android-toolchain:renjbs-gcc.git
7 years agoFix a bad interaction of inliner x return slot optimization x named value optimization master
Jim Huang [Mon, 15 Nov 2010 16:10:38 +0000 (00:10 +0800)]
Fix a bad interaction of inliner x return slot optimization x named value optimization

From b6b198cac722a8dcaa046f0589fe2c2c7caa6a01 Mon Sep 17 00:00:00 2001
From: Jing Yu <jingyu@google.com>
Date: Thu, 4 Nov 2010 17:50:54 -0700
Subject: [PATCH] Backport upstream 4.6 patches to fix 2 gcc bugs.

Fix a bad interaction of inliner x return slot optimization x
named value optimization

Replace tst instruction with lsls for a single bit test

8 years agoin preparation for Google style build system merge-gcc_4_4_4
Jim Huang [Sat, 14 Aug 2010 13:54:06 +0000 (21:54 +0800)]
in preparation for Google style build system

8 years agoMerge branch 'HEAD'; commit 'gcc_4_4_4_release' into merge-gcc_4_4_3
Jim Huang [Sat, 14 Aug 2010 13:49:52 +0000 (21:49 +0800)]
Merge branch 'HEAD'; commit 'gcc_4_4_4_release' into merge-gcc_4_4_3

Conflicts:
gcc/config/arm/constraints.md
gcc/ipa-inline.c
gcc/params.def
gcc/tree-inline.c

8 years agoDon't build plugins.texi
Jim Huang [Sat, 14 Aug 2010 07:12:05 +0000 (15:12 +0800)]
Don't build plugins.texi

8 years agoPost merge process
Jim Huang [Sat, 14 Aug 2010 04:59:24 +0000 (12:59 +0800)]
Post merge process

8 years agoMerge Google's changes in gcc-4.4.3
Jim Huang [Sat, 14 Aug 2010 04:57:03 +0000 (12:57 +0800)]
Merge Google's changes in gcc-4.4.3

HEAD:

commit a4d9421024ed3f70aede5b11f977d264eb69bf61
Author: Jing Yu <jingyu@google.com>
Date:   Tue Aug 10 10:08:15 2010 -0700

    Remove definition of getpagesize() in order to build c++ library
    for Android toolchain. This is a local work around. The long term
    solution is to change the getpagesize() definition in bionicC. However
    to be compatible with current bionicC and sysroot, we have to do
    this local workaround inside gcc. After getpagesize() in bionicC
    is changed and ndk constains that change, we can revert this CL.

    Change-Id: I14757b967e41e45c06224c1d9ce7752bc1a7e9cb

8 years agoin preparation for Google gcc-4.4.3 merge
Jim Huang [Sat, 14 Aug 2010 01:53:25 +0000 (09:53 +0800)]
in preparation for Google gcc-4.4.3 merge

8 years agoMark as release
gccadmin [Thu, 29 Apr 2010 06:53:13 +0000 (06:53 +0000)]
Mark as release

8 years agoMark ChangeLog
gccadmin [Thu, 29 Apr 2010 06:53:05 +0000 (06:53 +0000)]
Mark ChangeLog

8 years agoMark ChangeLog
gccadmin [Thu, 29 Apr 2010 06:52:58 +0000 (06:52 +0000)]
Mark ChangeLog

8 years agoMark ChangeLog
gccadmin [Thu, 29 Apr 2010 06:52:51 +0000 (06:52 +0000)]
Mark ChangeLog

8 years agoMark ChangeLog
gccadmin [Thu, 29 Apr 2010 06:52:45 +0000 (06:52 +0000)]
Mark ChangeLog

8 years agoMark ChangeLog
gccadmin [Thu, 29 Apr 2010 06:52:39 +0000 (06:52 +0000)]
Mark ChangeLog

8 years agoMark ChangeLog
gccadmin [Thu, 29 Apr 2010 06:52:33 +0000 (06:52 +0000)]
Mark ChangeLog

8 years agoMark ChangeLog
gccadmin [Thu, 29 Apr 2010 06:52:27 +0000 (06:52 +0000)]
Mark ChangeLog

8 years agoMark ChangeLog
gccadmin [Thu, 29 Apr 2010 06:52:20 +0000 (06:52 +0000)]
Mark ChangeLog

