On Tue, Apr 24, 2012 at 08:05:35AM +0200, Richard Guenther wrote:
> On Tue, Apr 24, 2012 at 1:01 AM, Jakub Jelinek wrote:
> I don't like using TYPE_MAX_VALUE in VRP - at least please use vrp_val_max.
> For enums this can be not what you expect(?)
>
> Please consider adding a double_int_max_value
Hi,
>> Great. I approved it this morning and it looks like your account was
>> created soon thereafter.
Thanks for the approval. My account has been created.
>> add yourself to the MAINTAINTERS file with write-after-approval
>> privileges.
Done.
http://gcc.gnu.org/ml/gcc-cvs/2012-04/msg00705.h
This fixes PR53098.
Committed as obvious.
Richard.
2012-04-24 Richard Guenther
PR tree-optimization/53098
* tree-vect-loop.c (vect_analyze_loop_operations): Fixup
comparison sign.
Index: gcc/tree-vect-loop.c
==
On Tue, Apr 24, 2012 at 9:10 AM, Jakub Jelinek wrote:
> On Tue, Apr 24, 2012 at 08:05:35AM +0200, Richard Guenther wrote:
>> On Tue, Apr 24, 2012 at 1:01 AM, Jakub Jelinek wrote:
>> I don't like using TYPE_MAX_VALUE in VRP - at least please use vrp_val_max.
>> For enums this can be not what you e
I tested my two patches independently and forgot to test together and
one used a variable which I removed in the other.
Anyways this patch changes the code to be correct and fixes the bootstrap issue.
Committed as obvious after building stage1.
Thanks,
Andrew Pinski
ChangeLog:
2012-04-24 Andrew
On Apr 18, 2012, at 8:23 PM, Bruce Korb wrote:
> Hi,
>
> When I approved a patch in 2008, there was a typo. I didn't
> notice and it was "fixed" by removing a formatting element.
> Your patch corrects the error.
>
> Please apply your changes to active branches. Thank you!
> Regards, Bruce
Th
I have two patches who also record the maximum number of loop iterations
for the epilogue loop the vectorizer creates. The first one reverts
an old patch that made us re-use that epilogue loop also for the
unvectorized case (when versioning for alignment or aliasing). Thus
it creates one more lo
Hi,
On Mon, Apr 23, 2012 at 03:30:19PM +0200, Richard Guenther wrote:
> On Mon, 23 Apr 2012, Martin Jambor wrote:
>
> > Hi,
> >
> > On Mon, Apr 23, 2012 at 12:50:51PM +0200, Richard Guenther wrote:
> > > On Fri, 20 Apr 2012, Martin Jambor wrote:
> > >
> > > > Hi,
> > > >
> > > > two days ago I
Richard Guenther wrote:
> On Sat, Apr 21, 2012 at 5:53 PM, Jim Meyering wrote:
>> I see that no one responded when I posted this in February.
>> Ok to commit, now?
>
> Ok.
Thanks. Pushed.
This implements HARD_REGNO_CALL_PART_CLOBBERED.
Ok to apply?
Johann
PR target/53065
* config/avr/avr.h (HARD_REGNO_CALL_PART_CLOBBERED): New hook define.
Index: config/avr/avr.h
===
--- config/avr/avr.h (revision 186
Hello,
here is a port of Android support patch
(http://gcc.gnu.org/ml/gcc-patches/2012-04/msg00944.html) for 4.7
branch. Is it OK?
Bootstrap and check passed on linux-x86_64.
Thanks,
Ilya
---
2012-04-24 Enkovich Ilya
* config/i386/linux-common.h: New.
* config.gcc: Add i386/
I've been carrying this patch for quite some while now and really
want to go forward with it - the problem is that while all default
languages work fine after this patch Ada shows some testsuite
regressions. I've had various hacks/workarounds throughout the
Ada frontend for them, but lost track o
Hi guys,
(coming back to an old patch proposed by FX some time ago ...)
