2013/1/24 Georg-Johann Lay :
> This patch adds the absolute value functions to stdfix.h.
>
> DEF_BUILTIN gets one more argument LIBNAME that sets the library name. If no
> folding is found, for some builtins it's more convenient to call libgcc
> support
> directly instead of expanding to an insn.
Hi,
Please find attached the patch that fixes following warning in aarch64:-
warning: "TARGET_FIXED_CONDITION_CODE_REGS" redefined
Thanks,
Naveen.H.S
2013-01-25 Naveen H.S
* config/aarch64/aarch64.c (TARGET_FIXED_CONDITION_CODE_REGS):
undef to avoid warning
--- gcc/config/a
I've updated my email address. KugelWorks is open for business.
--
Maxim Kuvyrkov
KugelWorks
0001-MAINTAINERS-Update-my-email.patch
Description: Binary data
From: Andi Kleen
The libitm TSX hardware transaction fast path currently does quite a bit of
unnecessary work (saving registers etc.) before even trying to start
a hardware transaction. This patch moves the initial attempt at a
transaction early into the assembler stub. Complicated work like retr
Hi Michael,
Please find attached updated patch for MicroBlaze Little Endian
support to reflect removal of microblaze*-*-* target.
Changelog updated as follows.
thanks,
David
gcc/Changelog
2013-01-25 Edgar E. Iglesias
* gcc/config.gcc (microblaze*-linux*): Add TARGET_BIG_ENDIAN_DEFAULT
gc
On 25 January 2013 01:38, Michael Eager wrote:
> On 01/23/2013 08:24 PM, David Holsgrove wrote:
>>
>> Hi Michael,
>>
>> On 24 January 2013 00:58, Michael Eager wrote:
>>>
>>> On 01/20/2013 09:42 PM, David Holsgrove wrote:
gcc/Changelog
2013-01-21 Edgar E. Iglesias
On 24 January 2013 19:03, Marc Glisse wrote:
> On Wed, 23 Jan 2013, Paul Pluzhnikov wrote:
>
>> This patch allows us to catch use of destructed strings.
>
>
> Hello,
>
> while a number of the google debug patches are just cheaper versions (that
> don't break the ABI) of what libstdc++'s debug mode
> Ok, inside the patch file now rather than at the end of the mail text.
> I'm still unsure if the patch should contain a real ChangeLog diff or
> a simple comment, as I doubt the ChangeLog diff would apply smoothly.
The general approach is insert the ChangeLog, as text, in the body of
the email m
gfortran.dg/bind_c_array_params_2.f90 fails for some mips*-linux-gnu
combinations because we output the linker hint:
.reloc 1f,R_MIPS_JALR,myBindC
as well as the call itself.
Fixed by passing the option that disables this feature. Tested on
mips64el-linux-gnu and applied.
Richard
gcc/t
octeon-pipe-1.c is a sched2 test, so make sure we generate some code.
Tested on mips64el-linux-gnu and applied. This fixes the UNRESOLVEDs
in the test_summary output.
Richard
gcc/testsuite/
* gcc.target/mips/octeon-pipe-1.c: Add -ffat-lto-objects
Index: gcc/testsuite/gcc.target/mips/o
On Thu, 24 Jan 2013, Iyer, Balaji V wrote:
> I have other patches that are also part of Cilk Plus. Can I send
> them out now also?
You can send them, but I don't propose to review them outside stage 1
(clearly such major new features are only suitable for mainline commit in
stage 1), and
Hello Joseph,
I have other patches that are also part of Cilk Plus. Can I send them
out now also?
Thanks,
Balaji V. Iyer.
> -Original Message-
> From: Joseph Myers [mailto:jos...@codesourcery.com]
> Sent: Thursday, January 24, 2013 3:22 PM
> To: Iyer, Balaji V
> Cc: 'gcc-patches
I intend to look at this once we're back in development stage 1 (for GCC
4.9).
--
Joseph S. Myers
jos...@codesourcery.com
This patch from minux changes libgo to use __USER_LABEL_PREFIX__ for all
asm symbols. This is a necessary step for Darwin support. Bootstrapped
and ran Go testsuite on x86_64-unknown-linux-gnu. Committed to
mainline.
https://codereview.appspot.com/7086062/
Ian
diff -r 246ddf77db61 libgo/go/by
Hurugalawadi, Naveen wrote:
> Hi,
>
> Please consider this as a reminder to review the patch posted at
> following link:-
> http://gcc.gnu.org/ml/gcc-patches/2012-12/msg00765.html
What is this good for?
