This patch updates libgo to the Go 1.0.3 release. Bootstrapped and ran
Go testsuite on x86_64-unknown-linux-gnu. Committed to mainline and 4.7
branch.
Ian
foo.patch.bz2
Description: patch
The Go := operator can declare new variables while also referring to
existing variables. The Go frontend was miscompiling a case like a, b,
c := b, a, 1 where a and b already exist. It was assigning to a and
then using the new value of to assign to b. This patch fixes this bug.
Bootstrapped and
On Tue, Oct 2, 2012 at 8:06 PM, Uros Bizjak wrote:
>
> 2012-10-03 Uros Bizjak
>
> * configure.ac (noexception_flags): Add -fasynchronous-unwind-tables.
> * configure: Regenerate.
Thanks for measuring. This is OK if nobody objects within 24 hours.
Ian
On Tue, Oct 2, 2012 at 12:10 PM, Lawrence Crowl wrote:
> On 10/2/12, Richard Guenther wrote:
>>
>> You are changing a hashtable used by fold checking, did you test
>> with fold checking enabled?
>
> I didn't know I had to do anything beyond the normal make check.
> What do I do?
Fold checking is
On Tue, Oct 2, 2012 at 8:31 PM, Ian Lance Taylor wrote:
> > On a related issue, it looks to me that the compiler itself should be
> > compiled with -funwind-tables, otherwise there are no backtraces
> > generated, even if libbacktrace is linked in and operational. Again,
> > x86_6
Tom, this is mainly a libcpp change. Would you mind taking a look?
Thanks,
Ollie
On Mon, Oct 1, 2012 at 9:56 AM, Simon Baldwin wrote:
>
> Ping, again.
>
>
> On 21 September 2012 12:45, Simon Baldwin wrote:
> >
> > Ping.
> >
> > http://gcc.gnu.org/ml/gcc-patches/2012-09/msg00459.html
> >
> > Fu
Hi,
This patch propagates the profile counts during RTL expansion. In
many cases, there is no way to determine the exact count of an edge
generated during the expansion. So this patch uses some simple
heuristics to estimate the edge counts but ensures that the counts of
the basic blocks correspond
Updated Patch
Add functions symtab_node_def::try_function and symtab_node_def::try_variable.
These function return a pointer to the more specific type (e.g. cgraph_node*)
if and only if the general type (symtab_node aka symtab_node_def*) is an
instance of the specific type. These functions are es
On 10/02/2012 06:42 PM, Bernd Schmidt wrote:
On 10/03/2012 12:29 AM, Vladimir Makarov wrote:
On 12-10-02 7:30 AM, Bernd Schmidt wrote:
On 09/28/2012 12:59 AM, Vladimir Makarov wrote:
+ We keep RTL code at most time in such state that the virtual
+ registers can be changed by just the corre
This patch to the Go frontend fixes the type reflection strings in a
couple of cases to match those generated by the gc compiler. For an
empty struct, this generates "struct {}" rather than "struct { }". For
a type defined within a function, this generates pkg.type rather than
pkg.function_name$
On 10/03/2012 12:29 AM, Vladimir Makarov wrote:
> On 12-10-02 7:30 AM, Bernd Schmidt wrote:
>> On 09/28/2012 12:59 AM, Vladimir Makarov wrote:
>>> + We keep RTL code at most time in such state that the virtual
>>> + registers can be changed by just the corresponding hard registers
>>> + (with
On Mon, Oct 1, 2012 at 1:11 AM, Steven Bosscher wrote:
> On Mon, Oct 1, 2012 at 12:44 AM, Vladimir Makarov wrote:
>> Actually, I don't see there is a problem with LRA right now. I think we
>> should first to solve a whole compiler memory footprint problem for this
>> test because cpu utilizati
On 12-10-02 7:30 AM, Bernd Schmidt wrote:
On 09/28/2012 12:59 AM, Vladimir Makarov wrote:
+ We keep RTL code at most time in such state that the virtual
+ registers can be changed by just the corresponding hard registers
+ (with zero offsets) and we have the right RTL code. To achieve this
http://code.google.com/p/go/issues/detail?id=4110 points out that the Go
frontend is misparsing some channel operations. The parsing in this
area is rather complex for a recursive descent parser (though
straightforward for an LALR parser). In any case, this patch fixes the
problem. Bootstrapped
This patch updates the Go testsuite to the current master testsuite.
