This was ... interesting. There were a couple of problems that triggered ICEs.
This patch fixes the reported file (I made sure this time) and causes no
regressions as far as I can tell.
Dominique ... merci de votre patience.
Louis
Index: gcc/fortran/ChangeLog
=
Hi!
The C FE apparently relies on c_parser_unary_expression never called
if the expression to be parsed is actually a cast expression, otherwise
it fails an assertion.
c_parser_cast_expression of course guarantees that and during
sizeof and __alignof__ parsing too, but c_parser_omp_atomic calls
c_
Hi!
The first 3 hunks are quite obvious, of course we have to complain if any
of those conditions are non-zero, rather than only when all of them are
non-zero (which never happens). The last hunk fixes ICE if
undeclared_variable is called during parsing of #pragma omp declare simd
clauses, where
Hi!
If some OpenMP clause is parsed including the argument, such that
OMP_CLAUSE tree is created, but there is some error afterwards,
the C FE fails to remove the clause if it is not allowed for the
current construct, which results in ICEs later on (e.g. during
clause splitting). While it would b
Hi!
The following testcase ICEs because the collapse argument is type dependent
and has NULL TREE_TYPE, so testing INTEGRAL_TYPE_P on it ICEs.
I hope all INTEGER_CSTs have non-NULL type and thus just swapping the
conditions is enough, if not, I could add !type_dependent_expression_p (num)
test.
B
On 2015/9/9 04:02 AM, Cesar Philippidis wrote:
> This patch forces GOACC_LOCK to use locks in global memory regardless if
> the lock us for a worker or a gang. We were using a shared memory for
> worker locks, but we ran into an issue with that would sporadically
> involve deadlocks in worker reduc
Ping.
On 2015/8/27 09:44 PM, Chung-Lin Tang wrote:
> We've discovered that, for several of the libgomp plugin interface routines,
> if the target specific routine calls exit() (usually upon a fatal condition),
> deadlock ensues. We found this using nvptx, but it's possible on intelmic as
> well.
On Tue, Sep 8, 2015 at 5:49 PM, Ilya Enkovich wrote:
> On 21 Aug 10:38, Jeff Law wrote:
>> On 08/21/2015 07:44 AM, Ilya Enkovich wrote:
>> >>Our of curiosity, what does LLVM do here in terms of costing
>> >>models?
>> >
>> >Unfortunately I have no idea where and how LLVM does this
>> >optimization
On Wed, Sep 9, 2015 at 10:12 AM, Uros Bizjak wrote:
> On Tue, Sep 8, 2015 at 5:49 PM, Ilya Enkovich wrote:
>> On 21 Aug 10:38, Jeff Law wrote:
>>> On 08/21/2015 07:44 AM, Ilya Enkovich wrote:
>>> >>Our of curiosity, what does LLVM do here in terms of costing
>>> >>models?
>>> >
>>> >Unfortunately
Hi,
This patch clears up some remaining confusion in the vector lane orderings
for the two intrinsics mentioned in the title.
Bootstrapped on aarch64-none-linux-gnu and regression tested for
aarch64_be-none-elf with no issues.
OK?
Thanks,
James
---
2015-09-09 James Greenhalgh
* co
On 09/09/15 01:15, Jonathan Roelofs wrote:
On 9/4/15 12:20 AM, Yury Gribov wrote:
On 09/03/2015 07:45 PM, Jonathan Roelofs wrote:
On 9/3/15 10:17 AM, Jakub Jelinek wrote:
On Thu, Sep 03, 2015 at 10:15:02AM -0600, Jonathan Roelofs wrote:
+kcc, mrs
Ping
On 8/27/15 4:44 PM, Jonathan Roelof
Hi Jeff,
On Tue, 8 Sep 2015 13:27:12, Jeff Law wrote:
>
> On 09/07/2015 07:46 AM, Bernd Edlinger wrote:
>> Hi,
>>
>> On Mon, 7 Sep 2015 12:07:00, Marek Polacek wrote:
>>>
>>> On Sun, Sep 06, 2015 at 07:21:13PM +0200, Bernd Edlinger wrote:
Hi,
we observed sporadic failures of the fol
Kyrill Tkachov writes:
> On 08/09/15 10:26, Rainer Orth wrote:
>> Hi Kyrill,
>>
>>> PR rtl-optimization/67481 is a testsuite regression on sparc-solaris that
>>> Rainer reported. I haven't tested
>>> that this patch fixes that, but I suspect that the root cause is the
>>> same. Rainer, could you
On 9 September 2015 at 10:31, James Greenhalgh wrote:
>
> Hi,
>
> This patch clears up some remaining confusion in the vector lane orderings
> for the two intrinsics mentioned in the title.
