Hi,
In PR71478, for vector negation of ssa produced by call stmt, we add
vector (-1) and ssa to the ops list. However, in the place where we
remove the (-1) from ops list, we failed to do this for vector integer.
As a result, rewrite_expr_tree wrongly assumes that it is working with
gimple_as
... and this version passes testing. For real.
Paolo.
Index: cp/decl.c
===
--- cp/decl.c (revision 237285)
+++ cp/decl.c (working copy)
@@ -6002,6 +6002,13 @@ check_initializer (tree decl, tree init,
This patch to the Go frontend fixes the quoting in an error message.
Bootstrapped and ran Go testsuite on x86_64-pc-linux-gnu. Committed
to mainline.
Ian
Index: gcc/go/gofrontend/MERGE
===
--- gcc/go/gofrontend/MERGE (revision 23
I submitted a patch for the website a few weeks ago, but received no
response in the negative or positive. Wanted to check if it had been
overlooked.
Original patch can be found here:
https://gcc.gnu.org/ml/gcc-patches/2016-05/msg00767.html
Thanks,
Stuart
I have applied the following patches for ISA 3.0 (power9) that were installed
on the trunk in May to the GCC 6.2 branch after doing a bootstrap and
regression testing. This patch combines the original patch, and the two
followup patches that fixed some bugs. It does remove the all_ones_predicate
Hi,
On 09/06/2016 23:38, Jason Merrill wrote:
I would think that when we see a duplicated base we should just drop
the duplicates and continue.
If the type of b1 is error_mark_node, why isn't the type of b2 also
error_mark_node?
Thanks Jason. Normally check_initializer massages the decl basing
On Tue, May 31, 2016 at 2:56 AM, James Greenhalgh
wrote:
> As you're proposing to have this on by default, I'd like to give a chance
> to hear whether there is consensus as to this being the right choice for
> the thunderx, xgene1, exynos-m1 and qdf24xx subtargets.
I tested this on qdf24xx using
Bruce,
The current AIX stdlib.h header file uses #define. This is exactly
what I am trying to remove with the fixincludes patch.
#define malloc __linux_malloc
needs to be replaced with
extern void *malloc(size_t) __asm__("__linux_malloc");
and so forth.
Thanks, David
On Thu, Jun 9, 2016 at
On Thu, Jun 9, 2016 at 3:39 AM, Richard Biener wrote:
> On Thu, 9 Jun 2016, Jakub Jelinek wrote:
>
>> On Thu, Jun 09, 2016 at 08:50:15AM +0200, Richard Biener wrote:
>> > On Wed, 8 Jun 2016, Jason Merrill wrote:
>> >
>> > > On Wed, Jun 8, 2016 at 11:16 AM, Marc Glisse
>> > > wrote:
>> > > > On W
On 06/06/2016 03:31 PM, David Malcolm wrote:
Jeff approved an earlier version of this (as
unittests/test-ggc.c):
https://gcc.gnu.org/ml/gcc-patches/2015-10/msg03306.html
Not terribly happy with that counter to used to create a big list
to detect recursion. But I'm not offhand sure how to avoid
This simply prevents valgrind from complaining about an invalid read when
pass_free_cfg::execute calls df_analyze for targets with delay slots, because
var_location instructions are present in the RTL stream at this point.
Tested on x86_64-suse-linux, applied on the mainline as obvious.
2016-0
On 06/09/2016 10:03 AM, Alan Hayward wrote:
This patch fixes PR 71407 by ensuring the BIT_FIELD_REF is created using
the
given vectype and then casted to the result type.
Tested on x86 and aarch64.
Note that the test vectorizes on x86 but does not vectorize on aarch64 or
power (due to a != state
I would think that when we see a duplicated base we should just drop
the duplicates and continue.
If the type of b1 is error_mark_node, why isn't the type of b2 also
error_mark_node?
