Hi!
On Wed, 23 Jul 2014 20:54:24 -0400, David Malcolm dmalc...@redhat.com wrote:
On Wed, 2014-07-23 at 15:00 +0200, Thomas Schwinge wrote:
In context of adding support for OpenACC, next to the existing
GIMPLE_OMP_TARGET (quoting from gcc/gimple.def and gcc/gimple.h):
As discussed in the audit trail both _Adaptor and generate_canonical are
both meant to use floating point values. Both are here given
static_asserts to that effect.
This would have prevented this error and might help future users.
The main issue is the use of value_type in _Adaptor and thus
Hi Cesar!
On Fri, 18 Jul 2014 14:07:00 -0700, Cesar Philippidis ce...@codesourcery.com
wrote:
This patch enables acc constructs to be used inside nested functions.
Thanks!
I
doubt nested functions will be used much in c, but some of the openacc
fortran tutorials I've seen online make use
On 29/07/14 04:11 -0400, Ed Smith-Rowland wrote:
As discussed in the audit trail both _Adaptor and generate_canonical
are both meant to use floating point values. Both are here given
static_asserts to that effect.
This would have prevented this error and might help future users.
The main
The following fixes the second bug you'll hit very fast when writing
testcases for -ftrapv (maybe not so often in real scenarios). There
are several places in the optimization pipeline where after constant
folding we fail to generate trapping instructions.
The simplest solution as implemented
Hi Cesar!
On Thu, 20 Mar 2014 15:42:48 +0100, I wrote:
On Tue, 18 Mar 2014 14:50:44 +0100, I wrote:
On Tue, 18 Mar 2014 16:37:24 +0400, Ilmir Usmanov i.usma...@samsung.com
wrote:
This patch introduces support of OpenACC loop directive (and combined
directives) in C front-end up to
On 07/28/2014 05:24 PM, Mike Stump wrote:
On Jul 28, 2014, at 3:29 AM, Gerald Pfeifer ger...@pfeifer.com wrote:
On Fri, 25 Jul 2014, Martin Liška wrote:
Yes, --enable-languages=all is the default, so it only builds the
frontends that are enabled by default, so it does not build the Go or
Ada
the patch athttps://gcc.gnu.org/ml/gcc-patches/2014-07/msg00961.html
actually has one glitch on if check.
thumb target is code size sensitive, the best solution is we set
prefer_callee_reg_p to
true if we know that shrink-wrap will be performanced if prefer_callee_reg_p is
true.
but
Hi all,
On 23 July 2014 11:26, Kyrill Tkachov kyrylo.tkac...@arm.com wrote:
On 22/07/14 17:29, Ramana Radhakrishnan wrote:
On Mon, Jul 14, 2014 at 11:11 AM, Jiong Wang jiong.w...@arm.com wrote:
currently the following testcases are disabled for arm target,
gcc.dg/ira-shrinkwrap-prep-1.c
On 29/07/14 10:21, Christophe Lyon wrote:
Hi all,
On 23 July 2014 11:26, Kyrill Tkachov kyrylo.tkac...@arm.com wrote:
Hi all,
I've committed this on Jiongs' behalf as r212927 with some minor ChangeLog
modifications to make it fit into the 80 character column limit:
2014-07-23 Jiong Wang
On Tue, Jul 29, 2014 at 10:19 AM, Jiong Wang jiong.w...@arm.com wrote:
the patch athttps://gcc.gnu.org/ml/gcc-patches/2014-07/msg00961.html
actually has one glitch on if check.
thumb target is code size sensitive, the best solution is we set
prefer_callee_reg_p to
true if we know that
I've tested Graphite with the ISL AST generator as the main code
generator and found out the following problem: in the current
implementation the gcc_expression_from_isl_ast_expr_id can return a
tree of a type, which doesn't correspond to the type chosen for
graphite expressions.