>
> Bootstrapped on aarch64-none-linux-gnu and regression tested for
> aarch64_be-none-elf with no issues.
>
Hmmm, hang on. I'm not quite sure what the actual issue/bug is here, but is this
the same issue as my patch 12 "with BE RTL fix"?
(https://gcc.gnu.org/ml/gcc-patches/2015-08/msg01482.html, explanation last at
https://gcc.gnu.org/ml/gcc-patches/2015-07/msg02365.html) I pushed this as
r227551 las
The Aarch64 instruction patterns for atomic operations on memory use
the same constraints for arithmetic and logical operations despite the
arithmetic operations requiring stronger restrictions. This causes an
ICE in some circumstances.
This patch backports the fix from trunk to the GCC-5 branch.
On Wed, Sep 9, 2015 at 12:12 AM, Ian Lance Taylor wrote:
> Mike Stump writes:
>
>> Not a big issue, but slightly better if (O_CLOEXEC>>32) != 0 is also
>> true. See, if AIX should ever define this to a sensible value, the
>> above would disappear the feature. However, if they did, then this
>>
On 08/24/2015 08:55 AM, Paolo Carlini wrote:
+ if (no_linkage_check (ftype, /*relaxed_p=*/false))
How about using the return value of no_linkage_check in the warning?
Jason
Hi,
with the new operand setter method, we've been attempting to set
definition of a register when we were adding it to the zeroth operand
of a phi node. Fortunately, there is an assert to detect the
situation. Fixed thusly, will commit to the branch after rudimentary
testing.
Thanks,
Martin
On 08/09/15 21:52 +0200, François Dumont wrote:
Very good approach, I will start moving light checks from _GLIBCXX_DEBUG
implementation to normal one then.
Great, thanks. I *think* my patch got most of the lightweight checks
already.
We could maybe include the irreflexivity checks, although it
We were crashing on this testcase in invert_tree_comparison because it
got BIT_AND_EXPR, but this function expects comparison codes only. In
this case pred_equal_p got two predicates: m != 1 and m & 1. By checking
the tcc_comparison first we don't ICE anymore and pred_equal_p correctly
says false
I noticed the atomic compare exchange was using 'and 1' to generate a bool. We
now have select, which makes it more obvious.
Sadly the middle end's expansion of of the builtin fails to make use of the
condition we just generated directly, instead generating a new one from the bool.
nathan
2
On Sep 8, 2015, at 9:41 PM, David Miller wrote:
> +#define TARGET_LRA_P hook_bool_void_true
Are we at the point there this should be the default, and old ports should just
define to false, if they really need to? I’m using nothing but LRA as well.
On 08/09/15 23:08, Jeff Law wrote:
On 09/08/2015 07:21 AM, Tom de Vries wrote:
[ was: Re: [RFC] Prevent unnecessary recompilation for trivial
params.def changes ]
On 08/09/15 14:03, Andreas Schwab wrote:
Tom de Vries writes:
After a subsequent rebuild I don't see anything being rebuild. So
On Sep 8, 2015, at 9:12 PM, Ian Lance Taylor wrote:
> Yes, I think this might be even better in code. How about something
> like
>
> /* On some versions of AIX O_CLOEXEC does not fit in int, so use a
> cast to force it. */
> descriptor = open (filename, (int) (O_RDONLY | O_BINARY | O_CLOE
Hi,
On 09/09/2015 03:00 PM, Jason Merrill wrote:
On 08/24/2015 08:55 AM, Paolo Carlini wrote:
+if (no_linkage_check (ftype, /*relaxed_p=*/false))
How about using the return value of no_linkage_check in the warning?
Agreed, more informative. The below, which I'm finishing testing, also
On Sep 3, 2015, at 9:45 AM, Jonathan Roelofs wrote:
> Moral of the story is: these tests fail in our environment, but only because
> the regexes do not expect the presence of the ansi color codes, and we can't
> trick the runtime into not emitting them.
