On Thu, May 10, 2018 at 07:53:15AM +0200, Richard Biener wrote:
> On May 9, 2018 10:52:05 PM GMT+02:00, Jakub Jelinek wrote:
> >Hi!
> >
> >glibc <= 2.23 has buggy nextafterl/nexttowardl as can be seen on the
> >nextafter-2.c testcase.
> >
> >Do we want to workaround this bug, e.g. with the followi
On Wed, May 09, 2018 at 04:53:19PM +0200, Allan Sandfeld Jensen wrote:
> > > @@ -2022,8 +2022,9 @@ simplify_vector_constructor (gimple_stmt_iterator
> > > *gsi)>
> > >elem_type = TREE_TYPE (type);
> > >elem_size = TREE_INT_CST_LOW (TYPE_SIZE (elem_type));
> > >
> > > - vec_perm_builder s
Hi.
It's removal of an assert at place where we calculate hash of a type.
For incomplete types, let's skip it.
Patch can bootstrap on x86_64-linux-gnu and survives regression tests.
Ready to be installed?
Martin
gcc/ChangeLog:
2018-05-09 Martin Liska
PR ipa/85607
* ipa-icf
(not a review)
On Thu, 10 May 2018, Hans-Peter Nilsson wrote:
Replacing a division feeding a division helps only when the
second division is the only user, and "fusing" the divisions is
Well, that's not quite true.
int x, y;
void f(int n){
int c = 3 << 20;
x = n / c;
y = n / c;
}
Here
Hi.
Parsing of no_sanitize attribute now supports
__attribute__((no_sanitize("address,undefined")))
which is wrong. And on the other hand this is not recognized:
__attribute__((no_sanitize("address", "undefined")))
Patch can bootstrap on x86_64-linux-gnu and survives regression tests. Then I
wou
This fixes an annyoing regression introduced in 2012 by this change:
https://codereview.appspot.com/5394041
The idea of the change is to interpolate the value passed to --with-sysroot in
the value passed to --with-gxx-include-dir, so that you can change the sysroot
at run time, i.e. pass --sys
On Thu, May 10, 2018 at 11:28:15AM +0200, Martin Liška wrote:
> Parsing of no_sanitize attribute now supports
> __attribute__((no_sanitize("address,undefined")))
Why is that wrong? I don't see why we shouldn't support it that way.
It matches how we handle other similar attributes, say target attr
On Mon, 23 Apr 2018, Alexander Monakov wrote:
> This rewrites global register vars doc to reflect that the register is no
> longer
> reserved exclusively, but in fact is available for general allocation, and
> also
> adds the requirement to properly inform the compiler where inline asms are
>
Hi Steve,
AFAICT the patch is missing.
Thanks for working on these PRs.
Dominique
https://gcc.gnu.org/ml/gcc-patches/2018-04/msg01172.html
Thanks in advance.
--
Eric Botcazou
This removes the -maltivec=be and -maltivec=le options. Those were
deprecated in GCC 8.
Altivec will keep working on both BE and LE; it is just the BE-vectors-
on-LE that is removed (the other way around was never supported).
The main change is replacing VECTOR_ELT_ORDER_BIG by BYTES_BIG_ENDIAN
Ping!
On 27/03/18 13:58, Sudakshina Das wrote:
Hi
This patch adds the no variant to -mstrict-align and the corresponding
function attribute. To enable the function attribute, I have modified
aarch64_can_inline_p () to allow checks even when the callee function
has no attribute. The need for thi
On 05/09/2018 10:44 AM, Kyrill Tkachov wrote:
On 09/05/18 13:30, Luis Machado wrote:
Hi Kyrill,
On 05/08/2018 11:09 AM, Kyrill Tkachov wrote:
Hi Luis,
On 07/05/18 15:28, Luis Machado wrote:
Hi,
On 02/08/2018 10:45 AM, Luis Machado wrote:
Hi Kyrill,
On 02/08/2018 09:48 AM, Kyrill Tkachov
Committed as obvious in revision 260113.
I will commence backporting as soon as I can.
