On Thu, Oct 16, 2014 at 05:25:57PM +1030, Alan Modra wrote:
> Ping?
> https://gcc.gnu.org/ml/gcc-patches/2014-09/msg00704.html
I think the simplification should be done when constructing the expressions,
i.e. if possible in the simplification callback or so if it isn't
performed at some level.
Bec
On 08/10/12 11:24, Richard Guenther wrote:
> On Sun, Oct 7, 2012 at 12:44 PM, Tom de Vries wrote:
>> Richard,
>>
>> attached patch checks that unlinked uses do not contain ssa-names when
>> renaming.
>>
>> This assert triggers when compiling (without the fix) the PR54735 example.
>>
>> AFAIU, it
On Thu, Oct 16, 2014 at 10:24:45AM +0400, Kirill Yukhin wrote:
> Hello,
> This patch extends insertion hook.
>
> AVX-512* tests on top of patch-set all pass
> under simulator.
>
>
> gcc/
> * config/i386/i386.c
> (ix86_expand_vector_set): Handle V8DF, V8DI, V16SF, V16SI, V32HI, V64QI
On Thu, Oct 16, 2014 at 07:37:18AM +0200, Marc Glisse wrote:
> Hello,
>
> the attached one-liner passed bootstrap+testsuite (really all languages) on
> x86_64-linux-gnu (I got an extra pass of unix/-m32: os but I assume that the
> failure with trunk was random).
>
> The current code is a bit weir
On Wed, Oct 15, 2014 at 12:08 PM, Marek Polacek wrote:
> On Tue, Oct 14, 2014 at 09:23:29AM +0200, Marek Polacek wrote:
>> The consensus seems to be to go forward with this change. I will
>> commit the patch in 24 hours unless I hear objections.
>
> I made the change. Please report any fallout t
On Thu, Oct 16, 2014 at 12:07 AM, DJ Delorie wrote:
>
> In the event that pointer sizes aren't powers of two, choose a more
> suitable alignment than (unsigned)(-1), which results in HUGE file
> sizes. Ok?
Ok.
Thanks,
Richard.
> Index: gcc/lto/lto-object.c
> ===
... the below completes the work, no -std=gnu++0x anymore.
Paolo.
/
2014-10-16 Paolo Carlini
* testsuite/lib/libstdc++.exp: Prefer -std=gnu++11.
* testsuite/20_util/add_lvalue_reference/requirements/
explicit_instantiation.cc: Likewise.
* t
On Thu, Oct 16, 2014 at 9:20 AM, Tom de Vries wrote:
> On 08/10/12 11:24, Richard Guenther wrote:
>> On Sun, Oct 7, 2012 at 12:44 PM, Tom de Vries wrote:
>>> Richard,
>>>
>>> attached patch checks that unlinked uses do not contain ssa-names when
>>> renaming.
>>>
>>> This assert triggers when co
On Thu, Oct 16, 2014 at 9:31 AM, Jakub Jelinek wrote:
> On Thu, Oct 16, 2014 at 07:37:18AM +0200, Marc Glisse wrote:
>> Hello,
>>
>> the attached one-liner passed bootstrap+testsuite (really all languages) on
>> x86_64-linux-gnu (I got an extra pass of unix/-m32: os but I assume that the
>> failur
Hi all,
As discussed in https://gcc.gnu.org/ml/gcc/2014-09/msg00234.html , this
patchset backports mainline patches necessary for Kernel ASan in GCC 4.9
(gcc-4_9-branch). The patcheset consists of
* Asan headers installation (1 patch)
* __asan_loadN/__asan_storeN support (3 patches)
* instrume
This patch adds support for asan_interface.h installation (it's required
by tests in other patches).
One change compared to mainline: asan_interface.h in 4.9 isn't C
friendly (it uses bool type) so I had to replace bool with unsigned char
in tests. The actual value isn't used by the test so I
Further work on __asan_loadN/__asan_storeN. I removed the tests
(misalign-1.c, misalign-2.c) because (as mentioned in comments for
preceeding patch) __asan_loadN/__asan_storeN are disabled for userspace.
2014-10-15 Yury Gribov
Backport from mainline
2014-05-30 Jakub Jelinek
* asan.c (
The following is a patch by Jakub that ought to fix gnu11 fallout
on i686.
Applying to trunk.
2014-10-16 Jakub Jelinek
Marek Polacek
* gcc.dg/20020122-2.c: Use dg-additional-options. Fix implicit
declarations.
