Hi,
libcilkrts is compiled with -nostdlib, that means we have to
explicitly specify the pthread library we should link with (e.g. we
don't have such problem with libgomp, because it is C). And, indeed,
-lpthread is hard-coded in the Makefile for cilkrts. For Android
this doesn't work, because
Those waw dependences seem to be correct. Should even the previous analysis
only mark the j-loop as parallel?
The previous and the current analysis mark the j-loop as
nonparallelizable. (Possibly, I don't fully understand the question.
Could you please reformulate it?)
--
Hi,
For some TARGET, like ARM THUMB1, the offset in load/store should be nature
aligned. But in function get_address_cost, when computing max_offset, it
only tries byte-aligned offsets:
((unsigned HOST_WIDE_INT) 1 i) - 1
which can not meet thumb_legitimate_offset_p check called from
On 08/01/2014 07:53 PM, Jakub Jelinek wrote:
I think we should use David Malcolm's approach i.e. add some --report-bug
flag
to driver. This could be enabled by default at configure time via
--with-spec.
-freport-bug or whatever we call it should not be, at least if it attempts
to communicate
On 04/08/2014 08:09, Roman Gareev wrote:
Those waw dependences seem to be correct. Should even the previous analysis
only mark the j-loop as parallel?
The previous and the current analysis mark the j-loop as
nonparallelizable. (Possibly, I don't fully understand the question.
Could you please
Thanks! My 2 (actually 4) cents below.
+if ($#ARGV == 1 ($ARGV[0] eq -i || $ARGV[0] eq --inline)) {
+ $diff = $ARGV[1];
Can we shift here and then just set $diff to $ARGV[0] unconditionally?
+ if ($diff eq -) {
+ die Reading from - and using -i are not compatible;
+ }
Hm,
LGTM.
Cheers,
Tobias
This patch modifies the implementation of AI05-0071 to allow several special
cases of equality to appear in instantiations where a formal type has unknown
discriminants, a defaulted equality and the actual type is class-wide.
-- Source --
-- equals_gen.ads
generic
This patch modifies the formal subprogram instantiation machinery to preserve
the source locations of all formal parameters when creating the corresponding
renaming declaration. No reproducer as this requires ASIS.
Tested on x86_64-pc-linux-gnu, committed on trunk
2014-08-04 Hristian Kirtchev
This change removes some unnecessary range checks. The following test
compiled with -gnatdt -gnatc:
1. procedure RCInteger (X : Integer) is
2.type YT is new Integer;
3.Y : constant YT := YT(X) * YT(X);
4. begin
5.null;
6. end;
generates a tree file that
A postcondition whose expression is a short-circuit is broken down into
individual aspects in order to provide better exception reporting. This
transformation is performed syntactically, before any analysis. The original
short-circuit expression is rewritten as its second operand, and an
This change makes sure that the Do_Range_Check flag is set in -gnatc or
GNATprove mode for type conversions from real to integer. This makes sure
that SPARK2014 programs properly verify that such conversions cannot raise
an exception due to an out of range value. The following test compiled
with
On 04-08-14 08:45, Yury Gribov wrote:
Thanks! My 2 (actually 4) cents below.
Hi Yuri,
thanks for the review.
+if ($#ARGV == 1 ($ARGV[0] eq -i || $ARGV[0] eq --inline)) {
+$diff = $ARGV[1];
Can we shift here and then just set $diff to $ARGV[0] unconditionally?
Done.
+if
On Aug 2, 2014, at 9:26 AM, Tom de Vries tom_devr...@mentor.com wrote:
I've made it an xfail, and added documentation in attached follow-up patch.
OK?
Ok.
Or do we go with the removal suggestion of Mike?
I’ll let others weigh in. I’m fine either way. Original author likely prefers
the
On Sat, Aug 2, 2014 at 8:36 PM, Prathamesh Kulkarni
bilbotheelffri...@gmail.com wrote:
* genmatch.c (dt_simplify::gen_gimple): Use dt_simplify::capture_max.
(dt_simplify::gen_generic): Likewise.
Thanks,
applied.
Richard.
