Hi Bill,
On Tue, Aug 30, 2016 at 08:23:46PM -0500, Bill Schmidt wrote:
> The ada bootstrap failure reported in
> https://gcc.gnu.org/bugzilla/show_bug.cgi?id=72827
> occurs because of a latent bug in the powerpc back end. The immediate cause
> is dead store
> elimination removing two stores rel
On 30/08/16 11:38, Bernd Edlinger wrote:
On 08/30/16 10:21, Tom de Vries wrote:
On 29/08/16 18:43, Bernd Edlinger wrote:
Thanks!
Actually my patch missed to fix one combination: -m32 with -fpic
make check-gcc-c++ RUNTESTFLAGS="ubsan.exp=object-size-9.c --tool_opts
'-m32 -fpic'"
FAIL: c-c++-c
Committed to trunk as revision 239880.
Many thanks to all those who tested, reviewed or commented upon the patch.
Paul and Jerry
PR77289 exposes a latent problem with LRA constraint matching. In the buggy
test cases, LRA performs a speculative register elimination before checking
operands for matching constraints. With the elimination, the operands
appear to match. However, when we call check_rtl() which attempts to
recog
Hi all,
The attached patch fixes the problem by adding a new helper function to
determine the buffer size needed for F0 editing depending on the kind. In this
new function there are some constants presented which document the limits needed
for each kind type.
As can be seen, the required buffers
Hi,
The ada bootstrap failure reported in
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=72827
occurs because of a latent bug in the powerpc back end. The immediate cause is
dead store
elimination removing two stores relative to the frame pointer that are not
dead; however,
DSE is tricked into d
On Tue, 2016-08-30 at 23:26 +0530, Prathamesh Kulkarni wrote:
[...]
> The attached (untested) patch silences the warning if parameter is
> const qualified.
> I will give it some testing after resolving a different issue.
I've now removed the hard-coded limit of 3 ranges per rich_location, so
you
This patch eliminates the hard-coded limits within rich_location
(up to 3 ranges, up to 2 fixits). The common case is still
handled by embedding the values inside rich_location - it only
uses dynamic allocation if these limits are exceeded, so
creation of rich_location instances on the stack shoul
On Mon, 2016-08-29 at 23:53 +0200, Bernd Schmidt wrote:
> On 08/25/2016 03:13 AM, David Malcolm wrote:
> > Split out a new base class for temp_source_file, named_temp_file,
> > moving the deletion to the base class dtor, so that we can write
> > out temporary files in other ways in selftests.
> >
On Aug 24, 2016, Richard Biener wrote:
> The issue with LTO / early debug is that we can't prune DIEs that refer
> to optimized out entities from early debug output thus we have to create
> those DIEs late only (in the above case, as tmpl_value_param_die_table
> is not streamed, not at all for LT
On Aug 26, 2016, Jason Merrill wrote:
> I suppose we can teach dwz to use the maximal set of friends...
*nod*
> This seems unnecessary; there is no semantic difference for a
> particular specialization depending on whether it became a friend
> directly or from its template.
'k, I took that out
On Aug 30, 2016, at 4:57 AM, Prathamesh Kulkarni
wrote:
>
> On 30 August 2016 at 17:11, Eric Gallager wrote:
>> On 8/29/16, Jason Merrill wrote:
>>> On Mon, Aug 29, 2016 at 10:28 AM, Marek Polacek wrote:
On Mon, Aug 29, 2016 at 09:20:53AM -0400, Eric Gallager wrote:
> I tried this pa
Usually a data clause would would have OMP_CLAUSE_SIZE set, but not all
do. In the latter case, lower_omp_target falls back to using size of the
type of the variable specified in the data clause. However, in the case
of assumed-size arrays, the size of the type may be NULL because its
undefined. My
Resending without the attachment as it was blocked as spam.
The actual change be seen in Subversion or at
https://groups.google.com/forum/?pli=1#!topic/gofrontend-dev/fT2-PlvVZ9o
.
Ian
-- Forwarded message --
From: Ian Lance Taylor
Date: Tue, Aug 30, 2016 at 2:07 PM
Subject: li
OK.
On Tue, Aug 30, 2016 at 10:43 AM, David Malcolm wrote:
> This patch adds fix-it hints to our warning for >>
> within a nested template argument list (for -std=c++98).