* gcc.dg/pr32176.c: Likewise.
* gcc.
I disabled __asan_loadN/__asan_storeN because 4.9's Asan runtime doesn't
support them. In a later patch I re-enable these functions specifically
for KAsan.
2014-10-15 Yury Gribov
Backport from mainline
2014-05-30 Jakub Jelinek
* sanitizer.def (BUILT_IN_ASAN_REPORT_LOAD_N,
BUILT_IN_ASA
4.9's Asan runtime library provides no support for these so I removed
the tests. I've also changed default value of threshold to INT_MAX to
completely disable outline instrumentation in userspace Asan.
New asan-instrumentation-with-call-threshold parameter.
2014-10-15 Yury Gribov
Backport f
Same as mainline.
2014-10-15 Yury Gribov
Backport from mainline
2014-06-16 Yury Gribov
* asan.c (build_check_stmt): Fix maybe-uninitialized warning.
diff --git a/gcc/asan.c b/gcc/asan.c
index 5c091d0..3729178 100644
--- a/gcc/asan.c
+++ b/gcc/asan.c
@@ -1635,6 +1635,13 @@ build_check_s
Difference from mainline: replaced non-C-friendly bool with unsigned
char (see patch 0001 for explanation).
2014-10-15 Yury Gribov
Backport from mainline
2014-06-18 Yury Gribov
PR sanitizer/61547
* asan.c (instrument_strlen_call): Fixed instrumentation of
trailing byte.
* c-c++-com
Same as mainline.
2014-10-15 Yury Gribov
Backport from mainline
2014-06-18 Yury Gribov
PR sanitizer/61530
* asan.c (build_check_stmt): Add condition.
* c-c++-common/asan/pr61530.c: New test.
diff --git a/gcc/asan.c b/gcc/asan.c
index 3729178..c838423 100644
--- a/gcc/asan.c
+++ b/g
Difference from mainline: replaced non-C-friendly bool with unsigned
char (see patch 0001 for explanation).
Do not instrument first byte in strlen if already instrumented.
2014-10-15 Yury Gribov
Backport from mainline
2014-06-24 Max Ostapenko
* asan.c (instrument_strlen_call): Do not i
The patch was slightly updated to take care of missing UBSan work
(SANITIZE_FLOAT_DIVIDE, SANITIZE_FLOAT_CAST, SANITIZE_BOUNDS).
2014-10-15 Yury Gribov
Backport from mainline
2014-07-31 Yury Gribov
* doc/cpp.texi (__SANITIZE_ADDRESS__): Updated description.
* doc/invoke.texi (-fsanitiz
The patch was slightly updated to take care of missing UBSan work
(UBSAN_BOUNDS).
Move inlining of Asan memory checks to sanopt pass.
Change asan-instrumentation-with-call-threshold to more closely match LLVM.
2014-10-15 Yury Gribov
Backport from mainline
2014-08-11 Yury Gribov
* asan.
The patch was slightly updated to take care of missing UBSan work
(UBSAN_BOUNDS).
Added fnspec to internal functions.
2014-10-15 Yury Gribov
Backport from mainline
2014-08-11 Yury Gribov
* gimple.c (gimple_call_fnspec): Support internal functions.
(gimple_call_return_fla
Same as mainline.
2014-10-15 Yury Gribov
Backport from mainline
2014-08-12 Yury Gribov
* internal-fn.c (init_internal_fns): Fix off-by-one.
diff --git a/gcc/internal-fn.c b/gcc/internal-fn.c
index 0749dce..d64e20d 100644
--- a/gcc/internal-fn.c
+++ b/gcc/internal-fn.c
@@ -62,7 +62,7 @@
On Thu, Oct 16, 2014 at 12:35:51PM +0400, Yury Gribov wrote:
> 2014-10-15 Yury Gribov
>
> Backport from mainline
> 2014-05-14 Yury Gribov
>
> PR sanitizer/61100
>
> * Makefile.am (nodist_saninclude_HEADERS): Install
> public headers.
> * Makefile.in: Reg
Difference from mainline: replaced non-C-friendly bool with unsigned
char (see patch 0001 for explanation).
2014-10-15 Yury Gribov
Backport from mainline
2014-08-18 Yury Gribov
PR sanitizer/62089
* asan.c (instrument_derefs): Fix bitfield check.
* c-c++-common/asan/pr62089.c: New te
Same as mainline.