Thanks,
Prathamesh
On 17 July 2014 11:10, Jeff Law l...@redhat.com wrote:
On 05/22/14 03:52, Zhenqiang Chen wrote:
On 21 May 2014 20:43, Steven Bosscher stevenb@gmail.com wrote:
On Wed, May 21, 2014 at 11:58 AM, Zhenqiang Chen wrote:
Hi,
The patch fixes the gcc.target/i386/pr49095.c FAIL in PR61225. The
On Mon, Aug 4, 2014 at 2:28 PM, Zhenqiang Chen zhenqiang.c...@arm.com wrote:
Hi,
For some TARGET, like ARM THUMB1, the offset in load/store should be nature
aligned. But in function get_address_cost, when computing max_offset, it
only tries byte-aligned offsets:
((unsigned HOST_WIDE_INT)
-Original Message-
From: Bin.Cheng [mailto:amker.ch...@gmail.com]
Sent: Monday, August 04, 2014 4:41 PM
To: Zhenqiang Chen
Cc: gcc-patches List
Subject: Re: [PATCH, ivopt] Try aligned offset when get_address_cost
On Mon, Aug 4, 2014 at 2:28 PM, Zhenqiang Chen
On 03/08/14 21:05 -0400, Samuel Bronson wrote:
Since Jonathan Wakely independantly committed some of the changes I was
preparing, I had to backport two commits, r210625 and r212453
(git commits 3a30bda and b0a6074).
2014-08-02 Samuel Bronson naes...@gmail.com
Backport r212453 from
On 25 July 2014 20:46, Siva Chandra wrote:
The attached patch is identical to v2 except that I rebased it over
the current head.
To recollect, GDB now supports xmethods in its Python API:
https://sourceware.org/gdb/current/onlinedocs/gdb/Xmethods-In-Python.html
This feature will be
Some of the project attributes have new defaults that are no longer the
empty string of the empty string list. These defaults are used as values
of attribute references when the attribute has not been declared.
Object_Dir has as default ..
Exec_Dir has as default Object_Dir, as previously
Some expression are only pre-analyzed in GNATprove mode, like default
expressions and assertions. These also need to be light-expanded for
formal verification. This is the second commit attempt, as the first one
failed to build GNAT due to a bug.
Tested on x86_64-pc-linux-gnu, committed on trunk
The compiler can generate a call to a type's invariant-checking procedure
on an uninitialized object, such as a temporary object created for holding
an aggregate. This is prevented by inhibiting the call when No_Initialization
is set on the object.
The following test must compile and execute
This patch implements aspect/pragma Default_Initial_Condition. The construct
has the following semantics and runtime behavior:
The Default_Initial_Condition aspect is introduced by an aspect_specification
where the aspect_mark is Default_Initial_Condition. The aspect may be specified
only as part
In -gnatc or GNATprove mode, there were cases in which the Do_Range_Check
flag was not set on an OUT parameter, causing e.g. GNAT prove to miss the
requirement for proving that an out parameter result was in range.
This is now corrected, the following test, compiled in -gnatd.F
(GNATprove mode)
This patch handles properly a function declared in package A that returns the
limited view of a type declared in package B, when the function is called from
a context that has with_clauses on A and B. Previous to this patch such a call
would crash the compiler because of a misplaced freeze node.
This patch removes the aarch64_simd_builtin_type enum. This is only really
used for indexing in to an array of strings and an array of machine_mode.
We don't need that. Given all the macro pasting we presently do in this
file, just add another. Then we can store a proper machine_mode for each
This PR is about bogus overflow warning that we issue for e.g.
int *q = i + 1;
q - (q - 1);
because pointer_diff receives p - (p + -1U) which gets simplified to
1U - with overflow. We could drop the overflow flag to suppress the
warning, but I think we should just remove the optimization
David Edelsohn wrote:
On Fri, Aug 1, 2014 at 2:03 PM, rohitarul...@freescale.com
rohitarul...@freescale.com wrote:
[libgcc]
2014-07-31 Rohit rohitarul...@freescale.com
* config/rs6000/linux-unwind.h (ppc_fallback_frame_state): Update
based on change in SPE high
On Mon, 4 Aug 2014, Marek Polacek wrote:
This PR is about bogus overflow warning that we issue for e.g.
int *q = i + 1;
q - (q - 1);
because pointer_diff receives p - (p + -1U) which gets simplified to
1U - with overflow. We could drop the overflow flag to suppress the
warning, but I
In Ada 2012, boolean aspects are True by default, but can specify explicitly
a value of False. In the case of an Import aspect, this indicates that the
corresponding entity must have a completion in Ada, and not in the language
specified in a Convention aspect.
