Attached is a patch that eliminates the unused stack frame
allocated by gcc 5 with -pg -mprofile-kernel on powepc64le
and brings the code into parity with previous gcc versions.
The patch doesn't do anything to change the emitted code
when -mprofile-kernel is used without -pg. Since the former
op
The gccgo parser had a bug: it rejected labels on fallthrough
statements, because those statements can only appear in specific
contexts (at the end of a switch block just before the next case
statement). This patch fixes the problem. I will add a test to the
master testsuite for this, which will
I am working on SHF_COMPRESSED support:
http://www.sco.com/developers/gabi/latest/ch4.sheader.html
I will import zlib from GCC source tree into binutils-gdb tree. But zlib
failed to build as a target library in binutils-gdb. There is no need to
build target libraries if not building gcc. OK fo
On Fri, Feb 27, 2015 at 8:37 AM, Lynn A. Boger
wrote:
> At a minimum I think it should be mentioned in libgo/README.
(Where "it" means that libgo should not be stripped.)
I've committed this minor update to libgo/README.
Ian
diff -r b0d3b9400c63 libgo/README
--- a/libgo/README Thu Mar 12 1
On Fri, Mar 13, 2015 at 6:55 PM, David Malcolm wrote:
>> As documented in [1] asprintf and vasprintf return:
> I assume that we can rely that any vasprintf implementation manages on
> failure to at least either write NULL to *ret or to return -1, even if
> some of them fail to do both?
Yes, thi
On Fri, 2015-03-13 at 17:53 +0100, Uros Bizjak wrote:
> Hello!
>
> As documented in [1] asprintf and vasprintf return:
>
> --quote--
> Return value:
>
> Both functions set *ret to be a pointer to a malloc()'d buffer
> sufficiently large to hold the formatted string. This pointer should
> be pass
On Wed, 2015-03-04 at 11:09 -0500, David Malcolm wrote:
> On Tue, 2015-03-03 at 11:49 +0100, Matthias Klose wrote:
> > Both gccjit and gnat now use sphinx to build the documentation. While not a
> > direct part of the build process, it would be nice to document the
> > requirements
> > on sphinx,
Hello!
As documented in [1] asprintf and vasprintf return:
--quote--
Return value:
Both functions set *ret to be a pointer to a malloc()'d buffer
sufficiently large to hold the formatted string. This pointer should
be passed to free() to release the allocated storage when it is no
longer needed.
Dear Mikael,
I have been on the road and have only had time for trivial bits and
pieces. The weather forecast from tomorrow afternoon onwards is not
good and so I think that this patch will get some attention :-)
Cheers
Paul
On 12 March 2015 at 19:04, Mikael Morin wrote:
> Hello Paul,
>
> have
On 13-03-15 13:36, Richard Biener wrote:
On Fri, Mar 13, 2015 at 1:07 PM, Jakub Jelinek wrote:
On Fri, Mar 13, 2015 at 01:04:57PM +0100, Richard Biener wrote:
Not really (I don't like -fdump-passes ...), but we need to make sure
that -fdump-passes doesn't crash (because it runs very early and
On Fri, Mar 13, 2015 at 04:24:09PM +0100, Thomas Schwinge wrote:
> Hi Jeff!
>
> On Tue, 10 Mar 2015 12:48:34 -0600, Jeff Law wrote:
> > On 03/10/15 07:36, James Norris wrote:
> > > Ping.
> > Note that the GCC trunk is in regression bugfix stage, so this patch may
> > (is likely?) be deferred unt
Hi Jeff!
On Tue, 10 Mar 2015 12:48:34 -0600, Jeff Law wrote:
> On 03/10/15 07:36, James Norris wrote:
> > Ping.
> Note that the GCC trunk is in regression bugfix stage, so this patch may
> (is likely?) be deferred until the next stage1 development cycle.
The decision is with you guys, but my un
Hi Jim!
Sorry for the delay. Please be a little more descriptive in submissions:
for example, if you had used »OpenACC deviceptr clause« instead of
»deviceptr clause«, my mail client's search/filter would have pointed me
to this email earlier, or, by using a more descriptive subject line:
»[PATCH
We started to reject this (IMHO valid) testcase with r214941 that did away with
try_move_mult_to_index -- meaning that we are no longer able to fold *(&s[0] +
1)
into s[1], while we are able to fold *(s + 1) into s[1].
I suppose cxx_fold_indirect_ref ought to be able to handle both cases, so I
a
On Fri, Mar 13, 2015 at 12:04:28PM -, Kyrill Tkachov wrote:
> > diff --git a/gcc/testsuite/gcc.target/arm/pr65358.c
> > b/gcc/testsuite/gcc.target/arm/pr65358.c
> > new file mode 100644
> > index 000..d663dcf
> > --- /dev/null
> > +++ b/gcc/testsuite/gcc.target/arm/pr65358.c
> > @@ -0,0 +1,
This change is an internal change to add extra fields to each
entity for future use. No functional effect. No test required.