2014-10-15 Yury Gribov
Backport from mainline
2014-08-28 Yury Gribov
* c-c++-common/asan/pr62089.c: Fix test on 32-bit platforms.
diff --git a/gcc/testsuite/c-c++-common/asan/pr62089.c b/gcc/testsuite/c-c++-common/asan/pr62089.c
index 6c25ea4..48167b1 100644
--- a/gcc
Same as mainline.
2014-10-15 Yury Gribov
Backport from mainline
2014-09-01 Yury Gribov
PR sanitizer/61897
PR sanitizer/62140
* asan.c (asan_mem_ref_get_end): Handle non-ptroff_t lengths.
(build_check_stmt): Likewise.
(instrument_strlen_call): Likewise.
(asan_expand_check_ifn): Lik
Same as mainline.
2014-10-15 Yury Gribov
Backport from mainline
2014-10-03 Yury Gribov
* asan.c (asan_finish_file): Disable __asan_init calls for KASan;
don't emit empty ctors.
diff --git a/gcc/asan.c b/gcc/asan.c
index 5c6d1c5..b8fbed3 100644
--- a/gcc/asan.c
+++ b/gcc/asan.c
@@ -238
Not a backport, specifically enables said instructions in KASan.
Instrument unaligned objects in KASan.
2014-10-15 Yury Gribov
* asan.c (instrument_derefs): Enable unaligned path for KASan.
diff --git a/gcc/asan.c b/gcc/asan.c
index b8fbed3..db4e3a0 100644
--- a/gcc/asan.c
+++ b/gcc/asan.c
On Thu, Oct 16, 2014 at 12:36:15PM +0400, Yury Gribov wrote:
> I disabled __asan_loadN/__asan_storeN because 4.9's Asan runtime doesn't
> support them. In a later patch I re-enable these functions specifically for
> KAsan.
> 2014-10-15 Yury Gribov
>
> Backport from mainline
> 2014-
On Thu, Oct 16, 2014 at 12:36:37PM +0400, Yury Gribov wrote:
> Further work on __asan_loadN/__asan_storeN. I removed the tests
> (misalign-1.c, misalign-2.c) because (as mentioned in comments for
> preceeding patch) __asan_loadN/__asan_storeN are disabled for userspace.
> 2014-10-15 Yury Gribov
On Thu, Oct 16, 2014 at 12:37:37PM +0400, Yury Gribov wrote:
> 4.9's Asan runtime library provides no support for these so I removed the
> tests. I've also changed default value of threshold to INT_MAX to completely
> disable outline instrumentation in userspace Asan.
> New asan-instrumentation-wi
On Thu, Oct 16, 2014 at 12:38:00PM +0400, Yury Gribov wrote:
> Same as mainline.
> 2014-10-15 Yury Gribov
>
> Backport from mainline
> 2014-06-16 Yury Gribov
>
> * asan.c (build_check_stmt): Fix maybe-uninitialized warning.
Ok.
Jakub
On Thu, Oct 16, 2014 at 12:38:19PM +0400, Yury Gribov wrote:
> Same as mainline.
> 2014-10-15 Yury Gribov
>
> Backport from mainline
> 2014-06-18 Yury Gribov
>
> PR sanitizer/61530
>
> * asan.c (build_check_stmt): Add condition.
>
> * c-c++-common/asan/pr615
On Thu, Oct 16, 2014 at 12:38:51PM +0400, Yury Gribov wrote:
> Difference from mainline: replaced non-C-friendly bool with unsigned char
> (see patch 0001 for explanation).
> 2014-10-15 Yury Gribov
>
> Backport from mainline
> 2014-06-18 Yury Gribov
>
> PR sanitizer/61547
On Thu, Oct 16, 2014 at 12:39:17PM +0400, Yury Gribov wrote:
> Difference from mainline: replaced non-C-friendly bool with unsigned char
> (see patch 0001 for explanation).
Where? I don't see unsigned char being used anywhere in the patch.
bool is certainly fine in GCC itself.
> Do not instrumen
On Thu, Oct 16, 2014 at 12:39:56PM +0400, Yury Gribov wrote:
> The patch was slightly updated to take care of missing UBSan work
> (SANITIZE_FLOAT_DIVIDE, SANITIZE_FLOAT_CAST, SANITIZE_BOUNDS).
> 2014-10-15 Yury Gribov
>
> Backport from mainline
> 2014-07-31 Yury Gribov
>
>
On Thu, Oct 16, 2014 at 12:41:03PM +0400, Yury Gribov wrote:
> Same as mainline.
