This fixes PR55525, when with LTO you have mismatching types
for decls we ICE when we have an ARRAY_REF of sth that is not
an array.
Fixup, similar as for component-refs.
LTO bootstrapped and tested on x86_64-unknown-linux-gnu, applied to trunk.
Richard.
2012-12-05 Richard Biener
There was a bug in syscall.Splice: it ignored the offsets that were
passed in. This patch fixes it. Bootstrapped and ran Go testsuite on
x86_64-unknown-linux-gnu. Committed to mainline and 4.7 branch.
Ian
diff -r 36bbf2143cba libgo/go/syscall/libcall_linux.go
---
On Tue, Dec 4, 2012 at 8:26 PM, Steven Bosscher stevenb@gmail.com wrote:
Hello,
Just comment updates, and a bit of documentation.
Not sure if the comment changes need a ChangeLog.
They do.
OK for trunk?
Ok.
Thanks,
Richard.
Ciao!
Steven
I've merged revision 194186 from trunk to the gccgo branch.
Ian
Committed to trunk.
Richard.
2012-12-05 Richard Biener rguent...@suse.de
PR tree-optimization/50222
* gcc.dg/torture/pr50222.c: New testcase.
Index: gcc/testsuite/gcc.dg/torture/pr50222.c
===
---
Ping~
On 22/11/12 16:49, Yufeng Zhang wrote:
Hi,
This patch implements the correct name mangling of AArch64 AdvSIMD
vector types in conformance to the AAPCS64 doc (Procedure Call Standard
for the ARM 64-bit Architecture, Appendix A).
OK for the trunk?
Thanks,
Yufeng
gcc/ChangeLog
OK
/Marcus
On 22 November 2012 16:49, Yufeng Zhang yufeng.zh...@arm.com wrote:
Hi,
This patch implements the correct name mangling of AArch64 AdvSIMD vector
types in conformance to the AAPCS64 doc (Procedure Call Standard for the ARM
64-bit Architecture, Appendix A).
OK for the trunk?
Which is fixed on trunk.
Committed.
Richard.
2012-12-05 Richard Biener rguent...@suse.de
PR tree-optimization/51238
* gcc.dg/torture/pr51238.c: New testcase.
Index: gcc/testsuite/gcc.dg/torture/pr51238.c
===
In order to be able to rely on frontend marks for checks in formal verification
(Alfa mode), we need to set the overflow mode to STRICT, thus forcing checking
formal verification when the user wants no run-time checks. Modes MINIMIZED
and ELIMINATED are respected if set by the user.
Tested on
Building the limited view of a package does not perform a semantic analysis of
the package. Nevertheless the limited-with clause must carry the entity of the
unit, so that warnings on unused with_clauses can be properly supported.
No short reproducer available.
Tested on x86_64-pc-linux-gnu,
This patch separates handling of overflow checks and overflow
modes. Internally several names are changed to reflect this.
The externally visible effect is that pragma Overflow_Checks
is changed to Overflow_Mode, CHECKED mode is renamed to STRICT,
and the Overflow_Mode pragma no longer affects
This patch remove discrepancies between the documentation on
implementation-defined
restrictions, and the internal data structure used for warnings and errors
on implementation-defined restriction violations.
Tested on x86_64-pc-linux-gnu, committed on trunk
2012-12-05 Ed Schonberg
This patch removes the setting of the overflow flag for MOD, REM,
and unary + operators, since overflow is not possible.
This does not have a functional effect for the compiler since
overflow is not possible anyway, but it did affect some other
tools (notably gnatprove, which had to do extra work
If the left operand of an equality operator is null, the visibility of the
operator must be determined from the interpretation of the right operand. This
processing was done for universal numeric types but not for Any_Access, the
internal representation of the type of the literal null.
Compiling
This patch removes a storage leak in gnatchop, and makes it slightly more
efficient. No noticeable change in behavior, so no test.
Tested on x86_64-pc-linux-gnu, committed on trunk
2012-12-05 Bob Duff d...@adacore.com
* gnatchop.adb (Read_File): Avoid storage leak, and in most cases
This patch provides the initial implementation of aspect/pragma Contract_Cases.
This construct is intended for formal verification proofs.
