> This patch consists intrinsics to properly create the bases and
> direct_bases of a class in the correct order (including multiple nested
> ambiguous virtual and non-virtual classes) for N2965
> (http://www.open-std.org/jtc1/sc22/wg21/docs/papers/2009/n2965.html).
> This allows you to create type
> exactly like the recently fixed c++/17212. Tested x86_64-linux.
c-family has its own ChangeLog file, all changes must be documented there.
--
Eric Botcazou
On Oct 17, 2011, at 3:16 PM, Tom Tromey wrote:
>> "Tristan" == Tristan Gingold writes:
>
> Tom> Another way to look at it is that there have been many changes to GCC's
> Tom> DWARF output in the last few years. Surely these have broken these
> Tom> DWARF consumers more than this change pos
On 10/18/2011 07:10 AM, Jeff Law wrote:
--- 467,475
if (gimple_code (use) != GIMPLE_PHI)
{
sinkbb = gimple_bb (use);
! sinkbb = select_best_block (frombb, gimple_bb (use), stmt);
! if (sinkbb == frombb)
return false;
*togsi =
> thread_prologue_and_epilogue_insns should detect all cases where a
> return insn can be created. So any CFG cleanup that runs before it does
> not need this functionality.
So we're left with CFG cleanups that run after it and could forward edges to an
edge from a return insn to the exit block i
On Mon, 17 Oct 2011, Eric Botcazou wrote:
> > 2011-10-17 Richard Guenther
> >
> > PR middle-end/50716
> > * expr.c (get_object_or_type_alignment): New function.
> > (expand_assignment): Use it.
> > (expand_expr_real_1): Likewise.
>
> Maybe move it to builtins.c alongside the ot
On Tue, 18 Oct 2011 09:24:08 +0200, Tristan Gingold wrote:
> What is wrong with my suggestion of adding a command line option to keep the
> siblings ? This option could be removed in a few years if nobody complained
> about sibling removal.
I find an extra option just a pollution of doc and every
On Tue, Oct 18, 2011 at 10:28:09AM +0200, Jan Kratochvil wrote:
> 2011-10-12 Jan Kratochvil
>
> Stop producing DW_AT_sibling without -gstrict-dwarf.
> * dwarf2out.c (dwarf2out_finish): Remove calls of
> add_sibling_attributes if !DWARF_STRICT. Extend the comment with
>
On 10/17/2011 11:37 PM, Janus Weil wrote:
here is another patch for PR47023, which takes care of comment #1,
i.e. rejecting polymorphic variables in a C-binding context.
Regtested on x86_64-unknown-linux-gnu. Ok for trunk?
OK. Thanks for the patch.
If I saw it correctly, one still needs to fi
On Tue, Oct 18, 2011 at 1:45 AM, Maxim Kuvyrkov wrote:
> On 13/10/2011, at 12:58 AM, Richard Guenther wrote:
>
>> On Wed, Oct 12, 2011 at 8:50 AM, Maxim Kuvyrkov
>> wrote:
>>> The following patch adds new knob to make GCC perform several iterations of
>>> early optimizations and inlining.
>>>
>
Jason Merrill writes:
> If you have a patch like this that fixes a major regression, go ahead
> and check it in without waiting for approval; we can adjust it as
> necessary after build is working again.
OK.
>> size_t num_expanded_macros;
>
>> - fprintf (stderr, "Number of expanded macros:
On Mon, 17 Oct 2011, Richard Guenther wrote:
> On Fri, 14 Oct 2011, Richard Guenther wrote:
>
> >
> > This follows up Michas testcase where we fail to handle the
> > conservatively propagated restrict tags properly. The following
> > patch simplifies handling of restrict in the oracle and thus
Gerald Pfeifer writes:
> On Mon, 17 Oct 2011, Dodji Seketeli wrote:
>>> this looks like it's causing the following bootstrap failure for me
>>> on i386-unknown-freebsd9.0?