Tested on x86_64-pc-linux-gnu, committed on trunk
2015-03-13 Robert Dewar
* einfo.adb, einfo.ads, atree.adb, atree.ads, atree.h: Add seventh
component t
On 12/03/15 18:49 +, Jonathan Wakely wrote:
I assumed that Pthreads was enough to ensure pthread_rwlock_t but
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=64847 shows that isn't
true for HPUX (seems it was optional prior to POSIX 1003.1-2001).
This adds an autoconf check to decide whether to
This is some preliminary checkins for aspect Unimplemented. No
functionality yet, so no test.
Tested on x86_64-pc-linux-gnu, committed on trunk
2015-03-13 Robert Dewar
* aspects.ads, aspects.adb: Add entries for aspect Unimplemented.
* einfo.ads, einfo.adb (Is_Unimplemented):
This is an internal change to the cleanup and simplify the
generation of FIRST/LAST temporaries for subtype bounds. No
test is needed since no functional effect on the behavior
of the compiler.
Tested on x86_64-pc-linux-gnu, committed on trunk
2015-03-13 Robert Dewar
* exp_util.ads, e
On 03/13/15 13:04, Kyrill Tkachov wrote:
Hi,
Hi Honggyu,
Thanks for helping out. I've got a couple of pointers for the testcase
inline.
I have wrote a testcase that reproduces argument overwriting bug during
arm code generation.
I wrote this testcase with the help of Mikael Pettersson.
If
Andre Vehreschild wrote:
> during debugging I found a segfault of gfortran, when it encounters an illegal
> fortran code. The mistake in Fortran is flagged correctly, but later gfortran
> crashes. This patch prevents the crash.
>
> Bootstraps and regtest ok on x86_64-linux-gnu.
> Ok for trunk?
OK.
Ville spotted this mistake, committed to trunk.
commit 6d2d6ae5c25e9cc71f2f3fa275a33b1ab805dfd9
Author: Jonathan Wakely
Date: Fri Mar 13 12:31:46 2015 +
* include/experimental/system_error: Fix include guard.
diff --git a/libstdc++-v3/include/experimental/system_error b/libstdc++-v3/
On Fri, Mar 13, 2015 at 1:07 PM, Jakub Jelinek wrote:
> On Fri, Mar 13, 2015 at 01:04:57PM +0100, Richard Biener wrote:
>> Not really (I don't like -fdump-passes ...), but we need to make sure
>> that -fdump-passes doesn't crash (because it runs very early and
>> with cfun == NULL I think)
>
> If
Trivial for loop header copying and vector lowering. Remaining
ones done for an empty function at -O1 are DOM and the related
phicprop, both a little more interesting to fix (we usually
defer to CFG cleanup to optimize control stmts with known target).
Bootstrap running on x86_64-unknown-linux-g
On Fri, Mar 13, 2015 at 01:04:57PM +0100, Richard Biener wrote:
> Not really (I don't like -fdump-passes ...), but we need to make sure
> that -fdump-passes doesn't crash (because it runs very early and
> with cfun == NULL I think)
If it runs with cfun == NULL, then supposedly the gates that are d
On Fri, Mar 13, 2015 at 12:12 PM, Tom de Vries wrote:
> On 13-03-15 11:36, Richard Biener wrote:
>>
>> On Fri, Mar 13, 2015 at 11:32 AM, Tom de Vries
>> wrote:
>>>
>>> Hi,
>>>
>>> this patch moves a bunch of early-out tests from the parloops pass to the
>>> gate function.
>>>
>>> The only effect
> Hi,
Hi Honggyu,
Thanks for helping out. I've got a couple of pointers for the testcase
inline.
>
> I have wrote a testcase that reproduces argument overwriting bug during
> arm code generation.
>
> I wrote this testcase with the help of Mikael Pettersson.
