Hello!
The only thing I can see is that libgfortran/config/fpu-sysv.h assumes that
FP_RM and others are macros, checking them with #ifdef FP_RM”. Is that the
reason?
It is.
If so, we might just want to use them unconditionally… unless it creates a
mess on some other SysV target!
FWIW,
Hi there,
Currently the insn type of DSP-kind instructions like QSUB8 is alu_reg which
is same as other normal instructions like SUB. In order to distinguish those
DSP-kind instructions, this patch intends to replace current alu_reg with
two sub categories alu_sreg and alu_dsp_reg. Meanwhile the
Hello!
Ping?
While I added the new define_insn_and_split for vec_merge, a bug is
exposed: in config/i386/sse.md, [ define_expand xop_vmfrczmode2 ]
only takes one input, but the corresponding builtin functions have two
inputs, which are shown in i386.c:
{ OPTION_MASK_ISA_XOP,
Hi all,
This patch adds rtx costing code for the round-to-integer instructions
in AArch64.
Some of them are implemented as UNSPECs, so I've added the appropriate
case to the big switch that can be
expanded upon to handle other unspecs in the future.
Tested aarch64-none-elf and bootstrapped on
On 02/07/14 08:59, Christophe Lyon wrote:
Hi,
It seems some of the scan-assembler directives fail:
http://cbuild.validation.linaro.org/build/cross-validation/gcc/trunk/212196/aarch64-none-elf/diff-gcc-rh50-aarch64-none-elf-default-default-default.txt
Christophe.
This patch should fix the
On 30/06/14 21:39, Jeff Law wrote:
On 06/27/14 02:29, Kyrill Tkachov wrote:
Hi all,
This patch generalises the TARGET_MACRO_FUSION_PAIR_P hook usage to work
on more than just
compares and conditional branches for which it was initially designed
for (for x86).
There are some instructions in
On 07/10/2014 06:30 AM, Tim Shen wrote:
Here's the sane patch :)
Ok, thanks!
Paolo.
Hi there,
As the second and final patch in this series, it intends to update alu_reg
and alus_reg types for AArch64 port. With this change, the gcc can be
successfully built for AArch64. Is it OK to trunk?
BR,
Terry
2014-07-10 Terry Guo terry@arm.com
* config/aarch64/aarch64.md
The title says it all.
I've been bootstrapping and testing with this on x86_64-linux for a month.
OK?
Ed
2014-07-10 Edward Smith-Rowland 3dw...@verizon.net
Add the logistic_distribution as an extension.
* include/ext/random: Add the logistic_distribution.
*
Except gcc now knows the size of partial int modes. In this case,
PSImode is 20 bits and TYPE_SIZE is 20 bits, so they match.
I don't understand. The problematic change is
- TYPE_SIZE (type) = bitsize_int (GET_MODE_BITSIZE (TYPE_MODE (type)));
+ TYPE_SIZE (type) = bitsize_int
Since all modes are supported, you can unconditionally return 1 in the
code above, as is the case with fpu-387.h.
Indeed. I’ve committed the attached patch, which performs this trivial cleanup,
adds some comments about unreachable cases (I don’t think we want to error out
at runtime, but a
Here are some C++ versioning changes reflecting C++14 status and adding
c++1z. It is a followup to Jason's patch on 2014-06-26 adding std=c++1z,
etc.
This will allow us to start making C++1z changes to the preprocessor
(n3981 remove trigraphs).
In fact, I made trigraphs opt-in for both
On Thu, Jul 10, 2014 at 4:07 AM, Ed Smith-Rowland 3dw...@verizon.net wrote:
The title says it all.
I've been bootstrapping and testing with this on x86_64-linux for a month.
OK?
Looks good to me.
The memory barriers generated for SPARC are targeting the weakest memory
model allowed for SPARC.
That's not quite true, they are targeting the sparc_memory_model, which is the
memory model selected for the architecture/OS pair by default and which can be
overridden by the user with
Hi!
On Mon, 23 Jun 2014 22:25:41 +0200, Jan Hubicka hubi...@ucw.cz wrote:
this patch makes DECL_VINDEX and DECL_SAVED_TREE to be FUNCTION_DECL only.
Bootstrapped/regtested x86_64-linux, OK?
... without --enable-checking=fold. ;-P
* class.c (check_methods, create_vtable_ptr,
On 1 July 2014 11:05, Christophe Lyon christophe.l...@linaro.org wrote:
* documentation (README)
* dejanu driver (neon-intrinsics.exp)
* support macros (arm-neon-ref.h, compute-ref-data.h)
* Tests for 3 intrinsics: vaba, vld1, vshl
Hi, The terminology in armv8 is advsimd rather than neon.