> 2014-10-15 Yury Gribov
>
> Backport from mainline
> 2014-08-12 Yury Gribov
>
> * internal-fn.c (init_internal_fns): Fix off-by-one.
Ok.
Jakub
On Thu, Oct 16, 2014 at 12:40:17PM +0400, Yury Gribov wrote:
> The patch was slightly updated to take care of missing UBSan work
> (UBSAN_BOUNDS).
> Added fnspec to internal functions.
>
> 2014-10-15 Yury Gribov
>
> Backport from mainline
> 2014-08-11 Yury Gribov
>
> *
On Thu, Oct 16, 2014 at 12:40:44PM +0400, Yury Gribov wrote:
> The patch was slightly updated to take care of missing UBSan work
> (UBSAN_BOUNDS).
> Move inlining of Asan memory checks to sanopt pass.
> Change asan-instrumentation-with-call-threshold to more closely match LLVM.
>
> 2014-10-15 Yu
On Thu, Oct 16, 2014 at 12:41:46PM +0400, Yury Gribov wrote:
> Same as mainline.
> 2014-10-15 Yury Gribov
>
> Backport from mainline
> 2014-08-28 Yury Gribov
>
> * c-c++-common/asan/pr62089.c: Fix test on 32-bit platforms.
Ok.
Jakub
On Thu, Oct 16, 2014 at 12:41:24PM +0400, Yury Gribov wrote:
> Difference from mainline: replaced non-C-friendly bool with unsigned char
> (see patch 0001 for explanation).
> 2014-10-15 Yury Gribov
>
> Backport from mainline
> 2014-08-18 Yury Gribov
>
> PR sanitizer/62089
On Thu, Oct 16, 2014 at 12:42:28PM +0400, Yury Gribov wrote:
> Same as mainline.
> 2014-10-15 Yury Gribov
>
> Backport from mainline
> 2014-10-03 Yury Gribov
>
> * asan.c (asan_finish_file): Disable __asan_init calls for KASan;
> don't emit empty ctors.
Ok.
On Thu, Oct 16, 2014 at 12:42:09PM +0400, Yury Gribov wrote:
> Same as mainline.
> 2014-10-15 Yury Gribov
>
> Backport from mainline
> 2014-09-01 Yury Gribov
>
> PR sanitizer/61897
> PR sanitizer/62140
>
> * asan.c (asan_mem_ref_get_end): Handle non-ptroff_t
On Thu, Oct 16, 2014 at 12:43:25PM +0400, Yury Gribov wrote:
> Not a backport, specifically enables said instructions in KASan.
> Instrument unaligned objects in KASan.
>
> 2014-10-15 Yury Gribov
>
> * asan.c (instrument_derefs): Enable unaligned path for KASan.
>
> diff --git a/gcc/as
On Thu, 16 Oct 2014, Richard Biener wrote:
Does this fix PR63537?
PR63537 is already fine for me with trunk, NRV replaces ret with retval
everywhere. It does so even if I add f(&ret); in the function with void
f(vec*);
I'd worry if both result and found are address taken before the pass,
On 10/16/2014 12:49 PM, Jakub Jelinek wrote:
On Thu, Oct 16, 2014 at 12:39:17PM +0400, Yury Gribov wrote:
Difference from mainline: replaced non-C-friendly bool with unsigned char
(see patch 0001 for explanation).
Where? I don't see unsigned char being used anywhere in the patch.
bool is cert
This fixes one test in the Ada testsuite (there may be more of these).
Applying to trunk.
2014-10-16 Marek Polacek
* ada/acats/tests/cd/cd300051.c: Fix defaulting to int.
diff --git gcc/testsuite/ada/acats/tests/cd/cd300051.c
gcc/testsuite/ada/acats/tests/cd/cd300051.c
index 5771fc8
Another one (ok, I should really sit down and move all the remaining
string related foldings).
Bootstrap / regtest running on x86_64-unknown-linux-gnu.
Richard.
2014-10-16 Richard Biener
PR middle-end/63554
* builtins.c (fold_builtin_4): Do not call fold_builtin_strncat_chk.
Uros Bizjak writes:
> Hello!
>
> Now that %ebx is no more fixed, we can remove all PIC related
> complications in atomic_compare_and_swap_doubleword pattern. The
> immediate consequence is, that we avoid "hidden" xchgs that clobbered
> unwinding state.