> If some format is not proper to ru
On Fri, Mar 6, 2015 at 7:46 AM, Bin Cheng wrote:
> Hi,
> This patch is the second part fixing memset-inline-{4,5,6,8,9}.c failures on
> cortex-a9. It adds a function checking CPU tuning information in dejagnu,
> it also uses that function to skip related testcase when we are compiling
> for corte
On Fri, Mar 6, 2015 at 7:42 AM, Bin Cheng wrote:
> Hi,
> This patch is the first part fixing memset-inline-{4,5,6,8,9}.c failures on
> cortex-a9. GCC/arm doesn't generate any tuning information in assembly, it
> can't tell whether we are compiling for cortex-a9 tune if the compiler is
> configure
Hi Guys,
A bug reported against the RL78 port has a uncovered a generic problem
with expand_expr_real_1() function - part of the code their computes
the offset of an address for an array reference, and if necessary
makes sure that it is in the correct mode:
if (offset)
{
On 13-03-15 11:36, Richard Biener wrote:
On Fri, Mar 13, 2015 at 11:32 AM, Tom de Vries wrote:
Hi,
this patch moves a bunch of early-out tests from the parloops pass to the
gate function.
The only effect is for functions that we don't consider at all for
parallelization in the parloops pass.
On 10/03/15 20:18 +, Jonathan Wakely wrote:
On 10/03/15 20:55 +0100, John Marino wrote:
On 3/10/2015 20:23, Jonathan Wakely wrote:
John, assuming I'm right that dragonfly supports all these features,
could you test this change? (You'll need the same change on line
19555 of the libstdc++-v3
On 13/03/15 11:43, Jakub Jelinek wrote:
On Fri, Mar 13, 2015 at 11:38:12AM +0100, Sebastian Huber wrote:
I would like to commit this patch to GCC 4.9 and 5.0.
libgomp/ChangeLog
2015-03-13 Sebastian Huber
* configure.tgt (*-*-rtems*): Use local-exec TLS model.
* configure.ac
On Fri, Mar 13, 2015 at 11:38:12AM +0100, Sebastian Huber wrote:
> I would like to commit this patch to GCC 4.9 and 5.0.
>
> libgomp/ChangeLog
> 2015-03-13 Sebastian Huber
>
> * configure.tgt (*-*-rtems*): Use local-exec TLS model.
> * configure.ac (*-*-rtems*): Assume Pthread is s
I would like to commit this patch to GCC 4.9 and 5.0.
libgomp/ChangeLog
2015-03-13 Sebastian Huber
* configure.tgt (*-*-rtems*): Use local-exec TLS model.
* configure.ac (*-*-rtems*): Assume Pthread is supported.
(pthread.h): Check for this header file.
* config
On Fri, Mar 13, 2015 at 11:32 AM, Tom de Vries wrote:
> Hi,
>
> this patch moves a bunch of early-out tests from the parloops pass to the
> gate function.
>
> The only effect is for functions that we don't consider at all for
> parallelization in the parloops pass. We no longer dump those in the
>
Hi all,
this is another patch preventing a segfault. This time the segfault occurred,
when -fdump-(fortran|tree)-original was given with the program having an
implicit class set. The issue is that the _data component is assumed to be
present in a BT_CLASS w/o checking and trying to access the unli
Hi all,
during debugging I found a segfault of gfortran, when it encounters an illegal
fortran code. The mistake in Fortran is flagged correctly, but later gfortran
crashes. This patch prevents the crash.
Bootstraps and regtest ok on x86_64-linux-gnu.
Ok for trunk?
Regards,
Andre
--
An
Hi,
this patch moves a bunch of early-out tests from the parloops pass to the gate
function.
The only effect is for functions that we don't consider at all for
parallelization in the parloops pass. We no longer dump those in the parloops
dump file.
Bootstrapped and reg-tested on x86_64.
O
On 3/13/2015 11:23, Jonathan Wakely wrote:
> On 13 March 2015 at 08:50, John Marino wrote:
>> Hi Jonathan,
>> Today I build gcc5 and looked at the config.h contained in
>> libstdc++-v3's build directory and none of those variables are enabled.
>>
>> The code that is patched doesn't seem to do anyth
On 13 March 2015 at 08:50, John Marino wrote:
> Hi Jonathan,
> Today I build gcc5 and looked at the config.h contained in
> libstdc++-v3's build directory and none of those variables are enabled.
>
> The code that is patched doesn't seem to do anything. I grepped the
> entire build directory for "
> From: Richard Biener [mailto:rguent...@suse.de]
> Sent: Friday, March 13, 2015 5:02 PM
> >
> > Is this ok for stage1? It's not a bug but it helps debuggability so is
> > this something we might consider backporting?
>
> It's ok now given you bootstrapped the change.
I did + regression testsuite
Ping.
Thanks,
bin
On Fri, Mar 6, 2015 at 3:46 PM, Bin Cheng wrote:
> Hi,
> This patch is the second part fixing memset-inline-{4,5,6,8,9}.c failures on
> cortex-a9. It adds a function checking CPU tuning information in dejagnu,
> it also uses that function to skip related testcase when we are c
Ping.
This is for case failures and it doesn't affect normal compilation, so
I suppose it's fine for this stage?