Thanks to Mirco Valentini for the initial fix, which he posted on 2017-02-28.
Cheers
Paul
2018-05-10 Paul Thomas
PR fortran/68846
PR fortran/70864
* resolve.c (get_temp_from_expr): The tempor
On 05/10/2018 11:45 AM, Jakub Jelinek wrote:
> On Thu, May 10, 2018 at 11:28:15AM +0200, Martin Liška wrote:
>> Parsing of no_sanitize attribute now supports
>> __attribute__((no_sanitize("address,undefined")))
>
> Why is that wrong? I don't see why we shouldn't support it that way.
> It matches
> Date: Thu, 10 May 2018 10:33:39 +0200 (CEST)
> From: Marc Glisse
> On Thu, 10 May 2018, Hans-Peter Nilsson wrote:
>
> > Replacing a division feeding a division helps only when the
> > second division is the only user, and "fusing" the divisions is
>
> Well, that's not quite true.
> int x, y;
And a couple of other doc improvements, and regenerated the HTML
pages.
Please read the proposed change to using.xml and let me know if it's
clear.
* doc/xml/faq.xml: Link to C++17 status. Add note to outdated answer.
* doc/xml/manual/debug_mode.xml: Add array and forward_list to
On Thu, May 10, 2018 at 12:59:39PM +0200, Martin Liška wrote:
> >From 9e2570eee9bb160b58075f6802d6ac1bb7b77341 Mon Sep 17 00:00:00 2001
> From: marxin
> Date: Thu, 10 May 2018 10:27:02 +0200
> Subject: [PATCH] Support LLVM style of no_sanitize attribute (PR
> sanitizer/85556).
>
> gcc/ChangeLog:
On Thu, 10 May 2018, Hans-Peter Nilsson wrote:
Date: Thu, 10 May 2018 10:33:39 +0200 (CEST)
From: Marc Glisse
On Thu, 10 May 2018, Hans-Peter Nilsson wrote:
Replacing a division feeding a division helps only when the
second division is the only user, and "fusing" the divisions is
Well, t
On Thu, May 10, 2018 at 01:51:29PM +0200, Marc Glisse wrote:
> > > There are probably more
> > > complicated transformations this disables.
> >
> > I'm providing an example from *real* code where the
> > transformation is bad (admittedly just for the div-div case).
Isn't the case of the posted re
On Thu, May 10, 2018 at 10:33:39AM +0200, Marc Glisse wrote:
> (not a review)
>
> On Thu, 10 May 2018, Hans-Peter Nilsson wrote:
>
> >Replacing a division feeding a division helps only when the
> >second division is the only user, and "fusing" the divisions is
>
> Well, that's not quite true.
>
This works around user error where our headers get included within an
extern "C" block.
PR libstdc++/85729
* include/bits/c++config.h (__replacement_assert): Add linkage
specification.
* include/bits/std_abs.h: Add comment to closing brace of block.
* inclu
Hi Uros,
Updated patch attached, please find comments below.
> -Original Message-
> From: Uros Bizjak [mailto:ubiz...@gmail.com]
> Sent: Wednesday, May 9, 2018 1:47 PM
> To: Peryt, Sebastian
> Cc: gcc-patches@gcc.gnu.org; Kirill Yukhin
> Subject: Re: [PATCH][i386] Adding WAITPKG instruc
> -Original Message-
> From: Uros Bizjak [mailto:ubiz...@gmail.com]
> Sent: Wednesday, May 9, 2018 1:53 PM
> To: Peryt, Sebastian
> Cc: gcc-patches@gcc.gnu.org; Kirill Yukhin
> Subject: Re: [PATCH][i386] Adding CLDEMOTE instruction
>
> On Tue, May 8, 2018 at 1:58 PM, Peryt, Sebastian
>
Hi,
just small nits:
On Tue, May 08 2018, Jan Hubicka wrote:
> Hi,
> this patch adds documentation of -flinker-output.