2012/3/3 FX :
>> I think that this approach is a mistake. The patch
>> starts us down a slippery slope. Why not export all
>> the nonstandard intrinsics? In additions, the
>> _gfortran_ prefix was used to separate the libg
On Mon, Apr 23, 2012 at 18:16, Ollie Wild wrote:
> If the trunk change isn't approved today, I'll probably go ahead and
> check in *that* patch to google/integration. The main difference
> relative to this is that the option name has been changed to
> -Wliteral-suffix. *That* part has been appr
2012/4/24 Georg-Johann Lay :
> This implements HARD_REGNO_CALL_PART_CLOBBERED.
>
> Ok to apply?
>
> Johann
>
> PR target/53065
> * config/avr/avr.h (HARD_REGNO_CALL_PART_CLOBBERED): New hook define.
Please commit.
Denis.
On Tue, Apr 24, 2012 at 2:54 PM, Ilya Enkovich wrote:
> Hello,
>
> here is a port of Android support patch
> (http://gcc.gnu.org/ml/gcc-patches/2012-04/msg00944.html) for 4.7
> branch. Is it OK?
This kind of re-org is not appropriate for the branch.
Thanks,
Richard.
> Bootstrap and check passed
On 24/04/12 14:03, Richard Guenther wrote:
> /* Size types *are* sign extended. */
> ! sign_extended_type = !TYPE_UNSIGNED (type);
looks like you failed to update the comment.
R.
On Tue, 24 Apr 2012, Richard Earnshaw wrote:
> On 24/04/12 14:03, Richard Guenther wrote:
> > /* Size types *are* sign extended. */
> > ! sign_extended_type = !TYPE_UNSIGNED (type);
>
> looks like you failed to update the comment.
Indeed. Fixed.
Thanks,
Richard.
On 23/04/12 22:36, Michael Hope wrote:
> On 24 April 2012 03:35, Richard Earnshaw wrote:
>> On 22/04/12 23:20, Michael Hope wrote:
>>> Change the dynamic linker path for ARM hard float executables.
>>> Matches the path discussed and agreed on last week[1]. Carlos will
>>> follow up with the match
> On Tue, Apr 24, 2012 at 2:54 PM, Ilya Enkovich wrote:
>> Hello,
>>
>> here is a port of Android support patch
>> (http://gcc.gnu.org/ml/gcc-patches/2012-04/msg00944.html) for 4.7
>> branch. Is it OK?
>
> This kind of re-org is not appropriate for the branch.
Could you please specify what exactl
On Tue, Apr 24, 2012 at 3:44 PM, Ilya Enkovich wrote:
>> On Tue, Apr 24, 2012 at 2:54 PM, Ilya Enkovich
>> wrote:
>>> Hello,
>>>
>>> here is a port of Android support patch
>>> (http://gcc.gnu.org/ml/gcc-patches/2012-04/msg00944.html) for 4.7
>>> branch. Is it OK?
>>
>> This kind of re-org is no
> On Tue, Apr 24, 2012 at 3:44 PM, Ilya Enkovich wrote:
>>> On Tue, Apr 24, 2012 at 2:54 PM, Ilya Enkovich
>>> wrote:
Hello,
here is a port of Android support patch
(http://gcc.gnu.org/ml/gcc-patches/2012-04/msg00944.html) for 4.7
branch. Is it OK?
>>>
>>> This kind of r
Committed as obvious.
Richard.
2012-04-12 Richard Guenther
* tree-if-conv.c (main_tree_if_conversion): Move bb under
ENABLE_CHECKING.
Index: gcc/tree-if-conv.c
===
--- gcc/tree-if-conv.c (revision 186760)
+++
On Tue, 24 Apr 2012, Richard Guenther wrote:
>
> I've been carrying this patch for quite some while now and really
> want to go forward with it - the problem is that while all default
> languages work fine after this patch Ada shows some testsuite
> regressions. I've had various hacks/workaround
This fixes redundant store elimination in FRE/PRE to preserve
volatile stores.
Bootstrapped and tested on x86_64-unknown-linux-gnu, applied to the
trunk and the branch.
Richard.