Compiling main.adb must yield:
Documentation change only, no test required
Tested on x86_64-pc-linux-gnu, committed on trunk
2014-08-04 Robert Dewar de...@adacore.com
* gnat_rm.texi: Add section on aspect Invariant'Class.
Index: gnat_rm.texi
===
---
On Mon, 4 Aug 2014, Marek Polacek wrote:
+/* { dg-do compile } */
+
+void
+foo (void)
+{
+ volatile __PTRDIFF_TYPE__ t;
+ int i;
+ int *p = i;
+ int *q = i + 1;
+ t = q - (q - 1);
+ t = (q - 1) - q; /* { dg-warning integer overflow in expression } */
+ t = p - (p - 1);
+ t = (p - 1) - p
The prefix of an entry call may be a limited view, in which case the expansion
of the call must use the non-limited view, which is available at the point
of an entry call.
The following must compile quietly:
gcc -c railway-train.adb
---
package Railway is
end Railway;
---
limited with
On Mon, 4 Aug 2014, Marc Glisse wrote:
On Mon, 4 Aug 2014, Marek Polacek wrote:
+/* { dg-do compile } */
+
+void
+foo (void)
+{
+ volatile __PTRDIFF_TYPE__ t;
+ int i;
+ int *p = i;
+ int *q = i + 1;
+ t = q - (q - 1);
+ t = (q - 1) - q; /* { dg-warning integer
This patch allows the uses of aspect/pragma SPARK_Mode in generic units. It
also implements the following rule concerning the interplay between instances
and SPARK_Mode off:
However, if an instance of a generic unit is enclosed by code where
SPARK_Mode is Off and if any SPARK_Mode
Below patch adds PTA_RDRND and PTA_MOVBE for bdver4.
Bootstrap passes.
Ok for upstream?
Regards
Ganesh
Index: gcc/ChangeLog
===
--- gcc/ChangeLog (revision 213568)
+++ gcc/ChangeLog (working copy)
@@ -24,6 +24,11 @@
Hi,
This patch makes sure that visibility-option for hidden-inlines
doesn't show any effect for pe-coff as for those targets no concept of
hidden-visiblitly isn't present.
ChangeLog
2014-08-04 Kai Tietz kti...@redhat.com
* decl2.c (determine_hidden_inline): Always fail for PE-COFF as
Two small changes to gcc code to support newer obstacks:
1) gcc currently calls _obstack_begin, which requires some ugly casts
on alloc and free functions, and the casts will change when/if 64-bit
obstacks are available. It's cleaner to use the convenience functions
provided in obstack.h.
2)
Hi,
This patch makes sure that we check for flag keep-inline-dllexport
that we operate on true inlines as we do it for keep-inline-functions.
ChangeLog
2014-08-04 Kai Tietz kti...@redhat.com
* semantics.c (expand_or_defer_fn_1): Check for keep-inline-dllexport
that we operate on a
On Mon, Aug 04, 2014 at 12:26:01PM +0200, Richard Biener wrote:
I think that tree-ssa-forwprop.c already simplifies this in
associate_plusminus with (T)(P + A) - (T)P - (T)A. Well,
maybe not - but then the code should be massages to handle it.
Can you double-check and do that?
Otherwise
On Mon, Aug 4, 2014 at 12:58 PM, Gopalasubramanian, Ganesh
ganesh.gopalasubraman...@amd.com wrote:
Below patch adds PTA_RDRND and PTA_MOVBE for bdver4.
Bootstrap passes.
Ok for upstream?
2014-08-04 Ganesh Gopalasubramanian ganesh.gopalasubraman...@amd.com
+*
On Mon, 4 Aug 2014, Jakub Jelinek wrote:
On Mon, Aug 04, 2014 at 12:26:01PM +0200, Richard Biener wrote:
I think that tree-ssa-forwprop.c already simplifies this in
associate_plusminus with (T)(P + A) - (T)P - (T)A. Well,
maybe not - but then the code should be massages to handle it.
Uros Bizjak ubiz...@gmail.com writes:
On Mon, Aug 4, 2014 at 12:58 PM, Gopalasubramanian, Ganesh
ganesh.gopalasubraman...@amd.com wrote:
Below patch adds PTA_RDRND and PTA_MOVBE for bdver4.
Bootstrap passes.