>
> * doc/invoke.texi (-flinker-output): Document
> Index: doc/invoke.texi
> ===
> --- doc/invoke.texi (re
On Thu, May 10, 2018 at 2:50 PM, Peryt, Sebastian
wrote:
> Hi Uros,
>
> Updated patch attached, please find comments below.
>
>> -Original Message-
>> From: Uros Bizjak [mailto:ubiz...@gmail.com]
>> Sent: Wednesday, May 9, 2018 1:47 PM
>> To: Peryt, Sebastian
>> Cc: gcc-patches@gcc.gnu.or
Prompted by PR go/85429 (gotools unconditionally using gas syntax for
setting SHF_EXCLUDE), I looked into what it takes to enable the flag
(SECTION_EXCLUDE in gcc) on Solaris with /bin/as. Here's what I found:
* Unlike most (all?) other section flags, the Solaris/x86 as doesn't
accept the "e" f
As described in the PR, there are a couple of jit testsuite failures on
Solaris when using /bin/as. The errors point to GC issues and indeed,
gcc/config/sol2.c lacked GTY markup. This patch fixes that, following
what darwin.c does for machopic_indirections. I confess I have no idea
why I had to
> -Original Message-
> From: Uros Bizjak [mailto:ubiz...@gmail.com]
> Sent: Thursday, May 10, 2018 3:26 PM
> To: Peryt, Sebastian
> Cc: gcc-patches@gcc.gnu.org; Kirill Yukhin
> Subject: Re: [PATCH][i386] Adding WAITPKG instructions
>
> On Thu, May 10, 2018 at 2:50 PM, Peryt, Sebastian
>
Hi,
bug 85655 happens because intersect_with_plats encounters a BOTTOM
lattice and does not check for it. On closer inspection, the real check
that should be performed is is_single_const, so this patch adds it.
It has passed bootstrap and testing, LTO bootstrap is underway. OK for
trunk if it p
On 05/09/2018 05:30 AM, Jonathan Wakely wrote:
On 07/05/18 12:39 -0400, Ed Smith-Rowland wrote:
All,
We were using a different convention for P_l^m assoc_legendre(int l,
int m, FloatTp x)
- the so-called Condon-Shortley convention which includes (-1)^m.
This unfortunately is common.
Thi
On Thu, May 10, 2018 at 11:48:24AM +0200, Dominique d'Humières wrote:
> Hi Steve,
>
> AFAICT the patch is missing.
>
> Thanks for working on these PRs.
>
Whoops. Looks like attached the log instead of diff.
Index: gcc/fortran/array.c
All,
We could consider dropping the TR1 support, and just provide these
functions for ISO/IEC 29124:2010 in C++11 (or later) and for C++17.
But that decision should be taken separately, and should only happen
on trunk anyway so we need to use _Tp(+1) here.
I am in favour of splitting new versi
Hello world,
I just commmitted the attached patch as obvious after checking that
it passes "make info", "make dvi" and "make pdf".
Regards
Thomas
2018-05-10 Thomas Koenig
PR fortran/54613
* intrinsic.texi: Document BACK for MINLOC and MAXLOC.
Index: intrinsic.texi
=
Hi Steve,
the patch looks OK to me. Is it possible to give a warning there, at least with
some higher warning-level? Or is there already one? I haven't tested it.
Regards,
Andre
On Thu, 10 May 2018 07:15:21 -0700
Steve Kargl wrote:
> On Thu, May 10, 2018 at 11:48:24AM +0200, Dominique
Am 10.05.2018 um 14:20 schrieb Thomas Koenig:
Am 10.05.2018 um 12:33 schrieb Jonathan Wakely:
Should the fix for PR 84073 be documented, so that users whose code is
now rejected understand why, and how to fix it?
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=84073
https://gcc.gnu.org/gcc-8/porti
OK, thanks.
On Wed, May 9, 2018 at 4:49 PM, Jakub Jelinek wrote:
> On Wed, May 09, 2018 at 11:01:18AM -0400, Jason Merrill wrote:
>> On Wed, May 9, 2018 at 10:47 AM, Jakub Jelinek wrote:
>> > On Wed, May 09, 2018 at 10:40:26AM -0400, Jason Merrill wrote:
>> >> On Wed, May 9, 2018 at 4:55 AM, Jak
Hi Steve,
ok for trunk.