When the user says, I don’t want color; i
On Wed, Sep 9, 2015 at 11:11 PM, Mike Stump wrote:
> On Sep 3, 2015, at 9:45 AM, Jonathan Roelofs
> wrote:
>> Moral of the story is: these tests fail in our environment, but only because
>> the regexes do not expect the presence of the ansi color codes, and we can't
>> trick the runtime into n
On 09/09/2015 08:02 AM, Marek Polacek wrote:
We were crashing on this testcase in invert_tree_comparison because it
got BIT_AND_EXPR, but this function expects comparison codes only. In
this case pred_equal_p got two predicates: m != 1 and m & 1. By checking
the tcc_comparison first we don't IC
On 09/09/2015 03:10 AM, Bernd Edlinger wrote:
Hi Jeff,
On Tue, 8 Sep 2015 13:27:12, Jeff Law wrote:
On 09/07/2015 07:46 AM, Bernd Edlinger wrote:
Hi,
On Mon, 7 Sep 2015 12:07:00, Marek Polacek wrote:
On Sun, Sep 06, 2015 at 07:21:13PM +0200, Bernd Edlinger wrote:
Hi,
we observed sporadic
On 09/04/2015 09:19 AM, Robert Suchanek wrote:
It appears that a possibly related issue is already reported to Bugzilla (bug
51513)
where the branch is not optimized away, leaving the compare and branch
instructions.
It would also appear that this should be fixed at the tree level, however,
On 09/08/2015 05:45 PM, Jonathan Wakely wrote:
>> I doubt we can achieve the complexity goals in all cases. I expect that
>>
>> for (int i = 0; i < 1; ++i) {
>>vector[i];
>> }
>>
>> is optimized away in default mode, but with _GLIBCXX_ASSERTIONS, it is
>> not.
>>
>> The last time I look
Tested on ppc64le.
OK for trunk and gcc-5?
PR sanitizer/67258
* ubsan/ubsan_type_hash.cc: Cherry pick upstream r244101.
Upstream patch:
commit 1d2477faafda9ad2cc19927b3c31efd22747f013
Author: Alexey Samsonov
Date: Wed Aug 5 19:35:46 2015 +
[UBSan] Fix UBSan-vptr false
On Thu, Jul 23, 2015 at 12:15 PM, Alex Velenko wrote:
> Hi,
>
> This patch prevents testcase pr63210.c from running with -march=armv4t.
> Object size check should be skipped with explicit -march=armv4t, because
> expected size is only correct using pop pc instruction which is unsafe for
> armv4t.
On Wed, Sep 09, 2015 at 06:18:25PM +0200, Markus Trippelsdorf wrote:
> Tested on ppc64le.
> OK for trunk and gcc-5?
>
> PR sanitizer/67258
> * ubsan/ubsan_type_hash.cc: Cherry pick upstream r244101.
Please add
-fno-sanitize-recover=vptr
to dg-options.
Ok with that change.
> --- /dev
OK, thanks.
Jason
This patch assigns any available gang, worker or vector level
parallelism to auto and independent loops inside acc parallel regions.
This is done in omplower for two reasons:
1. At the moment, it's too late to do this in oacc-xform because
ompexpand is responsible for partitioning loops. Th
Hi,
I tracked down this regression to r225621, a clean up committed by Jason
a while ago: unless we want to try something more aggressive, we can fix
it the regression by simply restoring a few lines in
cp_parser_template_parameter which consume the ellipsis. To clarify
implementation-wise, t
On Wed, 9 Sep 2015, Jonathan Wakely wrote:
> Gerald, I think we've had similar issues with these mirrors in the
> past as well, shall we just remove them from the list?
>
>> http://mirrors-ru.go-parts.com/gcc - Online Shop
>> ftp://mirrors-ru.go-parts.com/gcc - bad
>> rsync://mirrors-ru.go-parts.c
Hi Gerald,
It definitely was not. If it was, then we wouldn't keep the servers up
for a year+ (for some other mirrors, 2+ years)We just discontinued
our mirrors project a week or two ago, and haven't had time to contact
everyone yet to take the links down. We will not be returning. Tha
But we will keep the USA mirror up indefinitely.
On 9/9/15 10:00 AM, Dan Derebenskiy wrote:
Hi Gerald,
It definitely was not. If it was, then we wouldn't keep the servers
up for a year+ (for some other mirrors, 2+ years)We just
discontinued our mirrors project a week or two ago, and hav
I've applied this patch to port some cleanups, mainly formatting and loop idioms
from the gomp4 branch.
nathan
2015-09-09 Nathan Sidwell
* config/nvptx/nvptx.md (call_operation): Move bound out of loop.
(*cmp): Add assembler spacing.
(setcc_int, set_cc_float): Likewise.
* config/nvptx/nv
On 09/09/15 14:54 +0100, Jonathan Wakely wrote:
On 08/09/15 21:52 +0200, François Dumont wrote:
Very good approach, I will start moving light checks from _GLIBCXX_DEBUG
implementation to normal one then.
