This patch fixes some of the issues collected in the PR.
Some remarks to the changed in the attached patch:
- gcc_assert (*format++ == '$');
That code gets executed when
"some error message with value %d for string %s"
gets translated into
"some error message for string %2$s with va
Hello Michael,
On 12 Sep 2012, at 23:43, Michael Meissner wrote:
> It would be nice to know if this doesn't break the other ppc
> environments (AIX, Darwin) before I commit it. Are there any problems with
> this patch?
For powerpc-darwin9, there are a couple of issues which I've hacked aroun
Now with the patch properly attached.
Apologies.
Sofiane
> -Original Message-
> From: gcc-patches-ow...@gcc.gnu.org [mailto:gcc-patches-
> ow...@gcc.gnu.org] On Behalf Of Sofiane Naci
> Sent: 13 September 2012 13:18
> To: gcc-patches@gcc.gnu.org
> Subject: [AArch64, AArch64-4.7] Fix targe
On Sun, Sep 16, 2012 at 5:55 PM, Jakub Jelinek wrote:
> Hi!
>
> As discussed in the PR, this patch fixes the ICE by calling
> gimple_purge_dead_eh_edges if the last stmt has been replaced by the pass.
> Bootstrapped/regtested on x86_64-linux and i686-linux, ok for trunk?
>
> Still, I think it woul
On Sun, Sep 16, 2012 at 04:59:29PM -0400, Jack Howarth wrote:
> > All changes need explicit release manager approval until the final
> > release of GCC 4.7.2 which should happen roughly one week after
> > the release candidate if no issues show up with it.
>
>The backport of changes required t
On 09/05/2012 07:31 AM, Ian Lance Taylor wrote:
On Wed, Aug 29, 2012 at 10:32 AM, Florian Weimer wrote:
This patches fixes an integer overflow in libiberty, which leads to
crashes in binutils. The long version of the objalloc_alloc macro
would have needed another conditional, so I removed tha
On 08/21/2012 12:37 PM, Florian Weimer wrote:
I don't think there are any callers out there, but let's fix this for
completeness.
A compiler emitting code to call this function would still have to
perform overflow checks for the new T[n][m] case, so this interface is
not as helpful as it looks a
On 17/09/2012 09:47, Tobias Burnus wrote:
> This patch fixes some of the issues collected in the PR.
>
>
> Some remarks to the changed in the attached patch:
>
>
> - gcc_assert (*format++ == '$');
>
> That code gets executed when
> "some error message with value %d for string %s"
> ge
On 14/09/12 19:02, Ulrich Weigand wrote:
> Hello,
>
> this patch changes the ARM back-end to use vld1.64/vst1.64 instructions
> instead of vldm/vstm -where possible- to implement double-word moves.
>
> The main benefit of this is that it allows the compiler to provide
> appropriate alignment hi
On 14/09/12 21:34, Ben Cheng wrote:
> Hi,
>
> Recently we found out that the .interp section starts to show up in
> ARM executables compiled with "-shared -static" and the gold linker
> from binutils 2.22. We tracked down the origin of the dynamic linker
> commands and they are always explicitly s
Hi,
On 09/17/2012 11:51 AM, Florian Weimer wrote:
On 08/21/2012 12:37 PM, Florian Weimer wrote:
I don't think there are any callers out there, but let's fix this for
completeness.
A compiler emitting code to call this function would still have to
perform overflow checks for the new T[n][m] cas
On 14/09/12 19:03, Ulrich Weigand wrote:
> Hello,
>
> following up on the prior patch, this patch exploits more opportunities to
> generate the vld1 / vst1 family of instructions, this time to implement the
> vec_set and vec_extract patterns with memory scalar operands.
>
> Without the patch, vec
Dear Tobias,
> The following test case doesn't work; it should print "Overloaded" - and
> does so with crayftn. But with your patch, it doesn't.
For some reason, I guess, the attribute defined_assign_comp is not
getting passed along to type 'b'.