- Andre
On Wed, 9 May 2018 15:32:10 -0700
Steve Kargl wrote:
> I plan to commit the attach patch on Saturday unless someone objects.
>
> 2018-05-09 Steven G. Kargl
>
> PR fortran/70870
> * data.c (gfc_assign_data_value): Check that a data object does
>
Hi Thomas,
looks ok to me.
- Andre
On Thu, 10 May 2018 17:19:39 +0200
Thomas König wrote:
> Am 10.05.2018 um 14:20 schrieb Thomas Koenig:
> > Am 10.05.2018 um 12:33 schrieb Jonathan Wakely:
> >> Should the fix for PR 84073 be documented, so that users whose code is
> >> now rejected understand
It is certainly possible to give a warning, but it
would be odd (to me) to warn about technically
standard conforming code. gfortran doesn't warn
for zero-sized array references or zero-length
substrings in other context.
program foo
real a(4)
character(len=10) s
s = '12345'
a = 1
Hello.
This introduces a replacement for qsort() in GCC. The main selling point is
reproducibility (currently compiler output may change depending on how libc
qsort reorders not-bitwise-identical elements that compare equal) with a
small improvement speed-wise and small code growth (under 2K on x
* Makefile.in (OBJS-libcommon): Add sort.o.
(build/sort.o): New target. Use it...
(BUILD_RTL): ... here, and...
(build/gencfn-macros): ... here, and...
(build/genmatch): ... here.
---
gcc/Makefile.in | 9 +
1 file changed, 5 insertions(+), 4 deleti
* sort.cc: New file.
* system.h [!CHECKING_P] (qsort): Redirect to gcc_qsort.
* vec.c (qsort_chk): Use gcc_qsort.
---
gcc/sort.cc | 232 +++
gcc/system.h | 7 +-
gcc/vec.c| 2 +-
3 files changed, 238 insertio
On Thu, 2018-05-10 at 18:56 +0300, Alexander Monakov wrote:
> * sort.cc: New file.
> * system.h [!CHECKING_P] (qsort): Redirect to gcc_qsort.
> * vec.c (qsort_chk): Use gcc_qsort.
[...snip...]
I'm not a reviewer for this, but there's a lot of fiddly implementation
logic here
Ping: https://gcc.gnu.org/ml/gcc-patches/2018-04/msg00650.html
On 04/19/2018 10:03 AM, Martin Sebor wrote:
Ping: https://gcc.gnu.org/ml/gcc-patches/2018-04/msg00650.html
This just suppresses a duplicate warning. Please let me know
if it's preferable to defer it until GCC 9. Otherwise, I'll
be
On Thu, May 10, 2018 at 06:56:39PM +0300, Alexander Monakov wrote:
> Overall the new implementation is roughly 30% faster compared to Glibc qsort,
> with 2x or more speedup for cases with tiny element count. I see one instance
> where the new approach is significantly (1.5x) slower: it is ipa-icf.c
> As described in the PR, there are a couple of jit testsuite failures on
> Solaris when using /bin/as. The errors point to GC issues and indeed,
> gcc/config/sol2.c lacked GTY markup. This patch fixes that, following
> what darwin.c does for machopic_indirections. I confess I have no idea
> why
On May 10, 2018 5:56:39 PM GMT+02:00, Alexander Monakov
wrote:
>Hello.
>
>This introduces a replacement for qsort() in GCC. The main selling
>point is
>reproducibility (currently compiler output may change depending on how
>libc
>qsort reorders not-bitwise-identical elements that compare equal) w
On May 10, 2018 5:56:40 PM GMT+02:00, Alexander Monakov
wrote:
> * sort.cc: New file.
>* system.h [!CHECKING_P] (qsort): Redirect to gcc_qsort.
>* vec.c (qsort_chk): Use gcc_qsort.
Just a quick first remark - how about putting this into libiberty? And then
name it xqsort?