The syntax of the aspect is as follows:
ASPECT_DEFINITION ::= ( POST_CASE_LIST )
POST_CASE_LIST ::= POST_CASE {, POST_CASE_LIST}
POST_CASE ::=
An early version of the Ada 2012 RM made it illegal to declare the return object
aliased in an extended return statement. A later decision (AI05-0277) made it
legal in the confirming case where the return type is immutably limited, This
patch accordingly moves the application of this rule from the
The version of multi-precision division algorithm used was based on Algorithm D
published in 2nd edition of Donald Knuth's The Art of Computer
Programming. This algorithm was corrected twice, in 1995 and 2005, to protect
against a possible overflow. Although this problem may not be present in this
This patch splits pragma Loop_Assertion into two distinct pragmas. They have
the following syntax:
pragma Loop_Invariant ( boolean_EXPRESSION );
pragma Loop_Variant
( LOOP_VARIANT_ITEM {, LOOP_VARIANT_ITEM } );
LOOP_VARIANT_ITEM ::= CHANGE_DIRECTION = discrete_EXPRESSION
Issue more precise error messages when Contract_Cases aspects or pragmas are
given more than once on a given subprogram.
Tested on x86_64-pc-linux-gnu, committed on trunk
2012-12-05 Yannick Moy m...@adacore.com
* aspects.ads (No_Duplicates_Allowed): Forbid use of duplicate
This patch provides the initial implementation of attribute Update. This
construct is intended for formal verification proofs.
The syntax of the attribute is as follows:
X'Update ( RECORD_COMPONENT_ASSOCIATION_LIST )
X'Update ( ARRAY_COMPONENT_ASSOCIATION {, ARRAY_COMPONENT_ASSOCIATION} )
The patch implements properly the rule given in this Ada issue. The check is
purely syntactic and does not depend on whether the designated type is itself a
null excluding access type.
Compiling the following in Ada 2012 mode must yield:
alloc.adb:23:16:
an allocator with a subtype
LGTM
On Wed, Dec 5, 2012 at 2:59 PM, Konstantin Serebryany
konstantin.s.serebry...@gmail.com wrote:
Hi,
The attached patch is the libsanitizer mege from upstream r169371.
Automatically generated by libsanitizer/merge.sh
Tested with
rm -rf */{*/,}libsanitizer \
make -j 50 \
make -C
Jakub,
I hope I addressed all your comments in this revision.
I've fixed nand atomic operation, made atomic operations atomic again
and added visibility attribute to interface functions.
On Wed, Dec 5, 2012 at 3:11 PM, Dmitry Vyukov dvyu...@google.com wrote:
LGTM
On Wed, Dec 5, 2012 at 2:59
This change to the SCO generation circuitry adds support for Ada 2012 aspects.
SCO lines will be generated for precondition, postcondition, invariant, and
predicate aspects.
The following compilation must produce the indicated CA SCO:
$ gcc -c -fdump-scos -gnata -gnat12 main.adb
$ grep ^CA
Hi!
The c-c++-common/asan/global-overflow-1.c test fails because
we don't instrument local common vars (those that go into .lcomm).
Of course we can't instrument them if they are emitted using .lcomm,
because the linker can arbitrarily reorder the .lcomm vars for best packing,
but we can
Hi!
Another problem from the compiler side when working on the asan testsuite
is that at higher -O* levels the __asan_report_* noreturn calls are
cross-jumped, but the library relies on their locus to print accurrate
locations when symbolized. Without it, asan might report an error
completely
On Wed, Dec 5, 2012 at 3:39 PM, Jakub Jelinek ja...@redhat.com wrote:
Hi!
Another problem from the compiler side when working on the asan testsuite
is that at higher -O* levels the __asan_report_* noreturn calls are
cross-jumped, but the library relies on their locus to print accurrate
Hi!
On Tue, Dec 04, 2012 at 10:00:35AM -0800, Wei Mi wrote:
I updated the patch according to the comments. Please take a look. Thanks.
Rather than several further rounds of review, let me post an incremental
patch on top of yours. Your patch seems to be incomplete BTW, that
asan-special.exp
On Wed, Dec 05, 2012 at 03:13:20PM +0400, Dmitry Vyukov wrote:
I hope I addressed all your comments in this revision.