Why are you changing the semantics of an existing, global constraint?
On Wed, 23 Jan 2013, Paul Pluzhnikov wrote:
This patch allows us to catch use of destructed strings.
Hello,
while a number of the google debug patches are just cheaper versions (that
don't break the ABI) of what libstdc++'s debug mode provides, this
overwriting of classes on destruction is
On Thu, 2013-01-24 at 21:44 +0400, Dmitry Vyukov wrote:
> On Thu, Jan 24, 2013 at 9:29 PM, Torvald Riegel wrote:
> > On Mon, 2012-12-17 at 13:52 +0400, Dmitry Vyukov wrote:
> >> > I think the simplest way to solve it for now, it to use... well, single
> >> > global lock.
> >> > I.e. replace __txn
FYI.
$ svn merge -c 194864
svn+ssh://sim...@gcc.gnu.org/svn/gcc/branches/gcc-4_7-branch
...test, passes dejagnu
$ svn commit gcc/testsuite/g++.dg/cpp0x/constexpr-ctor11.C
gcc/cp/ChangeLog gcc/cp/semantics.c
Sendinggcc/cp/ChangeLog
Sendinggcc/cp/semantics.c
Adding gcc/testsu
I've committed this patch from minux to libgo to fix some typos in code
that is only used on non-GNU/Linux systems:
https://codereview.appspot.com/7158043
Bootstrapped on x86_64-unknown-gnu-linux, although that proves little.
Committed to mainline.
Ian
diff -r cbf555d07453 libgo/Makefile.am
---
Steven,
On 19/07/12 16:43, Steven Bosscher wrote:
> On Thu, Jul 19, 2012 at 3:43 PM, Tom de Vries wrote:
>>> I think you should compare your method to the one described in the
>>> paper, and at least reference the paper if it's somehow similar --
>>
>> Interesting, thanks. Will do.
>
Done.
> B
On 01/24/2013 08:45 AM, Andi Kleen wrote:
>Please keep this as-is and rather use the same approach as for
>gtm_thread::begin_transaction.
What approach?
Using __asm__ to set the assembler name.
r~
Like LWU, MIPS16 LBU and LHU only allow offset(base) addresses,
so they should use "W" rather than "o" as the memory constraint.
This fixes g++.dg/torture/vshuf-v8hi.C on mips64el-linux-gnu,
where the constraints matched a PC-relative address instead.
Tested on mips64el-linux-gnu and applied.
Ric
On Thu, Jan 24, 2013 at 4:18 AM, Michael Haubenwallner
wrote:
> Ok, inside the patch file now rather than at the end of the mail text.
> I'm still unsure if the patch should contain a real ChangeLog diff or
> a simple comment, as I doubt the ChangeLog diff would apply smoothly.
>
>> Also, is this
On Thu, Jan 24, 2013 at 9:29 PM, Torvald Riegel wrote:
> On Mon, 2012-12-17 at 13:52 +0400, Dmitry Vyukov wrote:
>> resend in plain text
>>
>> On Mon, Dec 17, 2012 at 1:50 PM, Dmitry Vyukov wrote:
>> >
>> > On Fri, Dec 14, 2012 at 5:43 PM, Torvald Riegel wrote:
>> > > On Thu, 2012-12-13 at 10:02
On Mon, 2012-12-17 at 13:52 +0400, Dmitry Vyukov wrote:
> resend in plain text
>
> On Mon, Dec 17, 2012 at 1:50 PM, Dmitry Vyukov wrote:
> >
> > On Fri, Dec 14, 2012 at 5:43 PM, Torvald Riegel wrote:
> > > On Thu, 2012-12-13 at 10:02 +0100, Jakub Jelinek wrote:
> > >> On Thu, Dec 13, 2012 at 10:
On Thu, Jan 24, 2013 at 11:52 AM, Caroline Tice wrote:
> Hello,
>
> I just created a new branch, based on GCC 4.7, for migrating the
> vtable verification work to GCC. 4.7. I've announced the new branch
> on the gcc mailing list. This patch updates the web patch docs to
> mention the new branch.
Hello,
I just created a new branch, based on GCC 4.7, for migrating the
vtable verification work to GCC. 4.7. I've announced the new branch
on the gcc mailing list. This patch updates the web patch docs to
mention the new branch.