>> Yes this is http://gcc.gnu.org/bugzilla/show_bug.cgi?id=50760, and I am
>> testing the patch attached to the bug at the mo
On 10/17/2011 01:51 PM, Richard Guenther wrote:
> On Sun, Oct 16, 2011 at 12:05 PM, Tom de Vries wrote:
>> On 10/14/2011 12:00 PM, Richard Guenther wrote:
>>> On Fri, Oct 14, 2011 at 1:12 AM, Tom de Vries
>>> wrote:
On 10/12/2011 02:19 PM, Richard Guenther wrote:
> On Wed, Oct 12, 2011
> In order to align with a standard naming convention, this patch renames
> all the __sync_mem references to __atomic.
>
> The only real addition is in builtins.c::is_builtin_name needed to know
> that __atomic_ is also a builtin prefix.
Are you sure about the renaming get_builtin_sync_mem -> get_
Hi,
>> exactly like the recently fixed c++/17212. Tested x86_64-linux.
>
> c-family has its own ChangeLog file, all changes must be documented there.
Oh my, I thought I had checked. I'll fix it, sorry.
Paolo
On Tue, Oct 18, 2011 at 11:39:22AM +0200, Ira Rosen wrote:
> On 2 October 2011 10:30, Ira Rosen wrote:
> > On 29 September 2011 17:30, Ramana Radhakrishnan
> > wrote:
> >> On 19 September 2011 08:54, Ira Rosen wrote:
> >>
> >>>
> >>> Bootstrapped on powerpc64-suse-linux, tested on powerpc64-suse
On 10/17/2011 07:02 PM, Steve Kargl wrote:
On Mon, Oct 17, 2011 at 12:22:03PM +0200, Tobias Burnus wrote:
I am not so happy with complete test replacements. How about adding it
as new test case?
Well, the old testcase is
[...]
The only line that would survive is the first line, which
is covere
I'm under the impression that some tests are written in C, wouldn't
better fit in c-c++-common?
Thanks,
Paolo.
On 18 October 2011 11:43, Jakub Jelinek wrote:
> On Tue, Oct 18, 2011 at 11:39:22AM +0200, Ira Rosen wrote:
>> On 2 October 2011 10:30, Ira Rosen wrote:
>> > On 29 September 2011 17:30, Ramana Radhakrishnan
>> > wrote:
>> >> On 19 September 2011 08:54, Ira Rosen wrote:
>> >>
>> >>>
>> >>> Boots
http://gcc.gnu.org/ml/gcc-patches/2011-09/msg01293.html
Thanks!
Dehao
Ping…
On Oct 13, 2011, at 5:11 PM, Tristan Gingold wrote:
> Hi,
>
> looks like an obvious typo. Ok for trunk ?
>
> Tristan.
>
> 2011-10-13 Tristan Gingold
>
> * config/rs6000/rs6000.c (rs6000_init_builtins): Fix typo.
>
> diff --git a/gcc/config/rs6000/rs6000.c b/gcc/config/rs6000/
>> here is another patch for PR47023, which takes care of comment #1,
>> i.e. rejecting polymorphic variables in a C-binding context.
>>
>> Regtested on x86_64-unknown-linux-gnu. Ok for trunk?
>
> OK. Thanks for the patch.
Thanks. Committed as r180130.
> If I saw it correctly, one still needs to
This fixes an oversight of my patch folding the added stmt in PRE -
failing to update it.
Bootstrapped on x86_64-unknown-linux-gnu, testing in progress.
Richard.
2011-10-18 Richard Guenther
PR tree-optimization/50767
* tree-ssa-pre.c (create_expression_by_pieces): Update th
Hello,
this patch adjusts __builtin_expect tests in tree-ssa so, that short-circuit
branch optimization don't lead to fallout. I've used here a multiplication, as
simple substraction/addition might get optimized away.
ChangeLog
2011-10-18 Kai Tietz
* gcc.dg/tree-ssa/builtin-expect
> The FSF has had my papers since Feb '11, but are stalling trying to decide
> whether to accept my employer's slightly non-standard disclaimer or not.
