Ping?
-Original Message-
From: Thomas Preud'homme [mailto:thomas.preudho...@arm.com]
Sent: Tuesday, January 27, 2015 1:38 PM
To: Thomas Preud'homme; Richard Earnshaw; gcc-patches@gcc.gnu.org
Cc: Ramana Radhakrishnan
Subject: RE: [PATCH 2/3, ARM, libgcc, ping6] Code size optimization
Ping?
-Original Message-
From: gcc-patches-ow...@gcc.gnu.org [mailto:gcc-patches-
ow...@gcc.gnu.org] On Behalf Of Thomas Preud'homme
Sent: Friday, January 23, 2015 4:23 PM
To: Ramana Radhakrishnan
Cc: gcc-patches
Subject: RE: [PATCH, ARM] Fix PR64453: live high register not
Ping?
-Original Message-
From: gcc-patches-ow...@gcc.gnu.org [mailto:gcc-patches-
ow...@gcc.gnu.org] On Behalf Of Thomas Preud'homme
Sent: Wednesday, January 28, 2015 6:02 PM
To: 'Andrew Pinski'
Cc: GCC Patches
Subject: RE: [PATCH, Aarch64] Add FMA steering pass for Cortex-A57
This is actually a long standing regression (though it's not marked that
way in the BZ database).
Basically there was a change from encoding the funcvec and eighbit_data
attributes in the symbol name to instead putting them in the function
decl's flags. This was, IMHO, a good thing.
On 2/5/2015 9:13 AM, Sandra Loosemore wrote:
In addition to fixing markup and capitalization, I've moved things
around a little bit to improve the flow, and rephrased a few things
that I thought were awkward or confusing. I propose to commit this in
a few days unless somebody tells me
On 02/05/15 08:53, Eric Botcazou wrote:
Hi,
this is the failure of gcc.dg/tree-ssa/ivopts-lt-2.c on SPARC 64-bit, but I
think that the problem is actually generic: get_shiftadd_cost only calls
shiftadd_cost to get the cost of a shift-add, which is actually a multiply
by a power-of-2 operation.
On 02/04/15 14:20, Sebastian Pop wrote:
Hi,
The attached patch stops the recursion in the detection of FSM jump-threads at
loop phi nodes after having visited a loop phi node. This avoids jump-threading
two iterations forward that were possible due to a flip-flop operation that
exchange the
Tested locally with a suitably hacked up script to work on this
machine; with this patch the Show Sources links work within the
files installed to /www/gcc/htdocs.
OK for trunk?
maintainer-scripts/ChangeLog:
* update_web_docs_svn: When copying up files for the built jit
The H8/300 port was mis-compiling 24-28 bit ASHIFTRTs for the H8/300.
These shifts are implemented by a copying the high byte to the low byte,
then extending the sign bit through the low middle, high middle and high
bytes (with compensation code for 25,26,27,28 bit shifts).
The problem was
On 02/05/15 04:22, Renlin Li wrote:
Hi all,
location information is missing when rewrite complex component-wise
load/store.
This patch add the location info to fix the recently gcc.dg/uninit-13.c
and gcc.dg/uninit-17.c failures on arm target. The compiler warns
uninitialised usage at the
A little bit of code went missing due to the constant expressions code
being factored out. Additionally, concepts are currently dependent on
changes to CPTK_IS_CONVERTIBLE_TO, so this patch restores that. I'm
told that this was removed from trunk and that the concepts branch will
be adapted
This patch from Lynn Boger adds memprofilerate as an option for the
GODEBUG environment variable, to match the gc compiler. Bootstrapped
and ran Go testsuite on x86_64-unknown-linux-gnu. Committed to
mainline.
Ian
diff -r 18b325ce33aa libgo/go/runtime/extern.go
--- a/libgo/go/runtime/extern.go
Users of GNATprove can now specify pragma Warnings for GNAT and GNATprove
separately, to selectively disable warnings from the compiler or the
formal verification tool. This also allows detecting useless pragma
Warnings with switch -gnatw.w
Tested on x86_64-pc-linux-gnu, committed on trunk
On Thu, Jan 29, 2015 at 04:38:21PM -0800, Segher Boessenkool wrote:
2015-01-29 Segher Boessenkool seg...@kernel.crashing.org
gcc/
PR target/64580
* config.rs6000/rs6000.c (compute_vrsave_mask): Reverse loop order.