Could also do the same in cpuid.h now
-And
We can fail to copy loop headers if the loop looks proper to us
(empty latch) even when the latch really isn't empty and just
contains a PHI node. In this case cfgcleanup should have merged
it with the previous block but it fails to do that because of
an overly strict check. Fixed by only "prese
Eric,
this patch is the first half of the fix for PR61605.
The problem it addresses is the following: Consider this copy cycle (a = b; b =
a):
...
(insn 2 18 3 2 (set (reg/v:SI 1 dx [orig:86 yD.1749 ] [86])
(reg:SI 5 di [ yD.1749 ])) test.c:9 90 {*movsi_internal}
(expr_list:REG_DEA
Eric,
this patch is the second half of the fix for PR61605.
It make sure in pass_cprop_hardreg that, if available we use the call-specific
information provided by fuse-caller-save, rather than the generic
regs_invalidated_by_call info.
The 2 patches together allow an insn to removed from the
On Thu, Oct 16, 2014 at 11:06 AM, Andi Kleen wrote:
>> Now that %ebx is no more fixed, we can remove all PIC related
>> complications in atomic_compare_and_swap_doubleword pattern. The
>> immediate consequence is, that we avoid "hidden" xchgs that clobbered
>> unwinding state.
>
> Could also do t
Hello!
Now that %ebx is also allocatable in PIC modes, we can cleanup
config/i386/cpuid considerably. I propose to remove all PIC related
specializations of __cpuid and __cpuid_count and protect the
compilation with "#if __GNUC__ >= 5".
The only drawback would be that non-bootstrapped build with
On Thu, Oct 16, 2014 at 11:19:49AM +0200, Uros Bizjak wrote:
> Hello!
>
> Now that %ebx is also allocatable in PIC modes, we can cleanup
> config/i386/cpuid considerably. I propose to remove all PIC related
> specializations of __cpuid and __cpuid_count and protect the
> compilation with "#if __GN
On Thu, Oct 16, 2014 at 11:03 AM, Marc Glisse wrote:
> On Thu, 16 Oct 2014, Richard Biener wrote:
>
>> Does this fix PR63537?
>
>
> PR63537 is already fine for me with trunk, NRV replaces ret with retval
> everywhere. It does so even if I add f(&ret); in the function with void
> f(vec*);
>
>>> I'd
On Thu, Oct 16, 2014 at 11:24 AM, Jakub Jelinek wrote:
>> Now that %ebx is also allocatable in PIC modes, we can cleanup
>> config/i386/cpuid considerably. I propose to remove all PIC related
>> specializations of __cpuid and __cpuid_count and protect the
>> compilation with "#if __GNUC__ >= 5".
On Thu, Oct 16, 2014 at 11:28:20AM +0200, Uros Bizjak wrote:
> On Thu, Oct 16, 2014 at 11:24 AM, Jakub Jelinek wrote:
>
> >> Now that %ebx is also allocatable in PIC modes, we can cleanup
> >> config/i386/cpuid considerably. I propose to remove all PIC related
> >> specializations of __cpuid and
On 15/10/14 22:06 +0200, François Dumont wrote:
On 15/10/2014 13:10, Jonathan Wakely wrote:
I find this much easier to read:
#if __cplusplus < 201103L
typedef _Is_contiguous_sequence<_Sequence> __tag;
#else
using __lvalref = std::is_lvalue_reference<
typename std::iterator_traits
Tested on ia64-suse-linux, installed as obvious.
Andreas.
* gcc.target/ia64/20090324-1.c: Fix implicit declarations and
implicit int.
* gcc.target/ia64/mfused-madd-vect.c: Likewise.
* gcc.target/ia64/mno-fused-madd-vect.c: Likewise.
* gcc.target/ia64/pr2968
Tested on m68k-suse-linux and installed as obvious.
Andreas.
* gcc.target/m68k/crash1.c: Fix implicit declaration.