2012-04-24 Richard Guenther
PR tree-optimization/53085
* tree-ssa-pre.c (eliminate): Do not elimi
This patch to the Go frontend adds a check to reject composite literals
that the language syntax forbids. The parser was able to parse some
cases unambiguously because it already knew that some name was a type,
but since that can not always be known that parse is not permitted.
Bootstrapped and ra
On 04/22/2012 03:49 PM, Manuel López-Ibáñez wrote:
On 22 April 2012 21:20, Jason Merrill wrote:
When we aren't printing the expression, please print the type instead; in a
template it might not be clear what the type of the expression we're
complaining about works out to be.
I can do that. Sh
On Tue, Apr 17, 2012 at 11:13 AM, Alan Modra wrote:
> This provides some protection against misuse of r0, r11 and r12. I
> found it useful when enabling out-of-line saves for large frames. ;-)
>
> * config/rs6000/rs6000.c (START_USE, END_USE, NOT_INUSE): Define.
> (rs6000_emit_prol
On Thu, Apr 19, 2012 at 11:36 AM, Alan Modra wrote:
> On Thu, Apr 19, 2012 at 08:00:15PM +0930, Alan Modra wrote:
>> On Wed, Apr 18, 2012 at 12:45:16AM +0930, Alan Modra wrote:
>> > This enables out-of-line save and restore for large frames, and for
>> > ABI_AIX when using the static chain.
>>
>>
OK, thanks!
Jason
2012-04-24 Bill Schmidt
PR target/47197
* config/rs6000/rs6000-c.c (fully_fold_convert): New function.
(altivec_build_resolved_builtin): Call fully_fold_convert.
2012-04-24 Bill Schmidt
PR target/47197
* g
This fixes an error wherein a nontrivial expression oassed to an Altivec
built-in results in an ICE, following Joseph Myers's suggested approach
in the bugzilla.
Bootstrapped and tested with no new regressions on
powerpc64-unknown-linux-gnu. Ok for trunk?
Thanks,
Bill
gcc:
2012-04-24 Bill Sc
On Fri, Apr 20, 2012 at 09:57:24AM +0200, Jakub Jelinek wrote:
> Looks good to me, but IMHO we should go further at this point, and
> default to dwarf_version 4 unless overridden (and override
> also on Darwin to too due to its tools not capable of handling it),
> probably together with making -fno
On 04/22/2012 03:38 PM, Manuel López-Ibáñez wrote:
Then, let's say we have one of such options. For example, let's call
it -Wx for this example. If the behaviour is consistent with other
"group" options like -Wall, then:
-Wx is enabled by default (like now)
-Wno-pedantic does not disable -Wx (li
On 24 April 2012 17:53, Jason Merrill wrote:
> On 04/22/2012 03:38 PM, Manuel López-Ibáñez wrote:
>>
>> Then, let's say we have one of such options. For example, let's call
>> it -Wx for this example. If the behaviour is consistent with other
>> "group" options like -Wall, then:
>>
>> -Wx is enabl
The following patches perform code reorganization in preparation for
epilogue generation in RTL.
[1/2] move the code of the special predicates load_multiple_operation and
store_multiple_operation into a separate function ldm_stm_operation_p
[2/2] generalize ldm_stm_operation_p
No regression on q
This patch to the Go frontend fixes the order of evaluation for a
statement like m[0] = len(m) when m is a map. I added a test case for
this to the master testsuite. Bootstrapped and ran Go testsuite on
x86_64-unknown-linux-gnu. Committed to mainline and 4.7 branch.
Ian
diff -r 5b5853844f00 go
Move the code of the special predicates load_multiple_operation and
store_multiple_operation into a separate function. No change in
functionality.
gcc/ChangeLog
2012-04-24 Ian Bolton
Sameera Deshpande
Greta Yorsh
* config/arm/arm-protos.h (ldm_stm_operatio
Generalize ldm_stm_operation_p with additional parameters that will be used
by epilogue patterns:
* machine mode to support both SImode and DFmode registers
* flag to request consecutive registers in the register list
* flag to indicate whether PC in the register list
gcc/ChangeLog
2012-04-
It seems an obvious cleanup to me.