Thanks,
bin
On Fri, Mar 6, 2015 at 3:42 PM, Bin Cheng wrote:
> Hi,
> This patch is the first part fixing memset-inline-{4,5,6,8,9}.c failures on
> cortex-a9. GCC/arm doesn't generate
Hi,
are there any more comments on this?
I would like to apply the patch as is, unless we find a
a way to get to a test case, maybe with a cross-compiler,
where the MODE_ALIGNMENT is different from MODE_BITSIZE.
Currently, I think that does not happen.
Thanks
Bernd.
> Date: Tue, 10 Mar 2015 14
On 03/03/15 17:59, Kyrylo Tkachov wrote:
-Original Message-
From: Kyrylo Tkachov
Sent: 27 February 2015 14:30
To: Kyrylo Tkachov; GCC Patches
Cc: Ramana Radhakrishnan; Richard Earnshaw
Subject: RE: [PATCH][ARM] PR target/64600 Fix another ICE with -
mtune=xscale: properly sign-extend
Hi!
On Wed, 25 Feb 2015 11:28:12 +0100, I wrote:
> Using [...] to forcefully disable -fvar-tracking (as done in
> nvptx_option_override), should then allow me to drop the following
> beautiful specimen of a patch (which I didn't commit anywhere, so far):
No progress yet with that, so for now, com
On Fri, 13 Mar 2015, Thomas Preud'homme wrote:
> Hi,
>
> The description is longer than the patch so you might want to skip directly
> to it.
>
> The dot file generated by -fdump-rtl-*-graph switches group basic blocks for
> a given function together in a subgraph and use the function name as
Hi!
Committed to gomp-4_0-branch in r221408:
commit 7543f50b1d9ff29d542fb63b5c77e0a96521756d
Merge: a167fe0 83ba0e6
Author: tschwinge
Date: Fri Mar 13 08:45:16 2015 +
svn merge -r 220892:221363 svn+ssh://gcc.gnu.org/svn/gcc/trunk
git-svn-id: svn+ssh://gcc.gnu.org/svn/gcc
The following fixes quadraticness observed in PR44563 related to
basic block splitting and merging (which both are O(n) in the size
of the "tail" due to adjusting stmts BB pointer). The splitting is induced by
the inliner which processes calls in a basic-block in a non-optimal
order (front-to-bac
On 3/10/2015 23:04, John Marino wrote:
> On 3/10/2015 21:18, Jonathan Wakely wrote:
>> On 10/03/15 20:55 +0100, John Marino wrote:
>>> On 3/10/2015 20:23, Jonathan Wakely wrote:
John, assuming I'm right that dragonfly supports all these features,
could you test this change? (You'll need
On March 13, 2015 7:30:43 AM GMT+01:00, Jan Hubicka wrote:
>Hi,
not commenting on the patch itself.
s/DELC_CXX_CONSTRUCTOR/DECL_CXX_CONSTRUCTOR/
s/consutrctors/constructors/
Thanks,
On Fri, 13 Mar 2015, Jan Hubicka wrote:
> > > CFG cleanup currently searches for calls that became noreturn and
> > > fixes them up (splitting block and removing the fallthru). Previously
> > > that was technically necessary as propagation may have turned an
> > > indirect call into a direct nore
On Fri, 13 Mar 2015, Jan Hubicka wrote:
> > >
> > > Honza - I think we performed this in CFG cleanup for the sake of CFG
> > > checking, not for the sake of prompt optimization, no?
>
> Aha, now I understand what you ask about. I guess it can be seen this way.
> NORETURN calls are CFG altering
Ping!
On Fri, 6 Mar 2015 13:38:21 +0100
Andre Vehreschild wrote:
> Hi everyone,
>
> please find attached the latest version of the patch with the comments from
> Dominique via IRC worked in. Those were mostly about clarifying comments and
> style. Nevertheless, thanks for your help Dominique.
Hi,
The description is longer than the patch so you might want to skip directly to
it.
The dot file generated by -fdump-rtl-*-graph switches group basic blocks for a
given function together in a subgraph and use the function name as the label.
However, when generating an image (for instance a
> >
> > Honza - I think we performed this in CFG cleanup for the sake of CFG
> > checking, not for the sake of prompt optimization, no?
Aha, now I understand what you ask about. I guess it can be seen this way.
NORETURN calls are CFG altering and veriy_flow_infowould indeed bomb.
Honza
Mikael Morin wrote:
this is about a pointer bounds remapping regression introduced at
http://gcc.gnu.org/r190641
[...]
The fix proposed clears the flag upon reuse, which should match exactly
the original behaviour, making it rather safe, and suitable also for the
branches.
I have to admit that I'
58 matches
Mail list logo