The Go frontend was using memcmp for struct equality if there was no
internal padding but there was trailing padding. This doesn't work when
the struct values are on the stack, because there is nothing that
explicitly zeroes out the trailing padding. This patch fixes the bug.
The patch is larger
On Sun, Nov 25, 2012 at 12:54 AM, Hans-Peter Nilsson wrote:
> On Sat, 24 Nov 2012, Steven Bosscher wrote:
>> Hello,
>>
>> The DELAY_SLOTS_FOR_EPILOGUE target macro, and the related
>> ELIGIBLE_FOR_EPILOGUE_DELAY macro, are unused. The attached patch
>> removes them.
>>
>> OK for trunk?
>
> Obviousl
On Sat, 24 Nov 2012, Steven Bosscher wrote:
> Hello,
>
> The DELAY_SLOTS_FOR_EPILOGUE target macro, and the related
> ELIGIBLE_FOR_EPILOGUE_DELAY macro, are unused. The attached patch
> removes them.
>
> OK for trunk?
Obviously; Jeff Law preapproved their removal long ago.
(I forgot about it.) Bu
On Thu, Nov 22, 2012 at 9:22 PM, Joern Rennecke wrote:
(nothing but a ChangeLog :-)
Looking at the ARC port a bit, and IMHO it doesn't look very messy...
First some general comments:
This target apparently wants to play tricks with reload (e.g.
arc_preserve_reload_p). IMHO at this point ports th
Hi.
This eighth patch adds new directives to manifests: @include, @remove.
These are useful when needing to handle a multi-dimensional matrix of test
configurations, and not wanting to maintain the otherwise necessary duplication
across all of the axes.
[@remove should be used sparingly, but I've f
Hi.
This seventh patch adds new options --manifest_subdir, --manifest_name.
Useful when using validate_failures.py with a different tool, instead of gcc.
Ok to check in?
2012-11-24 Doug Evans
* testsuite-management/validate_failures.py: New options
--manifest_subdir,
--manife
Hi.
This sixth patch simplifies calls to GetBuildData.
It never returns false and always terminates the process with an error message
(which is fine by me).
Ok to check in?
2012-11-24 Doug Evans
* testsuite-management/validate_failures.py: Remove pass/fail indicator
from
resu
Hi.
This fifth patch makes options a global variable.
As validate_failures.py becomes more complex, passing it around everywhere
becomes cumbersome with no real gain.
Ok to check in?
2012-11-24 Doug Evans
* testsuite-management/validate_failures.py: Make options a global
variable.
-
Hi.
This fourth patch renames option --manifest to --manifest_path.
I have a later patch that adds a --manifest_name option, making "--manifest"
too confusing/ambiguous.
Ok to check in?
[I still have to update invocations that use --manifest,
but that's a separate patch.]
2012-11-24 Doug Evans
Hi.
This third patch passes options.results to GetSumFiles when fetching the results
for the clean build.
It is useful in my use cases, but I'm not sure it's useful for upstream.
[An alternative is to add another option to specify the results file(s) for the
clean
build, but I'm being conservative
Hi.
This second patch uses "target_alias" instead of "target" in the help text,
and makes some minor whitespace changes.
Ok to check in?
2012-11-24 Doug Evans
* testsuite-management/validate_failures.py: Use instead
of
. Minor whitespace changes.
--- validate_failures.py=
Hi.
This set of 8 patch makes some minor tweaks to validate_failures.py,
and provides some new functionality.
They are relatively independent, but the patches are intended to be applied in
order.
This first patch adds an ordinal to TestResult so that the sorting
keeps the tests in order of appear
Hello,
The DELAY_SLOTS_FOR_EPILOGUE target macro, and the related
ELIGIBLE_FOR_EPILOGUE_DELAY macro, are unused. The attached patch
removes them.
OK for trunk?
Ciao!
Steven
dbr_sched_janitor.diff
Description: Binary data
This patch to libgo fixes the handling of Unix domain @ addresses, a
GNU/Linux feature. The library was computing the length incorrectly. I
actually fixed this in the master library in January 2011, but somehow
failed to propagate the fix into the gccgo library. Bootstrapped and
ran Go testsuite
On Wed, 21 Nov 2012, Mark Wielaard wrote:
> As mentioned in some bug reports it should be documented that DWARF4 is
> now the default for 4.8 when -g is used (and that one might need a newer
> version of debugger/profiling/tracing tools to use it). So I added the
> following:
I applied the followi
In http://gcc.gnu.org/ml/gcc-patches/2012-11/msg00257.html I said:
get_best_mode has various checks to decide what counts as an acceptable
bitfield mode. It actually has two copies of them, with slightly different
alignment checks:
MIN (unit, BIGGEST_ALIGNMENT) > align
vs.
unit
OK.