Ok for upstream?
2014-08-04 Ganesh Gopalasubramanian
On Mon, Aug 04, 2014 at 12:26:01PM +0200, Richard Biener wrote:
On Mon, 4 Aug 2014, Marek Polacek wrote:
This PR is about bogus overflow warning that we issue for e.g.
int *q = i + 1;
q - (q - 1);
because pointer_diff receives p - (p + -1U) which gets simplified to
1U - with
On 25/07/14 22:53 +0200, François Dumont wrote:
Hi
I think I never get feedback regarding this patch proposal. Note
I've been trying to weigh up the pros and cons and am unsure what's
best, but I think my preference is to have a noexcept default
constructor. Unless you hear any objections
thanks for the review.
Np, I'm personally happy to see that script is useful.
I've now interpreted it such that --inline prints to stdout what it
would print to the patch file otherwise, that is, both log and patch.
Printing just the log to stdout can be already be achieved by not using
On 04/08/14 12:49 +0100, Jonathan Wakely wrote:
On 25/07/14 22:53 +0200, François Dumont wrote:
Hi
I think I never get feedback regarding this patch proposal. Note
I've been trying to weigh up the pros and cons and am unsure what's
best, but I think my preference is to have a noexcept
On Mon, Aug 04, 2014 at 12:51:06PM +0200, Richard Biener wrote:
On Mon, 4 Aug 2014, Marc Glisse wrote:
On Mon, 4 Aug 2014, Marek Polacek wrote:
+/* { dg-do compile } */
+
+void
+foo (void)
+{
+ volatile __PTRDIFF_TYPE__ t;
+ int i;
+ int *p = i;
+ int *q = i
On Mon, 4 Aug 2014, Marek Polacek wrote:
On Mon, Aug 04, 2014 at 12:26:01PM +0200, Richard Biener wrote:
On Mon, 4 Aug 2014, Marek Polacek wrote:
This PR is about bogus overflow warning that we issue for e.g.
int *q = i + 1;
q - (q - 1);
because pointer_diff receives p - (p
2014-08-04 Richard Biener rguent...@suse.de
Merge from trunk r211444 through r213455.
* genmatch.c (operators): Adjust for hash_table interface change.
* tree-ssa-pre.c (eliminate_dom_walker::before_dom_children):
Also fold non-call stmts inplace to avoid
Hi all,
While working on another patch and looking at the rbit and clz patterns,
I noticed that in aarch32 the relevant patterns are not adjusted for
-mrestrict-it.
A program like:
int
foo (int a, int b)
{
if (a + 5 == b)
return __builtin_ctz (a);
return 1;
}
compiled with
Hi all,
Following up on Josephs' comments on PR 61876 this patch guards the rint
+ cast - lrint transformation on -fno-math-errno.
Bootstrapped and tested on aarch64-linux and x86.
Ok for trunk?
Thanks,
Kyrill
2014-08-04 Kyrylo Tkachov kyrylo.tkac...@arm.com
* convert.c
The following patch removes the only user of gimple_fold_builtin,
using the proper fold_stmt API.
Boostrapped on x86_64-unknown-linux-gnu, testing in progress.
Richard.
2014-08-01 Richard Biener rguent...@suse.de
* gimple-fold.h (gimple_fold_builtin): Remove.
* gimple-fold.c
On Mon, 4 Aug 2014, Marek Polacek wrote:
On Mon, Aug 04, 2014 at 12:51:06PM +0200, Richard Biener wrote:
On Mon, 4 Aug 2014, Marc Glisse wrote:
On Mon, 4 Aug 2014, Marek Polacek wrote:
+/* { dg-do compile } */
+
+void
+foo (void)
+{
+ volatile __PTRDIFF_TYPE__ t;
+ int i;
+ int *p = i;
This patch implements the following rules related to the discriminants of
derived types:
The type of a discriminant_specification shall be discrete.
A discriminant_specification shall not occur as part of a derived type
declaration whose parent type is discriminated.
--
This change introduces a new subprogram in the GNAT secure hash
framework to allow computing HMACs based on the secure hash functions.
This is achieved by initializing a context with a (non-empty) key.