>
> For example:
>
> double-greater-than-fixit.C:5:12: error: ‘>>’ should be ‘> >’ within a nested
> template argument list
>
Hi,
Just noticed a small nit - why not reuse ipa_vr in ipa_jump_func
instead of adding vr_known and m_vr ?
This is because we want to reuse vrp_intersect_ranges and vrp_meet
which are not trivial. On the other hand, in ipa_jump_func, since we
stream the data we have a simpler data structure
On Tue, Aug 30, 2016 at 11:53 AM, Alexander Ivchenko wrote:
> Would something like that count?
>
> I did not do the warning thing, cause the problem only appears when
> you provide the -Wl,-as-needed option to the linker.
> (In all other cases warning would be redundant). Are we able to check
> th
Would something like that count?
I did not do the warning thing, cause the problem only appears when
you provide the -Wl,-as-needed option to the linker.
(In all other cases warning would be redundant). Are we able to check
that on runtime?
diff --git a/gcc/config.in b/gcc/config.in
index fc3321
This patch changes the default for gang partitioned size to be determined at
runtime (and thus interrogate the hardware).
The auto-loop test is designed for num_gangs 32. The fortran nested function
test appears to also require that, but my be hiding another defect. Its use of
gang(static:1)
On 30 August 2016 at 10:50, Kugan Vivekanandarajah
wrote:
> Hi Honza,
>
> Here is a re-based version which also addresses the review comments.
>
> On 21/07/16 22:54, Jan Hubicka wrote:
>>> Maybe it is better to separate value range and alignment summary
>>> writing/reading to different functions.
On 30 August 2016 at 22:58, Tom de Vries wrote:
> On 30/08/16 17:34, Prathamesh Kulkarni wrote:
>>
>> On 30 August 2016 at 20:24, Tom de Vries wrote:
>>>
>>> On 26/08/16 13:39, Prathamesh Kulkarni wrote:
Hi,
The following patch adds option -Wrestrict that warns when an argumen
On 30/08/16 17:34, Prathamesh Kulkarni wrote:
On 30 August 2016 at 20:24, Tom de Vries wrote:
On 26/08/16 13:39, Prathamesh Kulkarni wrote:
Hi,
The following patch adds option -Wrestrict that warns when an argument
is passed to a restrict qualified parameter and it aliases with
another argume
On 08/30/2016 08:00 AM, Janne Blomqvist wrote:
> On Mon, Aug 29, 2016 at 1:15 PM, Paul Richard Thomas
> wrote:
>> Hi Janne, Andre, Jerry and All,
>>
>> I am perfectly happy to adopt Janne's suggestion for
>> st_set_(dtio_)nml_var. Do the changes to the IO structures have any
>> impact? These are i
On Tue, Aug 30, 2016 at 10:25 AM, Marek Polacek wrote:
>> > @@ -10585,15 +10610,26 @@ cp_parser_statement (cp_parser* parser, tree
>> > in_statement_expr,
>> > }
>> >/* Look for an expression-statement instead. */
>> >statement = cp_parser_expression_statement (parser,
>
Updated patch attached.
This one was updated with a new ultra simple fiber implementation that calls
ucontext.h API directly instead of requiring Pth in between. Plus cleanups.
On Mon, Aug 15, 2016 at 4:22 PM, Pekka Jääskeläinen wrote:
> Updated patch attached.
> Mostly comment cleanups and ren
On 30/08/16 17:11, Christophe Lyon wrote:
On 18 August 2016 at 11:15, Tamar Christina wrote:
Hi all,
This fixes a bug in the vector load functions in which they load the
vector in the wrong order for big endian systems. This patch flips the
order conditionally in the vec_concats.
No testcas
falign-functions=N is too simplistic.
Ingo Molnar ran some tests and it seems that on latest x86 CPUs, 64-byte
alignment
of functions runs fastest (he tried many other possibilites):
this way, after a call CPU can fetch a lot of insns in the first cacheline fill.
However, developers are less tha
This change drops forced alignment to 8 if requested alignment is higher
than 8: before the patch, -falign-functions=9 was generating
.p2align 4,,8
.p2align 3
which means: "align to 16 if the skip is 8 bytes or less; else align to 8".