> + build_assignment (gfc_exec_op op, gfc_expr *ex
> Yes - now cfg_cleanup does that (and it really shouldn't be its job). That
> was the improvement I talked about - reducing the number of BBs a lot.
OK, I removed the code and compiled the testcase of PR 43186 with --param max-
completely-peel-loop-nest-depth=32 and got back the explosion. Comp
On 09/17/2012 12:15 PM, Paolo Carlini wrote:
Hi,
On 09/17/2012 11:51 AM, Florian Weimer wrote:
On 08/21/2012 12:37 PM, Florian Weimer wrote:
I don't think there are any callers out there, but let's fix this for
completeness.
A compiler emitting code to call this function would still have to
p
Hi all,
here is another proc-ptr patch, which allows to submit calls to
procedure-pointer components which have a procedure-pointer result.
This sounds a bit twisted (and I guess one has to suffer from some
very special form of mental illness in order to feel the urge to do
something like this), b
The subject says it all:
Add CONST_FIXED_P predicate macro similar to CONST_INT_P to rtl.h.
Currently, CONST_FIXED_P is not used in the compiler, except in avr.c that does
#ifndef CONST_FIXED_P
#define CONST_FIXED_P(X) (CONST_FIXED == GET_CODE (X))
#endif
Ok for trunk?
If this change is fine,
structure constructures with proc-pointer arguments were either leading
to wrong results or to ICEs.
I use now the same condition as in gfc_trans_pointer_assignment.
Build and regtested on x86-64-linux.
OK for the trunk?
Tobias
2012-09-17 Tobias Burnus
PR fortran/54603
* trans-expr.c (gfc
On 09/17/2012 01:39 PM, Janus Weil wrote:
here is another proc-ptr patch, which allows to submit calls to
procedure-pointer components which have a procedure-pointer result.
This sounds a bit twisted (and I guess one has to suffer from some
Regtested on x86_64-unknown-linux-gnu. Ok for trunk?
2012/9/15 Georg-Johann Lay :
> Canadian cross build fails because the wrong compiler is used to compile
> gen-avr-mmcu-texi. This small tool must run on build and not on host, thus
> the right compiler is CC_FOR_BUILD.
>
> Just changing the compiler is not enough because files like system.h were
>
Hi Tobias,
> structure constructures with proc-pointer arguments were either leading to
> wrong results or to ICEs.
> I use now the same condition as in gfc_trans_pointer_assignment.
>
> Build and regtested on x86-64-linux.
> OK for the trunk?
looks good to me. Thanks!
The "scan-tree-dump-times"
On Mon, Sep 17, 2012 at 12:52 PM, Eric Botcazou wrote:
>> Yes - now cfg_cleanup does that (and it really shouldn't be its job). That
>> was the improvement I talked about - reducing the number of BBs a lot.
>
> OK, I removed the code and compiled the testcase of PR 43186 with --param max-
> compl
On Mon, Sep 17, 2012 at 1:56 PM, Georg-Johann Lay wrote:
> The subject says it all:
>
> Add CONST_FIXED_P predicate macro similar to CONST_INT_P to rtl.h.
>
> Currently, CONST_FIXED_P is not used in the compiler, except in avr.c that
> does
>
> #ifndef CONST_FIXED_P
> #define CONST_FIXED_P(X) (CO
>> Regtested on x86_64-unknown-linux-gnu. Ok for trunk?
>
>
> Looks OK. Thanks for the patch.
Thanks a bunch. Committed as r191383.
Cheers,
Janus
>> 2012-09-17 Janus Weil
>>
>> PR fortran/54285
>> * expr.c (gfc_check_pointer_assign): Correctly handle procedure
>> pointers
>>
On 05/09/12 23:14, Christophe Lyon wrote:
> Hello,
>
> Although the recent optimization I have committed to use Neon vext
> instruction for suitable builtin_shuffle calls does not support
> big-endian yet, I have written a patch to the existing testcases such
> they now support big-endian mode.