Jason
On Thu, Jun 9, 2016 at 1:01 PM, Paolo Carlini wrote:
> Hi again,
>
> thus today I had to revert my first
input.c's selftest::test_reading_source_line attempted to read from
__FILE__, which doesn't work if the binary is run from a different
location than the build dir.
Fix it by rewriting the test to write out a tempfile, and read from
that, rather than from __FILE__.
I used make_temp_file to create
On 06/09/2016 12:34 PM, Martin Sebor wrote:
Attached is an updated version of the original patch described
below to annotate with the nonnull attribute the Built-In Functions
to Perform Arithmetic with Overflow Checking.
Since the machinery that's in place doesn't handle the attribute
on type-ge
On 05/29/2016 08:37 AM, Bernd Edlinger wrote:
gcc:
2016-05-22 Bernd Edlinger
PR inline-asm/68843
* reg-stack.c (check_asm_stack_operands): Explicit input arguments
must be grouped on top of stack. Don't force early clobber
on ordinary reg outputs.
testsuit
On Thu, Jun 09, 2016 at 04:06:53PM -0400, Michael Meissner wrote:
> I'm including the global reviewers on the list. I just want to be sure that
> there is no problem installing these patches on the GCC 6.2 branch. While it
> is technically an enchancement, it is needed to be able to install the g
On 05/27/2016 11:44 AM, David Malcolm wrote:
On Fri, 2016-05-27 at 13:29 +0100, Thomas Preudhomme wrote:
Hi Rainer,
On Wednesday 25 May 2016 11:31:12 Rainer Orth wrote:
David Malcolm writes:
The following fixes the known failures of the must-tail-call
tests.
Tested with --target=
* aarch64-
On 05/29/2016 08:33 AM, JonY wrote:
On 5/23/2016 16:56, JonY wrote:
On 5/20/2016 06:36, JonY wrote:
On 5/20/2016 02:11, Jeff Law wrote:
So if we make this change (revert 227962), my understanding is that
cygwin bootstraps will fail because they won't find kernel32 and perhaps
other libraries.
On 06/09/2016 12:35 PM, Marcin Baczyński wrote:
2016-06-09 18:35 GMT+02:00 Jeff Law :
On 05/30/2016 05:19 PM, Marcin Baczyński wrote:
Hi,
this is my first GCC patch, so please bear with me if something is wrong
with
it in an obvious way.
I've found two unused macros in gcc/diagnostic.h. Is th
On 8 June 2016 at 20:38, Jan Hubicka wrote:
>> I think it would be nice to work towards transitioning
>> flag_section_anchors to a flag on varpool nodes, thereby removing
>> the Optimization flag from common.opt:fsection-anchors
>>
>> That would simplify the walk over varpool candidates.
>
> Makes
> On 8 June 2016 at 20:38, Jan Hubicka wrote:
> >> I think it would be nice to work towards transitioning
> >> flag_section_anchors to a flag on varpool nodes, thereby removing
> >> the Optimization flag from common.opt:fsection-anchors
> >>
> >> That would simplify the walk over varpool candidate
On Thu, 9 Jun 2016, Nathan Sidwell wrote:
> > (define_expand "restore_stack_block"
> >[(match_operand 0 "register_operand" "")
> > (match_operand 1 "register_operand" "")]
>
> you've not addressed my previous comments about this.
To be clear -- do you mean that "restore_stack_block" shou
I'm including the global reviewers on the list. I just want to be sure that
there is no problem installing these patches on the GCC 6.2 branch. While it
is technically an enchancement, it is needed to be able to install the glibc
support that is needed to complete the work to add IEEE 128-bit flo
[ Sorry I missed this ]
On Thu, Jun 02, 2016 at 02:51:37PM -0600, Kelvin Nilsen wrote:
> This patch adds the RS6000_BTM_MODULO flag to to the set of flags
> associated with the RS6000_BTM_COMMON variable.