OK?
2012-04-24 Manuel López-Ibáñez
gcc/
* tree-pretty-print.h (default_tree_printer): Do not declare.
* tree-diagnostic.c: Include tree-pretty-print.h, tree-pass.h and
intl.h.
(default_tree_diagnostic_starter): Make static.
Rename thumb_unexpanded_epilogue to thumb1_unexpanded_epilogue.
In preparation for epilogue generation in RTL and anyway it's the right name
for this function.
Ok for trunk?
Thanks,
Greta
gcc/ChangeLog
2012-04-24 Ian Bolton
Sameera Deshpande
Greta Yorsh
Hello,
PR 52633 is caused by bad interaction between two different vectorizer
pattern recognition passed. A minimal test case is:
void
test (unsigned short *x, signed char *y)
{
int i;
for (i = 0; i < 32; i++)
x[i] = (short) (y[i] << 5);
}
built with "cc1 -O3 -march=armv7-a -mfpu=neon -
Hello!
Back to converting x86 to post-reload compare elimination pass.
Arithmetic operations in x86_64 can implicitly zero extend the result,
and set flags according to the non-extended result. Following testcase
should exercise both features:
--cut here--
void foo (long int, int);
void test (u
Committed as obvious.
http://gcc.gnu.org/viewcvs?view=revision&revision=186775
Johann
PR testsuite/52641
PR tree-optimizations/52891
* gcc.c-torture/compile/pr52891-2.c: Fix test for 16-bit int.
On 04/13/12 03:46, Richard Guenther wrote:
On Fri, Apr 13, 2012 at 12:11 AM, Aldy Hernandez wrote:
Richard. Thanks so much for reviewing and providing an alternative
approach, which AFAICT provides superior results.
A similar effect could be obtained by keeping a flag whether we entered t
OK.
Jason
On Tue, Apr 24, 2012 at 6:56 PM, Uros Bizjak wrote:
> Back to converting x86 to post-reload compare elimination pass.
>
> Arithmetic operations in x86_64 can implicitly zero extend the result,
> and set flags according to the non-extended result. Following testcase
> should exercise both features
On Mon, 2012-04-23 at 12:29 -0700, Richard Henderson wrote:
> On 04/16/12 11:15, Torvald Riegel wrote:
> > - for acq/rel memorders, we don't need seq_cst fences whenever the
> > atomicity is implemented with a lock
> > (hostconfig.c:pre_barrier/post_barrier)
>
> Err.. where? This also seems t
On 04/24/12 09:56, Uros Bizjak wrote:
> 2012-04-24 Uros Bizjak
>
> * compare-elim.c (try_eliminate_compare): Also handle operands with
> implicit extensions.
Ok.
r~
Vladimir Makarov writes:
> On 04/23/2012 11:42 AM, Ulrich Weigand wrote:
>> Vladimir Makarov wrote:
>>
>>> I have a mixed feeling with the patch. I've tried it on SPEC2000 on
>>> x86/x86-64 and ARM. Model algorithm generates bigger code up to 3.5%
>>> (SPECFP on x86), 2% (SPECFP on 86-64), and 0.2
This patch to libgo only defines syscall.PathMax if PATH_MAX is defined
in the system header files. Bootstrapped on x86_64-unknown-linux-gnu.
Committed to mainline and 4.7 branch.
Ian
diff -r 4692bc350a0a libgo/mksysinfo.sh
--- a/libgo/mksysinfo.sh Tue Apr 24 09:23:40 2012 -0700
+++ b/libgo/mksy
On 19/04/2012 14:02, Tobias Burnus wrote:
> Updated patch enclosed.
>
> On 02/14/2012 12:42 PM, Tobias Burnus wrote:
>> in order to gain an overview for our code whether the recent RESHAPE
>> (and friends) bug affects us and to determine for which assignment a
>> reallocation happens, useful to mi
On 12/04/2012 17:23, Tobias Burnus wrote:
> This patch is a kind of follow up to the other one for the same PR -
> though this one is for a separate test case, it is not a regression and
> it's about actual/formal checks.