Hi Ed,
>>> 2018-05-07 Edward Smith-Rowland <3dw...@verizon.net>
>>>
>>> PR libstdc++/83140 - assoc_legendre returns negated value when m is
>>> odd
>>> * include/tr1/legendre_function.tcc (__assoc_legendre_p): Add
>>> __phase
>>> argument defaulted to +1. Doxy comments on same.
>>>
On Thu, 10 May 2018, Jakub Jelinek wrote:
> Have you gathered some statistics on the element sizes and how often they
> appear in qsort calls (perhaps weighted by n*log(n) of the element count)
> across bootstrap+regtest?
No, but Adhemerval Zanella collected stats on bootstrap, and they are simila
On Thu, May 10, 2018 at 3:44 PM, Peryt, Sebastian
wrote:
>> -Original Message-
>> From: Uros Bizjak [mailto:ubiz...@gmail.com]
>> Sent: Thursday, May 10, 2018 3:26 PM
>> To: Peryt, Sebastian
>> Cc: gcc-patches@gcc.gnu.org; Kirill Yukhin
>> Subject: Re: [PATCH][i386] Adding WAITPKG instru
On Thu, 10 May 2018, Richard Biener wrote:
> > - signature-compatible to C qsort, but relaxed contract:
> > - may apply the comparator to elements in a temporary buffer
>
> What consequences has this or rather how is this observable and makes
> comparators behave differently?
The only ser
GCC Maintainers:
The following patch fixes issues found with the instruction counts for
the vsx-vector-6.h test. It was found that on a BE system where GCC is
configured with --with-cpu=power6 the expected counts in vsx-vector-6-
be.c do not match. This patch fixes the expected counts when
confi
Am 10.05.2018 um 17:41 schrieb Steve Kargl:
It is certainly possible to give a warning, but it
would be odd (to me) to warn about technically
standard conforming code.
Maybe we could add
gfc_warning (0, "Standard-conforming code found, your code may run as
expected");
at the end of the comp
There doesn't seem to be any reason for _S_cache_size to be a function
rather than a variable. OK for trunk?
commit bf6843a87d3f64c9ffa29ce8d50aaa6fe8aabcfa
Author: Jason Merrill
Date: Wed Mar 14 21:39:06 2018 -0400
* include/bits/regex_compiler.h (_S_cache_size): Change from
On Thu, May 10, 2018 at 07:56:59PM +0200, Thomas Koenig wrote:
> Am 10.05.2018 um 17:41 schrieb Steve Kargl:
> > It is certainly possible to give a warning, but it
> > would be odd (to me) to warn about technically
> > standard conforming code.
>
> Maybe we could add
>
> gfc_warning (0, "Standard
On Thu, 10 May 2018, Richard Biener wrote:
>
> Just a quick first remark - how about putting this into libiberty? And then
> name it xqsort?
I'm not sure. It has a weaker contract compared to qsort, and I believe
functions in libiberty are understood to provide stronger/better replacements.
On Thu, May 10, 2018 at 7:47 PM, Uros Bizjak wrote:
> Please find attached the patch that fixes all the issues (plus some
> whitespace fixes). It is tested and generates acceptable code. Please
> update the ChangeLog (do not use past tense in the ChangeLog entry!)
> and, if it works for you, plea
GCC Maintainers:
I updated the patch so the dcbtt and dctstt instructions are only
generated if the supported ISA is 2.06 or newer. I addressed the use
of REG instead of REG_P and formatting issues as requested.
The regression testing of the patch was done on
powerpc64le-unknown-linux-gnu (
Running
./f951 -quiet
^D
crashes in gfc_create_decls:
/* Build our translation-unit decl. */
current_translation_unit
= build_translation_unit_decl (get_identifier (main_input_filename));
because main_input_filename is null in this case. We can fix it like
below, where main_input_filename
On Thu, May 10, 2018 at 02:17:15PM -0400, Marek Polacek wrote:
> ./f951 -quiet
> ^D
> crashes in gfc_create_decls:
> /* Build our translation-unit decl. */
> current_translation_unit
> = build_translation_unit_decl (get_identifier (main_input_filename));
> because main_input_filename is nu
I think this patch is causing a glibc testing error. The
tests math/bug-nextafter and math/bug-nexttoward are failing
due to underflow not getting set. Here is a test case that
should print nothing but is currently printing the
'did not underflow' message.