>
> This patch has bootstrapped with the trunk and the gcc-6-branch on
> powerpc64le-unknow
On Thu, Jun 09, 2016 at 10:12:53AM -0600, Jeff Law wrote:
> I'm going to largely let Bernd own the review on this. Just a few comments.
>
> I certainly like the concept. My mental model is that parts of the
> prologue might sink further than other parts of the prologue. It's not
> an exact ma
On 06/02/16 17:22, Alexander Monakov wrote:
On Wed, 25 May 2016, Nathan Sidwell wrote:
It seems like we should reject the combination of -msoft-stack -fopenacc?
Possibly; the doc text makes it explicit that the option is exposed only for
the purpose of testing the compiler, anyway.
It is alw
2016-06-09 18:35 GMT+02:00 Jeff Law :
> On 05/30/2016 05:19 PM, Marcin Baczyński wrote:
>>
>> Hi,
>> this is my first GCC patch, so please bear with me if something is wrong
>> with
>> it in an obvious way.
>>
>> I've found two unused macros in gcc/diagnostic.h. Is the patch okay as is?
>> Bootstra
Attached is an updated version of the original patch described
below to annotate with the nonnull attribute the Built-In Functions
to Perform Arithmetic with Overflow Checking.
Since the machinery that's in place doesn't handle the attribute
on type-generic built-ins changes to the (handle_nonnul
On Thu, Jun 09, 2016 at 04:25:32PM +, Joseph Myers wrote:
> On Thu, 9 Jun 2016, Jeff Law wrote:
>
> > On 05/26/2016 09:04 AM, David Malcolm wrote:
> > > When using config-list.mk to build all configurations, it's useful
> > > to filter out the configurations that are known to be broken.
> > >
gcc/ChangeLog:
* pretty-print.c (assert_pp_format_colored): Skip the test if
GCC_COLORS is set.
(test_pp_format): Remove comment about GCC_COLORS.
---
gcc/pretty-print.c | 7 ++-
1 file changed, 6 insertions(+), 1 deletion(-)
diff --git a/gcc/pretty-print.c b/gcc/prett
This patch introduce a selftest::location struct to wrap up __FILE__
and __LINE__ information (and __FUNCTION__) throughout the selftests,
allowing location information to be passed around.
It updates the helper functions in pretty-print.c to pass through
the precise location of each test, so that
Rework ASSERT_STREQ so that it prints the actual and expected values
to stderr when it fails (by moving it to a helper function).
gcc/ChangeLog:
* selftest.c (selftest::fail_formatted): New function.
(selftest::assert_streq): New function.
* selftest.h (selftests::fail_form
PR bootstrap/71471 highlighted the need for selftests to provide more
information when they fail. The report contained:
src/gcc/pretty-print.c:1246: FAIL: ASSERT_STREQ (expected, pp_formatted_text
(&pp))
This showed a string-equality failure within the helper function
assert_pp_format_va, but
On Thu, Jun 9, 2016 at 1:53 PM, David Malcolm wrote:
> On Thu, 2016-06-09 at 11:22 -0600, Jeff Law wrote:
>> On 06/09/2016 07:30 AM, David Edelsohn wrote:
>> >
>> > The self-tests specifically abort the build and break bootstrap
>> > upon
>> > failure. Most other changes that inadvertently have b
On Thu, 2016-06-09 at 11:22 -0600, Jeff Law wrote:
> On 06/09/2016 07:30 AM, David Edelsohn wrote:
> >
> > The self-tests specifically abort the build and break bootstrap
> > upon
> > failure. Most other changes that inadvertently have bugs or tickle
> > a
> > latent issue in a target will introd
On Thu, Jun 9, 2016 at 4:01 PM, Koval, Julia wrote:
> Hi,
>
> Here is a trivial patch, that fixes tests for these targets. Ok for trunk?
>
> gcc/testsuite/ChangeLog:
> * gcc.target/i386/interrupt-12.c: Fix test for -fpic and corei7.