Great, thanks. I *think* my patch got most of the lightweight checks
already.
We could ma
Hi,
On Wed, 9 Sep 2015 09:31:33, Jeff Law wrote:
> You could probably make the function static or change its visibility via
> a function attribute (there's a visibility attribute which can take the
> values default, hidden protected or internal). Default visibility
> essentially means the function
I've applied this to gomp4, which ports a bunch of cleanups and fixes from
trunk.
nathan
2015-09-09 Nathan Sidwell
Port from trunk:
* config/nvptx/nvptx.md (allocate_stack): Emit sorry.
(atomic_compare_and_swap_1): Use sel_truesi.
* config/nvptx/nvptx.c (nvptx_write_function_decl): Refor
We should fix the logic in cp_parser_parameter_declaration to handle
this, not work around it here.
Jason
I've applied this, where I'd incorrectly compared nodes for equality, rather
than the node value.
nathan
2015-09-09 Nathan Sidwell
* omp-low.c (execute_oacc_transform): Use integer_zerop, not
direct node comparison.
Index: gcc/omp-low.c
Hi,
On 09/09/2015 08:42 PM, Jason Merrill wrote:
We should fix the logic in cp_parser_parameter_declaration to handle
this, not work around it here.
Ok, that's what I meant for "more aggressive" ;) I didn't try do that
immediately when I noticed that the function is called from a few other
pla
Please find attached the promised patch, which bootstraps and regtests
on x86_64.
OK for trunk?
Paul
2015-07-23 Paul Thomas
* module.c : Add static no_module_procedures.
(mio_symbol_attribute): Reset above if module procedure is
encountered.
(gfc_dump_module): Set above and e
On 09/04/2015 05:03 AM, Marek Polacek wrote:
This patch fixes a few -Wshift-negative-value warnings in the code. It
oughtn't to cause any functional changes. This will be needed for the
bootstrap to pass after I commit some other patch.
Bootstrapped/regtested on x86_64-linux, ok for trunk?
20
On 07/30/2015 04:14 PM, Moore, Catherine wrote:
This patch implements a more compact format for exception handling data.
Although I don't have recent numbers for the amount of compression achieved, an
earlier measurement showed a 30% reduction in the size of EH data for libstdc++.
A design do
> 2015-07-23 Paul Thomas
>
>* module.c : Add static no_module_procedures.
>(mio_symbol_attribute): Reset above if module procedure is
>encountered.
>(gfc_dump_module): Set above and exit without writing smod file
>if it reset.
>
> 2015-07-23 Paul Thomas
>
>* gfortra
On 08/31/2015 05:30 AM, Richard Biener wrote:
On Mon, Aug 31, 2015 at 7:49 AM, Mikhail Maltsev wrote:
Hi, all!
This patch removes some conditional compilation from GCC. In this patch I define
a macro CHECKING_P, which is equal to 1 when ENABLE_CHECKING is defined and 0
otherwise. The reason fo
... what about something this simple? Passes testing...
Thanks,
Paolo.
///
Index: cp/parser.c
===
--- cp/parser.c (revision 227600)
+++ cp/parser.c (working copy)
@@ -19626,11 +19626,12 @@ cp_parser_parameter
On 9/9/15 9:20 AM, Andrew Pinski wrote:
On Wed, Sep 9, 2015 at 11:11 PM, Mike Stump
wrote:
On Sep 3, 2015, at 9:45 AM, Jonathan Roelofs
wrote:
Moral of the story is: these tests fail in our environment, but
only because the regexes do not expect the presence of the ansi
color codes, and we
On 09/03/2015 09:34 AM, Jiong Wang wrote:
As Rainer reported at
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=67421
Also, as described at
https://gcc.gnu.org/ml/gcc-patches/2015-08/msg01147.html
This patch relax the restriction on wide left shift. Previously we
always honor target priva
On 09/06/2015 03:57 AM, Chung-Lin Tang wrote:
Hi,
Currently most non-target specific options are skipped when crossing the
LTO/offload processing border, however since there are still quite a
number of warning calls in many target backends, it makes sense to
save and propagate the associated opti
On 08/26/2015 11:29 PM, Ajit Kumar Agarwal wrote:
Thanks. The following testcase testsuite/gcc.dg/tree-ssa/ifc-5.c
void dct_unquantize_h263_inter_c (short *block, int n, int qscale,
int nCoeffs) { int i, level, qmul, qadd;
qadd = (qscale - 1) | 1; qmul = qscale << 1;
for (i = 0; i <= nCoeffs;
The following found 14 bugs in my code base. I think it is useful to
warn about such usage since they are bugsr. If the argument is marked
as nonnull then passing in a NULL argument will produce bad results
even if the code checks against NULL.