(rs6000_stack_info): Add assert.
On Thu, Feb 5, 2015 at 9:50 AM, Segher Boessenkool
seg...@kernel.crashing.org wrote:
On Thu, Jan 29, 2015 at 04:38:21PM -0800, Segher Boessenkool wrote:
2015-01-29 Segher Boessenkool seg...@kernel.crashing.org
gcc/
PR target/64580
* config.rs6000/rs6000.c (compute_vrsave_mask):
On Thu, Feb 05, 2015 at 08:12:25AM +0100, Jakub Jelinek wrote:
On Thu, Feb 05, 2015 at 02:09:54PM +1030, Alan Modra wrote:
Jakub, was your suggestion to use get_last_insn_anywhere() based on
not wanting to expose details that should be internal to
emit-rtl.[ch]?
Yes. But if it doesn't
Just to ensure that this obvious commit gets noted... ( the other copy
of the patch is buried in this AArch64 specific thread:
https://gcc.gnu.org/ml/gcc-patches/2015-02/msg00328.html )
By the sort of exceptional bad luck that only comes around once
every thousand builtins, we now define enough
Hello!
When hard frame pointer isn't needed, the register for hard frame pointer
may be reused. This patch clears alignment on hard frame pointer if hard
frame pointer isn't needed. OK for trunk after bootstrap and test on
Linux/x86-64?
LRA can set up frame_pointer_needed in some
Hi!
On the following testcase we fail -fcompare-debug, because in .LASAN0
initializer we get a NOP_EXPR optimized away only in one of the cases.
The reason is that cgraph* is looking at DECL_ABSTRACT_P flag of
__PRETTY_FUNCTION__ and that flag is different between -g and -g0.
The DECL_ABSTRACT_P
Hi,
this is the failure of gcc.dg/tree-ssa/ivopts-lt-2.c on SPARC 64-bit, but I
think that the problem is actually generic: get_shiftadd_cost only calls
shiftadd_cost to get the cost of a shift-add, which is actually a multiply
by a power-of-2 operation. Since SPARC doesn't have a lea-like
On Thu, Feb 5, 2015 at 8:42 AM, H.J. Lu hjl.to...@gmail.com wrote:
On Wed, Jan 28, 2015 at 4:02 PM, H.J. Lu hjl.to...@gmail.com wrote:
On Wed, Jan 28, 2015 at 11:37 AM, H.J. Lu hjl.to...@gmail.com wrote:
On Wed, Jan 28, 2015 at 11:19 AM, Richard Biener
richard.guent...@gmail.com wrote:
On
Dear Tobias,
That is OK for trunk.
Thanks for the patch.
Paul
On 5 February 2015 at 22:04, Tobias Burnus bur...@net-b.de wrote:
A rather simple patch.
Build and reg-tested on x86-64-gnu-linux.
OK for the trunk?
Tobias
--
Outside of a dog, a book is a man's best friend. Inside of a
Hi Petar,
I've put your patch inline below and switched to plain text. I suspect
your post was bounced by gcc-patches.
I'm OK with this change but I'd like Catherine to comment before committing.
It seems a shame to duplicate the block of code but it is probably just as
ugly to define a macro
A rather simple patch.
Build and reg-tested on x86-64-gnu-linux.
OK for the trunk?
Tobias
2015-02-05 Tobias Burnus bur...@net-b.de
PR fortran/64943
* resolve.c (resolve_transfer): Also check structure
constructors.
2015-02-05 Tobias Burnus bur...@net-b.de
PR fortran/64943
*
On Thu, Feb 5, 2015 at 10:54 AM, Richard Henderson r...@redhat.com wrote:
On 02/04/2015 03:29 PM, H.J. Lu wrote:
+++ b/gcc/varasm.c
@@ -6826,11 +6826,17 @@ default_binds_local_p_1 (const_tree exp, int shlib)
(TREE_STATIC (exp) || DECL_EXTERNAL (exp)))
{
varpool_node
This is a follow-up to a change [1] in glibc. It fixes the issue [2]
when jal can not reach a target in different region.