After this change, ".p2align 3" is not emitted
On 18 August 2016 at 11:15, Tamar Christina wrote:
> Hi all,
>
> This fixes a bug in the vector load functions in which they load the
> vector in the wrong order for big endian systems. This patch flips the
> order conditionally in the vec_concats.
>
> No testcase given because plenty of existing
On Tue, Aug 30, 2016 at 07:38:18AM -0400, Eric Gallager wrote:
> On 8/30/16, Marek Polacek wrote:
> > On Mon, Aug 29, 2016 at 09:32:04AM -0400, Eric Gallager wrote:
> >> I tried v6 on my binutils-gdb fork, and it printed A LOT of
> >> warnings...
> >
> > BTW, why is that so? Does binutils-gdb not
On Thu, Aug 18, 2016 at 10:15:12AM +0100, Tamar Christina wrote:
> Hi all,
>
> This fixes a bug in the vector load functions in which they load the
> vector in the wrong order for big endian systems. This patch flips the
> order conditionally in the vec_concats.
>
> No testcase given because plen
On 30 August 2016 at 20:24, Tom de Vries wrote:
> On 26/08/16 13:39, Prathamesh Kulkarni wrote:
>>
>> Hi,
>> The following patch adds option -Wrestrict that warns when an argument
>> is passed to a restrict qualified parameter and it aliases with
>> another argument.
>>
>> eg:
>> int foo (const ch
On Tue, 2016-08-30 at 16:50 +0200, Marek Polacek wrote:
> On Tue, Aug 30, 2016 at 04:40:57PM +0200, Bernd Schmidt wrote:
> > On 08/30/2016 04:38 PM, David Malcolm wrote:
> >
> > > In conjunction with the not-yet-in-trunk -fdiagnostics-generate
> > > -patch,
> > > this can generate patches like thi
On 01/08/16 17:31 +0100, Jonathan Wakely wrote:
On 01/08/16 09:23 -0700, Mike Stump wrote:
On Jul 31, 2016, at 1:30 PM, Jonathan Wakely wrote:
-fno-show-column
is a good general option. If you guys want to add column number test cases,
they can avoid it, and test down to the column. Most
On Mon, Aug 29, 2016 at 1:15 PM, Paul Richard Thomas
wrote:
> Hi Janne, Andre, Jerry and All,
>
> I am perfectly happy to adopt Janne's suggestion for
> st_set_(dtio_)nml_var. Do the changes to the IO structures have any
> impact? These are in:
>
> fnode, st_parameter_dt & gfc_unit
>
> I don't thi
This patch interrogates the target device to determine default gemotry at
runtime. This has the greatest difference on gang partitioning, where there's a
noticeable sawtooth in the relationship between number of gangs and execution
time. Picking the number of gangs as an exact multiple of num
On 26/08/16 13:39, Prathamesh Kulkarni wrote:
Hi,
The following patch adds option -Wrestrict that warns when an argument
is passed to a restrict qualified parameter and it aliases with
another argument.
eg:
int foo (const char *__restrict buf, const char *__restrict fmt, ...);
void f(void)
{
On Tue, Aug 30, 2016 at 04:40:57PM +0200, Bernd Schmidt wrote:
> On 08/30/2016 04:38 PM, David Malcolm wrote:
>
> > In conjunction with the not-yet-in-trunk -fdiagnostics-generate-patch,
> > this can generate patches like this:
> >
> > --- modernize-named-inits.c
> > +++ modernize-named-inits.c
>
On 08/30/2016 04:38 PM, David Malcolm wrote:
In conjunction with the not-yet-in-trunk -fdiagnostics-generate-patch,
this can generate patches like this:
--- modernize-named-inits.c
+++ modernize-named-inits.c
@@ -16,7 +16,7 @@
/* Old-style named initializers. */
struct foo old_style_f = {
-
Committed as obvious as r239862, this fixes a "rebase-o" ("typo during
rebase") that found its way into the documentation on STRUCTURE which
is clearly incorrect:
Index: gcc/fortran/gfortran.texi
===
--- gcc/fortran/gfortran.texi (rev
On Mon, Aug 29, 2016 at 04:02:33PM -0400, Jason Merrill wrote:
> On Mon, Aug 29, 2016 at 7:58 AM, Marek Polacek wrote:
> > --- gcc/gcc/cp/parser.c
> > +++ gcc/gcc/cp/parser.c
> > @@ -5135,6 +5135,31 @@ cp_parser_primary_expression (cp_parser *parser,
> > case RID_AT_SELECTOR:
> >
This patch adds fix-it hints to our warning for >>
within a nested template argument list (for -std=c++98).