Ah. That seems a long time even for this kind of things.
Anyhow, this particular patch consisted of exactly 4 casts, so it seems to fall
in the
- Original Message -
From: "Paolo Carlini"
To: "Andy Gibbs"
Cc: ;
Sent: Tuesday, October 18, 2011 12:00 PM
Subject: Re: [C++ Patch] __builtin_choose_expr
> I'm under the impression that some tests are written in C, wouldn't better
> fit in c-c++-common?
Ok, I can do this but I think
This patch do some tweaks to addhi3 like adding QI scratch register.
The original *addhi3 insn is still there and located prior to new
addhi3_clobber insn because addhi3 is special to reload (thanks Danis for this
note) so that there is a version with and a version without scratch register.
Patch
The audit trail in the PR contains the detective work (mostly by Eric)
that concludes we have a long-standing bug in the Darwin x86-64
sigtramp unwind data.
This has been filed as radar #10302855, but we need a work-around
until that is resolved (possibly forever on older systems).
OK for t
In order to align with a standard naming convention, this patch renames
all the __sync_mem references to __atomic.
The only real addition is in builtins.c::is_builtin_name needed to know
that __atomic_ is also a builtin prefix.
Are you sure about the renaming get_builtin_sync_mem -> get_builtin
On 10/17/11 14:54, Richard Earnshaw wrote:
> On 14/10/11 14:31, Bernd Schmidt wrote:
>> On 07/13/11 16:03, Richard Earnshaw wrote:
* config/arm/arm.c (store_multiple_sequence): Avoid cases where
the base reg is stored iff compiling for Thumb1.
* gcc.target/arm/pr49641.c
On Tue, Sep 27, 2011 at 03:26:03PM +0100, Richard Sandiford wrote:
> This patch fixes a miscompilation of stage1 c-parser.o in an ARM bootstrap.
> When an access to an enum field was SRAed, a component ref used the type
> of the integer temporary variable instead of the type of the enum.
> It there
> The audit trail in the PR contains the detective work (mostly by Eric) that
> concludes we have a long-standing bug in the Darwin x86-64 sigtramp unwind
> data.
> This has been filed as radar #10302855, but we need a work-around until
> that is resolved (possibly forever on older systems).
>
> O
On 30 Sep 2011, at 00:33, Iain Sandoe wrote:
I'll re-jig with the typographical changes (sorry that were so
many ... )
I've addressed your points in :
http://gcc.gnu.org/ml/gcc-patches/2011-09/msg01974.html
specifically (other than typographical issues_
o we retain the ability to read an
2011/10/18 Georg-Johann Lay :
> This patch do some tweaks to addhi3 like adding QI scratch register.
>
> The original *addhi3 insn is still there and located prior to new
> addhi3_clobber insn because addhi3 is special to reload (thanks Danis for this
> note) so that there is a version with and a v
On 18 Oct 2011, at 13:22, Arnaud Charlet wrote:
The audit trail in the PR contains the detective work (mostly by
Eric) that
concludes we have a long-standing bug in the Darwin x86-64 sigtramp
unwind
data.
This has been filed as radar #10302855, but we need a work-around
until
that is res
On 18/10/11 13:19, Bernd Schmidt wrote:
> On 10/17/11 14:54, Richard Earnshaw wrote:
>> On 14/10/11 14:31, Bernd Schmidt wrote:
>>> On 07/13/11 16:03, Richard Earnshaw wrote:
> * config/arm/arm.c (store_multiple_sequence): Avoid cases where
> the base reg is stored iff compiling for Thu
> It's broken in all Libc versions that are in the wild (AFAICT from looking
> at the released sources).
>
> We will need to deal with
> configury/__ENVIRONMENT_MAC_OS_X_VERSION_MIN_REQUIRED__ stuff once
> there is a fixed Libc.
