.

This patch caches the information in the SSA name itself.  Several
flags seem to be free so the patch arbitrarily uses public_flag.

Tested on aarch64-linux-gnu and x86_64-linux-gnu.  It improves
compile time by a small (<1%) but reproducable amount on the
tests I've tried.

gcc/
* tree-core.h (tree_base): Document the meaning of public_flag
for SSA names.
* tree.h (SSA_NAME_IS_VIRTUAL_OPERAND): New macro.
(SET_SSA_NAME_VAR_OR_IDENTIFIER): Record whether the variable
is a virtual operand.
* gimple-expr.h (virtual_operand_p): Use SSA_NAME_IS_VIRTUAL_OPERAND.

From-SVN: r243916

7 years agorandom.tcc: Fix usage of _OutputIteratorConcept.
François Dumont [Sat, 24 Dec 2016 14:13:58 +0000 (14:13 +0000)]
random.tcc: Fix usage of _OutputIteratorConcept.

2016-12-24  François Dumont  <fdumont@gcc.gnu.org>

* include/ext/random.tcc: Fix usage of _OutputIteratorConcept.

From-SVN: r243915

7 years agore PR middle-end/78901 (ICE: verify_gimple failed (error: statement marked for throw...
Jakub Jelinek [Sat, 24 Dec 2016 11:15:44 +0000 (12:15 +0100)]
re PR middle-end/78901 (ICE: verify_gimple failed (error: statement marked for throw in middle of block))

PR middle-end/78901
* except.c (nothrow_libfn_p): Expect libc_name_p to return
const struct libc_name_struct *, if it returns NULL, return 0,
otherwise check c_ver and use flag_isoc99 or flag_isoc11.
* cfns.gperf: Add %struct-type and libc_name_struct definition.
For all C89 C library functions add , 89 after the name, add
C99 C library functions with , 99 and C11 C library functions
with , 11 suffix.
* cfns.h: Regenerated.

From-SVN: r243914

7 years agoDaily bump.
GCC Administrator [Sat, 24 Dec 2016 00:16:19 +0000 (00:16 +0000)]
Daily bump.

From-SVN: r243913

7 years agoclass_assign_1.f08: New test.
Andre Vehreschild [Fri, 23 Dec 2016 10:26:47 +0000 (11:26 +0100)]
class_assign_1.f08: New test.

gcc/testsuite/ChangeLog:

2016-12-23  Andre Vehreschild  <vehre@gcc.gnu.org>

* gfortran.dg/class_assign_1.f08: New test.

gcc/fortran/ChangeLog:

2016-12-23  Andre Vehreschild  <vehre@gcc.gnu.org>

* trans-expr.c (trans_class_assignment): Allocate memory of _vptr->size
        before assigning an allocatable class object.
(gfc_trans_assignment_1): Flag that (re-)alloc of the class object
shall be done.

From-SVN: r243909

7 years agoMIPS: Cleanup the forcing of assembly output in error tests.
Toma Tabacu [Fri, 23 Dec 2016 09:52:43 +0000 (09:52 +0000)]
MIPS: Cleanup the forcing of assembly output in error tests.

gcc/testsuite

* gcc.target/mips/oddspreg-2.c (dg-options): Remove dg-skip-if for
-fno-fat-lto-objects and add the -ffat-lto-objects option, along with
an explanation for its purpose.
* gcc.target/mips/oddspreg-3.c (dg-options): Likewise.
* gcc.target/mips/oddspreg-6.c (dg-options): Likewise.
* gcc.target/mips/no-dsp-1.c: Add an explanation for the purpose of
-ffat-lto-objects.
* gcc.target/mips/pr54240.c: Likewise.
* gcc.target/mips/r10k-cache-barrier-14.c: Likewise.
* gcc.target/mips/soft-float-1.c: Likewise.

From-SVN: r243908

7 years agoDaily bump.
GCC Administrator [Fri, 23 Dec 2016 00:16:16 +0000 (00:16 +0000)]
Daily bump.

From-SVN: r243904

7 years ago* es.po: Update.
Joseph Myers [Thu, 22 Dec 2016 23:55:31 +0000 (23:55 +0000)]
* es.po: Update.

From-SVN: r243900

7 years agocompiler: fix spurious redefinition error for anon struct
Ian Lance Taylor [Thu, 22 Dec 2016 23:05:02 +0000 (23:05 +0000)]
compiler: fix spurious redefinition error for anon struct

    Change Struct_type::do_mangled_name to incorporate the field
    names even for hidden symbols. This is needed in cases where
    a package imports a type "S" that has an anonymous struct, e.g.

      // imported from some other package
      type S struct {
        X struct{ _ struct{} }
      }

    and then defines a local type that uses a structurally identical
    anonymous struct, e.g.

      // defined locally
      type T struct {
        U struct{ _ struct{} }
      }

    In the case above both types triggered the creation of hash/equal
    methods, but the method names were clashing (since both structs
    had the same mangled name).

    Fixes golang/go#18414

    Reviewed-on: https://go-review.googlesource.com/34621

From-SVN: r243899

7 years agotree-pretty-print.c (dump_generic_node): Change dump format for REALPART_EXPR and...
Prathamesh Kulkarni [Thu, 22 Dec 2016 20:30:33 +0000 (20:30 +0000)]
tree-pretty-print.c (dump_generic_node): Change dump format for REALPART_EXPR and IMAGPART_EXPR with TDF_GIMPLE.