The following example shows how to compute the first HMAC-MD5 test from
RFC2104:
$ gnatmake -q
This patch removes a spurious elaboration warning on some calls to
postconditions. Such a call always appears within an enclosing body, and as
such is not subject to an elaboration check, but the call that is inserted
after the final (inserted) return statement in a procedure body is analyzed
in
On Mon, Aug 04, 2014 at 02:36:12PM +0200, Marc Glisse wrote:
On Mon, 4 Aug 2014, Marek Polacek wrote:
On Mon, Aug 04, 2014 at 12:51:06PM +0200, Richard Biener wrote:
On Mon, 4 Aug 2014, Marc Glisse wrote:
On Mon, 4 Aug 2014, Marek Polacek wrote:
+/* { dg-do compile } */
+
+void
+foo
From: Trevor Saunders tsaund...@mozilla.com
Hi,
It used to be that edge_var_maps held pointers to embedded vectors, but
now it holds vectors. This means that now instead of copying the
address of the embedded vector from the table we keep a pointer into the
table. However that's incorrect
This adds an internal abstraction for testing for standard string types.
Internal front end cleanup, no function effect, no test required.
Tested on x86_64-pc-linux-gnu, committed on trunk
2014-08-04 Robert Dewar de...@adacore.com
* einfo.ads, einfo.adb (Is_Standard_String_Type): New
LAST_UPDATED: Updated Sat Aug 2 10:33:58 CEST 2014 to revision 213515.
=== acats tests ===
=== acats Summary ===
# of expected passes2320
# of unexpected failures0
Native configuration is x86_64-apple-darwin13.3.0
=== g++
Wong list!-(
Sorry,
Dominique
At the moment, for two-element vectors, __builtin_shuffle (vector, (mask) {C,
C}) for identical constants C outputs a zip (with both argument vectors the
same) rather than a dup. Dup is more obvious and easier to read, so prefer it.
For big-endian, aarch64_evpc_dup always aborts; however tests
+if ($inline $diff ne -) {
+ $tmp = `mktemp`;
+ if ($? != 0) {
+ die Could not generate temp file;
+ }
IMHO better use consistent style: system() or ticks with $?.
Or let Perl itself create the temporary file:
open(my $tmp_fh, +, undef) or die cannot
This is a 99% complete implementation of the Filesystem TS as defined
by the N4099 draft,
http://open-std.org/jtc1/sc22/wg21/docs/papers/2014/n4099.html
The .cc files use C++14 but the headers are C++11 so it can be used by
more people. Rather than adding loads of new exports to libstdc++.so
the
Hi all,
Now that both backends have a way of generating CRC32 instructions this
patch adds the crc type to the scheduling information for the Cortex-A53
and Cortex-A57 cores.
For both Cortex-A53 and Cortex-A57 they behave similarly to shifted
arithmetic instructions
When scheduling for the
Added patterns in associate_pointerplus and associate_pointerplus_diff.
* genmatch.c (capture_max): Change value to 6.
(match-plusminus.pd): Add new patterns.
[gcc/testsuite/gcc.dg/tree-ssa]
* match-plusminus.c (plusminus_9): New test-case.
Thanks,
Prathamesh
Index: gcc/genmatch.c
Mike == Mike Stump mikest...@comcast.net writes:
Mike So the normal way to do this would be to make the plugin front-end
Mike non-default and then never gate any release decisions upon the state
Mike of the that front-end.
Not sure if this is overly pedantic, but the plugin is just a library,
I would expect the to mark the i loop as non-parallel, but the j-loop
as parallel. What is the partial schedule, the set of dependences and
the dimension you check for both the i and the j loop?
Yes, you are right. The i loop is non-parallel and j-loop is parallel.
I've found that this
Sorry for misprint. It passes all the tests from
libgomp/testsuite/libgomp.graphite
The attached patch contains the improved version of checking for the
loop parallelism, which passes all the tests from
libgomp/testsuite/libgomp.graphite except
graphite-isl-ast-to-gimple.c.
--
On 04/08/2014 16:23, Roman Gareev wrote:
I would expect the to mark the i loop as non-parallel, but the j-loop
as parallel. What is the partial schedule, the set of dependences and
the dimension you check for both the i and the j loop?
Yes, you are right. The i loop is non-parallel and j-loop
From: Andi Kleen a...@linux.intel.com
Add a new bootstrap-lto-slim config file that enables slim (non-fat) LTO
bootstrap. This improves the performance of the LTO bootstrap.
Speeds up the LTO bootstrap by ~18% on a 4 core system.