All the tests should pass. The complete patch is too large to include
here, and consists simply of copying files from the master repository.
I have included only the patch to the testsuite driver, go-test.exp.
The Go testsuite wa
Hi!
We ICE on the following testcase, because cxx_eval_constant_expression
on COMPOUND_EXPR uses STRIP_NOPS (op1) - needed for the check whether it
is an artificial COMPOUND_EXPR. Unfortunately that means it uses op1
without NOP_EXPRs even if it is a user comma expression.
Fixed thusly, bootstra
On Tue, Oct 2, 2012 at 11:01 PM, H.J. Lu wrote:
>
> This patch documents -mprefer-avx128. OK for trunk and 4.7?
>
> 2012-10-02 H.J. Lu
>
> PR target/54785
> * doc/invoke.texi: Document -mprefer-avx128.
OK everywhere.
Thanks,
Uros.
Hi all,
I have just committed as obvious a one-line patch to fix an
ICE-on-invalid OOP problem:
http://gcc.gnu.org/viewcvs?view=revision&revision=192005
Cheers,
Janus
Hi,
When I implemented the simplification of a truncate of a memory, I
did not think about the case where we would have a truncate of a
vector mode. This fixes this case.
Committed as obvious after a bootstrap and test on x86_64-linux-gnu
and also a build and test for arm-linux-gnueabi.
Thanks,
Hi,
This patch documents -mprefer-avx128. OK for trunk and 4.7?
Thanks.
H.J.
---
2012-10-02 H.J. Lu
PR target/54785
* doc/invoke.texi: Document -mprefer-avx128.
diff --git a/gcc/doc/invoke.texi b/gcc/doc/invoke.texi
index 7578dda..0e7e441 100644
--- a/gcc/doc/invoke.texi
+
Han,
Have you done full testsuite runs with and without this patch? I know
you mentioned you've done builds on both x86_64 and arm; it might be
nice to compare tests too.
Thanks,
-Kees
On Mon, Sep 10, 2012 at 5:26 PM, Matthias Klose wrote:
> On 08.09.2012 01:07, Kees Cook wrote:
>> Hi,
>>
>> I
On 2 October 2012 07:02, Ian Lance Taylor wrote:
>
> The libgcc part of this is OK.
Thanks, I've committed it and closed the PR.
On Tue, Oct 2, 2012 at 9:35 PM, H.J. Lu wrote:
> This patch checks SSE and YMM state support for -march=native. Tested
> on Linux/x86-64. OK to install?
>
> 2012-10-02 H.J. Lu
>
> PR target/54741
> * config/i386/driver-i386.c (XCR_XFEATURE_ENABLED_MASK): New.
> (XSTA
As promised, here's the patch to adjust the MIPS BADDU patterns for
the new (subreg (plus)) simplification. Tested on mipsisa32-elf
and mipsisa64-elf. Applied.
Richard
gcc/
* config/mips/mips.md (*baddu_si_eb, *baddu_si_el): Merge into...
(*baddu_si): ...this new pattern.
Inde
Hi,
This patch checks SSE and YMM state support for -march=native. Tested
on Linux/x86-64. OK to install?
Thanks.
H.J.
---
2012-10-02 H.J. Lu
PR target/54741
* config/i386/driver-i386.c (XCR_XFEATURE_ENABLED_MASK): New.
(XSTATE_FP): Likewise.