For example:
double-greater-than-fixit.C:5:12: error: ‘>>’ should be ‘> >’ within a nested
template argument list
foo> i;
^~
> >
In conjunction with the not-yet-i
This patch adds fix-it hints to our warning for old-style structure
member designators, showing how to modernize them to C99 form.
For example:
modernize-named-inits.c:19:5: warning: obsolete use of designated initializer
with ‘:’ [-Wpedantic]
foo: 1,
^
. =
In conjunction with the not
On Tue, 30 Aug 2016, Uros Bizjak wrote:
> IIUC, trying to solve this issue in the front-end would bring us to
> float->double promotion rules for variable arguments. The compiler
> doesn't convert va-args automatically, it only warns for undefined
> situation, e.g.:
No, those would be unchanged.
Adding a fix-it hint to a diagnostic usually follows one of these
patterns:
(a) an insertion fix-its, with the insertion at the primary caret location
(b) a removals/replacements, affecting the range of the primary location
(other cases are possible, e.g. multiple fix-its, and affecting other
loca
Ping.
On 19/08/16 15:47, Richard Earnshaw (lists) wrote:
On 19/08/16 15:06, Matthew Wahab wrote:
On 19/08/16 14:30, Richard Earnshaw (lists) wrote:
On 19/08/16 12:48, Matthew Wahab wrote:
2016-08-19 Matthew Wahab
PR target/77281
* config/arm/arm.c (neon_valid_immediate): Delet
On 10/08/16 17:31, Wilco Dijkstra wrote:
> Richard Earnshaw wrote:
>> OK. But please enhance the comment with some explanation as to WHY
>> you've chosen to use just two base pairings rather than separate bases
>> for each access size.
>
> OK here is the updated patch which also handles unaligned
On Tue, Aug 30, 2016 at 2:09 PM, Joseph Myers wrote:
> On Tue, 30 Aug 2016, Uros Bizjak wrote:
>
>> I didn't find a way around this limitation, so I propose to disable
>> _FloatN on alpha with the attached patch.
>
> If your ABI requires binary32 values to be promoted in variable arguments,
> I'd
On Tue, 2016-08-30 at 17:08 +0530, Prathamesh Kulkarni wrote:
> On 30 August 2016 at 05:34, David Malcolm
> wrote:
> > On Mon, 2016-08-29 at 20:01 -0400, David Malcolm wrote:
> > > On Mon, 2016-08-29 at 19:55 -0400, David Malcolm wrote:
> > > [...]
> > > > Assuming you have the location_t values a
On Tue, Aug 30, 2016 at 7:17 AM, Segher Boessenkool
wrote:
> The exit block (to which every return artificially jumps) already has
> a use of LR. The LR use in all returns and sibcalls is an anachronism,
> probably made unnecessary by the dataflow merge. The simple_returns
> that shrink-wrapping
Hi,
On Fri, 26 Aug 2016, Bernd Schmidt wrote:
> And that comment puzzles me. Surely prologue and epilogue are executed only
> once currently, so how does frequency come into it? Again - please provide an
> example.
int some_global;
int foo (void) {
if (!some_global) {
call_this();
call
On Tue, 30 Aug 2016, Uros Bizjak wrote:
> I didn't find a way around this limitation, so I propose to disable
> _FloatN on alpha with the attached patch.
If your ABI requires binary32 values to be promoted in variable arguments,
I'd think you could do that with a new target hook that specifies a
On 30 August 2016 at 17:11, Eric Gallager wrote:
> On 8/29/16, Jason Merrill wrote:
>> On Mon, Aug 29, 2016 at 10:28 AM, Marek Polacek wrote:
>>> On Mon, Aug 29, 2016 at 09:20:53AM -0400, Eric Gallager wrote:
I tried this patch on my fork of gdb-binutils and got a few warnings
from it.
On 30/08/16 13:19 +0200, Uros Bizjak wrote:
Hello!
There are several libstdc++ testsuite FAILs on systems where
de_DE.UTF-8 locale is missing.