If we try, for
On 29/07/2014 12:14, Roman Gareev wrote:
I've tested Graphite with the ISL AST generator as the main code
generator and found out the following problem: in the current
implementation the gcc_expression_from_isl_ast_expr_id can return a
tree of a type, which doesn't correspond to the type chosen
Hello Marc,
On 26 Jul 19:34, Marc Glisse wrote:
I did some AVX and AVX512F intrinsics, and it still passes the
testsuite (on my old pre-AVX x86_64-linux-gnu).
I've performed testing of your patch using functional simulator of
AVX*. And see no regressions as well.
--
Thanks, K
This reverts part of the earlier patch for PR52478 which turns out
to be not necessary. It also revealed that at least switch
expansion builds new arithmetic in signed char - going the -ftrapv
expansion way (where we obviously don't have __addvqi3 as C
promotes everything to int). We probably
This re-organizes the LTO streamer to do compression transparently
in the data-streamer routines (and disables section compression
by defaulting to -flto-compression-level=0). This avoids
keeping the whole uncompressed sections in memory, only retaining
the compressed ones.
The downside is that
This version of the patch should pass the entire test suite.
g++.dg/parse/error50.C noted that wasn't handling the case of empty
template parameters properly (accepted silently). As a result,
grokvardecl now calls check_explicit_specialization more liberally and
may return NULL_TREE.
The
An error occurring in a subexpression that is part of some construct in
general suppresses the reporting of further errors on the same construct,
to avoid noisy cascaded messages. This patch ensures that this is also
the case when named associations are present.
The following test case must be
In conversions of prefixed calls involving interfaces, the expression in the
conversion may have a limited view of a type obtained transitively through
several contexts. Use the non-limited view if available, to enable subsequent
interface membership tests.
The following must compile quietly:
This patch ensures that abort-related expansion generates the same amount of
internal entities when aborts are allowed or are being suppressed by pragma
Restriction (No_Abort_Statements).
Tested on x86_64-pc-linux-gnu, committed on trunk
2014-07-29 Hristian Kirtchev kirtc...@adacore.com
This patch corrects the transient object machinery to disregard aliasing when
the associated context is a Boolean expression with actions. This is because
the Boolean result is always known after the action list has been evaluated,
therefore the transient objects must be finalized at that point.
From: Trevor Saunders tsaund...@mozilla.com
Hi,
this adds a hash_set wrapper around hash_table, and then replaces usage of
pointer_set with it.
bootstrapped +regtested on x86_64-unknown-linux-gnu, ok?
Trev
ada/
* gcc-interface/trans.c: Use hash_set instead of pointer_set.
c-family/
A new pragma Unevaluated_Use_Of_Old (Error | Warn | Allow) is
implemented which controls the processing of attributes Old and
Loop_Entry. If either of these attributes is used in a potentially
unevaluated expression e.g. the then or else parts of an if
expression), then normally this usage is
Hi,
Please find the patches for pr61868. The problem is that even with
-frandom-seed gcc always uses 0 instead of specified number/string for places
such as lto section names. init_random_seed is never called with -frandom-seed.
Bootstrapped and tested on x86-64. OK for trunk?
Thanks,
This is just an internal cleanup, involving some name changes
and slightly cleaned up testing of flags etc. This is part of
the preparation for implementing static real predicates. No
functional effect.