> * gcc.target/i386/interrupt-13.c: Likewise.
>
On 06/09/2016 05:18 AM, Martin Liška wrote:
Hello.
The patch caused:
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=71466
THanks. I've got a few issues to address related to that change -- I've
just been swamped with some personal stuff the last week.
I'm seriously considering reverting the ch
Hi, Bruce!
I thought about a regex, but the aliases require a full function
signature and the original, narrow context does not provide a function
signature. If it was just alias XXXalloc as __linux_XXXalloc, it
would be more straight forward. If there's a convenient way to add
the other informa
On 06/09/2016 11:25 AM, David Edelsohn wrote:
AIX has added variants of malloc, realloc, calloc and valloc with
greater compatibility with Linux semantics, especially for NULL
addresses. The variants are declared in stdlib.h and use #define to
override the normal definition if _LINUX_SOURCE_COMP
AIX has added variants of malloc, realloc, calloc and valloc with
greater compatibility with Linux semantics, especially for NULL
addresses. The variants are declared in stdlib.h and use #define to
override the normal definition if _LINUX_SOURCE_COMPAT is defined,
e.g.,
#define malloc __linux_mal
On 06/09/2016 10:45 AM, David Malcolm wrote:
I was confused by the comment to pp_format:
/* The following format specifiers are recognized as being client independent:
...
%p: pointer
... */
into thinking that %p is printed in a host-independent manner, when
"client independent
On 06/09/2016 07:30 AM, David Edelsohn wrote:
The self-tests specifically abort the build and break bootstrap upon
failure. Most other changes that inadvertently have bugs or tickle a
latent issue in a target will introduce some additional testsuite
failures, not a bootstrap failure. x86 devel
> On 06/07/2016 09:27 PM, Jan Hubicka wrote:
> > There are bugs in few predictors - goto predictor is dead because the FE
> > code was dropped,
> > return predictor is bit random because CFG is optimized (it should probably
> > be done in FE),
> > loop iv compare seems bogus and fortran fail allo
On Thu, Jun 09, 2016 at 07:01:47PM +0200, Paolo Carlini wrote:
> Now the first comment I have about this, about which I'd like to have some
> feedback, is that the gcc_unreachable () at the end of
> build_simple_base_path without a return seems a bit weird to me, because the
> function is *supposed
Hi again,
thus today I had to revert my first try at resolving this error recovery
issue because it caused c++/71465. In retrospect it was a bit heavy
handed anyway, the zeroing of the type served us well for many years...
Today I tried to investigate the issue a bit more. I remind you that w
On Thu, Jun 09, 2016 at 07:53:52PM +0300, Alexander Monakov wrote:
> I'm sending updated patch series with backend prerequisites for OpenMP
> offloading to the NVIDIA PTX ISA. The first patch has already received some
> comments and this version reflects review feedback. The other patches have
>
On 06/02/2016 10:53 AM, James Greenhalgh wrote:
Hi,
When I was working in the area last year, I promised to revisit the cost
model for noce if-conversion and see if I could improve the modeling. This
turned out to be more tricky than I expected.
This patch set rewrites the cost model for noce i
This patch implements -muniform-simt code generation option, which is used to
emit code for OpenMP offloading. The goal is to emit code that can either
execute "normally", or can execute in a way that keeps all lanes in a given
warp active, their local state synchronized, and observable effects fr
This is a respin of the recently reviewed -msoft-stack patch that addresses
review feedback and adds required libgcc changes.
gcc/:
* config/nvptx/nvptx-protos.h (nvptx_output_set_softstack): Declare.
* config/nvptx/nvptx.c: (need_softstack_decl): New variable.