This requires using gcc-ar/ranlib in post stage 1 builds, so
This patchkit implements slim LTO bootstrap to speed up LTO
bootstrap by ~18%. Passes LTO and normal bootstrap on x86_64-linux.
Ok?
-Andi
From: Andi Kleen a...@linux.intel.com
To use gcc-{ar,ranlib} for boot strap we need to add a -B option
to the tool. Since ar has weird and unusual argument conventions
implement the code by hand instead of using any libraries.
gcc/:
2014-08-04 Andi Kleen a...@linux.intel.com
*
Hi,
The following patch disables stdio_va_list fix: stdio.h is already
good in Android and, since ndk gcc is indented to be used with
different Android sysroots, it is actually harmful, because without
this fix only the version of stdio.h from the sysroot the compiler was
built with will be used.
Hi!
I've tried to bootstrap with r213515 because later revisions broke because
of the hash_map and Ada changes, but unfortunately even that revision
failus to bootstrap, the new -Wreturn-local-addr warning rightfully
warns about get_ivts_expr possibly returning address of a local variable.
It
On Sun, Aug 3, 2014 at 11:12 AM, Iain Buclaw ibuc...@gdcproject.org wrote:
This adds a demangler for the D programming language to libiberty,
intended to be used in GDB and Binutils. GDB already has a trimmed
down implementation of this, but have been advised that here would be
a better
Iain == Iain Buclaw ibuc...@gdcproject.org writes:
Iain This adds a demangler for the D programming language to libiberty,
Iain intended to be used in GDB and Binutils. GDB already has a trimmed
Iain down implementation of this, but have been advised that here would be
Iain a better location to
Hi!
As can be seen on the depend-8.c testcase, only trying to schedule
taskgroup-children in GOMP_taskgroup_end can lead to deadlock
if depend clauses are involved - some task in the current taskgroup
can depend on a task with the same parent, but from another taskgroup
(one of the parent
Alan == Alan Modra amo...@gmail.com writes:
Alan Bootstrapped and regression tested x86_64-linux and powerpc-linux,
Alan with both the new obstack.{h,c} and the old versions. OK to apply?
Alan gcc/java/
Alan * mangle.c (finish_mangling): Cast result of obstack_base to (char *).
Alan *
Hi,
I suppose we can quickly resolve, one way or another, this rather old
issue. Considering:
extern void z();
void h() { if ( z != (void*)0 ); }
we -Waddress warn in C and we don't in C++, due to the rather subtle
differences between null_pointer_constant_p and null_ptr_cst_p. I
believe
Hi all,
As part of other intrinsics-related messing around due to the
float64x1_t changes I noticed these can be (re)implemented relatively
easily.
Tested on aarch64-none-elf and aarch64_be-none-elf to make sure the
lane-wise intrinsics do the right thing.
Ok for trunk?
2014-08-04
On 4 August 2014 16:52, Ian Lance Taylor i...@google.com wrote:
On Sun, Aug 3, 2014 at 11:12 AM, Iain Buclaw ibuc...@gdcproject.org wrote:
This adds a demangler for the D programming language to libiberty,
intended to be used in GDB and Binutils. GDB already has a trimmed
down implementation
On 4 August 2014 17:06, Tom Tromey tro...@redhat.com wrote:
Iain == Iain Buclaw ibuc...@gdcproject.org writes:
Iain This adds a demangler for the D programming language to libiberty,
Iain intended to be used in GDB and Binutils. GDB already has a trimmed
Iain down implementation of this, but
On Mon, 4 Aug 2014, Jakub Jelinek wrote:
I've tried to bootstrap with r213515 because later revisions broke because
of the hash_map and Ada changes, but unfortunately even that revision
failus to bootstrap, the new -Wreturn-local-addr warning rightfully
warns about get_ivts_expr possibly
Hi all,
This patch implements some saturating math *laneq_s* intrinsics.
The implementation is fairly straightforward, just use more general mode
iterators, add appropriate builtins etc.
Some execution tests are added with some scan-assembly parts to make
sure we generate the correct lane
Hi all,
Now that uint64x1_t and uint64_t are not interchangeable these
intrinsics have an incorrect type. The reason is that these intrinsics
should have had the scalar type for some of their arguments and results
according to the NEON intrinsics spec.
This patch fixes that and updates the
On 25/07/14 23:05, Jeff Law wrote:
On 07/23/14 02:53, Kyrill Tkachov wrote:
Darn, had forgotten to attach the patch...