8 years agoMark ChangeLog
gccadmin [Thu, 29 Apr 2010 06:52:14 +0000 (06:52 +0000)]
Mark ChangeLog

8 years agoMark ChangeLog
gccadmin [Thu, 29 Apr 2010 06:52:07 +0000 (06:52 +0000)]
Mark ChangeLog

8 years agoMark ChangeLog
gccadmin [Thu, 29 Apr 2010 06:52:00 +0000 (06:52 +0000)]
Mark ChangeLog

8 years agoMark ChangeLog
gccadmin [Thu, 29 Apr 2010 06:51:48 +0000 (06:51 +0000)]
Mark ChangeLog

8 years agoMark ChangeLog
gccadmin [Thu, 29 Apr 2010 06:51:40 +0000 (06:51 +0000)]
Mark ChangeLog

8 years agoMark ChangeLog
gccadmin [Thu, 29 Apr 2010 06:51:33 +0000 (06:51 +0000)]
Mark ChangeLog

8 years agoMark ChangeLog
gccadmin [Thu, 29 Apr 2010 06:51:26 +0000 (06:51 +0000)]
Mark ChangeLog

8 years agoMark ChangeLog
gccadmin [Thu, 29 Apr 2010 06:51:20 +0000 (06:51 +0000)]
Mark ChangeLog

8 years agoMark ChangeLog
gccadmin [Thu, 29 Apr 2010 06:51:13 +0000 (06:51 +0000)]
Mark ChangeLog

8 years agoMark ChangeLog
gccadmin [Thu, 29 Apr 2010 06:51:07 +0000 (06:51 +0000)]
Mark ChangeLog

8 years agoMark ChangeLog
gccadmin [Thu, 29 Apr 2010 06:51:00 +0000 (06:51 +0000)]
Mark ChangeLog

8 years agoMark ChangeLog
gccadmin [Thu, 29 Apr 2010 06:50:54 +0000 (06:50 +0000)]
Mark ChangeLog

8 years agoMark ChangeLog
gccadmin [Thu, 29 Apr 2010 06:50:43 +0000 (06:50 +0000)]
Mark ChangeLog

8 years agoMark ChangeLog
gccadmin [Thu, 29 Apr 2010 06:50:33 +0000 (06:50 +0000)]
Mark ChangeLog

8 years agoMark ChangeLog
gccadmin [Thu, 29 Apr 2010 06:50:25 +0000 (06:50 +0000)]
Mark ChangeLog

8 years agoMark ChangeLog
gccadmin [Thu, 29 Apr 2010 06:50:17 +0000 (06:50 +0000)]
Mark ChangeLog

8 years agoMark ChangeLog
gccadmin [Thu, 29 Apr 2010 06:50:09 +0000 (06:50 +0000)]
Mark ChangeLog

8 years agoMark ChangeLog
gccadmin [Thu, 29 Apr 2010 06:50:01 +0000 (06:50 +0000)]
Mark ChangeLog

8 years agoMark ChangeLog
gccadmin [Thu, 29 Apr 2010 06:49:48 +0000 (06:49 +0000)]
Mark ChangeLog

8 years agoMark ChangeLog
gccadmin [Thu, 29 Apr 2010 06:49:35 +0000 (06:49 +0000)]
Mark ChangeLog

8 years agoMark ChangeLog
gccadmin [Thu, 29 Apr 2010 06:49:29 +0000 (06:49 +0000)]
Mark ChangeLog

8 years agoMark ChangeLog
gccadmin [Thu, 29 Apr 2010 06:49:21 +0000 (06:49 +0000)]
Mark ChangeLog

8 years agoMark ChangeLog
gccadmin [Thu, 29 Apr 2010 06:49:15 +0000 (06:49 +0000)]
Mark ChangeLog

8 years agoMark ChangeLog
gccadmin [Thu, 29 Apr 2010 06:49:06 +0000 (06:49 +0000)]
Mark ChangeLog

8 years agoMark ChangeLog
gccadmin [Thu, 29 Apr 2010 06:48:53 +0000 (06:48 +0000)]
Mark ChangeLog

8 years agoMark ChangeLog
gccadmin [Thu, 29 Apr 2010 06:48:44 +0000 (06:48 +0000)]
Mark ChangeLog