(init_softst
Hi,
I'm sending updated patch series with backend prerequisites for OpenMP
offloading to the NVIDIA PTX ISA. The first patch has already received some
comments and this version reflects review feedback. The other patches have
been adjusted for clarity and re-cut in a more rigorous manner. All p
This patch implements emission of OpenMP target region entrypoints: the
compiler emits the target function with '$impl' appended to the name, and
under the original name it emits a short entry sequence that sets up shared
memory arrays and calls the target function via 'gomp_nvptx_main' (which is
i
This patch adds option -mgomp which enables -msoft-stack plus -muniform-simt,
and builds a multilib with it. This codegen convention is used for OpenMP
offloading.
* config/nvptx/nvptx.c (nvptx_option_override): Handle TARGET_GOMP.
* config/nvptx/nvptx.opt (mgomp): New option.
This patch implements in nvptx.md a few new instruction patterns that are used
for OpenMP SIMD code.
* config/nvptx/nvptx-protos.h (nvptx_shuffle_kind): Move enum
declaration from nvptx.c.
(nvptx_gen_shuffle): Declare.
* config/nvptx/nvptx.c (nvptx_shuffle_kind): Mo
This patch wires up generation of predicated instruction forms in nvptx.md and
fixes their handling in nvptx.c. This is a prerequisite for the following
patch. On its own it doesn't affect generated code because COND_EXEC
instructions are created by if-conversion only after register allocation,
w
This patch adds a new target hook and implements it in a straightforward
manner on NVPTX to indicate that the target is running in SIMT fashion with 32
threads in a synchronous group ("warp"). For use in OpenMP transforms.
* config/nvptx/nvptx.c (nvptx_simt_vf): New.
(TARGET_SIMT_
This patch wires up use of alternative -mgomp multilib for OpenMP offloading
via nvptx mkoffload. It makes OpenACC and OpenMP incompatible for
simultaneous offloading compilation, so I've added a diagnostic for that.
* config/nvptx/mkoffload.c (main): Check that either OpenACC or OpenMP
On Thu, Jun 09, 2016 at 06:43:04PM +0200, Jakub Jelinek wrote:
> Yes, I'm all in favor in disabling X constraint for inline asm.
> Especially if people actually try to print it as well, rather than make it
> unused. That is a sure path to ICEs.
Though, on the other side, even our documentation me
On Thu, Jun 09, 2016 at 10:30:13AM -0600, Jeff Law wrote:
> On 06/06/2016 01:40 PM, Jakub Jelinek wrote:
> >On Mon, Jun 06, 2016 at 09:27:56PM +0200, Marc Glisse wrote:
> >>The last one would miss floating point registers (no 2 platforms use the
> >>same letter for those, hence my quest for somethi
On 05/30/2016 05:19 PM, Marcin Baczyński wrote:
Hi,
this is my first GCC patch, so please bear with me if something is wrong with
it in an obvious way.
I've found two unused macros in gcc/diagnostic.h. Is the patch okay as is?
Bootstrapped on x86_64-pc-linux-gnu.
2016-05-31 Marcin Baczyński
Ping:
Any further comments to the patch set?
Thanks,
Pekka
On Wed, May 18, 2016 at 8:01 PM, Pekka Jääskeläinen wrote:
> Hi Joseph,
>
> Updated diffstat below:
> Makefile.def | 3 +
> Makefile.in | 489 +
> configu
On 06/03/2016 09:41 AM, Alan Hayward wrote:
check_GNU_style.sh fails to detect lines >80 chars on BSD / Mac OS X
systems.
This is becuase paste is being called with an empty delimiter list.
Instead \0 should be used.
Tested on Ubuntu 14.04 and OS X 10.9.5
contrib/
* check_GNU_style.sh:
On 06/06/2016 01:40 PM, Jakub Jelinek wrote:
On Mon, Jun 06, 2016 at 09:27:56PM +0200, Marc Glisse wrote:
The last one would miss floating point registers (no 2 platforms use the
same letter for those, hence my quest for something more generic).