Apparently, following "dg-do run" directive overrides
dg-require-namedlocale requirement.
Yes, a dg-do with a target selector will decide whether to r
On 8/29/16, Jason Merrill wrote:
> On Mon, Aug 29, 2016 at 10:28 AM, Marek Polacek wrote:
>> On Mon, Aug 29, 2016 at 09:20:53AM -0400, Eric Gallager wrote:
>>> I tried this patch on my fork of gdb-binutils and got a few warnings
>>> from it. Would it be possible to have the caret point to the arg
On 30 August 2016 at 05:34, David Malcolm wrote:
> On Mon, 2016-08-29 at 20:01 -0400, David Malcolm wrote:
>> On Mon, 2016-08-29 at 19:55 -0400, David Malcolm wrote:
>> [...]
>> > Assuming you have the location_t values available, you can create a
>> > rich_location for the primary range, and then
On 8/30/16, Marek Polacek wrote:
> On Mon, Aug 29, 2016 at 09:32:04AM -0400, Eric Gallager wrote:
>> I tried v6 on my binutils-gdb fork, and it printed A LOT of
>> warnings...
>
> BTW, why is that so? Does binutils-gdb not use various FALLTHRU comments?
>
> Marek
>
There are a lot of comm
Hi Jason,
> The testcase in 57728 demonstrates that we have been treating
> functions explicitly defaulted in the class body inconsistently with
> explicit instantiation: an extern instantiation causes them not to be
> generated, but a normal explicit instantiation doesn't cause them to
> be emitt
Hi Uros,
> There are several libstdc++ testsuite FAILs on systems where
> de_DE.UTF-8 locale is missing.
>
> Apparently, following "dg-do run" directive overrides
> dg-require-namedlocale requirement.
>
> Attached patch fixes testsuite failures by moving
> dg-require-namedlocale directive after dg
Hello!
There are several libstdc++ testsuite FAILs on systems where
de_DE.UTF-8 locale is missing.
Apparently, following "dg-do run" directive overrides
dg-require-namedlocale requirement.
Attached patch fixes testsuite failures by moving
dg-require-namedlocale directive after dg-do run.
2016-0
The exit block (to which every return artificially jumps) already has
a use of LR. The LR use in all returns and sibcalls is an anachronism,
probably made unnecessary by the dataflow merge. The simple_returns
that shrink-wrapping generates also do not have such a use. Newer
backends do not do th
Dear All,
Janne's proposed change to namelist transfer has been implemented.
This avoids ABI brekage.
Please find the ChangeLogs below and the new patch attached.
Bootstraps and regtests on FC21/x86_64.
I will commit tomorrow morning if there are no objections in the meantime.
Best regards
Pa
diff --git a/libstdc++-v3/testsuite/20_util/pair/cons/explicit_construct.cc
b/libstdc++-v3/testsuite/20_util/pair/cons/explicit_construct.cc
index 1525fef..7543397 100644
--- a/libstdc++-v3/testsuite/20_util/pair/cons/explicit_construct.cc
+++ b/libstdc++-v3/testsuite/20_util/pair/cons/explicit_c
On Sun, Aug 28, 2016 at 7:59 PM, Jan Hubicka wrote:
> Hi,
> I have noticed that ivopts build MEM_REF (NOP_EXPR (address))
> where NOP_EXPR converts one pointer type to another. Later it tries
> to expand it that on strict alignment targets punt on determining
> the alignment and winds expensive m
We ran into an ICE at -O3 on x86-64/Darwin for the attached testcase with our
GCC 6 compiler:
+===GNAT BUG DETECTED==+
| Pro 17.0w (20160830-62) (x86_64-apple-darwin14.5.0) GCC error: |
| in rtl_split_edge, at cfgrtl.c:1887
On 08/30/16 10:21, Tom de Vries wrote:
> On 29/08/16 18:43, Bernd Edlinger wrote:
>> Thanks!
>>
>> Actually my patch missed to fix one combination: -m32 with -fpic
>>
>> make check-gcc-c++ RUNTESTFLAGS="ubsan.exp=object-size-9.c --tool_opts
>> '-m32 -fpic'"
>>
>> FAIL: c-c++-common/ubsan/object-siz
On 29 August 2016 at 22:10, Jakub Jelinek wrote:
> On Mon, Aug 29, 2016 at 12:42:28PM -0400, Jason Merrill wrote:
>> Another missing ABI tag, sigh.