#include
#include
#include
#include
1) There's a function to count how many template headers we should
have, we should use it.
2) While working on something a while back I ran into trying to
instantiate a nested function while still in processing_template_decl
context, which doesn't work so well. Let's check for that.
3) A predica
Alexander Monakov writes:
> I'm not sure. It has a weaker contract compared to qsort, and I believe
> functions in libiberty are understood to provide stronger/better replacements.
The original intent of libiberty was to provide a stronger *portability*
contract, i.e. to work around differences
2310: We were crashing on this testcase; the proposed resolution of
2310 clarifies that we should reject it.
2267: brace and paren initialization should have the same semantics here.
Tested x86_64-pc-linux-gnu, applying to trunk.
commit ac97fad727f7ebcba2d2b345d95867331042a4af
Author: Jason Merri
On Sat, May 5, 2018 at 8:34 PM, sotrdg sotrdg wrote:
> https://github.com/euloanty/mingw-std-random_device/blob/master/random_device_gcc_withcxx11abi/random.cc
It's best to CC the gcc-patches and libstdc++ lists on library contributions.
Jason
On 10/05/18 13:59 -0400, Jason Merrill wrote:
There doesn't seem to be any reason for _S_cache_size to be a function
rather than a variable. OK for trunk?
I vaguely recall some problem with that constant, maybe it didn't work
as a variable once upon a time. If that was ever true it was fixed
l
Hi!
On Wed, May 09, 2018 at 12:42:11PM -0500, Kelvin Nilsen wrote:
> 1. Change the name of the first PowerPC built-in section from
>"PowerPC Built-in Functions" to "Basic PowerPC Built-in Functions".
>This section has never described all PowerPC built-in functions.
After the patch it loo
On 10/05/18 12:17 +0100, Jonathan Wakely wrote:
And a couple of other doc improvements, and regenerated the HTML
pages.
Please read the proposed change to using.xml and let me know if it's
clear.
* doc/xml/faq.xml: Link to C++17 status. Add note to outdated answer.
* doc/xml/man
GCC 8.1 warns for unbounded (and some bounded) string comparisons
involving arrays declared attribute nonstring (i.e., char arrays
that need not be nul-terminated). For instance:
extern __attribute__((nonstring)) char a[4];
int f (void)
{
return strncmp (a, "123", sizeof a);
}
wa
On Fri, Mar 16, 2018 at 8:38 AM, Jason Merrill wrote:
> On Thu, Mar 15, 2018 at 4:50 AM, Richard Biener
> wrote:
>> On Wed, Mar 14, 2018 at 8:57 PM, Jason Merrill wrote:
>>> Ping
>>>
>>> On Fri, Mar 2, 2018 at 1:23 PM, Jason Merrill wrote:
As I mentioned in the PR, the problem here is that
Hi,
On Thu, May 10, 2018 at 10:57:16AM -0700, Carl Love wrote:
> The following patch fixes issues found with the instruction counts for
> the vsx-vector-6.h test. It was found that on a BE system where GCC is
> configured with --with-cpu=power6 the expected counts in vsx-vector-6-
> be.c do not m
Hi Carl,
On Thu, May 10, 2018 at 11:10:29AM -0700, Carl Love wrote:
> * config/rs6000/rs6000.md (prefetch): Generate dcbtt and dcbtstt
> instructions if operands[2] is 0.
"and TARGET_POPCNTD"? Or "and generating code for ISA 2.06 or later".
Something like that.
Looks good, thanks!
> Date: Thu, 10 May 2018 14:03:10 +0200
> From: Jakub Jelinek
> On Thu, May 10, 2018 at 01:51:29PM +0200, Marc Glisse wrote:
> > > > There are probably more
> > > > complicated transformations this disables.