The goal of the experiment is described in PR5915
On 06/07/2016 11:58 AM, Bernd Edlinger wrote:
AFACT this is not the only place where overly complex RTL trees can
cause an ICE.
That wouldn't surprise me at all -- but the design of RTL is such that
it can be arbitrarily complex. Essentially, routines can not make
assumptions about the comple
On Thu, 9 Jun 2016, Jeff Law wrote:
> On 05/26/2016 09:04 AM, David Malcolm wrote:
> > When using config-list.mk to build all configurations, it's useful
> > to filter out the configurations that are known to be broken.
> >
> > This patch does so, adding a KNOWN_BROKEN variable.
> >
> > contrib/
I was confused by the comment to pp_format:
/* The following format specifiers are recognized as being client independent:
...
%p: pointer
... */
into thinking that %p is printed in a host-independent manner, when
"client independent" means in relation to different pretty-printer
On 06/07/2016 07:47 PM, Segher Boessenkool wrote:
This patch series introduces separate shrink-wrapping.
There are many things the prologue/epilogue of a function do, and most of
those things can be done independently. For example, most of the time,
for many targets, the save of callee-saved re
This patch fixes PR 71407 by ensuring the BIT_FIELD_REF is created using
the
given vectype and then casted to the result type.
Tested on x86 and aarch64.
Note that the test vectorizes on x86 but does not vectorize on aarch64 or
power (due to a != statement failing to vectorize)
Ok to commit?
g
On 06/07/2016 08:44 PM, Martin Sebor wrote:
Ping: This is a trivial patch to make the text of an error message
for an insufficient number of arguments to a built-in consistent
across different built-ins. Rather that sometimes saying "not
enough arguments" and others "too few arguments" the patch
On 06/09/2016 03:12 AM, Martin Liška wrote:
Hello.
I've prepared patch for the PR that is very similar to what was done in PR65044.
I've been currently testing the patch on x86_64-linux.
Ready to be installed after it finishes?
OK.
jeff
On 06/08/2016 10:47 AM, Jiong Wang wrote:
As discussed on the PR
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=70751,
here is the patch.
For this particular failure on arm, *arm_movsi_insn has the following
operand
constraints:
operand 0: "=rk,r,r,r,rk,m"
operand 1: "rk, I,K,j,mi,rk"
g
On Tue, 24 May 2016, Robert Suchanek wrote:
> The following changes the default behaviour of shift splitting
> for MIPS16 e.g. the shifts will be split only when used with
> undocumented -mno-debugd option that is now switched on by default.
>
> This appears to enable better optimization in certa
> On Jun 8, 2016, at 9:05 AM, Richard Biener wrote:
>
> On Wed, 8 Jun 2016, Bill Schmidt wrote:
>
>> Hi Richard,
>>
>>> On Jun 8, 2016, at 7:29 AM, Richard Biener
>>> wrote:
>>>
>>> On Wed, Jun 13, 2012 at 4:18 AM, William J. Schmidt
>>> wrote:
This patch is a follow-up to the discuss
Hello,
A number of tests were added to check for FP16 arguments and return
values being passed in registers. These require mfloat-abi=hard to be
selected but in some test configurations they were run with
-mfloat-abi=soft or -mfloat-abi=softfp.
Explict skip-if directives are added to the tests t
On Wed, 25 May 2016, Robert Suchanek wrote:
> > > diff --git a/gcc/doc/invoke.texi b/gcc/doc/invoke.texi
> > > index 73f1cb6..2f6195e 100644
> > > --- a/gcc/doc/invoke.texi
> > > +++ b/gcc/doc/invoke.texi
> > > @@ -837,6 +837,7 @@ Objective-C and Objective-C++ Dialects}.
> > > -mips16 -mno-mips
On 05/26/2016 09:04 AM, David Malcolm wrote:
config-list.mk currently requires the pwd to be in a sibling directory
of the source tree. However, building using config-list.mk can consume
over 400GB of disk space in this build directory (e.g. my machine
successfully built cc1 for 206 configuratio
On 05/26/2016 09:04 AM, David Malcolm wrote:
r233165 marked three deprecated rtems targets as obsolete.
r233887 marked mep-elf as obsolete.