OK, would be good to follow up with such patch when/if this is fixe
Denis Chertykov schrieb:
> 2011/10/18 Georg-Johann Lay :
>> This patch do some tweaks to addhi3 like adding QI scratch register.
>>
>> The original *addhi3 insn is still there and located prior to new
>> addhi3_clobber insn because addhi3 is special to reload (thanks Danis for
>> this
>> note) so
2011/10/18 Georg-Johann Lay :
> Denis Chertykov schrieb:
>> 2011/10/18 Georg-Johann Lay :
>>> This patch do some tweaks to addhi3 like adding QI scratch register.
>>>
>>> The original *addhi3 insn is still there and located prior to new
>>> addhi3_clobber insn because addhi3 is special to reload (t
On 10/18/11 14:30, Richard Earnshaw wrote:
> Well, if that's the case why do we need to test for Thumb1 at all? And
> why do we only enable write-back for Thumb1? other ISA variants can
> also do that (I know that Thumb1 requires write-back, but it's
> optionally available for the other ISA flavo
This patch has been approved by Janne off list - and has been committed
to the 4.6 branch only (Rev. 180138) after bootstrapping and regtesting it.
It is essentially my patch from
http://gcc.gnu.org/ml/fortran/2011-10/msg00120.html
minus the .texi change. And the inquire.c part of Janne's patc
On 18/10/11 13:47, Bernd Schmidt wrote:
> On 10/18/11 14:30, Richard Earnshaw wrote:
>> Well, if that's the case why do we need to test for Thumb1 at all? And
>> why do we only enable write-back for Thumb1? other ISA variants can
>> also do that (I know that Thumb1 requires write-back, but it's
>
- Original Message -
From: "Kai Tietz"
To: gcc-patches@gcc.gnu.org
Cc: "Richard Guenther"
Sent: Tuesday, October 18, 2011 1:33:17 PM
Subject: [patch testsuite]: Adjust tree-ssa/builtin-expect-*.c tests for high
cost-branching optimization
Hello,
this patch adjusts __builtin_expect test
On 17/10/11 14:24, Richard Earnshaw wrote:
There's a presumption in host_detect_local_cpu() that "CPU implementer"
will appear before "CPU part" in the output of /proc/cpuinfo. That's
probably a pretty safe assumption (and it appears that it will handle
that case relatively safely -- ie not cra
Tristan,
Mike fixed the typo already. Why are you still seeing a problem?
- David
On Tue, Oct 18, 2011 at 6:52 AM, Tristan Gingold wrote:
> Ping…
>
> On Oct 13, 2011, at 5:11 PM, Tristan Gingold wrote:
>
>> Hi,
>>
>> looks like an obvious typo. Ok for trunk ?
>>
>> Tristan.
>>
>> 2011-10-13
Greetings,
Here is a new revision of the tree portions of this patch. I moved the
pattern recognizer to expand, and added additional logic to look for the
same pattern in gimple form. I added two more tests to verify the new
logic.
I didn't run into any problems with the RTL CSE phases. I can'
On Oct 18, 2011, at 4:23 PM, David Edelsohn wrote:
> Tristan,
>
> Mike fixed the typo already. Why are you still seeing a problem?
Indeed, was fixed later.
Tristan.
>
> - David
>
> On Tue, Oct 18, 2011 at 6:52 AM, Tristan Gingold wrote:
>> Ping…
>>
>> On Oct 13, 2011, at 5:11 PM, Tristan
On 18/10/11 15:23, Andrew Stubbs wrote:
> + /* Detect arch/cpu. */
> + if (strncmp (buf, "CPU part", sizeof ("CPU part") - 1) == 0)
> + {
> + int i;
> +
> + if (cpu_table == NULL)
> + goto not_found;
> +
Which still jumps to not_found without closing f.
R.
http://gcc.gnu.org/ml/gcc-patches/2011-10/msg00291.html
configure: add target-libgloss to skipdirs when using --without-newlib.
Testing: Bootstrapped/tested on x86_64-linux and sparc-sun-solaris2.10
with a gcc+newlib+libgloss combined tree.