GCC might optimize such checks away so warn the user
On 09/02/2015 07:43 AM, Kyrill Tkachov wrote:
+ rtx_code code; + if (val_a == -val_b)
Do we have to worry about signed overflow here? I'm thinking
specifically when val_b is the smallest possible integer
representable by a HOST_WIDE_INT. I suspect you may be able to
avoid these problems wit
On 09/09/2015 03:44 PM, Mark Wielaard wrote:
The following found 14 bugs in my code base. I think it is useful to
warn about such usage since they are bugsr. If the argument is marked
as nonnull then passing in a NULL argument will produce bad results
even if the code checks against NULL.
GCC mi
prepare_call_address in calls.c is the wrong place to handle -fno-plt.
We shoudn't force function address into register and hope that load
function address via GOT and indirect call via register will be folded
into indirect call via GOT, which doesn't always happen. Also non-PIC
case can only be h
On Wed, Sep 09, 2015 at 04:01:07PM -0600, Jeff Law wrote:
> >* gcc.dg/nonnull-4.c: New test.
> >* g++.dg/warn/nonnull3.C: Likewise.
If the tests are the same, perhaps stick just one test into
c-c++-common/nonnull-1.c instead? Also, all the "cp1 compared to NULL"
strings mention cp
On Thu, 2015-09-10 at 00:03 +0200, Jakub Jelinek wrote:
> On Wed, Sep 09, 2015 at 04:01:07PM -0600, Jeff Law wrote:
> > >* gcc.dg/nonnull-4.c: New test.
> > >* g++.dg/warn/nonnull3.C: Likewise.
>
> If the tests are the same, perhaps stick just one test into
> c-c++-common/nonnull-1
On Wed, Sep 2, 2015 at 4:32 PM, Sriraman Tallam wrote:
>
> On Tue, Aug 18, 2015 at 9:46 PM, Cary Coutant wrote:
> >> Thanks, will make those changes. Do you recommend a different name
> >> for this flag like -fmake-comdat-functions-static?
> >
> > Well, the C++ ABI refers to this as "vague linka
The attached patch fixes a NULL pointer dereference. When
gfortran runs into an incomplete substring in an initialization
expression, she dereferences a NULL pointer. The patch checks
for NULL and returns false, which allows gfortran to issue a
sensible error message. Regression tested on x86_64
On 09/09/2015 01:35 PM, Jason Merrill wrote:
On 07/30/2015 04:14 PM, Moore, Catherine wrote:
This patch implements a more compact format for exception handling data.
Although I don't have recent numbers for the amount of compression achieved,
an earlier measurement showed a 30% reduction in the
Richard,
Here is a modified patch that takes your comments into account. Breaking
on depth == 0 with '>' does not work due to the code looking for whitespace.
2015-08-25 Michael Collison
PR other/57195
* read-md.c (read_name): Allow mode iterators inside angle
brackets in rtl e
On Tue, Sep 08, 2015 at 02:46:58PM +0200, Ulrich Weigand wrote:
> David Edelsohn wrote:
> > On Mon, Sep 7, 2015 at 11:47 PM, Alan Modra wrote:
> > > In https://gcc.gnu.org/bugzilla/show_bug.cgi?id=67378 analysis I show
> > > the reason for this PR is that insns emitted by secondary reload
> > > pa
On Wed, Sep 9, 2015 at 8:57 PM, Alan Modra wrote:
> On Tue, Sep 08, 2015 at 02:46:58PM +0200, Ulrich Weigand wrote:
>> David Edelsohn wrote:
>> > On Mon, Sep 7, 2015 at 11:47 PM, Alan Modra wrote:
>> > > In https://gcc.gnu.org/bugzilla/show_bug.cgi?id=67378 analysis I show
>> > > the reason for t
This patch by Chris Manghane fixes the Go frontend to not try to do a
constant shift of non-integer values. This fixes
https://golang.org/issue/11616 . Bootstrapped and ran Go testsuite on
x86_64-unknown-linux-gnu. Committed to mainline.
Ian
Index: gcc/go/gofrontend/MERGE
==
> 2015-09-09 Steven G. Kargl
>
> PR fortran/67526
> * gfortran.dg/pr67526.f90: New test.
>
> 2015-09-09 Steven G. Kargl
>
> PR fortran/67526
> * expr.c (gfc_check_init_expr): Do not dereference a NULL pointer.
OK.
Thanks for the patch!
72 matches
Mail list logo