> > >
> > > I'm providing an example from *real* code where the
> > > transformation is
> Date: Thu, 10 May 2018 07:23:05 -0500
> From: Segher Boessenkool
> On Thu, May 10, 2018 at 10:33:39AM +0200, Marc Glisse wrote:
> > int x, y;
> > void f(int n){
> > int c = 3 << 20;
> > x = n / c;
> > y = x / c;
> > }
> Without the replacement we have two dependent divisions; with the
>
Hello!
The missing return in _xgetbv is a bug and will be backported to
release branches.
2018-05-10 Uros Bizjak
* config/i386/i386.c (ix86_expand_builtin) :
Generate SImode target register for null target.
: Ditto.
: Optimize LSHIFTRT generation.
* config/i386/xsaveintrin
Hi Marc,
Can you please comment/respond to Jeff's question below and
confirm whether my understanding of the restriction (below)
is correct?
Thanks
Martin
On 04/30/2018 11:50 AM, Jeff Law wrote:
On 01/12/2018 02:30 PM, Martin Sebor wrote:
A failure in a test for the recently enhanced -Warray-
On Donnerstag, 10. Mai 2018 09:57:22 CEST Jakub Jelinek wrote:
> On Wed, May 09, 2018 at 04:53:19PM +0200, Allan Sandfeld Jensen wrote:
> > > > @@ -2022,8 +2022,9 @@ simplify_vector_constructor
> > > > (gimple_stmt_iterator
> > > > *gsi)>
> > > >
> > > >elem_type = TREE_TYPE (type);
> > > >
On Wed, May 09, 2018 at 06:14:27PM -0500, Segher Boessenkool wrote:
> On Thu, May 03, 2018 at 01:22:10PM -0400, Michael Meissner wrote:
> > 2018-05-03 Michael Meissner
> >
> > * config/rs6000/rs6000.c (mode_supports_d_form): Rename
> > mode_supports_vmx_dform to mode_supports_d_form. A
Hi,
On Thu, May 03, 2018 at 01:23:24PM -0400, Michael Meissner wrote:
> -/* Helper function to say whether a mode supports PRE_INC or PRE_DEC. */
> +/* Helper function to say whether a mode supports PRE_INC or PRE_DEC in a
> given
> + reload register class or if some reload register class supp
On Wed, May 09, 2018 at 05:54:53PM -0500, Segher Boessenkool wrote:
> Hi Mike,
>
> On Thu, May 03, 2018 at 01:17:03PM -0400, Michael Meissner wrote:
> > 2018-05-03 Michael Meissner
> >
> > * config/rs6000/rs6000.c (mode_supports_dq_form): Rename
> > mode_supports_vsx_dform_quad to mode
On Thu, May 03, 2018 at 02:10:50PM -0400, Michael Meissner wrote:
> 2018-05-03 Michael Meissner
>
> * config/rs6000/rs6000.md (mov_hardfloat32, FMOVE64):
> Reformat alternatives and attributes so it is easier to identify
> which constraints/attributes go with which instruction
On Thu, May 10, 2018 at 04:55:02PM -0500, Segher Boessenkool wrote:
> Hi,
>
> On Thu, May 03, 2018 at 01:23:24PM -0400, Michael Meissner wrote:
> > -/* Helper function to say whether a mode supports PRE_INC or PRE_DEC. */
> > +/* Helper function to say whether a mode supports PRE_INC or PRE_DEC i
On Thu, May 03, 2018 at 02:12:05PM -0400, Michael Meissner wrote:
> 2018-05-03 Michael Meissner
>
> * config/rs6000/rs6000.md (mov_softfloat32, FMOVE64):
> Reformat alternatives and attributes so it is easier to identify
> which constraints/attributes go with which instruction
On Thu, May 03, 2018 at 02:13:42PM -0400, Michael Meissner wrote:
> 2018-05-03 Michael Meissner
>
> * config/rs6000/rs6000.md (mov_softfloat64, FMOVE64):
> Reformat alternatives and attributes so it is easier to identify
> which constraints/attributes go with which instruction
On Thu, 10 May 2018, Martin Sebor wrote:
Can you please comment/respond to Jeff's question below and
confirm whether my understanding of the restriction (below)
is correct?