>
> When trying to fix the rejects-valid bug, I realized that one function
>
This patch to the libgo testsuite driver removes a race in the use of
the ../testdata directory. This is PR go/52462. Ran Go testsuite on
x86_64-unknown-linux-gnu. Committed to mainline and 4.7 branch.
Ian
diff -r 0b5618e33848 libgo/testsuite/gotest
--- a/libgo/testsuite/gotest Tue Apr 24 12:1
Hi,
I'm trying to resolve this rather old PR, where we don't reject uses of
the dot operator in default template arguments, eg:
class blah { int member; };
blah global;
template
class template_blah;
I'm enforcing the check by adding to cp_parser a new
in_template_default_argument_p which c
On Tue, 17 Apr 2012, Terry Guo wrote:
> 2012-04-17 Terry Guo
>
> * Makefile.in (s-mlib): Add new argument MULTILIB_REQUIRED.
> * genmultilib (MULTILIB_REQUIRED): New.
You need to add documentation in fragments.texi for MULTILIB_REQUIRED.
--
Joseph S. Myers
jos...@codesourcer
On Apr 24, 2012, at 17:49 , Jakub Jelinek wrote:
> Talked to Jason about this on IRC last night, he is ok with that, so your
> patch is ok for trunk.
Just committed - Thanks :-)
On 17/04/12 14:24, Richard Guenther wrote:
> On Sat, Apr 14, 2012 at 9:26 AM, Tom de Vries wrote:
>> On 27/01/12 21:37, Tom de Vries wrote:
>>> On 24/01/12 11:40, Richard Guenther wrote:
On Mon, Jan 23, 2012 at 10:27 PM, Tom de Vries
wrote:
> Richard,
> Jakub,
>
> the f
Hi,
We should already reject that because it isn't a constant expression;
I don't think we want to bother checking specifically for this case in
the parser as well.
Sure we reject this kind of code at instantiation time. I suspect
submitter bothered filing a PR because other front-end also
This patch adds heuristics to limit unrolling in loops with branches that may
increase
branch mispredictions. It affects loops that are not frequently iterated, and
that are
nested within a hot region of code that already contains many branch
instructions.
Performance tested with both internal
On Tue, Apr 24, 2012 at 2:26 PM, Teresa Johnson wrote:
> This patch adds heuristics to limit unrolling in loops with branches that may
> increase
> branch mispredictions. It affects loops that are not frequently iterated, and
> that are
> nested within a hot region of code that already contains
On Fri, 20 Apr 2012, Jakub Jelinek wrote:
> 2012-04-19 Jakub Jelinek
>
> PR c/52880
> * c-typeck.c (set_nonincremental_init,
> set_nonincremental_init_from_string): Pass true instead of false
> as IMPLICIT to add_pending_init.
>
> * gcc.dg/pr52880.c: New test.
O
On Fri, 20 Apr 2012, Ian Lance Taylor wrote:
> In Go, a division overflow is supposed to wrap. On x86 processors it
> would raise a SIGFPE signal leading to a panic. This patch fixes that
> problem as well.
It's a pity that this change is Go-specific. For -fwrapv, INT_MIN / -1
and INT_MIN % -
Resending my response in plain text so it will go through to gcc-patches...
On Tue, Apr 24, 2012 at 2:36 PM, Teresa Johnson wrote:
>
>
>
> On Tue, Apr 24, 2012 at 2:30 PM, Andrew Pinski wrote:
>>
>> On Tue, Apr 24, 2012 at 2:26 PM, Teresa Johnson wrote:
>> > This patch adds heuristics to limit
On Sat, 21 Apr 2012, Manuel L?pez-Ib??ez wrote:
> 2012-04-19 Manuel L?pez-Ib??ez
>
> * c-typeck.c (pop_init_level): Improve diagnostics.
> testsuite/
> * gcc.dg/m-un-2.c: Update.
> * gcc.dg/20011021-1.c: Update.