Hi,
On 07/10/2014 10:07 AM, Ed Smith-Rowland wrote:
The title says it all.
I've been bootstrapping and testing with this on x86_64-linux for a
month.
OK?
Look OK to me too, but I would move both operator() out of line, and
definitely operator()(_UniformRandomNumberGenerator, const
.. I have another comment: are we sure the usual strategy:
templatetypename _UniformRandomNumberGenerator
result_type
operator()(_UniformRandomNumberGenerator __urng)
{ return this-operator()(__urng, this-_M_param); }
doesn't make sense here too?
Paolo.
Hi!
For the record:
On Sat, 05 Apr 2014 12:34:30 +0200, I wrote:
In r209148, I have committed a merge from trunk r209072 (2014-04-04) into
gomp-4_0-branch.
Compared to trunk, there is one now regression:
PASS: gcc.dg/lto/save-temps c_lto_save-temps_0.o assemble, -O -flto
On Wed, Jul 9, 2014 at 6:02 PM, Ulrich Weigand uweig...@de.ibm.com wrote:
Hello,
the implementation of homogenous float aggregates for the ELFv2 ABI has
unfortunately shown to have a bug in a corner case.
The problem is that because such aggregates are packed in the argument
save area, but
The optimization was almost immediately disabled accidentally by:
https://gcc.gnu.org/ml/gcc-cvs/2012-06/msg00577.html
Of course I should have added a testcase in the first place...
Tested on x86_64-suse-linux, applied on all active branches as obvious (this
only affects the Ada compiler).
On Jul 10, 2014, at 2:12 AM, Ed Smith-Rowland 3dw...@verizon.net wrote:
Rename C++1y to C++14.
So generally when we do this we change most of the uses in the test suite of
the c++1y flag to use the standard spelling. Some might appreciate that we
have one test case testing the 1y spelling for
Hi,
since displaying the caret was made the default, the -Wstack-usage warning has
a strange-looking final part:
stack-usage-2.c: In function 'foo2':
stack-usage-2.c:16:1: warning: stack usage is 920 bytes [-Wstack-usage=]
}
^
Tested on x86_64-suse-linux, OK for all active branches?
On Wed, Jul 09, 2014 at 02:19:36PM -0400, David Edelsohn wrote:
This is an ABI change for the affected corner cases of the ELFv2
ABI. However, those cases should be extremely rare; the full
compat.exe and struct-layout-1.exp ABI compatibility test suite
passed, with the exception of two
On Thu, Jul 10, 2014 at 11:50 AM, Thomas Schwinge
tho...@codesourcery.com wrote:
Hi!
On Mon, 23 Jun 2014 22:25:41 +0200, Jan Hubicka hubi...@ucw.cz wrote:
this patch makes DECL_VINDEX and DECL_SAVED_TREE to be FUNCTION_DECL only.
Bootstrapped/regtested x86_64-linux, OK?
... without
Hi!
In r212427, I have committed a merge from trunk r212389 (2014-07-09) into
gomp-4_0-branch.
Grüße,
Thomas
pgppgDa3bbJhH.pgp
Description: PGP signature
I'm seeing ICEs: »tree check: expected function_decl, have type_decl in
fold_checksum_tree, at fold-const.c:14861«. Is the following the correct
fix, or should this be done differently?
No, it seems fine to me. I added it to my local patch, thanks!
Honza
--- gcc/fold-const.c
+++
Jason Merrill ja...@redhat.com writes:
On 07/04/2014 05:13 AM, Dodji Seketeli wrote:
// preprocessed output
# 3 test.cpp 3 4
sys_token
# 3 test.cpp
3
# 3 test.cpp 3 4
sys_token
Yeah. For Built-in tokens that are expanded like that we only do track
their the location of their
On Mon, Jul 7, 2014 at 8:55 AM, Kugan kugan.vivekanandara...@linaro.org wrote:
For -fwrapv I don't see why you'd get into trouble ever, the VRP computation
should be well aware of the -fwrapv semantics and the value ranges should
reflect that.
For -fno-strict-overflow, I have no idea since it
Noticed by David, applied as obvious (stupid C language!).
Richard.
2014-07-10 Richard Biener rguent...@suse.de
PR c-family/61741
* c-c++-common/torture/pr61741.c: Use signed char.
Index: gcc/testsuite/c-c++-common/torture/pr61741.c
On Wed, 9 Jul 2014, Tom G. Christensen wrote:
Patch applies to trunk, gcc 4.9 and gcc 4.8 branches.