-- Caroline Tice
cmt...@google.com
wwwdocs.patch
Description: B
On Thu, 24 Jan 2013, Jakub Jelinek wrote:
> Fixed thusly, bootstrapped/regtested on x86_64-linux and i686-linux,
> ok for trunk? The 20030305-1.c testcase is now rejected with hard
> errors instead of resulting just in warnings and ignored initializers,
> but I believe the error is right, initial
> > // See gtm_thread::begin_transaction.
> > -uint32_t GTM::htm_fastpath = 0;
> > +extern "C"
> > +{
> > + uint32_t __gtm_htm_fastpath = 0;
>
> Please keep this as-is and rather use the same approach as for
> gtm_thread::begin_transaction.
What approach?
I don't want to hardcode C++ mangling
This patch adds the absolute value functions to stdfix.h.
DEF_BUILTIN gets one more argument LIBNAME that sets the library name. If no
folding is found, for some builtins it's more convenient to call libgcc support
directly instead of expanding to an insn.
gcc's folding is not very good, thus im
On 01/23/2013 08:24 PM, David Holsgrove wrote:
Hi Michael,
On 24 January 2013 00:58, Michael Eager wrote:
On 01/20/2013 09:42 PM, David Holsgrove wrote:
gcc/Changelog
2013-01-21 Edgar E. Iglesias
* config.gcc (microblaze*-*-elf): Add new target
libgcc/Changelog
2013-01-21 Ed
Hi,
the testcase reproduce ICE in a very corner case of inlining. What happens is
1) early inlining is tricked to produce very complex self recursive callgraph
graph
by somewhat bogus inlining of self recursive cycles
2) late inliner decide to continue in the recursive inlining
3) ipa-pure-con
Hi!
As discussed in the PR, while most of c-typeck.c handles
constructor_max_index == NULL the same as when
tree_int_cst_lt (constructor_max_index, some_index)
returns false, i.e. some_index is within bounds, there is one spot
that handles this incorrectly and one which doesn't consider NULL
const
Committed the following change:
http://gcc.gnu.org/r195424
* config/avr/avr.c (avr_out_fract): Make register numbers that
might be outside of source operand signed.
The problem occurs with fixed-point conversions to a wider mode and where the
input operand is in a small regno li
On Wed, Jan 23, 2013 at 8:08 PM, Uros Bizjak wrote:
> All things equal, we would like to avoid x87 registers to move DFmode
> immediates to a memory.
Attached patch is much better fix for the above problem. The patch
conditionally disables x87 register preferences when -mfpmath=sse is
in effect.
Hi,
tested x86_64-linux, committed mainline and 4_7-branch.
Thanks,
Paolo.
///
2013-01-24 Paolo Carlini
PR libstdc++/56085
* include/std/complex (pow(const complex<>&, int)): Avoid __n
signed overflow.
Index: include/std/complex
==
On Thu, Jan 24, 2013 at 1:23 AM, Paul Pluzhnikov wrote:
> This patch allows us to catch use of destructed strings.
>
> Google ref: b/5430313
>
> Ok for google/gcc-4_7 and google/integration?
OK.
Diego.
Hi,
PR 55927 is about failing scan dumps in g++.dg/ipa/devirt-10.C on
ia64. The problem turned out to be that early inliner was a bit more
active on ia64 and left nothing for IPA to figure out and so the
strings were not there in the dumps. Fixed below by running the test
with early inliner disa
On Wed, Jan 23, 2013 at 7:43 PM, Vladimir Makarov wrote:
> The error occurs because a pseudo in asm insn is not changed into hard
> register as the pseudo info is incorrect after info updating.
>
> You should just use lra_update_insn_regno_info. The patch (and the patch is
> ok to commit) should lo
On Sat, 2013-01-12 at 13:03 -0800, Andi Kleen wrote:
> From: Andi Kleen
>
> The libitm TSX hardware transaction fast path currently does quite a bit of
> unnecessary work (saving registers etc.) before even trying to start
> a hardware transaction. This patch moves the initial attempt at a
> tran
On 01/23/13 20:55, David Edelsohn wrote:
> This patch looks okay, although it needs a ChangeLog entry.
Ok, inside the patch file now rather than at the end of the mail text.
I'm still unsure if the patch should contain a real ChangeLog diff or
a simple comment, as I doubt the ChangeLog diff woul
Eric Botcazou writes:
> ERROR: gcc.dg/vect/vect-multitypes-12.c: error executing dg-final: bad index
> "18-1": must be integer or end?-integer?
Does that help? Perhaps the M-N feature isn't supported by your version
of tcl.
* lib/target-supports-dg.exp (dg-process-target): Use expr to
42 matches
Mail list logo