8 years agoMark ChangeLog
gccadmin [Thu, 29 Apr 2010 06:48:38 +0000 (06:48 +0000)]
Mark ChangeLog

8 years agoMark ChangeLog
gccadmin [Thu, 29 Apr 2010 06:48:31 +0000 (06:48 +0000)]
Mark ChangeLog

8 years agoMark ChangeLog
gccadmin [Thu, 29 Apr 2010 06:48:24 +0000 (06:48 +0000)]
Mark ChangeLog

8 years agoMark ChangeLog
gccadmin [Thu, 29 Apr 2010 06:48:16 +0000 (06:48 +0000)]
Mark ChangeLog

8 years agoMark ChangeLog
gccadmin [Thu, 29 Apr 2010 06:48:06 +0000 (06:48 +0000)]
Mark ChangeLog

8 years agoDaily bump.
gccadmin [Wed, 28 Apr 2010 23:16:42 +0000 (23:16 +0000)]
Daily bump.

8 years agoDaily bump.
gccadmin [Tue, 27 Apr 2010 23:18:02 +0000 (23:18 +0000)]
Daily bump.

8 years agoDaily bump.
gccadmin [Mon, 26 Apr 2010 23:19:44 +0000 (23:19 +0000)]
Daily bump.

8 years agoDaily bump.
gccadmin [Sun, 25 Apr 2010 23:17:17 +0000 (23:17 +0000)]
Daily bump.

8 years agoDaily bump.
gccadmin [Sat, 24 Apr 2010 23:17:03 +0000 (23:17 +0000)]
Daily bump.

8 years agoDaily bump.
gccadmin [Fri, 23 Apr 2010 23:16:46 +0000 (23:16 +0000)]
Daily bump.

8 years agoDaily bump.
gccadmin [Thu, 22 Apr 2010 23:16:46 +0000 (23:16 +0000)]
Daily bump.

8 years agoDaily bump.
gccadmin [Wed, 21 Apr 2010 23:17:21 +0000 (23:17 +0000)]
Daily bump.

8 years ago PR fortran/43836
jakub [Wed, 21 Apr 2010 20:26:11 +0000 (20:26 +0000)]
PR fortran/43836
* f95-lang.c (gfc_define_builtin): Set TREE_NOTHROW on
the decl.

* gfortran.dg/gomp/pr43836.f90: New test.

8 years ago2010-04-19 Andrew Haley <aph@redhat.com>
aph [Wed, 21 Apr 2010 16:04:42 +0000 (16:04 +0000)]
2010-04-19  Andrew Haley  <aph@redhat.com>

        PR libgcj/40860
        * configure.ac: Handle --no-merge-exidx-entries.

8 years ago PR middle-end/43570
jakub [Wed, 21 Apr 2010 13:02:39 +0000 (13:02 +0000)]
PR middle-end/43570
* omp-low.c (scan_sharing_clauses): Don't scan_omp_op
OMP_CLAUSE_DECL for OMP_CLAUSE_COPYPRIVATE.
(lower_copyprivate_clauses): Use private var in outer
context instead of original var.  Make sure the types
are correct for VLAs.

* testsuite/libgomp.fortran/vla8.f90: New test.

8 years ago PR libgomp/43706
jakub [Wed, 21 Apr 2010 13:00:10 +0000 (13:00 +0000)]
PR libgomp/43706
* config/linux/affinity.c (gomp_init_affinity): Decrease
gomp_available_cpus if affinity mask confines the process to fewer
CPUs.
* config/linux/proc.c (get_num_procs): If gomp_cpu_affinity is
non-NULL, just return gomp_available_cpus.

8 years ago PR libgomp/43569
jakub [Wed, 21 Apr 2010 12:58:59 +0000 (12:58 +0000)]
PR libgomp/43569
* sections.c (gomp_sections_init): Initialize ws->mode.

8 years agoDaily bump.
gccadmin [Tue, 20 Apr 2010 23:16:51 +0000 (23:16 +0000)]
Daily bump.

8 years ago2010-04-20 Harald Anlauf <anlauf@gmx.de>
kargl [Tue, 20 Apr 2010 20:29:39 +0000 (20:29 +0000)]
2010-04-20  Harald Anlauf  <anlauf@gmx.de>

* intrinsic.c (sort_actual): Remove 'is' in error message.