diff --git a/gcc/testsuite/gcc.target/m68k/crash1.c
b/gcc/testsuite/gcc.target/m68k/crash1.c
index fdd737a..2554d62 100644
--- a/gcc/testsuite/gcc.target/m68k/crash1.c
+++
On 09.10.2014 16:42, Kai Tietz wrote:
>> Not approved:
>> * gcc/cp/vtable-class-hierarchy.c
>
> Index: gcc/cp/vtable-class-hierarchy.c
> ===
> --- gcc/cp/vtable-class-hierarchy.c(Revision 214408)
> +++ gcc/cp/vtable-class-hierarch
On 10 Oct 18:37, Uros Bizjak wrote:
> On Fri, Oct 10, 2014 at 5:47 PM, Ilya Tocar wrote:
>
>
> Please recode that horrible first switch statement to:
>
> --cut here--
> rtx (*gen) (rtx, rtx, rtx, rtx) = NULL;
>
> switch (mode)
> {
> case V8HImode:
> if (TARGET_AVX512VL && TAR
On Thu, Oct 16, 2014 at 11:36 AM, Jakub Jelinek wrote:
>> >> Now that %ebx is also allocatable in PIC modes, we can cleanup
>> >> config/i386/cpuid considerably. I propose to remove all PIC related
>> >> specializations of __cpuid and __cpuid_count and protect the
>> >> compilation with "#if __GN
FAIL: g++.dg/init/enum1.C -std=gnu++11 (test for errors, line 12)
FAIL: g++.dg/init/enum1.C -std=gnu++1y (test for errors, line 12)
FAIL: g++.dg/init/enum1.C -std=gnu++98 (test for errors, line 12)
That used to complain about "enum1.C:12:1: error: no integral type can
represent all of the en
Hi,
As discussed in the PR, the __fpscr_values based FPU mode switching is
replaced by other means, which preserve the other FPU mode and status
bits across mode switches. For details please refer to the coments in
the PR.
Tested on sh-sim with -m4 -ml and -m4 -mb without new failures, except
fo
This patch (also applicable to trunk) makes us canoncialize operand
order for comparisons at the same time we canonicalize other
operand order, in particular before dispatching to generic_simplify.
It also adds operand canonicalization to ternary ops and adds
FMA_EXPR and DOT_PROD_EXPR to the list
On Thu, Oct 16, 2014 at 02:23:16PM +0400, Ilya Tocar wrote:
> On 10 Oct 18:37, Uros Bizjak wrote:
> > On Fri, Oct 10, 2014 at 5:47 PM, Ilya Tocar wrote:
> >
> >
> > Please recode that horrible first switch statement to:
> >
> > --cut here--
> > rtx (*gen) (rtx, rtx, rtx, rtx) = NULL;
> >
> >
On 15 Oct 16:38, Jakub Jelinek wrote:
> > Done. But it turned out that the gcc_GAS_CHECK_FEATURE from
> > gcc/configure.ac:
> >
> > gcc_GAS_CHECK_FEATURE([.section with e], gcc_cv_as_section_has_e,
> > [2,22,51],,
> > [.section foo1,"e"
> > .byte 0,0,0,0])
> >
> > does not work properly.
On Thu, Oct 16, 2014 at 8:28 AM, Kirill Yukhin wrote:
> Hello,
> This patch extends expand_mul_widen_hilo to 512-bit QI,SI,HI modes.
>
> Bootstrapped and regtested
>
> gcc/
> * config/i386/i386.c
> (ix86_expand_mul_widen_hilo): Handle V32HI, V16SI, V64QI modes.
>
> Is it ok for tru
On Thu, Oct 16, 2014 at 9:28 AM, Jakub Jelinek wrote:
> On Thu, Oct 16, 2014 at 10:24:45AM +0400, Kirill Yukhin wrote:
>> Hello,
>> This patch extends insertion hook.
>>
>> AVX-512* tests on top of patch-set all pass
>> under simulator.
>>
>>
>> gcc/
>> * config/i386/i386.c
>> (ix86_ex
On Wed, Oct 15, 2014 at 6:08 PM, Jeff Law wrote:
> On 10/15/14 08:35, Marc Glisse wrote:
>>>
>>>
>>> Would that extra pass be acceptable?
Ugh, rather not. We have too many passes ;)
>>> Otherwise, what do you think should be responsible for cleaning up the
>>> dead assignments?
>>
>>
>> Does an
On Thu, Oct 16, 2014 at 03:17:36PM +0400, Ilya Verbin wrote:
> On 15 Oct 16:38, Jakub Jelinek wrote:
> > > Done. But it turned out that the gcc_GAS_CHECK_FEATURE from
> > > gcc/configure.ac:
> > >
> > > gcc_GAS_CHECK_FEATURE([.section with e], gcc_cv_as_section_has_e,
> > > [2,22,51],,
> > >
On Wed, Oct 15, 2014 at 6:15 PM, David Malcolm wrote:
> Back in March I posted an 89-patch kit to expand and make use of the
> gimple statement class hierarchy to move much of the type-checking of
> statement accessors to be at compile-time rather than run-time:
> https://gcc.gnu.org/ml/gcc-pat
On Wed, Oct 15, 2014 at 6:26 PM, Martin Liška wrote:
> Hello.