(XSTATE_SSE):
Andrew Pinski writes:
> On Thu, Sep 27, 2012 at 11:13 AM, Uros Bizjak wrote:
>> 2012-09-27 Uros Bizjak
>>
>> PR rtl-optimization/54457
>> * simplify-rtx.c (simplify_subreg):
>> Simplify (subreg:M (op:N ((x:N) (y:N)), 0)
>> to (op:M (subreg:M (x:N) 0) (subreg:M (
Latest results for 4.7.x
-tgc
Testresults for 4.7.2
alphaev68-dec-osf5.1a (2)
hppa2.0w-hp-hpux11.00
hppa2.0w-hp-hpux11.11
hppa64-hp-hpux11.11
i386-apple-darwin10.8.0
i686-pc-linux-gnu
powerpc-apple-darwin8.11.0
x86_64-apple-darwin10.8.0
x86_64-apple-darwin12.2.0
Testresults for
Latest results for 4.6.x
-tgc
Testresults for 4.6.3
alphaev68-dec-osf5.1a
Testresults for 4.6.2
alphaev68-dec-osf5.1a
Testresults for 4.6.1
alphaev68-dec-osf5.1a (2)
Testresults for 4.6.0
alphaev68-dec-osf5.1a
Index: buildstat.html
=
On 10/2/12, Richard Guenther wrote:
> On Mon, 1 Oct 2012, Lawrence Crowl wrote:
> > Change more non-GTY hash tables to use the new type-safe
> > template hash table. Constify member function parameters that
> > can be const. Correct a couple of expressions in formerly
> > uninstantiated template
On Sep 25, 2012, Jakub Jelinek wrote:
> On Sun, Sep 23, 2012 at 07:59:37AM -0300, Alexandre Oliva wrote:
>> This patch introduces a global mode of dead_debug tracking for use in
>> fast DCE. If a debug use reaches the top of a basic block before
>> finding its death point, the pending and subseq
This patch fixes a crash in dwarf2out because of a too-large debug
expression. Jakub approved it for trunk and 4.7 branches in bugzilla.
I'm installing it in the trunk momentarily, and later today on 4.7 after
I give it a spin there. Regstrapped on x86_64-linux-gnu and
i686-linux-gnu.
I'm keepin
Latest results for 4.5.x
-tgc
Testresults for 4.5.4:
alphaev68-dec-osf5.1a
hppa2.0w-hp-hpux11.00
hppa64-hp-hpux11.00
Testresults for 4.5.3:
alphaev68-dec-osf5.1a
i386-pc-solaris2.8
Testresults for 4.5.2:
alphaev68-dec-osf5.1a
Testresults for 4.5.1:
alphaev68-dec-osf5.1a
Testresu
Uros has already taken care of the main patch for the problem, but I
feel it's appropriate to protect vt_add_function_parameter should
val_lowpart actually return NULL.
I'm checking this in as obvious. Regstrapped on x86_64-linux-gnu and
i686-linux-gnu.
Deal with var_lowpart failure in function
On Tue, 2 Oct 2012, Gabriel Dos Reis wrote:
Whining on this list about libstdc++ internal macros and your dislike
of them is not going to produce anything today or tomorrow.
Other compilers using libstdc++ was just an extra argument. Even if g++
was the only compiler on earth, I would still c
On Tue, Oct 2, 2012 at 10:48 AM, Uros Bizjak wrote:
> On Tue, Oct 2, 2012 at 7:44 PM, Gabriel Dos Reis
> wrote:
>
> On a related issue, it looks to me that the compiler itself should be
> compiled with -funwind-tables, otherwise there are no backtraces
> generated, even if libbackt
Latest results for 4.4.x
-tgc
Testresults for 4.4.7:
alphaev68-dec-osf5.1a
hppa2.0w-hp-hpux11.00
hppa2.0w-hp-hpux11.11
hppa64-hp-hpux11.00
hppa64-hp-hpux11.11
i386-pc-solaris2.8
Testresults for 4.4.1:
alphaev68-dec-osf5.1a
Index: buildstat.html
Forgot to attach...
On 10/2/2012 2:11 PM, rbmj wrote:
Patch 2: [fixincludes] Clean up fixincludes test machinery
TODO Prior to commit:
* fixincl.x: Regenerate
ChangeLog
2012-09-23 Bruce Korb
* check.tpl: export TEST_MODE=true for testing
* fixincl.c (te_verbose): extract to fix
Hi, any one got a chance to take look at this patch?
It seems that some other guys are also interested in this patch, the
clang developer is also proposing implement this
"-fstack-protect-strong" option.
Patch has just been merged with newest trunk and fixed a bug reported by Kees.