Tested on x86_64-pc-linux-gnu, committed on trunk
2014-07-29 Robert Dewar de...@adacore.com
This implements static predicates for string and real types,
as defined in the RM. There is one exception, which is that
the RM allows X ABC as being predicate static, but since
ABC ABA is not static, that's peculiar, so we assume
that this is a mistake in the RM, and that string comparisons
On Tue, 29 Jul 2014, Richard Biener wrote:
This re-organizes the LTO streamer to do compression transparently
in the data-streamer routines (and disables section compression
by defaulting to -flto-compression-level=0). This avoids
keeping the whole uncompressed sections in memory, only
Normally the default scalar storage order is the native order of the
target. This pragma, which can be either a configuration pragma, or
appear in a package spec or declarative part, can provide a default
value that overrides this normal default. If used in a package spec
or declarative part, it
Hello,
Ping for https://gcc.gnu.org/ml/gcc-patches/2014-07/msg00237.html
Thanks in advance for your feedback,
With Kind Regards,
Olivier
On Jul 3, 2014, at 16:57 , Olivier Hainque hain...@adacore.com wrote:
From gcc/i386/config/mingw32.h, STARTFILE_SPEC and ENDFILE_SPEC include
crtbegin.o
The Error option for pragma Unevaluated_Use_Of_Old was not properly
recognized, due to an internal problem with the generation of the
names table for the Snames package. This is now corrected, and
the following program compiles as shown with -gnatld7 -gnatj60:
1. package Uneval_Old is
This patch implements the following set of rules related to shared variables:
1. A volatile representation aspect may only be applied to an
object_declaration or a full_type_declaration.
2. A component of a non-volatile type declaration shall not be volatile.
3. A discriminant
If a procedure or entry has an formal out-parameter of a null-excluding access
type, there is no check applied to the actual before the call. This patch
removes a spurious access check on such parameters on entry calls.
Compiling and executing p.adb must yield;
Procedure version did not
On Mon, Jul 28, 2014 at 7:30 PM, David Malcolm dmalc...@redhat.com wrote:
On Mon, 2014-07-28 at 02:35 +0530, Prathamesh Kulkarni wrote:
- if (o-type == operand::OP_CAPTURE)
+ if (is_acapture * (o))
{
- capture *c = static_castcapture * (o);
- fprintf (f, @%s,
Hi,
in this diagnostic issue Jon noticed that for testcases like:
templateclass T1, class... Tn
void foo(T1, Tn...);
int main()
{
foo();
}
we provide diagnostic saying candidate expects 2 arguments, 0
provided, whereas of course we want to say something like candidate
expects at least 1
Hi,
As discussed here in https://gcc.gnu.org/ml/gcc/2014-01/msg00189.html
Gcc should complain about pragma omp declare target without
corresponding pragma omp end declare target. This patch adds a warning
for those cases.
Bootstraps/passes make-check.
Ok for trunk?
ChangeLog:
2014-07-29 Ilya
On Linux, s-oscons-tmplt.c needs to define _BSD_SOURCE in order for
CRTSCTS to be visible. Otherwise the macro is undefined, and defaulted
to -1.
Tested on x86_64-pc-linux-gnu, committed on trunk
2014-07-29 Thomas Quinot qui...@adacore.com
PR ada/60652
* s-oscons-tmplt.c: For
On 07/28/2014 07:22 AM, Jan Hubicka wrote:
We decide what we can devirtualize in can_refer_decl_in_current_unit_p.
So the problem is that we have functions that should be COMDAT but they are not
because we decide to not produce their body (believing it is not reachable).
Can't
we arrange them
this patch optimize copysign/copysignf for -mfloat-abi=soft on arm when BFI
instruction is available.
before this patch, we do copysign (a, b) by three steps:
* fetch the sign bit of b to A
* fetch all non-sign bits of a to B
* or A and B
while these three steps could be finished by one
For formal verification with GNATprove, frontend inlining can be used to
relieve users from the need to add contracts to local subprograms. Thus,
we adopt here a simple policy for inlining in GNATprove mode, which consists
in inlining all local subprograms which can be inlined, as soon as they
..after cleaning up the conflicting long double builtins and fixing a
few bugs in the test cases, I'd like to resubmit the following:
Currently, the IBM long double routines in libgcc use a union to construct
a long double from two double values. This causes horrific code generation
that copies
On 07/29/2014 04:29 AM, Jonathan Wakely wrote:
On 29/07/14 04:11 -0400, Ed Smith-Rowland wrote:
As discussed in the audit trail both _Adaptor and generate_canonical
are both meant to use floating point values. Both are here given
static_asserts to that effect.