>
> Following patch introduces a new class called callgraph_annotation. Idea
> behind the patch is to provide a generic interface one can use to register
> custom info related to a cgraph_node. As you know, symbol_table provides
> hooks
On Wed, Oct 15, 2014 at 4:47 PM, Ilya Palachev wrote:
> Hi all,
>
> The attached patch fixes PR lto/61048.
>
> The basic idea is to write option -fsanitize to existing ELF section
> .gnu.lto_.opts in object files and then read it in lto-wrapper.
>
> On 15.10.2014 12:46, Richard Biener wrote:
>>
>>
>
> I don't like "generic annotation" facilities at all. Would it be possible
Why?
> to make cgraph UIDs not sparse? (keep a free-list of cgraph nodes
cgraph nodes are already kept "dense" via freelist. However in WPA you usually
have a lot
of different nodes prior merging and unreachable c
On 10/16/2014 01:31 PM, Richard Biener wrote:
On Wed, Oct 15, 2014 at 6:26 PM, Martin Liška wrote:
Hello.
Following patch introduces a new class called callgraph_annotation. Idea
behind the patch is to provide a generic interface one can use to register
custom info related to a cgraph_node. As
On Thu, Oct 16, 2014 at 1:40 PM, Jan Hubicka wrote:
>>
>> I don't like "generic annotation" facilities at all. Would it be possible
>
> Why?
Because it's the way to hell if the IL has "magic" things only one
pass can understand. It can't ever know if it may invalidate some
of that data.
Same r
On Thu, Oct 16, 2014 at 1:42 PM, Martin Liška wrote:
> On 10/16/2014 01:31 PM, Richard Biener wrote:
>>
>> On Wed, Oct 15, 2014 at 6:26 PM, Martin Liška wrote:
>>>
>>> Hello.
>>>
>>> Following patch introduces a new class called callgraph_annotation. Idea
>>> behind the patch is to provide a gene
This fixes fallout of graphite tests.
Applying to trunk.
2014-10-16 Marek Polacek
* gcc.dg/graphite/id-1.c: Fix defaulting to int.
* gcc.dg/graphite/id-2.c: Likewise.
* gcc.dg/graphite/id-4.c: Likewise.
* gcc.dg/graphite/id-8.c: Likewise.
* gcc.dg/graph
> On Thu, Oct 16, 2014 at 1:40 PM, Jan Hubicka wrote:
> >>
> >> I don't like "generic annotation" facilities at all. Would it be possible
> >
> > Why?
>
> Because it's the way to hell if the IL has "magic" things only one
> pass can understand. It can't ever know if it may invalidate some
> of
Hi All,
Here is a simple enhancement for predicate computation in if-convert phase:
We use notion of cd equivalence to get simpler predicate for
join block, e.g. if join block has 2 predecessors with predicates
p1 & p2 and p1 & !p2, we'd like to get p1 for it instead of
p1 & p2 |
Hello,
This patch extends expand_sse2_mulvxdi3.
Bootstrapped.
AVX-512* tests on top of patch-set all pass
under simulator.
Is it ok for trunk?
gcc/
* config/i386/i386.c (ix86_expand_sse2_mulvxdi3): Extend
expand_sse2_mulvxdi3.
--
Thanks, K
diff --git a/gcc/config/i386/i386.c b/
Markus reported on IRC that gnu11 default broke these two tests
on powerpc.
Applying to trunk.
2014-10-16 Marek Polacek
* gcc.target/powerpc/pr58673-1.c: Fix defaulting to int.
* gcc.target/powerpc/pr58673-2.c: Add declarations.
diff --git gcc/testsuite/gcc.target/powerpc/pr5
> > Hello.
> >
> > If I recall correctly, we recycle cgraph_nodes and it's possible that an UID
> > is given to different nodes:
> > symbol_table::allocate_cgraph_symbol (void). Such uid is problematic from
> > perspective that it cannot be used as a index to a vector.
> >
> > It was also Honza's n
On 10/16/2014 02:01 PM, Jan Hubicka wrote:
Hello.
If I recall correctly, we recycle cgraph_nodes and it's possible that an UID
is given to different nodes:
symbol_table::allocate_cgraph_symbol (void). Such uid is problematic from
perspective that it cannot be used as a index to a vector.