Tested fox x8
Patch 4: Minor changes to fix compilation on VxWorks
ChangeLog [gcc]:
* gcov-io.c (gcov_open): Pass third argument to open() unconditionally
ChangeLog [libstdc++-v3]:
* libstdc++-v3/config/os/vxworks/os_defines.h: Define NOMINMAX
>From 420bf6c2b0bde5f1689663b477add8fc9df2a6f0 Mon Sep 17 00:00:00
Patch 3: Add --enable-libstdcxx option at top level configure
TODO prior to commit:
* configure: regenerate
ChangeLog:
* configure.ac: Add --enable-libstdcxx option
>From 3f0d38b7b7b70659a57ac4266701a71a5f948860 Mon Sep 17 00:00:00 2001
From: rbmj
Date: Tue, 2 Oct 2012 13:54:21 -0400
Subject
Patch 2: [fixincludes] Clean up fixincludes test machinery
TODO Prior to commit:
* fixincl.x: Regenerate
ChangeLog
2012-09-23 Bruce Korb
* check.tpl: export TEST_MODE=true for testing
* fixincl.c (te_verbose): extract to fixlib.h
(run_compiles): in test mode, if the
Forgot to attach.
On 10/2/2012 2:09 PM, rbmj wrote:
Patch 1: [fixincludes] Fixes for VxWorks
TODO Prior to commit:
* fixincl.x: Regenerate
ChangeLog [fixincludes]:
2012-06-19 Robert Mason
* fixinc.in: Check to see if the machine_name fix needs to be
disabled.
viz. vxworks must
Patch 1: [fixincludes] Fixes for VxWorks
TODO Prior to commit:
* fixincl.x: Regenerate
ChangeLog [fixincludes]:
2012-06-19 Robert Mason
* fixinc.in: Check to see if the machine_name fix needs to be disabled.
viz. vxworks must not check the machine name for fix applicabilit
On 9/23/2012 7:19 PM, Bruce Korb wrote:
The attached patch needs to be split into two and I will do that before
I actually push the thing. Since I have run out of play time this weekend
and since I will be in the Ukraine in two weeks for two weeks, this patch
is unlikely to get pushed before th
On Tue, Oct 2, 2012 at 7:44 PM, Gabriel Dos Reis
wrote:
>>> > On a related issue, it looks to me that the compiler itself should be
>>> > compiled with -funwind-tables, otherwise there are no backtraces
>>> > generated, even if libbacktrace is linked in and operational. Again,
>>> > x86_64-linux-
On Tue, Oct 2, 2012 at 12:14 PM, Jakub Jelinek wrote:
> On Tue, Oct 02, 2012 at 10:12:38AM -0700, Ian Lance Taylor wrote:
>> On Tue, Oct 2, 2012 at 8:22 AM, Uros Bizjak wrote:
>> >
>> > On a related issue, it looks to me that the compiler itself should be
>> > compiled with -funwind-tables, other
On Mon, Oct 1, 2012 at 7:11 PM, Michael Meissner
wrote:
> 2012-10-01 Michael Meissner
>
> * config/rs6000/rs6000.c (rs6000_option_override_internal): If
> -mcpu= is not specified and the compiler is not configured
> using --with-cpu=, use the bits from the TARGET_DEFAULT
On Tue, Oct 02, 2012 at 10:12:38AM -0700, Ian Lance Taylor wrote:
> On Tue, Oct 2, 2012 at 8:22 AM, Uros Bizjak wrote:
> >
> > On a related issue, it looks to me that the compiler itself should be
> > compiled with -funwind-tables, otherwise there are no backtraces
> > generated, even if libbacktr
On Tue, Oct 2, 2012 at 8:22 AM, Uros Bizjak wrote:
>
> On a related issue, it looks to me that the compiler itself should be
> compiled with -funwind-tables, otherwise there are no backtraces
> generated, even if libbacktrace is linked in and operational. Again,
> x86_64-linux-gnu host defaults to
On Tue, Oct 2, 2012 at 8:07 AM, Marc Glisse wrote:
> Or do you mean:
> always call __builtin_llabs (whether we have an llabs or not), and let the
> compiler replace it with either (x<0)?-x:x or a library call (I assume it
> never does that unless it has seen a corresponding declaration)?