Is it not a problem that can be solved with rearranging the order of
sections in output?
Is not this more clean solution to it? Otherwise we need to make sure that
On 4 February 2015 at 12:38, Marcus Shawcroft
marcus.shawcr...@gmail.com wrote:
On 12 January 2015 at 15:52, Kyrill Tkachov kyrylo.tkac...@arm.com wrote:
Hi all,
As raised in https://gcc.gnu.org/ml/gcc-patches/2014-12/msg01237.html and
discussed in that thread, using __builtin_sqrt for
On Thu, Feb 5, 2015 at 5:29 AM, Alan Modra amo...@gmail.com wrote:
On Thu, Feb 05, 2015 at 08:12:25AM +0100, Jakub Jelinek wrote:
On Thu, Feb 05, 2015 at 02:09:54PM +1030, Alan Modra wrote:
Jakub, was your suggestion to use get_last_insn_anywhere() based on
not wanting to expose details that
If an inherited subprogram is implemented by a protected function,
then the first parameter of the inherited subprogram shall be of
mode in, but not an access-to-variable parameter (RM 9.4(11/9)
After this patch the error is reported in the following example:
procedure by30015_01p is
package
On Thu, Feb 05, 2015 at 08:21:46AM +0100, Richard Biener wrote:
On February 4, 2015 10:20:06 PM CET, Sebastian Pop seb...@gmail.com wrote:
The attached patch stops the recursion in the detection of FSM
jump-threads at
loop phi nodes after having visited a loop phi node. This avoids
Hi!
Long time ago, probably all FEs or most of them emitted an extra BLOCK
around the BLOCK corresponding to user scope of the function, then
the C FE has been changed not to and no_body_blocks langhook has been added
for it (and Java/Ada/Fortran were mishandled), later on r144474
removed the
On Thu, Feb 5, 2015 at 11:59 AM, Richard Henderson r...@redhat.com wrote:
On 02/05/2015 11:01 AM, H.J. Lu wrote:
Can you elaborate why it depends on COPY relocation? There
is no COPY relocation on x86-64.
Ho hum, we appear to have switched topics mid-thread.
I agree that we cannot override
:
.text
.LHOTE0:
.weak
.bss
.align 4
.type , @object
.size , 4
:
.zero 4
.ident GCC: (GNU) 5.0.0 20150205 (experimental)
.section .note.GNU-stack,,@progbits
.file initweak.i
.section .text.unlikely,ax,@progbits
.LCOLDB0:
.text
.LHOTB0
GCC: (GNU) 5.0.0 20150205 (experimental)
.section .note.GNU-stack,,@progbits
.file initweak.i
.section .text.unlikely,ax,@progbits
.LCOLDB0:
.text
.LHOTB0:
.p2align 4,,15
.globl foo
.type foo, @function
foo:
.LFB0:
.cfi_startproc
movl (%rip), %eax
ret
.cfi_endproc
.LFE0:
.size foo, .-foo
On Thu, Feb 05, 2015 at 08:59:01PM +1030, Alan Modra wrote:
Thanks, I'll use it directly now and have a patch in the works to tidy
m32c.c and rs6000.c.
As threatened, the emit-rtl tidy. Besides adding a couple of
accessors, I've moved x_first_insn and x_last_insn into a struct
sequence_stack.
The m32c parts are OK.
It looks to me like that, when the rewrite of the inline asm sections
was committed last May (r210273), there was general agreement that it
needed additional copy-editing to fix markup problems and the like.
But, nobody ever followed up to do that.
Here's my attempt. In addition to fixing
On Wed, Jan 28, 2015 at 4:02 PM, H.J. Lu hjl.to...@gmail.com wrote:
On Wed, Jan 28, 2015 at 11:37 AM, H.J. Lu hjl.to...@gmail.com wrote:
On Wed, Jan 28, 2015 at 11:19 AM, Richard Biener
richard.guent...@gmail.com wrote:
On January 28, 2015 7:12:43 PM CET, H.J. Lu hongjiu...@intel.com wrote:
As outlined in the PR, the h8300 U constraint handling uses
REG_OK_FOR_BASE_P, which, most of the time is safe. However, once
reload has started we must use the strict variant
REG_OK_FOR_BASE_STRICT_P.Patch is from Joern, and I've included a
freshly reduced testcase.