Thanks,
Jeremy
Hi,
in a few places in libgfortran we have some code for handling special
and/or non-seekable files differently. The problem is that special
files don't all have some nice consistent behavior. E.g. wrt. seeking,
some allow seeking just fine, others allow some seeks and not others,
others allow the
On 18/10/11 15:34, Richard Earnshaw wrote:
On 18/10/11 15:23, Andrew Stubbs wrote:
+ /* Detect arch/cpu. */
+ if (strncmp (buf, "CPU part", sizeof ("CPU part") - 1) == 0)
+ {
+ int i;
+
+ if (cpu_table == NULL)
+ goto not_found;
+
Which still jumps to
Hi,
the attached patch introduces a new "member function" ssize in order
to determine the size of a unit. The benefit is that with unbuffered
IO, when getting the size of a file we can replace 3 lseek() calls
with one fstat() call, and also that for unseekable file we get a size
of 0 rather than -
Hey, Dodji,
Your patch broke bootstrap on AIX because of the typedef "loc_t"
introduced in tree-diagnostics.c. The typedef conflicts with a
typedef in an AIX 5.3 header file for locales. AIX should not be
using that namespace, but the failure occurs before fix-includes, so
it is not possible to
Hi,
libgfortran maintains a position flag which is used by the
INQUIRE(POSITION=...) statement. Currently we update this flag after
every IO statement. For unbuffered IO this is somewhat tedious, as
figuring out whether we're at the beginning of a file or the end
requires at least two syscalls. Th
On Tue, 18 Oct 2011, David Edelsohn wrote:
> Hey, Dodji,
>
> Your patch broke bootstrap on AIX because of the typedef "loc_t"
> introduced in tree-diagnostics.c. The typedef conflicts with a
> typedef in an AIX 5.3 header file for locales. AIX should not be
> using that namespace, but the failu
On Tue, Oct 18, 2011 at 11:09:04AM -0400, David Edelsohn wrote:
> Your patch broke bootstrap on AIX because of the typedef "loc_t"
> introduced in tree-diagnostics.c. The typedef conflicts with a
> typedef in an AIX 5.3 header file for locales. AIX should not be
> using that namespace, but the fa
On Tue, Oct 18, 2011 at 10:19 AM, Joseph S. Myers
wrote:
> On Tue, 18 Oct 2011, David Edelsohn wrote:
>
>> Hey, Dodji,
>>
>> Your patch broke bootstrap on AIX because of the typedef "loc_t"
>> introduced in tree-diagnostics.c. The typedef conflicts with a
>> typedef in an AIX 5.3 header file for
Hi,
The attached patch (by Maxim Kuvyrkov) enables building GCC for Linux
when the --with-arch=cf configuration option is in effect (i.e. when
building a compiler targeting ColdFire rather than legacy m68k).
Without the patch the build fails with an error such as:
.../gcc/config/m68k/t-mlibs:45:
On Sunday 09 October 2011 18:25:25 Tobias Burnus wrote:
> On 07.10.2011 16:38, Mikael Morin wrote:
> > The full patchset has passed the fortran testsuite successfully.
> > OK for trunk?
>
> OK for the whole patch set. Thanks for finding and fixing the issue!
>
Committed as follows.
I committed 8
Hi!
Now that there is a better testsuite for constant reshuffling, this patch
fixes various issues I found plus improves various permutations.
Bootstrapped/regtested on x86_64-linux and i686-linux, additionally
tested with
GCC_TEST_RUN_EXPENSIVE=1 make check-gcc
RUNTESTFLAGS='--target_board=unix\
On Mon, Oct 17, 2011 at 08:49:34AM +0200, Richard Guenther wrote:
> On Sun, Oct 16, 2011 at 5:47 PM, Jakub Jelinek wrote:
> I think this should be exactly the other way around, using create_tmp_var,
> copying TREE_ADDRESSABLE and setting DECL_GIMPLE_REG_P if it is not
> addressable.