This would have prevented this
This fix is inspired by ACATS test C324002, which tests that
predicate tests for OUT and IN OUT parameters are properly
applied. They were missed in some cases, and applied when
they should not be to Finalize procedures.
The following three tests (cutdown versions of C324002) compile
and execute
Ping for
https://gcc.gnu.org/ml/gcc-patches/2014-06/msg02400.html
Thanks in advance,
With Kind Regards,
Olivier
On Jun 30, 2014, at 15:29 , Olivier Hainque hain...@adacore.com wrote:
Hello,
The vxworks_override_option code is general enough to apply to both regular
VxWorks and
on aarch64, we are using load register pair post-writeback instruction in
epilogue.
for example, for the following instruction:
ldp, x0, x1, [sp], #16
what it's doing is:
x0 - MEM(sp + 0)
x1 - MEM(sp + 8)
sp sp + 16
while there is a glitch in our loadwb_pair* pattern, the restore of the
Add a pridicate to check whether the const offset is valid.
For 32-bit variant: -256 = offset = 252
For 64-bit variant: -512 = offset = 504
no regression on aarch64-none-elf bare-metal.
2014-07-29 Renlin Li renlin...@arm.com
2014-07-29 Jiong Wang jiong.w...@arm.com
gcc/
*
This removes a subprogram which serves no useful purpose and changes the
affected case to use the common code path. No functional changes.
Tested on x86_64-pc-linux-gnu, committed on trunk
2014-07-29 Eric Botcazou ebotca...@adacore.com
* exp_aggr.adb (Safe_Slice_Assignment): Remove.
On 07/29/2014 10:28 AM, Paolo Carlini wrote:
+unify_arity (bool explain_p, int have, int wanted, bool lb_p = false)
I don't understand lb_p.
@@ -16598,6 +16608,8 @@ type_unification_real (tree tparms,
tree argvec;
tree parmvec = make_tree_vec (1);
+ remaining_pack_p =
On Mon, Jul 28, 2014 at 7:02 PM, Richard Henderson r...@redhat.com wrote:
On 07/26/2014 05:35 AM, Uros Bizjak wrote:
On Mon, May 2, 2011 at 9:21 AM, Uros Bizjak ubiz...@gmail.com wrote:
It looks that GP relative relocations do not fit anymore into GPREL16
reloc, so bootstrap on alpha hosts
Hi,
On 07/29/2014 06:01 PM, Jason Merrill wrote:
On 07/29/2014 10:28 AM, Paolo Carlini wrote:
+unify_arity (bool explain_p, int have, int wanted, bool lb_p = false)
I don't understand lb_p.
lower_bound ;) The first name which came to my mind...
@@ -16598,6 +16608,8 @@
On 07/29/14 16:04, Olivier Hainque wrote:
The attached patch fixes this by defining VXWORKS_OVERRIDE_OPTIONS for AE as
for the regular vxworks targets.
Tested by verifying that the port still builds after the change, and
that -fPIC without -mrtp is rejected as it should.
OK to commit ?
On Jul 29, 2014, at 18:18 , Nathan Sidwell nat...@acm.org wrote:
2014-05-30 Olivier Hainque hain...@adacore.com
* config/vxworksae.h (VXWORKS_OVERRIDE_OPTIONS): Define.
Ok,
Great, thanks!
thanks for the ping.
Sure, no problem of course. Thanks for your prompt feedback :)
On 07/29/2014 06:11 AM, Uros Bizjak wrote:
Perhaps even better solution for mainline would be to detect a recent
enough linker and skip the workaround in that case? I guess that 2.25
will have this issue fixed?
Certainly 2.25 will have this fixed. If you want to add a check for binutils
On 07/29/2014 12:23 PM, Paolo Carlini wrote:
On 07/29/2014 06:01 PM, Jason Merrill wrote:
Why would we get here in the too few args case? Won't we only hit
this code if we had enough args for the non-pack parms?