I've fixed nand atomic operation, made atomic operations atomic again
and added visibility attribute to interface functions.
Mostly, still, __sync_lock_test_and_set isn't full
On Wed, Dec 05, 2012 at 02:59:48PM +0400, Konstantin Serebryany wrote:
The attached patch is the libsanitizer mege from upstream r169371.
Automatically generated by libsanitizer/merge.sh
Tested with
rm -rf */{*/,}libsanitizer \
make -j 50 \
make -C gcc check-g{cc,++}
On Wed, Dec 5, 2012 at 4:42 PM, Jakub Jelinek ja...@redhat.com wrote:
On Wed, Dec 05, 2012 at 03:13:20PM +0400, Dmitry Vyukov wrote:
I hope I addressed all your comments in this revision.
I've fixed nand atomic operation, made atomic operations atomic again
and added visibility attribute to
Yeah, but then there will be all that additional __sync_synchronize(),
that are not needed if we use __atomic. And on the other hand, if we
have that __sync_synchronize(), then we do not need __atomic...
On Wed, Dec 5, 2012 at 4:57 PM, Jakub Jelinek ja...@redhat.com wrote:
On Wed, Dec 05, 2012
http://llvm.org/viewvc/llvm-project/compiler-rt/trunk/lib/tsan/rtl/tsan_interface_atomic.cc?view=diffr1=169378r2=169379pathrev=169379
We will integrate it later.
On Wed, Dec 5, 2012 at 5:03 PM, Dmitry Vyukov dvyu...@google.com wrote:
Yeah, but then there will be all that additional
On Wed, Dec 5, 2012 at 4:43 PM, Jakub Jelinek ja...@redhat.com wrote:
On Wed, Dec 05, 2012 at 02:59:48PM +0400, Konstantin Serebryany wrote:
The attached patch is the libsanitizer mege from upstream r169371.
Automatically generated by libsanitizer/merge.sh
Tested with
rm -rf
There is a proposal to add several user-defined literals to the standard
library that looks like it will probably go through:
http://www.open-std.org/jtc1/sc22/wg21/docs/papers/2012/n3468.pdf
The proposed string literals
std::string hw = Hello, world!s;
will not parse as a literal in gcc.
This
Is this patch OK to commit?
OK
/Marcus
Thanks Marcus,
I've also back-ported this to AArch64-4.7-branch.
The back-port was clean and passed a regression run on
aarch64-none-elf with no regressions.
Thanks,
James
2012-12-04 Marc Glisse marc.gli...@inria.fr
PR target/54855
gcc/
* simplify-rtx.c (simplify_binary_operation_1) VEC_CONCAT: Replace
with VEC_MERGE.
* config/i386/sse.md (sse_vmplusminus_insnmode3): Rewrite
pattern.
*
Hi!
This patch fixes
-FAIL: g++.dg/asan/asan_test.C -O2 AddressSanitizer_LongJmpTest execution test
-FAIL: g++.dg/asan/asan_test.C -O2 AddressSanitizer_BuiltinLongJmpTest
execution test
-FAIL: g++.dg/asan/asan_test.C -O2 AddressSanitizer_SigLongJmpTest execution
test
(both x86_64 and
Hi gentlemen.
Just in case you missed this post. This only affects Darwin, so it's
technically not just an Objective C problem. This problem came about
with Ian Sandoe's patch of rev@170260. The CPP builtin __OBJC2__ is
only set for Darwin.
FYI, in this patch I followed the way we
On Wed, Dec 05, 2012 at 03:49:52PM +0100, Jakub Jelinek wrote:
+FAIL: g++.dg/asan/asan_test.C -O2 AddressSanitizer_PthreadExitTest
execution test
but that looks like a library problem to me:
AddressSanitizer CHECK failed:
../../../../../libsanitizer/asan/asan_rtl.cc:271
r194221
It breaks bootstrap on x86_64-apple-darwin10:
/opt/gcc/build_a/./gcc/xg++ -B/opt/gcc/build_a/./gcc/ -nostdinc++ -nostdinc++
-I/opt/gcc/build_a/x86_64-apple-darwin10.8.0/libstdc++-v3/include/x86_64-apple-darwin10.8.0
-I/opt/gcc/build_a/x86_64-apple-darwin10.8.0/libstdc++-v3/include
* Tom Tromey, 2012-11-29 :
Thomas Now, an alternative might be to have an array of instance IDs stored
at
Thomas the struct line_maps level, with the same indices at the set of
ordinary
Thomas maps, which would be allocated only when flag_debug_instances is used;
Thomas when it is not
Yea, sorry.