See wha
Hi Richi,
(Snip)
> + (!cst_and_fits_in_hwi (step))
> +{
> + if( loop->inner != NULL)
> +{
> + if (dump_file && (dump_flags & TDF_DETAILS))
> +{
> + fprintf (dump_file, "Reference %p:\n", (void *) ref);
> + fprintf (dump_file, "(base "
On Tue, Oct 2, 2012 at 9:34 AM, Daniel Krügler
wrote:
> 2012/10/2 Marc Glisse :
>> Here I am talking of a library issue: the wording that says that there are
>> sufficient overloads such that integer types call the double version of math
>> functions. It is fairly obvious that it doesn't apply to
On Tue, Oct 2, 2012 at 8:07 AM, Marc Glisse wrote:
>>> The library installed by the system was compiled with g++, and is then
>>> used
>>> with clang++. If we can avoid installing 2 config.h files to make that
>>> work...
>>
>>
>> Two things:
>> 1. that is clearly a clang problem. I don't think
Hi,
I've backported Ulrich's reload fix(attached)
http://gcc.gnu.org/ml/gcc-patches/2012-07/msg01421.html
to aarch64-4.7-branch and committed it.
SendingChangeLog.aarch64
Sendingreload.c
Transmitting file data ..
Committed revision 191987.
Thanks,
Tejas.
diff --git a/gcc/rel
On Tue, Oct 02, 2012 at 10:13:25AM +0200, Gunther Nikl wrote:
> Michael Meissner wrote:
> > Segher Boessenkool asked me on IRC to break out the fix in the last change.
> > This patch is just the change to set the default options if the user did not
> > use -mcpu= and the compiler was not configured
On 10/01/2012 02:51 PM, Richard Sandiford wrote:
Vladimir Makarov writes:
+/* Return register bank of given hard regno for the current target. */
+DEFHOOK
+(register_bank,
+ "A target hook which returns the register bank number to which the\
+ register @var{hard_regno} belongs to. The smalle
Similar to PR43742, the ARCompact port gets an ICE from the current mainline
version of web.c:union_match_dups for its zero overhead loop pattern.
My first attempt at rectifying this was equivalent in effect to the patch
from comment #1 from this patch; that seemed to work well enough.
Later I stu
On Tue, Oct 2, 2012 at 3:08 PM, Ian Lance Taylor wrote:
>> 2012-10-02 Uros Bizjak
>>
>> PR other/54761
>> * configure.ac (EXTRA_FLAGS): New.
>> * Makefile.am (AM_FLAGS): Add $(EXTRA_FLAGS).
>> * configure, Makefile.in: Regenerate.
>
> This is OK.
Thanks, committ
On 10/02/2012 01:01 AM, Jeff Law wrote:
On 09/27/2012 04:59 PM, Vladimir Makarov wrote:
The following patch adds a code neccessary for correct work of LRA
(function ira_setup_eliminable_regset) and for correct work of the
compiler when LRA is used (see file dwarf2out.c).
2012-09-27 Vladimir
On 10/02/2012 12:22 AM, Jeff Law wrote:
On 10/01/2012 07:14 PM, Vladimir Makarov wrote:
Analogous live ranges are used in IRA as intermidiate step to build a
conflict graph. Actually, the first approach was to use IRA code to
assign hard registers to pseudos (e.g. Jeff Law tried this appro
2012/10/2 Marc Glisse :
> Here I am talking of a library issue: the wording that says that there are
> sufficient overloads such that integer types call the double version of math
> functions. It is fairly obvious that it doesn't apply to abs(long) for
> instance which has an explicit overload. For
On Mon, Oct 01, 2012 at 02:05:50PM +0200, Richard Guenther wrote:
> 2012-10-01 Richard Guenther
>
> PR lto/47788
> * tree-streamer-out.c (write_ts_block_tree_pointers): For
> inlined functions outer scopes write the ultimate origin
> as BLOCK_ABSTRACT_ORIGIN and BLOCK_SO
Richard Sandiford writes:
>> +/* Merge ranges R1 and R2 and returns the result. The function
>> + maintains the order of ranges and tries to minimize size of the
>> + result range list. */
>> +lra_live_range_t
>> +lra_merge_live_ranges (lra_live_range_t r1, lra_live_range_t r2)
>> +{
>> +
Aaron,
I'm currently fixing other issues with gengtype and I needed this
patch on top of them. I will be rolling both patches into a single
one and commit them today/tomorrow. If you were working on further
fixes to this, please give me a chance to commit this one first.