Yeah, that is exactly for the cases I was mentioning at the end of the
my first
On Jul 29, 2014, at 2:10 AM, Martin Liška mli...@suse.cz wrote:
I am not a configure script guru, but what do you think about suggested
patch?
Looks exactly like what I would do.
Hi,
On 07/29/2014 06:56 PM, Jason Merrill wrote:
On 07/29/2014 12:23 PM, Paolo Carlini wrote:
On 07/29/2014 06:01 PM, Jason Merrill wrote:
Why would we get here in the too few args case? Won't we only hit
this code if we had enough args for the non-pack parms?
Yeah, that is exactly for the
On Jul 29, 2014, at 7:56 AM, Peter Bergner berg...@vnet.ibm.com wrote:
Currently, the IBM long double routines in libgcc use a union to construct
a long double from two double values. This causes horrific code generation
that copies the two double from the FP registers over to GPRs and back
Suppose the following shift:
T A = (T) B C
which met the following conditions:
* T is a signed type with 2 x word_size.
* B is with signed type with the size = word_size.
* C + sizeof (B) word_size.
| T |
| high | low |
| B shifted to |
| E | D
This has been broken since 4.3.0
Tested x86_64-linux, committed to trunk.
commit 35ba1889a1652eece0b4a3ab07290fcd8dc8f120
Author: Jonathan Wakely jwak...@redhat.com
Date: Tue Jul 29 10:06:39 2014 +0100
PR libstdc++/61946
* include/ext/rope (rope::rope(char_producer_CharT*, size_t,
This adds a tag parameter to the tuple base class constructors that
are used to implement uses-allocator construction but currently only
have a single parameter, ensuring they are not viable candidates when
calling a single-argument constructor.
Tested x86_64-linux, committed to trunk. I plan to
On Sun, 27 Jul 2014, Richard Sandiford wrote:
Marc Glisse marc.gli...@inria.fr writes:
Hello,
I followed the advice in this discussion:
https://gcc.gnu.org/ml/gcc-patches/2014-04/msg00269.html
and here is a new patch. I made an effort to isolate a path in at least
one subcase so it doesn't
On Tue, 2014-07-29 at 19:36 +0200, Marc Glisse wrote:
On Sun, 27 Jul 2014, Richard Sandiford wrote:
Marc Glisse marc.gli...@inria.fr writes:
Hello,
I followed the advice in this discussion:
https://gcc.gnu.org/ml/gcc-patches/2014-04/msg00269.html
and here is a new patch. I made an
On Tue, Jul 29, 2014 at 02:58:03PM -0400, David Malcolm wrote:
This is possibly a dumb question, but what happens for a static local,
rather than an auto local? e.g.
int *f (void)
{
static int i;
return i;
}
This is fine. The variable i has a static storage duration, so is
On Tue, 29 Jul 2014, David Malcolm wrote:
On Tue, 2014-07-29 at 19:36 +0200, Marc Glisse wrote:
On Sun, 27 Jul 2014, Richard Sandiford wrote:
Marc Glisse marc.gli...@inria.fr writes:
Hello,
I followed the advice in this discussion:
https://gcc.gnu.org/ml/gcc-patches/2014-04/msg00269.html
On 23/07/14 22:33 +0200, François Dumont wrote:
On 27/06/2014 21:48, Paolo Carlini wrote:
Hi,
On 06/27/2014 07:33 PM, Jonathan Wakely wrote:
I didn't see an obvious fix (I'm not sure if the templated constructor
can deduce its argument since the change) but have been out all day
and not had a
On Tue, 2014-07-29 at 21:13 +0200, Marek Polacek wrote:
On Tue, Jul 29, 2014 at 02:58:03PM -0400, David Malcolm wrote:
This is possibly a dumb question, but what happens for a static local,
rather than an auto local? e.g.
int *f (void)
{
static int i;
return i;
}
This
On Sep 12, 2013, at 1:47 AM, Richard Biener rguent...@suse.de wrote:
! #if 0
! FILE *file = fopen (/tmp/rdg.dot, w);
! gcc_assert (file != NULL);
!
dot_rdg_1 (file, rdg);
! fclose (file);
!