OK. The strange "near initialization" formulation came in with
On Mon, Apr 23, 2012 at 2:07 AM, Jan Hubicka wrote:
> Hi,
> this patch removes remaining traces of the code handling needed and
> reachability flags and replaces last use of it in cgraph_mark_reachable_node
> by
> pointer set. It is nice to have 60k of patch that essentially just removes
> thing
Hi,
In split_function() (ipa-split.c), for the newly created call stmt,
its block is set to the outermost scope, i.e.
DECL_INITIAL(current_function_decl). When we inline this
partially outlined function, we create the new block based on the
block for the call stmt (in expand_call_inline()).
So th
On Sun, 22 Apr 2012, Manuel L?pez-Ib??ez wrote:
> Wshadow warns whenever any declaration shadows a global function
> declaration. This is almost always noise, since most (always?) of the
> time one cannot mistakenly replace a function by another variable. The
> false positives are too common (Linu
On Tue, Apr 24, 2012 at 2:57 PM, Rong Xu wrote:
> Hi,
>
> In split_function() (ipa-split.c), for the newly created call stmt,
> its block is set to the outermost scope, i.e.
> DECL_INITIAL(current_function_decl). When we inline this
> partially outlined function, we create the new block based on t
please the find test case in the attachment. It shows the issue in
google-4_6 branch
-c -O2 -fno-strict-aliasing ss.C -fdump-rtl-expand-all
in the rtl-expand dump, trianglevertices and one the gtest_ar are in
the same partition.
the problem is found in arm compiler, but we manager to reproduce
Olivier Hainque writes:
> *** /tmp/rkQ7Ep_emit-rtl.c2012-04-12 11:19:51.830104512 +0200
> --- gcc/emit-rtl.c2012-04-11 22:39:11.323103686 +0200
> *** set_unique_reg_note (rtx insn, enum reg_
> *** 4955,4960
> --- 4955,4975
> if (GET_CODE (datum) == ASM_OPERAND
Ping?
Here is the formatted ChangeLog in case you want a summary of what I did:
* gcc/ipa-inline.c (edge_badness): Make sure profile is valid before
using it to compute badness.
* gcc/predict.c (maybe_hot_frequency_p): Ditto.
(cgraph_maybe_hot_edge_p): Ditto.
(maybe_hot_edge_p): Ditto.
(probably_
On Sat, Apr 21, 2012 at 2:48 AM, Alan Modra wrote:
> This patch adds out-of-line vector saves and restores. To do this I
> made some infrastructure changes to various functions like
> rs6000_emit_savres_rtx that currently take boolean parameters (savep,
> gpr, and lr). Rather than add yet anothe
On Tue, Apr 24, 2012 at 11:26 PM, Teresa Johnson wrote:
> * params.def (PARAM_MIN_ITER_UNROLL_WITH_BRANCHES): New param.
> (PARAM_UNROLL_OUTER_LOOP_BRANCH_BUDGET): Ditto.
You should add documentation for these new PARAMs to doc/invoke.texi.
I don't really like these new PARAMs: Al
On 04/24/2012 05:24 PM, Paolo Carlini wrote:
Perhaps we aren't checking default arguments unless they're actually
used; a could change that if they aren't dependent.
Your reply reached the mailing list a bit mangled, could you please
clarify?
If the default argument isn't dependent on other t
Hi,
Thanks for all the comments. I have made all the changes as
mentioned and submiited the patch. Summary of changes made:
* Add support for AVX
* Fix documentation in extend.texi
* Make it thread-safe according to H.J.'s comments.