Thanks. Diego.
Vladimir Makarov writes:
> This is the major patch containing all new files. The patch also adds
> necessary calls to LRA from IRA.As the patch is too big, it continues in
> the next email.
>
> 2012-09-27 Vladimir Makarov
>
> * Makefile.in (LRA_INT_H): New.
> (OBJS): Add lra.o, lra
This fixes PR54735 - a bad interaction of non-up-to-date virtual
SSA form, update-SSA and cfg cleanup. Morale of the story:
cfg cleanup can remove blocks and thus release SSA names - SSA
update is rightfully confused when such released SSA name is
still used at update time.
The following patch f
On Tue, Oct 2, 2012 at 3:01 PM, Jakub Jelinek wrote:
> Hi!
>
> As discussed in the PR and on IRC, this patch verifies that vector
> CONSTRUCTOR in GIMPLE is either empty CONSTRUCTOR, or contains scalar
> elements of type compatible with vector element type (then the verification
> is less strict,
On Tue, Oct 2, 2012 at 2:01 AM, Uros Bizjak wrote:
>
> 2012-10-02 Uros Bizjak
>
> PR other/54761
> * configure.ac (EXTRA_FLAGS): New.
> * Makefile.am (AM_FLAGS): Add $(EXTRA_FLAGS).
> * configure, Makefile.in: Regenerate.
This is OK.
Thanks.
Ian
On Tue, 2 Oct 2012, Gabriel Dos Reis wrote:
I understand that it is originally a library issue, but I don't think
it makes sense to resolve it in isolation of that core issue.
They seem mostly orthogonal to me, since the library only uses an informal
language describing the desired outcome an
Hi!
As discussed in the PR and on IRC, this patch verifies that vector
CONSTRUCTOR in GIMPLE is either empty CONSTRUCTOR, or contains scalar
elements of type compatible with vector element type (then the verification
is less strict, allows less than TYPE_VECTOR_SUBPARTS elements and allows
non-NUL
This patch cleans up some documentation issues for eliminated mode, and
fixes some errors for marginal cases. Not worth trying to concoct tests
for these cases, which were found by code review, not from any reported
bugs. Also forbid use of Eliminated mode if Long_Long_Integer'Size is
not 64. Also
On 09/29/2012 12:37 AM, Bin Cheng wrote:
Hi Steven,
This is the updated patch according to your comments. Please review.
I also re-collected code size data and found it is improved by about 0.24%
for mips, which is better than previous data. I believe this should be
caused by recent changes in t
On Mon, Oct 1, 2012 at 5:57 PM, Marc Glisse wrote:
> [merging both threads, thanks for the answers]
>
>
> On Mon, 1 Oct 2012, Richard Guenther wrote:
>
optabs should be fixed instead, an is_gimple_val condition is
implicitely
val != 0.
>>>
>>>
>>> For vectors, I think it should be v
When the type of an object is a CPP type and the object initialization
requires calling its default C++ constructor, the Ada compiler did not
generate the call to a C++ constructor which has all parameters with
defaults (and hence it covers the default C++ constructor). The
following test must now
This patch goes almost all the way in removing N_Return_Statement,
and replacing it by N_Simple_Return_Statement. No test, since no
functional effect.
Tested on x86_64-pc-linux-gnu, committed on trunk
2012-10-02 Robert Dewar
* sinfo.adb, sinfo.ads, sem_util.adb, sem_util.ads, types.h,
On Mon, 1 Oct 2012, venkataramanan.ku...@amd.com wrote:
> Hi,
>
> The below patch fixes the FFT/Scimark regression caused by useless prefetch
> generation.
>
> This fix tries to make prefetch less aggressive by prefetching arrays in the
> inner loop, when the step is invariant in the entire loo
Oleg Endo wrote:
> This is the patch as proposed in the PR to make
> libgcc/config/sh/linux-atomic use the appropriate compiler generated
> atomic built-in functions depending on the currently selected
> atomic-model.