I've noticed this upstream and committed a fix there.
Feel free to remove these lines (or wait until I do another merge
first thing tomorrow).
--kcc
On Wed, Dec 5, 2012 at 7:13 PM, Dominique Dhumieres domi...@lps.ens.fr wrote:
r194221
It breaks bootstrap on x86_64-apple-darwin10:
Hi!
On Wed, Dec 05, 2012 at 04:06:51PM +0100, Jakub Jelinek wrote:
FAIL: g++.dg/asan/asan_test.C -O2 AddressSanitizer_GlobalStringConstTest
Ident(p[15]) execution test
So, this one fails because we don't instrument string literals, in GCC
they aren't anything close to global variables (like
Found by Coverity's scanner.
Committed as Rev. 194225 after building with the patch.
Tobias
Index: ChangeLog
===
--- ChangeLog (Revision 194224)
+++ ChangeLog (Arbeitskopie)
@@ -1,3 +1,7 @@
+2012-12-05 Tobias Burnus
On 12/03/2012 08:27 AM, Dodji Seketeli wrote:
+ - HAS_EXPANSION_ARG_P: Set to TRUE iff at least one parameter
+ pack has got an argument that is an expansion.
The got is unnecessary, just has an argument is better.
+ Setup APS, which is an instance of an ARGUMENT_PACK_SELECT
First, I would like to *ping* the patch
http://gcc.gnu.org/ml/fortran/2012-12/msg00022.html
The patch below is part of my FINAL patch. The NULL - NULL_TREE change
is obvious. The expr to al-expr change is to pass a BT_CLASS instead of
a BT_DERIVED. And the NULL to gfc_lval_expr_from_sym
Hi!
This patch tries to handle bitfield accesses as if it were writes/reads
from the corresponding DECL_BIT_FIELD_REPRESENTATIVE.
Richard, does it make sense this way (the code will take ADDR_EXPR
of the COMPONENT_REF with the DECL_BIT_FIELD_REPRESENTATIVE and use
the size of the representative
On Wed, Dec 5, 2012 at 4:43 PM, Jakub Jelinek ja...@redhat.com wrote:
Hi!
On Wed, Dec 05, 2012 at 04:06:51PM +0100, Jakub Jelinek wrote:
FAIL: g++.dg/asan/asan_test.C -O2 AddressSanitizer_GlobalStringConstTest
Ident(p[15]) execution test
So, this one fails because we don't instrument
On Wed, 5 Dec 2012, Jakub Jelinek wrote:
Hi!
This patch tries to handle bitfield accesses as if it were writes/reads
from the corresponding DECL_BIT_FIELD_REPRESENTATIVE.
Richard, does it make sense this way (the code will take ADDR_EXPR
of the COMPONENT_REF with the
Can the report builtins be marked with certain attribute such as
'no-return' etc?
David
On Wed, Dec 5, 2012 at 3:48 AM, Konstantin Serebryany
konstantin.s.serebry...@gmail.com wrote:
On Wed, Dec 5, 2012 at 3:39 PM, Jakub Jelinek ja...@redhat.com wrote:
Hi!
Another problem from the compiler
On Wed, Dec 05, 2012 at 05:10:56PM +0100, Richard Biener wrote:
This patch tries to handle bitfield accesses as if it were writes/reads
from the corresponding DECL_BIT_FIELD_REPRESENTATIVE.
Richard, does it make sense this way (the code will take ADDR_EXPR
of the COMPONENT_REF with the
On Wed, Dec 05, 2012 at 08:24:59AM -0800, Xinliang David Li wrote:
Can the report builtins be marked with certain attribute such as
'no-return' etc?
They are marked as noreturn obviously (and leaf, nothrow etc.).