Update config-list.mk to add OPT-enable-obsolete to these 4
targets.
contrib/ChangeLog:
* config-list.mk (FULL_LIST): Add OPT-enable-obsolete to
On 05/26/2016 09:04 AM, David Malcolm wrote:
When using config-list.mk to build all configurations, it's useful
to filter out the configurations that are known to be broken.
This patch does so, adding a KNOWN_BROKEN variable.
contrib/ChangeLog:
* config-list.mk (LIST): Rename to...
On 06/07/2016 09:27 PM, Jan Hubicka wrote:
> There are bugs in few predictors - goto predictor is dead because the FE code
> was dropped,
> return predictor is bit random because CFG is optimized (it should probably
> be done in FE),
> loop iv compare seems bogus and fortran fail alloc does not s
On Thu, Jun 9, 2016 at 1:43 AM, Kyrill Tkachov
wrote:
> On 03/06/16 21:32, Evandro Menezes wrote:
>> Shouldn't this read "The Qualcomm QDF24xx processors are now supported via
>> the"?
>>
> I used Jim's suggestion at
> https://gcc.gnu.org/ml/gcc-patches/2016-04/msg01423.html.
> I don't have a stro
Hi,
I have applied this change, as obvious.
gcc/
* config/mips/mips.c (mips_output_jump): Fix formatting.
Maciej
gcc-mips-output-jump-format.diff
Index: gcc/config/mips/mips.c
===
--- gcc/config/mips/mips.c
On Jun 8, 2016, at 6:14 AM, Senthil Kumar Selvaraj
wrote:
>
> This patch requires int32plus support for a few more tests - these
> were failing for the avr target.
> If this is ok, could someone commit please? I don't have commit access.
Ok.
Committed revision 237266.
> 2016-06-08 Senthi
On Jun 8, 2016, at 4:20 AM, Senthil Kumar Selvaraj
wrote:
>
> I forgot to send this testcase modification with that patch - here's
> the fix for making gcc.dg/stack-usage-1.c pass again for avr.
>
> If this is ok, could someone commit please? I don't have commit
> access.
Ok.
Committed re
Hi,
Here is a trivial patch, that fixes tests for these targets. Ok for trunk?
gcc/testsuite/ChangeLog:
* gcc.target/i386/interrupt-12.c: Fix test for -fpic and corei7.
* gcc.target/i386/interrupt-13.c: Likewise.
* gcc.target/i386/interrupt-15.c: Likewise.
* gcc.t
Hi,
the patch below is the main part of the series that enhances the
existing gridification code to pattern-match sequences of constructs
in which the distribute, parallel and a loop constructs are not in one
simple combined construct but the step in the distribute loop exactly
matches the iterati
Hi,
the patch below allows emit_insn_operands to instructions with no
operands gracefully. Apparently so far we have not produced any.
I'll commit this to the hsa branch in a few moments and then to trunk
at some point in summer.
Martin
2016-06-02 Martin Jambor
* hsa-brig.c (emit_i
this patch reorganizes the class hierarchy we use to represent HSA
branching and synchronization instructions. The ultimate goal is to
find a way of representing a barrier instruction which shares the same
kind with branches.
It basically renames hsa_insn_br, which we have used to represent only
Hi,
the following patch adds the capability to put local HSA variables
into the group or global segment as indicated by new declaration
attributes. In the process I had to fix how we differentiate between
local and global attributes because context and allocation properties
now can diverge.