>
> Tested on 191894 with 'make all-gcc' and by compiling code to see if the
>
Oleg Endo wrote:
> This handles the case where the T bit is stored to a reg as the value
> 0x7FFF or 0x8000.
> Tested on rev 191894 with
> make -k check RUNTESTFLAGS="--target_board=sh-sim
> \{-m2/-ml,-m2/-mb,-m2a/-mb,-m4/-ml,-m4/-mb,-m4a/-ml,-m4a/-mb}"
>
> and no new failures.
> OK?
OK.
On 09/28/2012 12:59 AM, Vladimir Makarov wrote:
> + We keep RTL code at most time in such state that the virtual
> + registers can be changed by just the corresponding hard registers
> + (with zero offsets) and we have the right RTL code. To achieve this
> + we should add initial offs
Hi Vlad,
Vladimir Makarov writes:
> This is the major patch containing all new files. The patch also adds
> necessary calls to LRA from IRA.As the patch is too big, it continues in
> the next email.
>
> 2012-09-27 Vladimir Makarov
>
> * Makefile.in (LRA_INT_H): New.
> (OBJS): Add
Hello All,
As I observed in http://gcc.gnu.org/ml/gcc/2010-07/msg00248.html and in
http://gcc.gnu.org/ml/gcc/2012-10/msg3.html the mark_hook GTY annotation is
sometimes incorrectly ingored by gengtype.
The example in http://gcc.gnu.org/ml/gcc/2012-10/msg3.html demonstrates
that incorrect
This avoids applying the NRV optimization for small structures and creating
useless elaboration variables for loops.
Tested on x86_64-suse-linux, applied on the mainline and 4.7 branch.
2012-10-02 Eric Botcazou
* gcc-interfaces/decl.c (elaborate_expression_1): Use the variable for
Hello,
This is the patch as proposed in the PR to make
libgcc/config/sh/linux-atomic use the appropriate compiler generated
atomic built-in functions depending on the currently selected
atomic-model.
Tested on 191894 with 'make all-gcc' and by compiling code to see if the
__SH_ATOMIC_MODEL_*__ de
On Tue, Oct 2, 2012 at 12:08 PM, Vladimir Yakovlev wrote:
> Will we wait for LRA commit or is it possiple to commit to trank
> vzeroupper patch now?
Since we can emit vzeroupper now, we will wait for LRA.
Uros.
Will we wait for LRA commit or is it possiple to commit to trank
vzeroupper patch now?
2012/10/2 Uros Bizjak :
> On Tue, Oct 2, 2012 at 11:35 AM, Vladimir Yakovlev
> wrote:
> The compiler with the patch and without post_reload.patch is built and
> works
> successfully. It has the on
On Tue, Oct 2, 2012 at 11:35 AM, Vladimir Yakovlev wrote:
The compiler with the patch and without post_reload.patch is built and
works
successfully. It has the only failure with avx-vzeroupper-3 test because of
post reload problem.
>>>
>>> Ok, can you please elaborate a bit on
On Tue, Oct 2, 2012 at 4:31 AM, Richard Guenther
wrote:
> On Mon, Oct 1, 2012 at 8:39 PM, Gabriel Dos Reis
> wrote:
>> On Mon, Oct 1, 2012 at 1:27 PM, Michael Meissner
>> wrote:
>>> On Mon, Oct 01, 2012 at 02:02:26PM -0400, Michael Meissner wrote:
Your change on September 30th, breaks the p
2012/9/30 Uros Bizjak :
> On Thu, Sep 20, 2012 at 8:35 AM, Uros Bizjak wrote:
>> On Thu, Sep 20, 2012 at 8:06 AM, Vladimir Yakovlev
>> wrote:
>>> The compiler with the patch and without post_reload.patch is built and works
>>> successfully. It has the only failure with avx-vzeroupper-3 test beca
Il 02/10/2012 10:49, Steven Bosscher ha scritto:
> On Tue, Oct 2, 2012 at 10:29 AM, Paolo Bonzini wrote:
>> Il 02/10/2012 09:28, Steven Bosscher ha scritto:
My experience shows that these lists are usually 1-2 elements. Although
in
> this case, there are pseudos with huge number e
On Tue, Oct 2, 2012 at 4:21 AM, Marc Glisse wrote:
> On Tue, 2 Oct 2012, Gabriel Dos Reis wrote:
>
>> On Tue, Oct 2, 2012 at 3:57 AM, Marc Glisse wrote:
>>>
>>> (Forgot libstdc++...)