But noreturn isn't attribute that prevents cross-jumping. It is fine
to
Il 05/12/2012 15:22, Marc Glisse ha scritto:
+
+ /* The x86 back-end uses VEC_CONCAT to set an element in a V2DF, but
+VEC_MERGE for scalar operations that preserve the other elements
+of a vector. */
+ if (GET_CODE (trueop1) == VEC_SELECT
+ GET_MODE (XEXP
Ok. Tail/head merging optimization also happens after PRE. Though it
happens before asan instrumentation, in theory it can trigger similar
bogus loc problem, but less likely.
David
On Wed, Dec 5, 2012 at 8:35 AM, Jakub Jelinek ja...@redhat.com wrote:
On Wed, Dec 05, 2012 at 08:24:59AM -0800,
On Wed, Dec 5, 2012 at 7:43 PM, Jakub Jelinek ja...@redhat.com wrote:
Hi!
On Wed, Dec 05, 2012 at 04:06:51PM +0100, Jakub Jelinek wrote:
FAIL: g++.dg/asan/asan_test.C -O2 AddressSanitizer_GlobalStringConstTest
Ident(p[15]) execution test
So, this one fails because we don't instrument
On Wed, Dec 05, 2012 at 09:42:08PM +0400, Konstantin Serebryany wrote:
On Wed, Dec 5, 2012 at 7:43 PM, Jakub Jelinek ja...@redhat.com wrote:
So, this one fails because we don't instrument string literals, in GCC
they aren't anything close to global variables (like they are in LLVM?)
In
On Wed, Dec 5, 2012 at 9:46 PM, Jakub Jelinek ja...@redhat.com wrote:
On Wed, Dec 05, 2012 at 09:42:08PM +0400, Konstantin Serebryany wrote:
On Wed, Dec 5, 2012 at 7:43 PM, Jakub Jelinek ja...@redhat.com wrote:
So, this one fails because we don't instrument string literals, in GCC
they
This is a division by zero ICE.
In the testcase in the PR, both `n' and `lra_live_max_point' are zero.
I have opted to inhibit the dump when lra_live_max_point is zero, but I
can just as easily avoiding printing the percentage in this case.
Let me know what you prefer.
OK for trunk?
commit
Hi,
The attached patch is the libsanitizer merge from upstream r169392.
Among other things It fixes the darwin build.
Automatically generated by libsanitizer/merge.sh
Tested with
rm -rf */{*/,}libsanitizer \
make -j 50 \
make -C gcc check-g{cc,++}
Adjusted ChangeLog entry as suggested by Jakub on IRC.
OK?
commit ddc4494409331e5a75814a6c8a57aa0646257fb2
Author: Aldy Hernandez al...@redhat.com
Date: Wed Dec 5 08:38:10 2012 -0600
PR target/54160
* config/t-darwin (darwin-c.o): Depend on $(CPP_INTERNAL_H).
*
On 12/05/12 12:13, Aldy Hernandez wrote:
This is a division by zero ICE.
In the testcase in the PR, both `n' and `lra_live_max_point' are zero. I
have opted to inhibit the dump when lra_live_max_point is zero, but I
can just as easily avoiding printing the percentage in this case.
Let me know
On Wed, Dec 5, 2012 at 7:13 PM, Aldy Hernandez wrote:
This is a division by zero ICE.
In the testcase in the PR, both `n' and `lra_live_max_point' are zero. I
have opted to inhibit the dump when lra_live_max_point is zero, but I can
just as easily avoiding printing the percentage in this
Hi!
Here is instrumentation of STRING_CSTs. Using
TREE_TYPE (shadow_ptr_types[0]) instead of plain char_type_node is
a hack to avoid instrumenting the asan_pp_string created STRING_CSTs
that describe the protected automatic vars in functions and protected
global vars, no need to waste space for
The transition of libasan on darwin from using mach_override to the
replacement mac function
imposition code results in a non-functional static libasan (PR55599). The
method now used needs a
dynamic shared library and retaining a static libasan would require restoring
the deprecated
and
On Wed, Dec 5, 2012 at 8:11 PM, Steven Bosscher wrote:
The function that triggers this, has this RTL after IRA:
At which point, of course, I wanted to show after substituting hard
regs for the pseudos, so that:
1: NOTE_INSN_DELETED
3: NOTE_INSN_BASIC_BLOCK 2
2:
On 12-12-05 2:11 PM, Steven Bosscher wrote:
On Wed, Dec 5, 2012 at 7:47 PM, Vladimir Makarov wrote:
On 12-12-05 1:13 PM, Aldy Hernandez wrote:
This is a division by zero ICE.