Jason
For x86_64-linux-gnux32 I used 'x32-linux-gnu' as the multiarch name, which is
wrong, should be 'x86_64-linux-gnux32' instead (see [1]). Pointed out by Daniel
Schepler.
Committed as obvious.
Matthias
[1] http://bugs.debian.org/667037
* gcc/config/i386/t-linux64 (MULTILIB_OSDIRNAMES):
Hi Thomas,
> the attached patch fixes a 4.6/4.7/4.8 rejects-valid regression.
> OK for trunk?
looks good to me. Thanks for the patch. Ok for trunk and also for the
4.6/4.7 branches ...
Cheers,
Janus
> 2012-11-24 Thomas Koenig
>
> PR fortran/55314
> * resolve.c (resolve_all
Hello world,
the attached patch fixes a 4.6/4.7/4.8 rejects-valid regression.
OK for trunk?
Thomas
2012-11-24 Thomas Koenig
PR fortran/55314
* resolve.c (resolve_allocate_deallocate): Compare all
subscripts when deciding if to reject a (de)allocate
s
Quoting Joern Rennecke :
Quoting Richard Sandiford :
1) As Richard B says, having "locked lengths" with the comment "care must
be taken to avoid cycles" doesn't sound like good design. So the
question was: without this, why would the length be going up and down
"arbitrarily", even thou
Quoting Richard Sandiford :
[responding because you kept me on cc:]
Joern Rennecke writes:
This uses the same interface as my previous patch:
http://gcc.gnu.org/ml/gcc-patches/2012-11/msg00473.html ,
but I refined the algorithm for the get_insn_variants
mechanism to work properly with the rew
On Fri, Nov 23, 2012 at 8:39 PM, Jakub Jelinek wrote:
> On Fri, Nov 23, 2012 at 08:10:39PM +0400, Dmitry Vyukov wrote:
>> > This patch attempts to instrument __atomic_* and __sync_* builtins.
>> > Unfortunately for none of the builtins there is 1:1 mapping to the tsan
>> > replacements, tsan uses
On Fri, 23 Nov 2012, H.J. Lu wrote:
> On Fri, Nov 23, 2012 at 9:38 AM, Jakub Jelinek wrote:
> > On Fri, Nov 23, 2012 at 09:23:37AM -0800, H.J. Lu wrote:
> >> This patch allocates extra 16 bytes for -fsanitize=address so that
> >> asan won't report read beyond memory buffer. It is used by
> >> boot
On Fri, 23 Nov 2012, Georg-Johann Lay wrote:
> Here are some more fixes for 16-bit int and similar.
> * gcc.c-torture/execute/20120919-1.x: New file (int32plus).
No, you should be able to use dg-directives in the main file these days.
(The .x files are obsolete since a few years, IIRC.)
br
Without this, I got weird ERRORs and those Tcl backtraces you come
to hate, instead of the expected FAILs. Committed as obvious after
running a failing guality test and obvserving the intended change.
(Well ok, *adding an explanatory comment* is apparently not obvious,
but I'll take that liberty..
[responding because you kept me on cc:]
Joern Rennecke writes:
> This uses the same interface as my previous patch:
> http://gcc.gnu.org/ml/gcc-patches/2012-11/msg00473.html ,
> but I refined the algorithm for the get_insn_variants
> mechanism to work properly with the reworked ARC port -
> http:
On Fri, Nov 23, 2012 at 9:07 PM, Xinliang David Li wrote:
> On Fri, Nov 23, 2012 at 8:39 AM, Jakub Jelinek wrote:
>> On Fri, Nov 23, 2012 at 08:10:39PM +0400, Dmitry Vyukov wrote:
>>> > This patch attempts to instrument __atomic_* and __sync_* builtins.
>>> > Unfortunately for none of the builtin
On Sat, Nov 24, 2012 at 12:11 PM, Jakub Jelinek wrote:
> On Fri, Nov 23, 2012 at 08:10:39PM +0400, Dmitry Vyukov wrote:
>> I've just committed a patch to llvm with failure_memory_order (r168518).
>
> Shouldn't it be something like this instead?
>
> --- tsan_interface_atomic.cc.jj 2012-11-23 17:20:
On Fri, Nov 23, 2012 at 08:10:39PM +0400, Dmitry Vyukov wrote:
> I've just committed a patch to llvm with failure_memory_order (r168518).
Shouldn't it be something like this instead?
--- tsan_interface_atomic.cc.jj 2012-11-23 17:20:49.0 +0100
+++ tsan_interface_atomic.cc2012-11-23 19:
30 matches
Mail list logo