This
Hi,
this patch adds two hsa builtins and code for emission of HSAIL for
them and for BUILT_IN_GOMP_BARRIER. These built-ins are going to be
introduced to GIMPLE IL at OMP expansion time by the subsequent patch
in the series. I plan to experiment with adding more builtins for
special HSAIL instru
Hi,
this patch series, currently intended just for the branch but
eventually also for trunk in time for gcc 7, enables gridification,
that is the expansion of OpenMP loops for HSA GPUs, to work with
separate distribute and loop construct, provided that the step size of
the distribute loop is equal
Hi,
when I added HSA-only builtins to the HSA branch, I added them only
conditionally if HSA was enabled, which however broke non-HSA
compilation. This patch fixes this be removing the ifdefs.
Now that we have reorganized HSA so that all GTY stuff is in one file
(hsa.c), the right thing to do is
On Thu, Jun 9, 2016 at 9:10 AM, Jakub Jelinek wrote:
> On Thu, Jun 09, 2016 at 09:02:27AM -0400, David Edelsohn wrote:
>> On Thu, Jun 9, 2016 at 8:48 AM, Bernd Schmidt wrote:
>> > On 06/09/2016 02:21 PM, David Edelsohn wrote:
>> >
>> >> This is a completely unacceptable way to introduce these sel
Hi,
This patch teaches cgraphunit.c:output_in_order to output undefined external
variables via assemble_undefined_decl. At the moment that is only done for
-ftoplevel-reorder in varpool.c:symbol_table::output_variables. This patch
makes both behave the same way. I've also made handling of varia
On Thu, Jun 09, 2016 at 09:02:27AM -0400, David Edelsohn wrote:
> On Thu, Jun 9, 2016 at 8:48 AM, Bernd Schmidt wrote:
> > On 06/09/2016 02:21 PM, David Edelsohn wrote:
> >
> >> This is a completely unacceptable way to introduce these self-tests.
> >> Please stop adding self-tests that only are te
On Thu, Jun 9, 2016 at 8:48 AM, Bernd Schmidt wrote:
> On 06/09/2016 02:21 PM, David Edelsohn wrote:
>
>> This is a completely unacceptable way to introduce these self-tests.
>> Please stop adding self-tests that only are tested on x86 Linux and
>> cause bootstrap failures.
>
>
> We have no requir
On 06/09/2016 02:21 PM, David Edelsohn wrote:
This is a completely unacceptable way to introduce these self-tests.
Please stop adding self-tests that only are tested on x86 Linux and
cause bootstrap failures.
We have no requirement to test patches on more than one target. I think
your request
On Thu, Jun 09, 2016 at 02:40:43PM +0200, Christophe Lyon wrote:
> > Bet it depends if this happens before the signal(SIGILL, sig_ill_handler);
> > call or after it. If before, then I guess you'd better rewrite the
> > long long a = 0, b = 1;
> > asm ("vorr %P0, %P1, %P2"
> > : "=
On 9 June 2016 at 14:31, Jakub Jelinek wrote:
> On Thu, Jun 09, 2016 at 02:18:44PM +0200, Christophe Lyon wrote:
>> On 8 June 2016 at 16:50, Jakub Jelinek wrote:
>> > On Wed, Jun 08, 2016 at 04:44:00PM +0200, Christophe Lyon wrote:
>> >> I've tried the attached patch (which does only dg-options -
> On Thu, 9 Jun 2016, Jan Hubicka wrote:
>
> > Hi,
> > after we read the profile, we know expected number of iterations.
>
> We know the average ;) It may make sense to add some histogram
> value profiling for niter now that we should easily able to do so.
I always interpreted the estimated num
On Thu, Jun 09, 2016 at 02:18:44PM +0200, Christophe Lyon wrote:
> On 8 June 2016 at 16:50, Jakub Jelinek wrote:
> > On Wed, Jun 08, 2016 at 04:44:00PM +0200, Christophe Lyon wrote:
> >> I've tried the attached patch (which does only dg-options ->
> >> dg-additional-options).
> >> For GCC, it's be
1 - 100 of 121 matches
Mail list logo