>
> Ok with
Hi!
Apparently a recent change (2011-10-1[23] or so) regressed
gcc.dg/guality/asm-1.c test on x86_64-linux on the 4.6 branch.
This turned out to be a bug in loc_descriptor which has been passing wrong
mode on the recursive call, so when we had var_location note containing
SImode subreg of some DIm
On Mon, Oct 17, 2011 at 7:49 AM, Kirill Yukhin wrote:
> Thanks, guys, could anybody please commit that?
>
I checked it in for you.
--
H.J.
Hello,
current mainline miscompiles -fPIC code on SPU, which leads to wrong code
in libgcc, causing a large number of test failures.
The problem is in the back-end's get_pic_reg routine which implements an
optimization: in a leaf function, if register 74 is unused so far, use that
register instea
Thank you!
K
On Tue, Oct 18, 2011 at 7:42 PM, H.J. Lu wrote:
> On Mon, Oct 17, 2011 at 7:49 AM, Kirill Yukhin
> wrote:
>> Thanks, guys, could anybody please commit that?
>>
>
> I checked it in for you.
>
>
> --
> H.J.
>
David Edelsohn writes:
> Your patch broke bootstrap on AIX because of the typedef "loc_t"
> introduced in tree-diagnostics.c. The typedef conflicts with a
> typedef in an AIX 5.3 header file for locales. AIX should not be
> using that namespace, but the failure occurs before fix-includes, so
>
Hi,
This patch fixes computed gotos on m68k, and probably other targets too
(the fix is in the middle end). Several tests fail at present with "-O3
-fomit-frame-pointer".
One example of erroneous behaviour is as follows: the
function 'x' in comp-goto-1.c compiles to:
x:
lea (-104,%sp),%s
When libiberty defines psignal, it doesn't use the canonical signature.
This came up as a problem in a configuration where libiberty wants to
define psignal itself, but the build environment's declares
it too.
This was a --with-newlib configuration using the newlib trunk, where newlib
does defi
On Tue, Oct 18, 2011 at 9:49 AM, Roland McGrath wrote:
> When libiberty defines psignal, it doesn't use the canonical signature.
> This came up as a problem in a configuration where libiberty wants to
> define psignal itself, but the build environment's declares
> it too.
>
> This was a --with-ne
On Tue, Oct 18, 2011 at 9:50 AM, Andrew Pinski wrote:
> libiberty is no longer compiled for the target so this should never
> happen really.
I see. I was indeed using an older source base, and just noticed that all
the offending configure logic was still the same. Perhaps all the
--with-newlib
Denis Chertykov schrieb:
> 2011/10/18 Georg-Johann Lay :
>> Denis Chertykov schrieb:
>>> 2011/10/18 Georg-Johann Lay :
This patch do some tweaks to addhi3 like adding QI scratch register.
The original *addhi3 insn is still there and located prior to new
addhi3_clobber insn becau
On Tue, Oct 18, 2011 at 9:49 AM, Roland McGrath wrote:
>
> libiberty/
> 2011-10-18 Roland McGrath
>
> * strsignal.c (psignal): Use const second in parameter type.
> * functions.texi: Updated.
This is OK.
Thanks.
Ian
Julian Brown writes:
> Hi,
>
> This patch fixes computed gotos on m68k, and probably other targets too
> (the fix is in the middle end). Several tests fail at present with "-O3
> -fomit-frame-pointer".
This is the same bug as PR47918. As described in that PR trail, the bug
can be triggered
On Tue, Oct 18, 2011 at 10:18 AM, Ian Lance Taylor wrote:
> On Tue, Oct 18, 2011 at 9:49 AM, Roland McGrath wrote:
>>
>> libiberty/
>> 2011-10-18 Roland McGrath
>>
>> * strsignal.c (psignal): Use const second in parameter type.
>> * functions.texi: Updated.
>
> This is OK.