It was
On 10/16/2014 02:05 PM, Martin Liška wrote:
On 10/16/2014 02:01 PM, Jan Hubicka wrote:
Hello.
If I recall correctly, we recycle cgraph_nodes and it's possible that an UID
is given to different nodes:
symbol_table::allocate_cgraph_symbol (void). Such uid is problematic from
perspective that it c
On 16-10-14 10:14, Richard Biener wrote:
On Thu, Oct 16, 2014 at 9:20 AM, Tom de Vries wrote:
On 08/10/12 11:24, Richard Guenther wrote:
On Sun, Oct 7, 2012 at 12:44 PM, Tom de Vries wrote:
Richard,
attached patch checks that unlinked uses do not contain ssa-names when renaming.
This asser
On Thu, Oct 16, 2014 at 2:11 PM, Tom de Vries wrote:
> On 16-10-14 10:14, Richard Biener wrote:
>>
>> On Thu, Oct 16, 2014 at 9:20 AM, Tom de Vries
>> wrote:
>>>
>>> On 08/10/12 11:24, Richard Guenther wrote:
On Sun, Oct 7, 2012 at 12:44 PM, Tom de Vries
wrote:
>
> Richard
Hi,
This is not entirely related to PR 59401, but the discussion came up
there. The attached patch makes the GBR on SH call preserved by
default. Tested by Kaz on sh4-linux, committed as r216314.
Cheers,
Oleg
gcc/ChangeLog:
PR target/59401
* config/sh/sh.h (CALL_REALLY_USED_REG
On Thu, Oct 16, 2014 at 1:55 PM, Kirill Yukhin wrote:
> Hello,
> This patch extends expand_sse2_mulvxdi3.
>
> Bootstrapped.
> AVX-512* tests on top of patch-set all pass
> under simulator.
>
> Is it ok for trunk?
>
> gcc/
> * config/i386/i386.c (ix86_expand_sse2_mulvxdi3): Extend
>
Hi,
I've committed the attached obvious patch as r216317.
Cheers,
Oleg
gcc/testsuite/ChangeLog:
PR target/63260
* gcc.target/sh/pr63260.c: Fix typo __builtin_fabs vs. __builtin_abs.
Index: gcc/testsuite/gcc.target/sh/pr63260.c
=
(I've taken Alan's patch and added two new testcases.)
This patch cures the linux kernel boot failure when compiled using
trunk gcc.
At its heart, the problem is caused by merge_decls merging from the
old decl to the new decl, then copying back to the old decl and
discarding the new. When Jan m
Hello.
Following patch fixes i586 bootstrap failure:
https://gcc.gnu.org/ml/gcc-testresults/2014-10/msg01713.html
../../src-trunk/gcc/ipa-icf.c:2096:23: error: format ‘%lu’ expects argument of type
‘long unsigned int’, but argument 3 has type
‘std::list::size_type {aka unsigned int}’
[-Werro
> Hello.
>
> Following patch fixes i586 bootstrap failure:
> https://gcc.gnu.org/ml/gcc-testresults/2014-10/msg01713.html
>
> ../../src-trunk/gcc/ipa-icf.c:2096:23: error: format ‘%lu’ expects argument
> of type ‘long unsigned int’, but argument 3 has type
> ‘std::list::size_type {aka unsigned
On Thu, Oct 16, 2014 at 02:41:36PM +0200, Martin Liška wrote:
> Hello.
>
> Following patch fixes i586 bootstrap failure:
> https://gcc.gnu.org/ml/gcc-testresults/2014-10/msg01713.html
>
> ../../src-trunk/gcc/ipa-icf.c:2096:23: error: format ‘%lu’ expects argument
> of type ‘long unsigned int’,
On 10/16/2014 02:45 PM, Jakub Jelinek wrote:
On Thu, Oct 16, 2014 at 02:41:36PM +0200, Martin Liška wrote:
Hello.
Following patch fixes i586 bootstrap failure:
https://gcc.gnu.org/ml/gcc-testresults/2014-10/msg01713.html
../../src-trunk/gcc/ipa-icf.c:2096:23: error: format ‘%lu’ expects argum
Hi,
The fldi_ok function in sh.c has been returning 'true' for while. Thus
it can as well be removed. Tested with 'make all-gcc', committed as
r216320.
Cheers,
Oleg
gcc/ChangeLog:
* config/sh/sh-protos.h (fldi_ok): Remove.
* config/sh/sh.c (fldi_ok): Likewise.
(sh_secon
1 - 100 of 194 matches
Mail list logo