>
Add __builtin_ppc_get_timebase and __builtin_ppc_mftb to read the Time
Base Register on PowerPC.
They are required by applications that measure time at high frequencies
with high precision that can't afford a syscall.
__builtin_ppc_get_timebase returns the 64 bits of the Time Base Register
while __
> Looks OK to me, though I have no authority to approve it
> except SH specific part.
Is there any more comments? Can it be committed in trank?
Regards,
Vladimir
2012/9/14 Kaz Kojima :
> Vladimir Yakovlev wrote:
>> I reproduced the failure and found reason of it. I understood haw it
>> resolve
This makes use of the new builtin functions for FILE, LINE and FUNCTION
to turn gimple_build_assign_with_ops/gimple_build_assign_with_ops3 into
two overloads of gimple_build_assign_with_ops (in theory the _stats
function can go and we could move the inlines to gimple.c instead,
eventually removing
On 2012-09-17 09:25 , Richard Guenther wrote:
This makes use of the new builtin functions for FILE, LINE and FUNCTION
to turn gimple_build_assign_with_ops/gimple_build_assign_with_ops3 into
two overloads of gimple_build_assign_with_ops (in theory the _stats
function can go and we could move the
On Mon, 17 Sep 2012, Diego Novillo wrote:
> On 2012-09-17 09:25 , Richard Guenther wrote:
> >
> > This makes use of the new builtin functions for FILE, LINE and FUNCTION
> > to turn gimple_build_assign_with_ops/gimple_build_assign_with_ops3 into
> > two overloads of gimple_build_assign_with_ops (
OK for 4.7 as well?
> -Original Message-
> From: Richard Earnshaw
> Sent: 14 September 2012 18:18
> To: Ian Bolton
> Cc: gcc-patches@gcc.gnu.org
> Subject: Re: [PATCH, AArch64] Implement fnma, fms and fnms standard
> patterns
>
> On 14/09/12 18:05, Ian Bolton wrote:
> > The following stan
OK for aarch64-4.7-branch as well?
> -Original Message-
> From: Richard Earnshaw
> Sent: 14 September 2012 18:31
> To: Ian Bolton
> Cc: gcc-patches@gcc.gnu.org
> Subject: Re: [PATCH, AArch64] Implement ffs standard pattern
>
> On 14/09/12 16:26, Ian Bolton wrote:
> > I've implemented the
On 17/09/12 15:12, Ian Bolton wrote:
> OK for 4.7 as well?
>
NO! But it is ok for the aarch64 branch of 4.7 :-)
R.
>> -Original Message-
>> From: Richard Earnshaw
>> Sent: 14 September 2012 18:18
>> To: Ian Bolton
>> Cc: gcc-patches@gcc.gnu.org
>> Subject: Re: [PATCH, AArch64] Implement
On 17/09/12 15:18, Ian Bolton wrote:
> OK for aarch64-4.7-branch as well?
>
yes.
R.
>> -Original Message-
>> From: Richard Earnshaw
>> Sent: 14 September 2012 18:31
>> To: Ian Bolton
>> Cc: gcc-patches@gcc.gnu.org
>> Subject: Re: [PATCH, AArch64] Implement ffs standard pattern
>>
>> On
On 09/15/2012 09:01 AM, Marc Glisse wrote:
+ /* For vector types, pick a non-opaque variant. */
+ if (TREE_CODE (type) == VECTOR_TYPE)
+ type = cp_build_qualified_type (TYPE_MAIN_VARIANT (type),
+ cp_type_quals (type));
I bel
On 17 September 2012 14:56, Richard Earnshaw wrote:
> On 05/09/12 23:14, Christophe Lyon wrote:
>> Hello,
>>
>> Although the recent optimization I have committed to use Neon vext
>> instruction for suitable builtin_shuffle calls does not support
>> big-endian yet, I have written a patch to the exi
On 17/09/12 16:13, Christophe Lyon wrote:
> On 17 September 2012 14:56, Richard Earnshaw wrote:
>> On 05/09/12 23:14, Christophe Lyon wrote:
>>> Hello,
>>>
>>> Although the recent optimization I have committed to use Neon vext
>>> instruction for suitable builtin_shuffle calls does not support
>>>
The attached patch is required so that plug-ins can include gimple.h. I
tested that cfg-flags.def is actually installed after this change.