Thank
On 10/18/2011 08:30 AM, Jakub Jelinek wrote:
> * config/i386/i386.c (ix86_expand_vec_perm): In merge_two use
> mode SUBREG of operands[0] as target.
> (valid_perm_using_mode_p): Don't ignore higher bits of d->perm.
> (expand_vec_pshufb): For V8SImode vmode emit avx2_permvarv
A using-declaration adds a TREE_LIST to the list of names on a binding
level, so we need to cope with that here.
Tested x86_64-pc-linux-gnu, applying to trunk.
commit 052e893fe307f33ca3d3c2ead06248e0ef738f16
Author: Jason Merrill
Date: Mon Oct 17 22:04:08 2011 -0400
PR c++/50742
*
Dodji Seketeli writes:
> David Edelsohn writes:
>
>> Your patch broke bootstrap on AIX because of the typedef "loc_t"
>> introduced in tree-diagnostics.c.
I have committed the patch below. Tested on x86_64-unknown-linux-gnu
only, unfortunately. I hope it fixes the build on AIX.
From: Dodji S
2011/10/18 Georg-Johann Lay :
> Denis Chertykov schrieb:
>> 2011/10/18 Georg-Johann Lay :
>>> Denis Chertykov schrieb:
2011/10/18 Georg-Johann Lay :
> This patch do some tweaks to addhi3 like adding QI scratch register.
>
> The original *addhi3 insn is still there and located prior
Just a note.
Instead of `!reload_completed && !reload_in_progress'you can use
`can_create_pseudo_p()' declared in rtl.h
Denis.
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
On 10/18/11 01:44, Paolo Bonzini wrote:
> On 10/18/2011 07:10 AM, Jeff Law wrote:
>> --- 467,475 if (gimple_code (use) != GIMPLE_PHI) { sinkbb =
>> gimple_bb (use); ! sinkbb = select_best_block (frombb,
>> gimple_bb (use), stmt);
>>
>> !
On Oct 18, 2011, at 4:58 AM, Iain Sandoe wrote:
> The audit trail in the PR contains the detective work (mostly by Eric) that
> concludes we have a long-standing bug in the Darwin x86-64 sigtramp unwind
> data.
Yeah, well, it is a bug, if they ever fix it. If they burn it in and define it
as c
Denis Chertykov schrieb:
> 2011/10/18 Georg-Johann Lay :
>> Denis Chertykov schrieb:
>>> 2011/10/18 Georg-Johann Lay :
Denis Chertykov schrieb:
> 2011/10/18 Georg-Johann Lay :
>> This patch do some tweaks to addhi3 like adding QI scratch register.
>>
>> The original *addhi3 ins
On Oct 18, 2011, at 5:23 AM, Iain Sandoe wrote:
> o we retain the ability to read an old-style file.
> - I wonder if there should be an argument for removing this from the trunk
> version, but retaining in 4.6.
> - or, at the least, can we remove it from 4.8?
I'd be fine with simply requirin
Here we were failing to recognize that a defaulted fn had already been
instantiated, and crashed trying to instantiate it again.
Tested x86_64-pc-linux-gnu, applying to trunk and 4.6.
commit 10ff48cb9f7817ae8e95e71bd4b22989dafb2be2
Author: Jason Merrill
Date: Tue Oct 18 14:25:42 2011 -0400
http://codereview.appspot.com/5272048/diff/18001/tree-asan.c
File tree-asan.c (right):
http://codereview.appspot.com/5272048/diff/18001/tree-asan.c#newcode325
tree-asan.c:325: base = build_addr (t, current_function_decl);
You need to create a temp var and build as gimple assignment. See
init_tmp
Several ICEs in the TM tests on C++, but I think they are old. C TM
tests work except some missing optimizations (old failures too).
c-c++-common/tm/wrap-[12].c fail on C++, but not on C:
wrap-1.c:5:57: error: 'transaction_wrap' argument not an identifier
I believe this is an old error too?
(I d
On 17/10/11 11:43, Richard Guenther wrote:
Ok.
Committed, thanks.