! system (dotty /tmp/rdg.dot );
#else
dot_rdg_1 (stderr, rdg);
#endif
---
On Tue, Jul 29, 2014 at 6:39 PM, Richard Henderson r...@redhat.com wrote:
On 07/29/2014 06:11 AM, Uros Bizjak wrote:
Perhaps even better solution for mainline would be to detect a recent
enough linker and skip the workaround in that case? I guess that 2.25
will have this issue fixed?
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=61948
The lshrdi3_neon,ashrdi3_neon,ashldi3_neon patterns can call
gen_arm_shiftdi3_1bit without checking that the register allocation
constraints of the resulting insn are satisfied. This results in an
ICE:
$ arm-unknown-linux-gnueabihf-gcc -O2
On 07/29/2014 07:01 PM, Mike Stump wrote:
On Jul 29, 2014, at 2:10 AM, Martin Liška mli...@suse.cz wrote:
I am not a configure script guru, but what do you think about suggested
patch?
Looks exactly like what I would do.
All right, I'll send the patch, as well as documentation
A complaint I heard at Cauldron with the C++ification of GCC passes is
that it's become much more difficult to set breakpoints on the execute
hooks of a pass, now that the passes are classes within anonymous
namespaces.
When this was first done, the execute methods were trivial
implementations
On Tue, 2014-07-29 at 16:44 -0400, David Malcolm wrote:
A complaint I heard at Cauldron with the C++ification of GCC passes is
that it's become much more difficult to set breakpoints on the execute
hooks of a pass, now that the passes are classes within anonymous
namespaces.
When this was
Hi,
The following patch moves dwarf unwind information from the data
section to the read-only/text section on AIX. This means the memory
for the unwind information can be shared across multiple processes.
The frame tables are still registered through collect2 scanning for
_GLOBAL_F symbol
On 25/07/14 00:19 +0100, Jonathan Wakely wrote:
On 24 July 2014 21:11, François Dumont wrote:
Yes I have tested with no other changes in my tree and got only those pretty
printers errors which are unrelated I think:
Python Exception class 'TypeError' iter() returned non-iterator of type
On Tue, Jul 29, 2014 at 04:44:29PM -0400, David Malcolm wrote:
A complaint I heard at Cauldron with the C++ification of GCC passes is
that it's become much more difficult to set breakpoints on the execute
hooks of a pass, now that the passes are classes within anonymous
namespaces.
Well, its
On 29/07/14 22:33 +0100, Jonathan Wakely wrote:
Tested x86_64-linux (thanks to Samual Bronson for testing with
Python3) and committed to trunk.
Oops, Samuel not Samual ... no idea why my fingers typed that wrong!
On Tue, 1 Jul 2014, Tom Tromey wrote:
2014-07-01 Tom Tromey tro...@redhat.com
PR c/59855
* doc/invoke.texi (Warning Options): Document -Wdesignated-init.
* doc/extend.texi (Type Attributes): Document designated_init
attribute.
2014-07-01 Tom Tromey
With r213117 we are seeing additional failures while testing a bare-iron build:
FAIL: gcc.dg/torture/ftrapv-1.c -O0 (test for excess errors)
...
This newly added test does a fork, though it has a guard
to prevent testing targets that don't support fork, but the
guard needs to be tweaked:
On 2014.07.29 at 15:10 +0200, Richard Biener wrote:
On Tue, 29 Jul 2014, Richard Biener wrote:
This re-organizes the LTO streamer to do compression transparently
in the data-streamer routines (and disables section compression
by defaulting to -flto-compression-level=0). This avoids
78 matches
Mail list logo