Okay for trunk?
--
Florian Weimer / Red Hat Product Security Team
2012-09-17 Florian Weimer
* Makefile.in (BASIC_BLOCK_H): Add cfg-flags.def.
Index:
Hi,
I have just merged upstream trunk on the aarch64-branch up to r191370.
Thanks
Sofiane
http://gcc.gnu.org/viewcvs?view=revision&revision=191393
Committed as obvious.
CONST_FIXED_P is promoted to rtl.h now.
Johann
* config/avr/avr.c (CONST_FIXED_P): Remove define.
* config/avr/avr.md (add3): Use CONST_FIXED_P.
Index: config/avr/avr.md
==
On Sun, 2012-09-16 at 09:22 +0100, Richard Sandiford wrote:
> > +/* This is idential to sde.h except for the ABI setting which defaults
> > + to 32 instead of n32 on 32 bit architectures and the addition of
> > + MIPS_ISA_SYNCI_SPEC. */
>
> SDE defaults to o32 (-mabi=32) for 32-bit targets a
On 17 September 2012 17:21, Richard Earnshaw wrote:
> On 17/09/12 16:13, Christophe Lyon wrote:
>> On 17 September 2012 14:56, Richard Earnshaw wrote:
>>> On 05/09/12 23:14, Christophe Lyon wrote:
Hello,
Although the recent optimization I have committed to use Neon vext
instru
Currently the description of -feliminate-unused-debug-types says that it is off
by default. In fact, it is on by default. The attached patch corrects the
documentation to reflect that.
Ok to commit?
paul
2012-09-17 Paul Koning
* doc/invoke.text (-feliminate-unused-debug-t
On Mon, Sep 17, 2012 at 2:49 AM, Florian Weimer wrote:
> On 09/05/2012 07:31 AM, Ian Lance Taylor wrote:
>>
>> On Wed, Aug 29, 2012 at 10:32 AM, Florian Weimer wrote:
>>>
>>>
>>> This patches fixes an integer overflow in libiberty, which leads to
>>> crashes in binutils. The long version of the
When determining whether a conversion from an unscoped enum is a
narrowing conversion, we check whether the values of the enum fit into
the target type. [dcl.enum] defines the values of an enum with
unspecified underlying type to be the values of the smallest bit-field
that could hold it, not
http://gcc.gnu.org/viewcvs?view=revision&revision=191394
Committed as preapproved in http://gcc.gnu.org/PR54598#c1
Johann
PR lto/54598
* tree-streamer-in.c (unpack_ts_real_cst_value_fields): Use ggc'ed
FIXED_VALUE_TYPE instead of struct fixed_value.
Index: tree-streame
On Thu, Sep 13, 2012 at 1:00 PM, Diego Novillo wrote:
> On 2012-09-11 18:53 , Ian Lance Taylor wrote:
>
>> 2012-09-11 Ian Lance Taylor
>>
>> * Initial implementation.
>
> OK.
Thanks for all the reviews. I have committed the libbacktrace library
to trunk. I will follow up with a patch
On Sat, Sep 15, 2012 at 06:49:19PM +0200, Andreas Tobler wrote:
> Tested on powerpc64-freebsd10.0, results here:
>
> http://gcc.gnu.org/ml/gcc-testresults/2012-09/msg01403.html
>
> I ran the tests with and w/o patch. No diffs seen.
Great, thanks.
--
Michael Meissner, IBM
5 Technology Place Dri
Richard Earnshaw wrote:
> On 14/09/12 19:02, Ulrich Weigand wrote:
> > * config/arm/arm.c (output_move_neon): Update comment.
> > Use vld1.64/vst1.64 instead of vldm/vstm where possible.