Andrew
Georg-Johann Lay schrieb:
Denis Chertykov schrieb:
What difference in size of avr-libc ?
I have no tool for smart size analysis, so here is just a diff:
After rebuilding avr-libc with respective compiler version, did respectively:
$ find . -name 'lib[mc].a' -exec avr-size {} ';' > size-orig
Here's the last of the missing intrinsics. compare_exchange is
slightly different than the others since it doesn't map directly to an
rtl pattern. Its format is :
bool __atomic_compare_exchange (T* mem, T* expected, T desired, bool
weak, memory_order success, memory_order failure)
note
On Tue, Oct 18, 2011 at 06:03:16PM -0400, Andrew MacLeod wrote:
>
> Here's the last of the missing intrinsics. compare_exchange is
> slightly different than the others since it doesn't map directly to
> an rtl pattern. Its format is :
>
> bool __atomic_compare_exchange (T* mem, T* expected, T
On 10/18/2011 06:07 PM, Jakub Jelinek wrote:
On Tue, Oct 18, 2011 at 06:03:16PM -0400, Andrew MacLeod wrote:
Here's the last of the missing intrinsics. compare_exchange is
slightly different than the others since it doesn't map directly to
an rtl pattern. Its format is :
bool __atomic_comp
Suppress verbose notes/warnings printed in FDO-use compilation.
(1) Add option -fprofile-use-verbose. When this option is on, FDO-use
compilation prints out all the notes as that of today. When this
option is off (the default), all notes are suppressed.
(2) Make several unconditional warnings unde
http://codereview.appspot.com/5272048/diff/18001/tree-asan.c
File tree-asan.c (right):
http://codereview.appspot.com/5272048/diff/18001/tree-asan.c#newcode325
tree-asan.c:325: base = build_addr (t, current_function_decl);
There are issues with creating address expressions from TARGET_MEM_REF
in
On Tue, Oct 18, 2011 at 3:56 PM, wrote:
> On 2011/10/18 22:52:33, davidxl wrote:
>>
>> http://codereview.appspot.com/5272048/diff/18001/tree-asan.c
>> File tree-asan.c (right):
>
>
> http://codereview.appspot.com/5272048/diff/18001/tree-asan.c#newcode325
>>
>> tree-asan.c:325: base = build_addr (
On Wed, Oct 5, 2011 at 9:58 AM, Mike Stump wrote:
>
> On Oct 5, 2011, at 12:47 AM, Sharad Singhai wrote:
> > This patch adds an intermediate coverage format (enabled via 'gcov
> > -i'). This is a compact format as it does not require source files.
>
> I don't like any of the tags, I think if you s
It will be weird to put the instrumentation pass inside loop opt,
besides memory loads which are loop invariants and redundant stores in
loop should be handled by pre/pde.
+cc Richard Guenther
You may want to ask the middle-end maintainer to review your code at
this point if you want it to be in
On Tue, Oct 18, 2011 at 19:31, Xinliang David Li wrote:
> It will be weird to put the instrumentation pass inside loop opt,
> besides memory loads which are loop invariants and redundant stores in
> loop should be handled by pre/pde.
>
> +cc Richard Guenther
>
> You may want to ask the middle-end
On 29/09/2011, at 10:23 PM, Richard Guenther wrote:
> On Thu, Sep 29, 2011 at 10:27 AM, Tom de Vries wrote:
> + && SSA_NAME_IS_DEFAULT_DEF (expr)
> + && TREE_CODE (var) == PARM_DECL
> + && POINTER_TYPE_P (TREE_TYPE (var))
> + && TYPE_USER_ALIGN
On Tue, Oct 18, 2011 at 3:48 PM, Rong Xu wrote:
> Suppress verbose notes/warnings printed in FDO-use compilation.
> (1) Add option -fprofile-use-verbose.
Gcc currently does not emit informational messages on high level
transformations such as inlining, value profiling transformations (ic
promotio
1 - 100 of 105 matches
Mail list logo