In the testcase in the PR, both `n' and `lra_live_max_point' are zero. I
have opted to inhibit the dump when
On 12-12-05 1:13 PM, Aldy Hernandez wrote:
This is a division by zero ICE.
In the testcase in the PR, both `n' and `lra_live_max_point' are zero. I
have opted to inhibit the dump when lra_live_max_point is zero, but I can
just as easily avoiding printing the percentage in this case.
Other than David approving of the powerpc stuff, and Gerald Pfeifer asking a
question, I didn't get any response for my patch to add new configuration
options to enable additional executable/startfile/shared library prefixes:
http://gcc.gnu.org/ml/gcc-patches/2012-11/msg00144.html
Can I get a
On Wed, 5 Dec 2012, Paolo Bonzini wrote:
Il 05/12/2012 15:22, Marc Glisse ha scritto:
+
+ /* The x86 back-end uses VEC_CONCAT to set an element in a V2DF, but
+ VEC_MERGE for scalar operations that preserve the other elements
+ of a vector. */
+ if (GET_CODE
Hello!
Following aarch64 example, we can define FP_TRAPPING_EXCEPTIONS for x86, too.
2012-12-05 Uros Bizjak ubiz...@gmail.com
* config/i386/sfp-machine.h (FP_EX_ALL): Define.
(FP_TRAPPING_EXCEPTIONS): Define.
* config/i386/32/sfp-machine.h (FP_EX_SHIFT): Define.
It could work too. Either code is ok to me. The testscase is very rare
and can be achieved only for a small program. So the code speed is not
important. I favor a bit your code as it does not change the code
presented in IRA and LRA which might be unified someday. So if you are
ready to
On Dec 5, 2012, at 10:22 AM, Aldy Hernandez al...@redhat.com wrote:
Adjusted ChangeLog entry as suggested by Jakub on IRC.
OK?
Ok. Please watch for hate mail from cpp folks that don't want internal being
used in this way…
could you take a look at the small simplify-rtx bit of this patch to see
if the general approach makes sense to you?
(this targets 4.9 and passes bootstrap+testsuite on x86_64-linux)
The point of this transformation is to avoid writing a second define_insn
in config/i386/sse.md as in the
As mentioned in the PR, this seems to no longer be an issue after we
redesigned the TM abnormal edges (in the uninstrumented code path patchset).
Tested with the STAMP benchmark for both 1 and 4 threads, as well as
with the GCC regression suite.
OK for trunk?
commit
On Dec 5, 2012, at 10:43 AM, Jack Howarth howa...@bromo.med.uc.edu wrote:
The transition of libasan on darwin from using mach_override to the
replacement mac function
imposition code results in a non-functional static libasan (PR55599).
Ok unless the asan people have a better idea...
Look good to me. Please check with David if google-main branch is
currently opened for check-in.
-Rong
https://codereview.appspot.com/6878045/
It (a trunk backport) is not needed for google/main.
David
On Wed, Dec 5, 2012 at 1:46 PM, x...@google.com wrote:
Look good to me. Please check with David if google-main branch is
currently opened for check-in.
-Rong
https://codereview.appspot.com/6878045/
On Thu, Nov 29, 2012 at 2:27 PM, Kyrylo Tkachov kyrylo.tkac...@arm.com wrote:
Hi all,
This patch adds support for the vrint builtins. It also gathers the unspec
definitions in the various .md files in one file: unspecs.md.
A new iterator is defined that iterates over some new unspecs, in a
On Thu, Nov 29, 2012 at 2:27 PM, Kyrylo Tkachov kyrylo.tkac...@arm.com wrote:
Hi all,
This patch adds an effective target check to the testsuite for ARMv8 NEON
support. A corresponding add_options procedure is added.
This is used by the AArch32 NEON intrinsics tests. An entry in the
Below are the port changes to enable building ASAN on powerpc*-linux.
The libsanitizer changes required for powerpc*-linux have already been
committed. This passes bootstrap and regtesting with no regressions
and we also pass the two ASAN test suite test cases.
Ok for mainline?