> > (neon_vector_mem_operand): Support double-word modes.
> > * config/arm/neon.md (*neon_mov VD
On Mon, Sep 17, 2012 at 09:21:09AM +0100, Iain Sandoe wrote:
> Hello Michael,
>
> On 12 Sep 2012, at 23:43, Michael Meissner wrote:
>
> > It would be nice to know if this doesn't break the other ppc
> > environments (AIX, Darwin) before I commit it. Are there any problems with
> > this patch?
This patch to the diagnostic code uses the new backtrace library to
print a backtrace on an ICE. For example, here is the output of a test
case I took from a C++ PR:
/home/iant/foo2.cc:6:6: internal compiler error: in cp_lexer_new_from_tokens,
at cp/parser.c:638
0xec549f internal_error(char cons
Richard Earnshaw wrote:
> You confused me for a bit with the reference to "line numbers", but I
> think you must mean "lane numbers".
Indeed, sorry for the confusion.
> FTR, I see no reason why GCC would have problems with 64-bit vectors in
> big-endian mode, it's only 128-bit (and larger) vecto
On Mon, Sep 17, 2012 at 8:59 AM, wrote:
>
> 2012-09-17 Paul Koning
>
> * doc/invoke.text (-feliminate-unused-debug-types): Update to
> reflect that this is enabled by default.
This is OK.
Thanks.
Ian
On Mon, Sep 17, 2012 at 8:34 AM, Florian Weimer wrote:
> The attached patch is required so that plug-ins can include gimple.h. I
> tested that cfg-flags.def is actually installed after this change.
> -BASIC_BLOCK_H = basic-block.h $(PREDICT_H) $(VEC_H) $(FUNCTION_H) cfghooks.h
> +BASIC_BLOCK_H =
Eric Botcazou wrote:
> Reload 1 of insn #85 inherits the reload reg from reload 1 of insn #84, the
> bug being that the same reload reg is also used for reload 0 of insn #85.
>
> This is supposed to work like so: the inheritance code in choose_reload_regs
> calls free_for_value_p with IGNORE_ADDR
gfortran shouldn't simplify SCAN/VERIFY if the BACK= argument is present
but not a constant.
Build and regtested on x86-64-linux.
OK for the trunk?
Tobias
2012-09-17 Tobias Burnus
PR fortran/54608
* simplify.c (gfc_simplify_scan, gfc_simplify_verify):
Fix handling of BACK=variable.
2012
It didn't take long to find a build problem with libbacktrace. I forgot
to back and retest the unknown cases after changing the way backtrace
state was handled. This patch should fix bootstrap. Tested on
x86_64-unknown-linux-gnu, which proves little since the file is not used
on that platform.
On Mon, Sep 17, 2012 at 07:34:12PM +0200, Tobias Burnus wrote:
> gfortran shouldn't simplify SCAN/VERIFY if the BACK= argument is present
> but not a constant.
>
> Build and regtested on x86-64-linux.
> OK for the trunk?
>
OK.
--
Steve
On 17/09/12 16:50, Christophe Lyon wrote:
> On 17 September 2012 17:21, Richard Earnshaw wrote:
>> On 17/09/12 16:13, Christophe Lyon wrote:
>>> On 17 September 2012 14:56, Richard Earnshaw wrote:
On 05/09/12 23:14, Christophe Lyon wrote:
> Hello,
>
> Although the recent optimiza
On Fri, Sep 14, 2012 at 9:39 AM, Carrot Wei wrote:
>
> Hi Jakub
>
> I have run it on 4.6, it passes the following testing:
>
> x86-64 bootstrap
> x86-64 regression test
> regression test on arm qemu
>
> Is it OK for gcc4.6?
>
> Ahmad, is it OK for google/gcc-4_6/ and google/gcc-4_6-mobile ?
Yes,
On 09/17/2012 11:35 AM, Dodji Seketeli wrote:
>And I wonder if we want to offer this as an optional warning for GNU
attribute syntax.
What option would be used to control this optional feature? Would
you accept this a separate patch?