Verified the
Hi,
my previous fix to ipa-prop actually causes ipa-cp to ICE during FDO build of
firefox.
What happens is that we produce speculative call that is off and leads to a
method with
wrong number of parameters.
It would probably make a lot of sense to validate devirtualization with the
actual type
On February 5, 2015 12:29:40 AM GMT+01:00, H.J. Lu hjl.to...@gmail.com
wrote:
On Wed, Feb 4, 2015 at 3:10 PM, H.J. Lu hjl.to...@gmail.com wrote:
On Wed, Feb 4, 2015 at 2:47 PM, Bernhard Reutner-Fischer
rep.dot@gmail.com wrote:
On February 4, 2015 11:37:01 PM GMT+01:00, H.J. Lu
And what about namespace-scope variables? Maybe instead of checking for
TRANSLATION_UNIT_DECL context this should check non-function context.
Done.
+ /* This is the second iteration through the global symbols. Here we
+ pick up function statics that have been discovered by the call
On 4 February 2015 at 17:34, Gerald Pfeifer ger...@pfeifer.com wrote:
On Wednesday 2015-02-04 16:19, Matthew Wahab wrote:
This patch documents in gcc-5/changes.html the addition of support for the
Cortex-A72 to the ARM and the AArch64 backends.
Looks good to me, but you may want to wait a
On Thu, Feb 05, 2015 at 08:53:00AM +, Christophe Lyon wrote:
On 4 February 2015 at 12:38, Marcus Shawcroft
marcus.shawcr...@gmail.com wrote:
On 12 January 2015 at 15:52, Kyrill Tkachov kyrylo.tkac...@arm.com wrote:
Hi all,
As raised in
In some unusual cases, the index check on a subscript of an
unconstrained array was omitted.
The following test program
1. package Out_Constraint_Violation is
2.type Arr is array (Natural range ) of Integer;
3.procedure Violate
4. (Input : Arr; Modifier : Arr;
On 26-01-15 15:47, Richard Biener wrote:
Index: gcc/testsuite/gcc.dg/uninit-19.c
===
--- gcc/testsuite/gcc.dg/uninit-19.c(revision 0)
+++ gcc/testsuite/gcc.dg/uninit-19.c(working copy)
@@ -0,0 +1,23 @@
+/* { dg-do compile }
From: Marcus Shawcroft [mailto:marcus.shawcr...@gmail.com]
Sent: Thursday, February 05, 2015 5:17 PM
OK but wait for stage-1 to open for general development before you
commit it please.
/Marcus
Duly noted.
Best regards,
Thomas
This patch improves the error recovery when WHEN is used in place
of WITH for aspect specifications:
1. package WhenWith is
2.X : integer when Size = 4;
|
when should be with
3.Y : integer when Rubbish;
|
On Thu, Feb 05, 2015 at 11:31:53AM +0100, Jakub Jelinek wrote:
On Thu, Feb 05, 2015 at 08:59:01PM +1030, Alan Modra wrote:
On Thu, Feb 05, 2015 at 08:12:25AM +0100, Jakub Jelinek wrote:
On Thu, Feb 05, 2015 at 02:09:54PM +1030, Alan Modra wrote:
Jakub, was your suggestion to use
Moore, Catherine catherine_mo...@mentor.com writes:
Hi Matthew,
I made a few edits. I removed the markup in the process, so that will
need to be added back.
See the text at the end.
Thanks Catherine. Good call to remove the markup while reviewing. I've
done one more pass on this to have
On Thu, Feb 05, 2015 at 08:20:20AM +0100, Richard Biener wrote:
On February 5, 2015 12:28:14 AM CET, Jan Hubicka hubi...@ucw.cz wrote:
Hi,
this patch avoids ICE in programs that uses multiple copies of libgcov
at once.