8 years ago PR fortran/43339
jakub [Tue, 20 Apr 2010 07:41:02 +0000 (07:41 +0000)]
PR fortran/43339
* openmp.c (gfc_resolve_do_iterator): Only make iteration vars for
sequential loops private in the innermost containing task region.

* gfortran.dg/gomp/sharing-2.f90: Adjust for iteration vars
of sequential loops being private only in the innermost containing
task region.

8 years ago PR middle-end/43337
jakub [Tue, 20 Apr 2010 07:37:12 +0000 (07:37 +0000)]
PR middle-end/43337
* tree-nested.c (convert_nonlocal_omp_clauses): OMP_CLAUSE_PRIVATE
with non-local decl doesn't need chain.

* gfortran.dg/gomp/pr43337.f90: New test.

8 years agoDaily bump.
gccadmin [Mon, 19 Apr 2010 23:18:14 +0000 (23:18 +0000)]
Daily bump.

8 years ago PR target/43662
jiez [Mon, 19 Apr 2010 09:06:13 +0000 (09:06 +0000)]
PR target/43662
* reginfo.c (reinit_regs): Set caller_save_initialized_p
to false.

testsuite/
PR target/43662
* gcc.target/i386/pr43662.c: New test.

8 years ago PR tree-optimization/43771
irar [Mon, 19 Apr 2010 05:17:56 +0000 (05:17 +0000)]
PR tree-optimization/43771
* tree-vect-analyze.c (vect_supported_load_permutation_p): Check
that load permutation doesn't have gaps.

8 years agoDaily bump.
gccadmin [Sun, 18 Apr 2010 23:16:55 +0000 (23:16 +0000)]
Daily bump.

8 years ago PR tree-optimization/43769
ebotcazou [Sun, 18 Apr 2010 14:56:32 +0000 (14:56 +0000)]
PR tree-optimization/43769
* tree-sra.c (bitfield_overlaps_p): If the length of the element is
self-referential, try to compute an upper bound.

8 years agoDaily bump.
gccadmin [Sat, 17 Apr 2010 23:16:41 +0000 (23:16 +0000)]
Daily bump.

8 years agoDaily bump.
gccadmin [Fri, 16 Apr 2010 23:16:57 +0000 (23:16 +0000)]
Daily bump.

8 years agoRemove previous ChangeLog entry. It was committed to the
kargl [Fri, 16 Apr 2010 14:54:38 +0000 (14:54 +0000)]
Remove previous ChangeLog entry.  It was committed to the
wrong branch.

8 years ago2010-04-16 Steven G. Kargl <kargl@gcc.gnu.org>
kargl [Fri, 16 Apr 2010 14:48:40 +0000 (14:48 +0000)]
2010-04-16  Steven G. Kargl  <kargl@gcc.gnu.org>

    PR fortran/30073
    * trans-array.c (gfc_trans_array_bound_check): Eliminate a redundant
    block of code.  Set name to the variable associated with the descriptor.

8 years agoDaily bump.
gccadmin [Thu, 15 Apr 2010 23:17:17 +0000 (23:17 +0000)]
Daily bump.

8 years agoDaily bump.
gccadmin [Wed, 14 Apr 2010 23:17:11 +0000 (23:17 +0000)]
Daily bump.

8 years agoDaily bump.
gccadmin [Tue, 13 Apr 2010 23:16:42 +0000 (23:16 +0000)]
Daily bump.

8 years agoDaily bump.
gccadmin [Mon, 12 Apr 2010 23:16:43 +0000 (23:16 +0000)]
Daily bump.

8 years ago PR tree-optimization/43560
jakub [Mon, 12 Apr 2010 09:25:11 +0000 (09:25 +0000)]
PR tree-optimization/43560
* gcc.c-torture/execute/pr43560.c: New test.

8 years agoDaily bump.
gccadmin [Sun, 11 Apr 2010 23:16:43 +0000 (23:16 +0000)]
Daily bump.

8 years agoDaily bump.
gccadmin [Sat, 10 Apr 2010 23:16:39 +0000 (23:16 +0000)]
Daily bump.

8 years ago PR target/43417
jiez [Sat, 10 Apr 2010 14:14:13 +0000 (14:14 +0000)]
PR target/43417
* gcc.target/sh/pr43417.c: New test.