Peter
This is the file that causes the boostrap to fail if --enable-decimal-float is
used as a configuration option on the LRA branch. You need to compile this in
32-bit mode with either -mcpu=power7, -mcpu=power6, or -mhard-dfp to enable the
decimal instructions. I used -m32 -O2 -O3 -mcpu=power6.
Reviewers: xur, davidxl,
Message:
Please take a look at this patch.
Description:
2012-03-12 Richard Guenther rguent...@suse.de
* gthr.h (__GTHREAD_MUTEX_INIT_FUNCTION): Adjust specification.
* gthr-posix.h (__GTHREAD_MUTEX_INIT_FUNCTION): Define.
Hi!
On Tue, Dec 04, 2012 at 10:00:35AM -0800, Wei Mi wrote:
I updated the patch according to the comments. Please take a look. Thanks.
I've been testing your patch plus all the patches from me posted yesterday
just with asan.exp testing. While doing full pair of bootstraps/regtests,
I've
On Wed, Dec 5, 2012 at 9:47 PM, Aldy Hernandez wrote:
Steven, go right ahead. Let me know when you commit so I can close the PR.
Thanks, committed as r194230 (patch) and r194231 (test case) after
bootstraptest on x86_64-unknown-linux-gnu.
Ciao!
Steven
On Dec 5, 2012, at 3:28 PM, Jakub Jelinek ja...@redhat.com wrote:
Here is an attempt to fix that up.
Ok for trunk?
Ok.
The following tiny patch is to fix ARM bootstrap failure.
The patch was successfully bootstrapped on ARM.
Committed as rev. 194228.
2012-12-05 Vladimir Makarov vmaka...@redhat.com
* lra-constraints.c (check_and_process_move): Add ATTRIBUTE_UNUSED
to parameter sec_mem_p.
On Mon, 12 Nov 2012, John David Anglin wrote:
The attached change fixes PR target/55195.
The attached change updates the above fix. It results in slightly better
code being generated by genattrab. However, the main reason for the change
is that eliminating the use of match_test makes the fix
This patch to libgo cleans up the directory reading code. It now calls
Entersyscall and Exitsyscall around calls to pathconf and closedir,
which are necessary when using a fused file system. It avoids a minor
race condition calling pathconf. It generally neatens the code.
Bootstrapped and ran
On 12-12-05 5:51 PM, Michael Meissner wrote:
This is the file that causes the boostrap to fail if --enable-decimal-float is
used as a configuration option on the LRA branch. You need to compile this in
32-bit mode with either -mcpu=power7, -mcpu=power6, or -mhard-dfp to enable the
decimal
SRA (and perhaps other passes?) introduce “addressable” MEM_REFs to
variables that were not addressable at the time mudflap1 ran registering
valid memory ranges; when mudflap2 ran to emit the checks, it found such
MEM_REFs, marking the variables as addressable and emitting checks that
will fail at
mudflap's free overrider, when verbose traces enabled, calls output
functions that, at least on the version of glibc I'm using, calls free
recursively. Since it happens to be free(NULL), it's quite easy to fix
this problem with the following simple patch. Ok to install?
Avoid infinite recursion
My patch of a couple of days ago to prohibit a parenthesized argument to
a go or defer statement was faulty. I tested whether the expression
starts with a parenthesis, but of course there are valid function calls
that start with a parenthesis but for which the entire expression is not
I merged revision 194240 from trunk to the gccgo branch.
Ian
In fact nothing prevents GCC from having a linkable static ASan runtime.
The reason for the static library being unusable is not in the dynamic
runtime itself, but in mach_override not working well with other GCC
libraries (it is unable to parse some function prologues). Because
ASan is
Hi!
With Marc's recent changes to fold vector CONSTRUCTORs into VECTOR_CSTs
if possible, we now ICE in adjust_temp_type, because the type of VECTOR_CST
is not SCALAR_TYPE_P. Either we can handle VECTOR_CSTs directly
(in that case ignore_overflows is also called on each VECTOR_CST constant,
but
Hi!
We ICE on the following testcase, because perform_implicit_conversion_flags
doesn't guarantee the type of the returned value is boolean_type_node,
if it is some other type compatible with it (in the same_type_p sense),
then simple == boolean_true_node and == boolean_false_node comparisons
98 matches
Mail list logo