Let's not worry about this for now.
+ found_a
Hello,
On 10/09/2012 20:58, Paul Richard Thomas wrote:
> Dear All,
>
> Please find attached a new attempt at the patch for PR46897. It now
> uses temporaries to overcome the side effects that Mikael pointed out.
And here comes the next round of comments.
> Index: gcc/fortran/resolve.c
> =
On 17 September 2012 20:04, Richard Earnshaw wrote:
> On 17/09/12 16:50, Christophe Lyon wrote:
>> On 17 September 2012 17:21, Richard Earnshaw wrote:
>>> On 17/09/12 16:13, Christophe Lyon wrote:
On 17 September 2012 14:56, Richard Earnshaw wrote:
> On 05/09/12 23:14, Christophe Lyon w
On Sep 17, 2012, at 1:20 PM, Ian Lance Taylor wrote:
> On Mon, Sep 17, 2012 at 8:59 AM, wrote:
>>
>> 2012-09-17 Paul Koning
>>
>>* doc/invoke.text (-feliminate-unused-debug-types): Update to
>>reflect that this is enabled by default.
>
> This is OK.
>
> Thanks.
>
> Ian
On Mon, Sep 17, 2012 at 8:53 AM, Tulio Magno Quites Machado Filho
wrote:
> Add __builtin_ppc_get_timebase and __builtin_ppc_mftb to read the Time
> Base Register on PowerPC.
> They are required by applications that measure time at high frequencies
> with high precision that can't afford a syscall.
Steve Ellcey writes:
> On Sun, 2012-09-16 at 09:22 +0100, Richard Sandiford wrote:
>
>> > +/* This is idential to sde.h except for the ABI setting which defaults
>> > + to 32 instead of n32 on 32 bit architectures and the addition of
>> > + MIPS_ISA_SYNCI_SPEC. */
>>
>> SDE defaults to o32 (
On Mon, 17 Sep 2012, Ian Lance Taylor wrote:
> Please let me know about any build problems.
i386-unknown-freebsd10.0 now fails as follows. Others presumably
have not run into this since "This source file is compiled if the
unwind library is not available".
/scratch2/tmp/gerald/gcc-HEAD/libbacktr
On Mon, Sep 17, 2012 at 2:00 PM, Gerald Pfeifer wrote:
> On Mon, 17 Sep 2012, Ian Lance Taylor wrote:
>> Please let me know about any build problems.
>
> i386-unknown-freebsd10.0 now fails as follows. Others presumably
> have not run into this since "This source file is compiled if the
> unwind l
On Mon, Sep 17, 2012 at 2:00 PM, Gerald Pfeifer wrote:
> On Mon, 17 Sep 2012, Ian Lance Taylor wrote:
>> Please let me know about any build problems.
>
> i386-unknown-freebsd10.0 now fails as follows. Others presumably
> have not run into this since "This source file is compiled if the
> unwind l
If the only reference in a source file is to a static method of a class, then
GCC would output debug information for the class name but not any of its
members or base classes. The attached patch fixes this by having
"prune_unused_types_mark" mark all of the parent's children if the parent DIE
On Mon, 2012-09-17 at 21:36 +0100, Richard Sandiford wrote:
> It's a hosted vs. embedded thing. Hosted targets like *-linux-gnu have
> dynamic ABI requirements and so are keyed off an ABI rather than an
> architecture. The only effect of -march= and --with-arch= should be to
> extend the choice
The change to make SCOPE_REFs value-dependent revealed a bug in our
handling of alias-template instantiations; we weren't pushing into the
context of the alias before substituting the type, and then we were
remembering the previously instantiated one when instantiating in a
different context.
On Mon, Sep 17, 2012 at 10:39 PM, Jason Merrill wrote:
> The change to make SCOPE_REFs value-dependent revealed a bug in our handling
> of alias-template instantiations; we weren't pushing into the context of the
> alias before substituting the type, and then we were remembering the
> previously i
73 matches
Mail list logo