8 years ago2010-04-10 Jonathan Wakely <jwakely.gcc@gmail.com>
redi [Sat, 10 Apr 2010 12:46:25 +0000 (12:46 +0000)]
2010-04-10  Jonathan Wakely  <jwakely.gcc@gmail.com>

Backport:
2009-06-23  Paolo Carlini  <paolo.carlini@oracle.com>

PR libstdc++/40518
* include/bits/basic_string.h (basic_string<>::_Rep::
_M_set_length_and_sharable): Do not write the empty rep.
(basic_string<>::erase(iterator, iterator)): Likewise,
move out of line...
* include/bits/basic_string.tcc: ... here.

8 years agoDaily bump.
gccadmin [Fri, 9 Apr 2010 23:16:38 +0000 (23:16 +0000)]
Daily bump.

8 years ago Backport from mainline:
danglin [Fri, 9 Apr 2010 00:40:47 +0000 (00:40 +0000)]
Backport from mainline:
2009-12-05  John David Anglin  <dave.anglin@nrc-cnrc.gc.ca>

PR ada/41912
* pa/linux-unwind.h (pa32_fallback_frame_state): Set fs->signal_frame
for signal frames.
* pa/hpux-unwind.h (pa32_fallback_frame_state): Likewise.

8 years ago Backport:
danglin [Fri, 9 Apr 2010 00:37:42 +0000 (00:37 +0000)]
Backport:
2009-12-09  John David Anglin  <dave.anglin@nrc-cnrc.gc.ca>

* include/pa-signal.h: Do not include ucontext.h.
(SIGNAL_HANDLER): Add __attribute__ ((unused)) to signal handler
arguments.
(MAKE_THROW_FRAME): Do not adjust program counter.
* include/hppa-signal.h: Do not include sys/types.h, sys/syscall.h and
unistd.h.
(MAKE_THROW_FRAME): Do not adjust program counter.
(HANDLE_FPE): Define.
(INIT_SEGV, INIT_FPE): Revise indentation.

8 years ago PR target/43458
danglin [Thu, 8 Apr 2010 23:26:41 +0000 (23:26 +0000)]
PR target/43458
* testsuite/26_numerics/headers/cmath/c99_classification_macros_c.cc:
Xfail hppa*-*-hpux*.

8 years agoDaily bump.
gccadmin [Thu, 8 Apr 2010 23:16:45 +0000 (23:16 +0000)]
Daily bump.

8 years ago gcc:
ro [Thu, 8 Apr 2010 14:09:17 +0000 (14:09 +0000)]
gcc:
PR target/43643
* config/i386/gmon-sol2.c [__x86_64__]: Properly restore %rcx.

gcc/testsuite:
PR target/43643
* gcc.dg/pr43643.c: New test.

gcc:
Backport from mainline:
2010-03-22  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>

PR target/38085
* config/i386/i386.c (x86_function_profiler)
[!NO_PROFILE_COUNTERS]: Fix typo.
* config/i386/gmon-sol2.c (_mcleanup) [__x86_64__]: Use call
instead of callq.

libgcc:
Backport from mainline:
2010-03-22  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>

PR target/38085
* config/i386/t-sol2 ($(T)gmon.o): Use CFLAGS instead of
MULTILIB_CFLAGS.
($(T)gcrt1.o): Likewise.
($(T)crt1.o): Likewise.
($(T)crti.o): Likewise.
($(T)crtn.o): Likewise.

8 years ago2010-04-08 Richard Guenther <rguenther@suse.de>
rguenth [Thu, 8 Apr 2010 10:47:13 +0000 (10:47 +0000)]
2010-04-08  Richard Guenther  <rguenther@suse.de>

PR middle-end/42956
* gimplify.c (gimple_fold_indirect_ref): Avoid generating
new ARRAY_REFs on variable size element or minimal index arrays.
* tree-ssa-loop-ivopts.c (find_interesting_uses_address): Use
gimple_fold_indirect_ref.

* gcc.c-torture/compile/pr42956.c: New testcase.

8 years ago2010-04-08 Wolfgang Gellerich <gellerich@de.ibm.com>
krebbel [Thu, 8 Apr 2010 10:42:59 +0000 (10:42 +0000)]
2010-04-08  Wolfgang Gellerich  <gellerich@de.ibm.com>

* config/s390/s390.c (override_options): Adjust the z10
          defaults for max-unroll-times, max-completely-peeled-insns
          and max-completely-peel-times.