>>>
>>>
>>> Hello,
>>>
>>> here is the patch from PR54686. Several notes:
>>>
>>> * I'll have to ask experts if st
On Mon, Oct 1, 2012 at 8:39 PM, Gabriel Dos Reis
wrote:
> On Mon, Oct 1, 2012 at 1:27 PM, Michael Meissner
> wrote:
>> On Mon, Oct 01, 2012 at 02:02:26PM -0400, Michael Meissner wrote:
>>> Your change on September 30th, breaks the powerpc port because the
>>> REPORT_DETAILS value in the enumerati
On Tue, Oct 2, 2012 at 1:31 AM, Sharad Singhai wrote:
> Here is a patch to fix test breakage caused by r191883. Bootstrapped
> on x86_64 and tested with
> make -k check RUNTESTFLAGS="--target_board=unix/\{,-m32\}".
>
> Okay for trunk?
Ok.
Thanks,
Richard.
> Thanks,
> Sharad
>
> 2012-10-01 Shar
On Tue, Oct 2, 2012 at 1:11 AM, Xinliang David Li wrote:
> On Mon, Oct 1, 2012 at 4:05 PM, Sharad Singhai wrote:
>> Thanks for tracking down and fixing the powerpc port.
>>
>> The "dump_kind_p ()" check is redundant but canonical form here. I
>> think blocks of dump code guarded by "if dump_kind_
On Mon, 1 Oct 2012, Lawrence Crowl wrote:
> Change more non-GTY hash tables to use the new type-safe template hash table.
> Constify member function parameters that can be const.
> Correct a couple of expressions in formerly uninstantiated templates.
>
> The new code is 0.362% faster in bootstrap
On Tue, 2 Oct 2012, Gabriel Dos Reis wrote:
On Tue, Oct 2, 2012 at 3:57 AM, Marc Glisse wrote:
(Forgot libstdc++...)
Hello,
here is the patch from PR54686. Several notes:
* I'll have to ask experts if std::abs(unsigned) (yes, a weird thing to do,
but still) is meant to return a double...
Hi,
I have just merged upstream gcc-4_7-branch on the aarch64-4.7-branch up to
r191881.
Thanks
Sofiane
On Mon, 1 Oct 2012, Jan Hubicka wrote:
> > >
> > > So the unvectorized cost is
> > > SIC * niters
> > >
> > > The vectorized path is
> > > SOC + VIC * ((niters-PL_ITERS-EP_ITERS)/VF) + VOC
> > > The scalar path of vectorizer loop is
> > > SIC * niters + SOC
> >
> >
On Tue, Oct 2, 2012 at 3:57 AM, Marc Glisse wrote:
> (Forgot libstdc++...)
>
>
> Hello,
>
> here is the patch from PR54686. Several notes:
>
> * I'll have to ask experts if std::abs(unsigned) (yes, a weird thing to do,
> but still) is meant to return a double...
don't we have a core issue about p
On Tue, Oct 2, 2012 at 10:08 AM, Andreas Schwab wrote:
>> +if test "x$GCC" = "xyes"; then
>> + CFLAGS="$CFLAGS -funwind-tables"
>> +fi
>> +
>
> Don't modify CFLAGS, instead you should substitute a new variable that
> is added to AM_CFLAGS. CFLAGS is reserved for the user to override.
Thanks, a
(Forgot libstdc++...)
Hello,
here is the patch from PR54686. Several notes:
* I'll have to ask experts if std::abs(unsigned) (yes, a weird thing to do, but
still) is meant to return a double...
* I still don't like the configure-time _GLIBCXX_USE_INT128, I think it should
use defined(__SIZEOF
1 - 100 of 120 matches
Mail list logo