I have attached the patch. Boot-strapped and checked for tes
On Tue, Apr 24, 2012 at 5:10 PM, Sriraman Tallam wrote:
> Hi,
>
> Thanks for all the comments. I have made all the changes as
> mentioned and submiited the patch. Summary of changes made:
>
> * Add support for AVX
> * Fix documentation in extend.texi
> * Make it thread-safe according to H.J.'s c
tejohn...@google.com (Teresa Johnson) writes:
> This patch adds heuristics to limit unrolling in loops with branches that may
> increase
> branch mispredictions. It affects loops that are not frequently iterated, and
> that are
> nested within a hot region of code that already contains many bran
Noticed that this testcase wasn't put in as part of the patch. Fixed as
follows.
tested x86/linux
-benjamin2012-04-24 Benjamin Kosnik
PR libstdc++/52689
* testsuite/17_intro/static.cc: New.
diff --git a/libstdc++-v3/testsuite/17_intro/static.cc b/libstdc++-v3/testsuite/17_intro/static.cc
On 04/25/2012 01:41 AM, Jason Merrill wrote:
On 04/24/2012 05:24 PM, Paolo Carlini wrote:
Perhaps we aren't checking default arguments unless they're actually
used; a could change that if they aren't dependent.
Your reply reached the mailing list a bit mangled, could you please
clarify?
If t
On Tue, Apr 24, 2012 at 5:24 PM, H.J. Lu wrote:
> On Tue, Apr 24, 2012 at 5:10 PM, Sriraman Tallam wrote:
>> Hi,
>>
>> Thanks for all the comments. I have made all the changes as
>> mentioned and submiited the patch. Summary of changes made:
>>
>> * Add support for AVX
>> * Fix documentation in
On Tue, Apr 24, 2012 at 7:06 PM, Sriraman Tallam wrote:
> On Tue, Apr 24, 2012 at 5:24 PM, H.J. Lu wrote:
>> On Tue, Apr 24, 2012 at 5:10 PM, Sriraman Tallam wrote:
>>> Hi,
>>>
>>> Thanks for all the comments. I have made all the changes as
>>> mentioned and submiited the patch. Summary of cha
This patch to libgo fixes the net package to work correctly on Solaris.
The main fix is to the select code, to fix the case where one goroutine
closes a file descriptor while another goroutine is waiting for I/O on
the descriptor. Bootstrapped and ran Go testsuite on
sparc-sun-solaris2.11. Commit
This patch to libgo fixes the crypto/rand test to read all the data that
it needs. This should fix PR 52341. Bootstrapped and ran crypto/rand
test on x86_64-unknown-linux-gnu. Committed to mainline and 4.7 branch.
Ian
diff -r c4931780d05a libgo/go/crypto/rand/rand_test.go
--- a/libgo/go/crypto
On Tue, Apr 24, 2012 at 6:13 PM, Andi Kleen wrote:
> tejohn...@google.com (Teresa Johnson) writes:
>
>> This patch adds heuristics to limit unrolling in loops with branches that
>> may increase
>> branch mispredictions. It affects loops that are not frequently iterated,
>> and that are
>> nested
On Tue, Apr 24, 2012 at 07:19:42PM -0400, David Edelsohn wrote:
> This patch is okay with the macro usage fix.
Thanks, series 2 to 6 committed as 186796, 186797, 186798, 186799,
186800. I noticed after I committed the lot that 186797 has some
duplicated lines (harmless), corrected in 186798, and
http://codereview.appspot.com/6099055/diff/1/loop-unroll.c
File loop-unroll.c (right):
http://codereview.appspot.com/6099055/diff/1/loop-unroll.c#newcode156
loop-unroll.c:156: static bool
An empty line here.
http://codereview.appspot.com/6099055/diff/1/loop-unroll.c#newcode182
loop-unroll.c:182
On 18/04/2012, at 9:17 PM, Richard Guenther wrote:
> On Wed, Apr 18, 2012 at 4:15 AM, Maxim Kuvyrkov
> wrote:
>> Steven,
>> J"orn,
>>
>> I am looking into fixing performance regression on EEMBC's bitmnp01, and a
>> version of your combined patch attached to PR38785 still works very well.
>>
Hello Mikael,
thanks for the review. Regarding:
Mikael Morin wrote:
is there a reason to guard the class_pointer condition with
attr.class_ok in the first conditional and with CLASS_DATA(...) !=
NULL in the two other ones? Not that it matters much, and in fact, I
think the patch as is is good
89 matches
Mail list logo