8 years ago2010-04-08 Wolfgang Gellerich <gellerich@de.ibm.com>
krebbel [Thu, 8 Apr 2010 10:36:14 +0000 (10:36 +0000)]
2010-04-08  Wolfgang Gellerich  <gellerich@de.ibm.com>

* config/s390/s390.c (override_options): Set
default of max-pending-list-length to 256

8 years ago Backport from mainline:
jakub [Thu, 8 Apr 2010 10:31:00 +0000 (10:31 +0000)]
Backport from mainline:
2010-03-29  Richard Guenther  <rguenther@suse.de>

PR tree-optimization/43560
* tree-ssa-loop-im.c (ref_always_accessed_p): Add store_p
parameter.
(can_sm_ref_p): Treat stores to readonly locations as
trapping.

* gcc.dg/torture/pr43560.c: New testcase.

8 years ago Backport from mainline:
jakub [Thu, 8 Apr 2010 10:29:28 +0000 (10:29 +0000)]
Backport from mainline:
2010-04-01  Richard Guenther  <rguenther@suse.de>

PR tree-optimization/43607
* ipa-type-escape.c (check_call): Do not access non-existing
arguments.

8 years ago Backport from mainline:
jakub [Thu, 8 Apr 2010 10:28:06 +0000 (10:28 +0000)]
Backport from mainline:
2010-04-01  Richard Guenther  <rguenther@suse.de>

PR middle-end/43614
* tree-ssa-address.c (copy_mem_ref_info): Copy TREE_SIDE_EFFECTS
and TREE_THIS_VOLATILE.
(copy_ref_info): Likewise.
* tree-ssa-operands.c (get_tmr_operands): Check TREE_THIS_VOLATILE.
* tree.c (build7_stat): Ignore side-effects of all but arg5
for TARGET_MEM_REF.  Set TREE_THIS_VOLATILE from arg5 of
TARGET_MEM_REF.

* gcc.c-torture/compile/pr43614.c: New testcase.

8 years ago2010-04-08 Richard Guenther <rguenther@suse.de>
rguenth [Thu, 8 Apr 2010 09:46:46 +0000 (09:46 +0000)]
2010-04-08  Richard Guenther  <rguenther@suse.de>

PR tree-optimization/43186
* params.def (PARAM_MAX_UNROLL_ITERATIONS): New param.
* doc/invoke.texi (max-completely-peel-loop-nest-depth): Document.
* tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely): Limit
unroller iterations.

* gcc.c-torture/compile/pr43186.c: New testcase.

8 years agoDaily bump.
gccadmin [Wed, 7 Apr 2010 23:17:08 +0000 (23:17 +0000)]
Daily bump.

8 years agoFix ChangeLog entries.
hjl [Wed, 7 Apr 2010 20:59:26 +0000 (20:59 +0000)]
Fix ChangeLog entries.

8 years agoAlign stack to 16byte for FP register save area.
hjl [Wed, 7 Apr 2010 20:48:51 +0000 (20:48 +0000)]
Align stack to 16byte for FP register save area.

gcc/

2010-04-06  H.J. Lu  <hongjiu.lu@intel.com>

PR target/43668
* config/i386/i386.c (setup_incoming_varargs_64): Align stack to
16byte for FP register save area.

gcc/testsuite/

2010-04-06  H.J. Lu  <hongjiu.lu@intel.com>

PR target/43668
* gcc.target/i386/pr43668.c: New.

8 years agoPR fortran/40539 Document LOGICAL representation
jb [Wed, 7 Apr 2010 17:07:48 +0000 (17:07 +0000)]
PR fortran/40539 Document LOGICAL representation

8 years ago2010-04-07 Richard Guenther <rguenther@suse.de>
rguenth [Wed, 7 Apr 2010 14:40:43 +0000 (14:40 +0000)]
2010-04-07  Richard Guenther  <rguenther@suse.de>

PR tree-optimization/43629
* tree-ssa-ccp.c (likely_value): Properly look for constant
values.  Reset all_undefined_operands if we have seen a
constant value.

* gcc.c-torture/execute/pr43629.c: New testcase.