It is just a temporary workaround as discussed in the PR log.
Tested on
Hi Andrew,
On Wed, Feb 4, 2015 at 9:07 AM, Mike Stump mikest...@comcast.net wrote:
On Feb 4, 2015, at 2:28 AM, Rainer Orth r...@cebitec.uni-bielefeld.de
wrote:
Rainer Orth r...@cebitec.uni-bielefeld.de writes:
On Jan 28, 2015, Mike Stump mikest...@comcast.net wrote:
On Jan 28, 2015, at
This patch modifies the mode conformance checks of pragma SPARK_Mode to account
for the case where the pragma appears without an argument.
-- Source --
-- pack.ads
package Pack is
procedure Error;
end Pack;
-- pack.adb
package body Pack is
procedure Error
Hi all,
location information is missing when rewrite complex component-wise
load/store.
This patch add the location info to fix the recently gcc.dg/uninit-13.c
and gcc.dg/uninit-17.c failures on arm target. The compiler warns
uninitialised usage at the variable's declaration location
This internal modification changes the representation of the
Fatal_Error field in the unit record to record the presence
of fatal errors even if -gnatq/Q is set. No functional effect
for the compiler itself so no test now. This should allow an
improvement in ASIS processing which will be
This patch adds an h2-level section about libgccjit between the
sections on languages and targets.
It has various links to the documentation (now that that's on the
website), and to a couple of 3rd party repos (the Python and D
bindings). It also mentions the license of the library.
OK to
Hi,
in this (mild) ICE on invalid 4.9/5 regression things go wrong in
various ways when we try to use cp_parser_iteration_statement at
top-level. I think the simplest thing to do is just returning early,
then cp_parser_declaration will emit an appropriate diagnostic (the same
produced when
On 02/04/2015 03:29 PM, H.J. Lu wrote:
+++ b/gcc/varasm.c
@@ -6826,11 +6826,17 @@ default_binds_local_p_1 (const_tree exp, int shlib)
(TREE_STATIC (exp) || DECL_EXTERNAL (exp)))
{
varpool_node *vnode = varpool_node::get (exp);
- if (vnode (resolution_local_p
Move Darwin case before default one. Hopefully should be the last iteration
on this PR, sorry about the confusion!
Committed on trunk.
2015-02-05 Tristan Gingold ging...@adacore.com
PR ada/64349
* env.c: Fix thinko: handle Darwin case before default one.
Index: env.c
On 28 January 2015 at 10:01, Thomas Preud'homme
thomas.preudho...@arm.com wrote:
From: gcc-patches-ow...@gcc.gnu.org [mailto:gcc-patches-
ow...@gcc.gnu.org] On Behalf Of Thomas Preud'homme
Hi Andrew,
cortex-a57-fma-steering.c is really meant to be autosufficient with
On Thu, Feb 05, 2015 at 08:59:01PM +1030, Alan Modra wrote:
On Thu, Feb 05, 2015 at 08:12:25AM +0100, Jakub Jelinek wrote:
On Thu, Feb 05, 2015 at 02:09:54PM +1030, Alan Modra wrote:
Jakub, was your suggestion to use get_last_insn_anywhere() based on
not wanting to expose details that
Jakub Jelinek wrote:
On Thu, Feb 05, 2015 at 08:21:46AM +0100, Richard Biener wrote:
On February 4, 2015 10:20:06 PM CET, Sebastian Pop seb...@gmail.com wrote:
The attached patch stops the recursion in the detection of FSM
jump-threads at
loop phi nodes after having visited a loop phi
On 02/05/2015 11:01 AM, H.J. Lu wrote:
Can you elaborate why it depends on COPY relocation? There
is no COPY relocation on x86-64.
Ho hum, we appear to have switched topics mid-thread.
I agree that we cannot override a weak symbol in the executable with even a
non-weak symbol in a shared
On 02/04/2015 07:27 PM, Aldy Hernandez wrote:
+ /* Anything who's context is not a TRANSLATION_UNIT_DECL will
+ be handled by either handling reachable functions below, or
+ by generating DIEs for types. */
whose
And what about namespace-scope variables? Maybe
64 matches
Mail list logo