2016-12-22  Prathamesh Kulkarni  <prathamesh.kulkarni@linaro.org>

* tree-pretty-print.c (dump_generic_node): Change dump format for
REALPART_EXPR and IMAGPART_EXPR with TDF_GIMPLE.

testsuite/
* gcc.dg/gimplefe-20.c: New test-case.

From-SVN: r243898

7 years agoPR c++/78906 - ICE with member variable template
Jason Merrill [Thu, 22 Dec 2016 20:29:07 +0000 (15:29 -0500)]
PR c++/78906 - ICE with member variable template

* pt.c (finish_template_variable): Use most_general_template.

From-SVN: r243897

7 years agoFeature-test macro for P0522R0, matching of template template arguments.
Jason Merrill [Thu, 22 Dec 2016 20:29:02 +0000 (15:29 -0500)]
Feature-test macro for P0522R0, matching of template template arguments.

* c-cppbuiltin.c (c_cpp_builtins): Define
__cpp_template_template_args.

From-SVN: r243896

7 years agoPR c++/78898 - ICE on constructor with TTP
Jason Merrill [Thu, 22 Dec 2016 15:19:54 +0000 (10:19 -0500)]
PR c++/78898 - ICE on constructor with TTP

PR c++/42329
* pt.c (unify): Don't look for a class template from a non-class.

From-SVN: r243890

7 years agore PR testsuite/52641 (Test cases fail for 16-bit int targets)
Georg-Johann Lay [Thu, 22 Dec 2016 15:02:43 +0000 (15:02 +0000)]
re PR testsuite/52641 (Test cases fail for 16-bit int targets)

gcc/testsuite/
PR testsuite/52641
* gcc.dg/fold-and-rshift-2.c [int <= 2]: Use 32-bit int as needed.

From-SVN: r243889

7 years agovarasm: Propagate litpool decl alignment to generated RTX.
Andreas Krebbel [Thu, 22 Dec 2016 14:50:29 +0000 (14:50 +0000)]
varasm: Propagate litpool decl alignment to generated RTX.

When pushing a value into the literal pool the resulting decl might
get a higher alignment than the original expression depending on how a
target defines CONSTANT_ALIGNMENT.  Generating an RTX for the constant
pool access we currently use the alignment from the original
expression.  Changed with the attached patch.

This fixes a GCC 6 regression for S/390.  For arrays of string
constants as in the attached testcase encode_section_info is not able
to figure out that the constant pool slot is already properly aligned
since the mem_align field in the rtx is not set properly.

gcc/ChangeLog:

2016-12-22  Andreas Krebbel  <krebbel@linux.vnet.ibm.com>

* varasm.c (build_constant_desc): Use the alignment of the var
decl instead of the original expression.

gcc/testsuite/ChangeLog:

2016-12-22  Andreas Krebbel  <krebbel@linux.vnet.ibm.com>

* gcc.target/s390/litpool-str-1.c: New test.

From-SVN: r243888

7 years agoRun tests only if the machine supports the instruction set.
Dominik Vogt [Thu, 22 Dec 2016 14:46:46 +0000 (14:46 +0000)]
Run tests only if the machine supports the instruction set.

gcc/ChangeLog:

2016-12-22  Dominik Vogt  <vogt@linux.vnet.ibm.com>

* config/s390/s390-c.c (s390_cpu_cpp_builtins_internal): Define
__S390_ARCH_LEVEL__.

gcc/testsuite/ChangeLog:

2016-12-22  Dominik Vogt  <vogt@linux.vnet.ibm.com>

* gcc.target/s390/md/setmem_long-1.c: Use "s390_useable_hw".
* gcc.target/s390/md/rXsbg_mode_sXl.c: Likewise.
* gcc.target/s390/md/andc-splitter-1.c: Likewise.
* gcc.target/s390/md/andc-splitter-2.c: Likewise.
* lib/gcc-dg.exp (gcc-dg-runtest): Export torture_current_flags.
* gcc.target/s390/s390.exp: Import torture_current_flags.
(check_effective_target_s390_useable_hw): New.
(check_effective_target_s390_z900_hw): New.
(check_effective_target_s390_z990_hw): New.
(check_effective_target_s390_z9_ec_hw): New.
(check_effective_target_s390_z10_hw): New.
(check_effective_target_s390_z196_hw): New.
(check_effective_target_s390_zEC12_hw): New.
(check_effective_target_s390_z13_hw): New.
(check_effective_target_z10_instructions): Removed.
(torture tests): Add optimization level without -march=.
Reorder torture tests for good cache usage.

From-SVN: r243887

7 years agoFix tree-optimization/78886.
Martin Liska [Thu, 22 Dec 2016 13:09:11 +0000 (14:09 +0100)]
Fix tree-optimization/78886.

PR tree-optimization/78886
* gcc.dg/tree-ssa/pr78886.c: New test.
PR tree-optimization/78886
* tree-ssa-strlen.c (handle_builtin_malloc): Return when LHS
is equal to NULL.

From-SVN: r243886