>>
>> Tested x86_64-pc-linux-gnu, applying to trunk.
>
>> commit 1337a943a2d3926537b63d6e1f0d7f46ef10a06d
>> Author: Jason Merrill
>> Date: Fri Aug
On Mon, 29 Aug 2016, Andreas Schwab wrote:
> On Aug 26 2016, Richard Biener wrote:
>
> > Index: gcc/testsuite/gcc.dg/pr69047.c
> > ===
> > --- gcc/testsuite/gcc.dg/pr69047.c (revision 0)
> > +++ gcc/testsuite/gcc.dg/pr69047.c (wor
On Tue, 30 Aug 2016, Jakub Sejdak wrote:
> > Ok.
> >
> > Thanks,
> > Richard.
>
> Sorry for silly question, but does it mean that I have a green light
> to commit it to SVN or did you just acknowledge the fact and will take
> a look at this later?
It was an approval for the backport.
> For futu
Hello!
The _FloatN patch uncovers a problem on alpha OSF/1 ABI when _Float32
variable arguments are passed to a function. As seen from the source,
32bit _Float32 arguments (SFmode) are *not* promoted to 64bit DFmode
when passing to as variable arguments, and this uncovers following
limitation:
Va
On Mon, Aug 29, 2016 at 09:32:04AM -0400, Eric Gallager wrote:
> I tried v6 on my binutils-gdb fork, and it printed A LOT of
> warnings...
BTW, why is that so? Does binutils-gdb not use various FALLTHRU comments?
Marek
On 29/08/16 18:43, Bernd Edlinger wrote:
make check-gcc-c++ RUNTESTFLAGS="ubsan.exp=object-size-9.c --tool_opts
'-m32 -fpic'"
FAIL: c-c++-common/ubsan/object-size-9.c -O2 execution test
FAIL: c-c++-common/ubsan/object-size-9.c -O2 -flto
-fno-use-linker-plugin -flto-partition=none execution
Bump
From: gcc-patches-ow...@gcc.gnu.org on behalf
of Tamar Christina
Sent: Thursday, August 18, 2016 10:15:12 AM
To: GCC Patches
Cc: James Greenhalgh; Richard Earnshaw; Marcus Shawcroft; nd
Subject: [PATCH][Aarch64][gcc] Fix vld2/3/4 on big endian systems
Hi all,
This fixes a bug i
On 29/08/16 18:43, Bernd Edlinger wrote:
Thanks!
Actually my patch missed to fix one combination: -m32 with -fpic
make check-gcc-c++ RUNTESTFLAGS="ubsan.exp=object-size-9.c --tool_opts
'-m32 -fpic'"
FAIL: c-c++-common/ubsan/object-size-9.c -O2 execution test
FAIL: c-c++-common/ubsan/object-
> Ok.
>
> Thanks,
> Richard.
Sorry for silly question, but does it mean that I have a green light
to commit it to SVN or did you just acknowledge the fact and will take
a look at this later?
For future: I'm planning to commit it to GCC-5 also later. Do I have
to send another patch or maybe it is g
On 11/08/16 15:13, Andre Vieira (lists) wrote:
> On 25/07/16 11:52, Andre Vieira (lists) wrote:
>> On 11/07/16 17:56, Andre Vieira (lists) wrote:
>>> On 07/07/16 13:30, mickael guene wrote:
Hi Andre,
Another feedback on your purecode patch.
You have to disable casesi pattern s
First, pardons all around if I'm completely repeating what Marc
already tried to do.
I think I'm taking a different approach.
I'm not adding any defaulted or deleted functions, so I don't think I'm changing
triviality. Chances are I need to actually delete the copy operations I want to
get rid of,
On Mon, 29 Aug 2016, Jakub Sejdak wrote:
> Hi!
>
> I would like to ping a couple of unreviewed patches for GCC-6 branch
> (they are already in trunk):
>
> - Backport new Phoenix-RTOS OS name to config.sub
> https://gcc.gnu.org/ml/gcc-patches/2016-07/msg01441.html
>
> - Backport support for Phoe
78 matches
Mail list logo