I don't remember it at all, I really should have expanded that comment...
The documentation of nonzero_chars seems to in
Hi,
On Thu, May 03, 2018 at 02:12:55PM -0400, Michael Meissner wrote:
> ; ld/std require word-aligned displacements -> 'Y' constraint.
> ; List Y->r and r->Y before r->r for reload.
> +
> +;; STFD LFD FMR LXSDSTXSD
> +;; LXSDXSTXSDX
On Thu, May 03, 2018 at 02:14:34PM -0400, Michael Meissner wrote:
> 2018-05-03 Michael Meissner
>
> * config/rs6000/rs6000.md (mov_softfloat, FMOVE32):
> Reformat alternatives and attributes so it is easier to identify
> which constraints/attributes go with which instruction.
On Thu, May 10, 2018 at 05:49:12PM -0400, Michael Meissner wrote:
> > > -/* Return true if we have D-form addressing in altivec registers. */
> > > +/* Return true if we have D-form addressing (register+offset) in either a
> > > + specific reload register class or whether some reload register cl
On Thu, May 10, 2018 at 05:20:52PM -0500, Segher Boessenkool wrote:
> On Thu, May 10, 2018 at 05:49:12PM -0400, Michael Meissner wrote:
> > > > -/* Return true if we have D-form addressing in altivec registers. */
> > > > +/* Return true if we have D-form addressing (register+offset) in
> > > > e
The attached patch removed an unused function.
OK to commit?
2018-05-10 Steven G. Kargl
* gfortran.h: Remove prototype.
* symbol.c (gfc_new_undo_checkpoint): Remove unused function.
--
Steve
Index: gcc/fortran/gfortran.h
=
> @@ -31325,7 +31329,21 @@ ix86_init_mmx_sse_builtins (void)
> continue;
>
>ftype = (enum ix86_builtin_func_type) d->flag;
> - def_builtin2 (d->mask, d->name, ftype, d->code);
> + decl = def_builtin2 (d->mask, d->name, ftype, d->code);
> +
> + /* Avoid edges for ptwrite
On 05/10/2018 01:44 PM, Rainer Orth wrote:
Hi Ed,
2018-05-07 Edward Smith-Rowland <3dw...@verizon.net>
PR libstdc++/83140 - assoc_legendre returns negated value when m is
odd
* include/tr1/legendre_function.tcc (__assoc_legendre_p): Add
__phase
argument defaulted to +1. Doxy
Normally we defer instantiation of static data member initializers
until they are needed, but we were doing it immediately for auto
variables.
Tested x86_64-pc-linux-gnu, applying to trunk.
commit a16c8170ac1c148087ee7798b9783656e53ab490
Author: Jason Merrill
Date: Wed Mar 14 20:11:38 2018 -040
On Fri, May 11, 2018 at 2:44 AM, Andi Kleen wrote:
>> @@ -31325,7 +31329,21 @@ ix86_init_mmx_sse_builtins (void)
>> continue;
>>
>>ftype = (enum ix86_builtin_func_type) d->flag;
>> - def_builtin2 (d->mask, d->name, ftype, d->code);
>> + decl = def_builtin2 (d->mask, d->name,
From: Vladimir Mezentsev
When weakref_targets is not empty a target cannot be removed from the weak list.
A small example is below when 'wv12' is removed from the weak list on aarch64:
static vtype Wv12 __attribute__((weakref ("wv12")));
extern vtype wv12 __attribute__((weak));
Bootstrapped
On Thu, 10 May 2018, Jakub Jelinek wrote:
> On Wed, May 09, 2018 at 04:53:19PM +0200, Allan Sandfeld Jensen wrote:
> > > > @@ -2022,8 +2022,9 @@ simplify_vector_constructor (gimple_stmt_iterator
> > > > *gsi)>
> > > >elem_type = TREE_TYPE (type);
> > > >elem_size = TREE_INT_CST_LOW (TYPE_
98 matches
Mail list logo