On 16/07/14 12:30, Kyrill Tkachov wrote:
Hi all,
This fixes the PR mentioned in the subject. When expanding
atomic_test_and_set we try the corresponding sync optabs and if
Committed on trunk, revision 213596
Committed on 4.9 branch, revision 213597
I made an omission on the first commit. I did not add
the test case and corresponding ChangeLog entry.
Committed as obvious on trunk, revision 213598
Thanks
Edmar
On 08/04/2014 05:25 AM, Ulrich Weigand wrote:
David
Hi,
A years ago there was a discussion
(https://gcc.gnu.org/ml/gcc-patches/2004-01/msg02437.html) about
debugging compiler ICEs that resulted in a patch from Jakub, which dumps
useful information into temporary file, but for some reasons this patch
wasn't applied to trunk.
This is the
Thanks Jeff and Jakub, I've reposted ICE debugging patch into
gcc-patches mailing list
(https://gcc.gnu.org/ml/gcc-patches/2014-08/msg00285.html).
-Maxim
On 08/01/2014 11:29 PM, Jeff Law wrote:
On 08/01/14 02:00, Jakub Jelinek wrote:
On Thu, Jul 24, 2014 at 04:39:28PM +0400, Maxim Ostapenko
On 16 July 2014 17:22, Uros Bizjak ubiz...@gmail.com wrote:
On Fri, Jun 6, 2014 at 4:31 PM, Jonathan Wakely jwak...@redhat.com wrote:
On 06/06/14 12:40 +0200, Uros Bizjak wrote:
On Fri, Jun 6, 2014 at 11:19 AM, Jonathan Wakely jwak...@redhat.com
wrote:
On 06/06/14 10:27 +0200, Uros Bizjak
On 08/04/14 10:54, Kyrill Tkachov wrote:
On 25/07/14 23:05, Jeff Law wrote:
On 07/23/14 02:53, Kyrill Tkachov wrote:
Darn, had forgotten to attach the patch...
On 16/07/14 12:30, Kyrill Tkachov wrote:
Hi all,
This fixes the PR mentioned in the subject. When expanding
atomic_test_and_set
On 08/04/14 11:15, Trevor Saunders wrote:
On Sat, Aug 02, 2014 at 03:44:17PM -0700, Andrew Pinski wrote:
On Tue, Jul 29, 2014 at 5:50 AM, tsaund...@mozilla.com wrote:
From: Trevor Saunders tsaund...@mozilla.com
Hi,
this adds a hash_set wrapper around hash_table, and then replaces usage of
On Tue, Jul 22, 2014 at 1:14 AM, Rainer Orth
r...@cebitec.uni-bielefeld.de wrote:
The recent Go 1.3 update (which I couldn't find posted to gcc-patches)
broke Solaris bootstrap:
/vol/gcc/src/hg/trunk/local/libgo/runtime/mem.c:50:20: error: pointer targets
in passing argument 1 of 'mincore'
On Mon, Aug 4, 2014 at 7:29 PM, Jonathan Wakely jwakely@gmail.com wrote:
These two tests timeout on alpha-linux-gnu:
FAIL: 30_threads/async/forced_unwind.cc execution test
WARNING: program timed out.
FAIL: 30_threads/packaged_task/forced_unwind.cc execution test
WARNING: program timed
On Aug 4, 2014, at 10:46 AM, Trevor Saunders tsaund...@mozilla.com wrote:
I had a conflict which is why I noticed this issue in the first place. And I
was using git merge :).
I bet git log or blaim would get you better data faster.
So, git merging I would recommend:
[merge]
On 01/06/14 18:24 +0100, Jonathan Wakely wrote:
Tested x86_64-linux, committed to trunk.
This should probably go on the 4.9 branch too, although we could leave
the old default cosntructor semantics and just fix the conversion
operator.
Now committed to the 4.9 branch too.
PR
On 04/08/14 19:59 +0200, Uros Bizjak wrote:
On Mon, Aug 4, 2014 at 7:29 PM, Jonathan Wakely jwakely@gmail.com wrote:
Uros, would backporting this to the 4.9 branch cause problems for alpha?
No, I don't think so. The problem was identified as glibc bug that was
fixed almost a year ago for
1 - 100 of 117 matches
Mail list logo