2014-07-09 Tom G. Christensen t...@jupiterrise.com
* doc/install.texi: Remove links to defunct package providers for
Solaris
Thanks, Tom! I was just going to create the same
On Wed, Jul 09, 2014 at 05:28:34PM -0400, Jason Merrill wrote:
On 07/09/2014 10:40 AM, Manuel López-Ibáñez wrote:
All of these warnings (-Wsizeof-pointer-memaccess, -Wsizeof-array-argument
and -Wmemset-transposed-args) are implemented in a hackish way, because we
fold everything too early.
Original Message
Subject:[PING][PATCH] Fix for PR 61561
Date: Mon, 30 Jun 2014 19:21:49 +0400
From: Marat Zakirov m.zaki...@samsung.com
To: gcc-patches@gcc.gnu.org gcc-patches@gcc.gnu.org
CC: Ramana Radhakrishnan ramana.radhakrish...@arm.com, Richard
Earnshaw
On 30/06/14 16:21, Marat Zakirov wrote:
Thank for your attention.
This is OK for trunk - Sorry about the delayed response.
Ramana
Marat.
On Thu, Jun 19, 2014 at 9:19 PM, Yuri Gribov tetra2...@gmail.com wrote:
Thirdly, we also need to fix movhi_bytes (for pre-v4) thumb2_movhi_insn
(for thumb2) and, quite possibly, thumb1_movhi_insn (for thumb1). There
may well be additional changes for movqi variants as well.
A general
On 09-07-14 23:06, Tom de Vries wrote:
On 09-07-14 20:32, Richard Sandiford wrote:
Tom de Vries tom_devr...@mentor.com writes:
Richard,
during testing the gcc.target/mips/fuse-caller-save*.c test-cases with more
combinations of -march, -mabi, -fpic etc, I found that the checks for amount of
Tom de Vries tom_devr...@mentor.com writes:
The mips16e save/restore enabling is controlled by this code in mips.h:
...
/* Generate mips16e code. Default 16bit ASE for mips32* and mips64* */
#define GENERATE_MIPS16E(TARGET_MIPS16 mips_isa = 32)
/* Generate mips16e register
Here is a preprocessor patch to make error messages show C++11 and other
relevant C++ language instead of C99.
Built and tested on x86_64-linux.
OK?
libcpp/
2014-07-09 Edward Smith-Rowland 3dw...@verizon.net
PR CPP/61389
* macro.c (_cpp_arguments_ok, parse_params,
On Fri, 27 Jun 2014, Richard Biener wrote:
On Fri, 27 Jun 2014, Jakub Jelinek wrote:
On Fri, Jun 27, 2014 at 01:49:38PM +0200, Richard Biener wrote:
I'm going to go for a single load/store and MOVE_MAX for now - I
have quite some fallout to deal with anyway (analyzed strlenopt-1.c
The following patch partially reverts r212352 to restore bootstrap
on i585-linux (where I reproduced and verified the fix). It makes
phicprop honor that loop-depth condition again which doing not
seems to cause a latent bug to pop up which miscompiles the stage2
compiler which in turn
On Fri, Jul 4, 2014 at 2:11 PM, Jan-Benedict Glaw jbg...@lug-owl.de wrote:
On Wed, 2014-06-25 08:55:57 -0700, augustine.sterl...@gmail.com
augustine.sterl...@gmail.com wrote:
On Tue, Jun 24, 2014 at 10:20 PM, Jan Hubicka hubi...@ucw.cz wrote:
This is the xtensa fallout, see
Jakub Jelinek wrote:
On Wed, Jul 09, 2014 at 02:19:36PM -0400, David Edelsohn wrote:
This is an ABI change for the affected corner cases of the ELFv2
ABI. However, those cases should be extremely rare; the full
compat.exe and struct-layout-1.exp ABI compatibility test suite
passed,
On Mon, Jun 30, 2014 at 1:38 AM, Marc Glisse marc.gli...@inria.fr wrote:
Hello,
with this patch on top of
https://gcc.gnu.org/ml/gcc-patches/2014-06/msg02315.html
we finally warn for the testcase of PR 60517.
The new function is copied from init_subtree_with_zero right above. I guess
it
On Thu, Jul 10, 2014 at 04:54:53PM +0200, Richard Biener wrote:
+ else if (access-grp_to_be_debug_replaced)
+{
Why would we care to create clobbers for debug stmts?! Are those
even valid?
It is not valid. Though, the fields supposedly live nowhere after the
clobber, so perhaps one
On Sun, Jun 29, 2014 at 12:33 AM, Marc Glisse marc.gli...@inria.fr wrote:
Hello,
we currently drop clobbers on variables whose address is not taken anymore.
However, rewrite_stmt has code to replace them with an SSA_NAME with a
default definition (an uninitialized variable), and I believe
On Thu, Jul 10, 2014 at 04:30:13PM +0200, Richard Biener wrote:
Compromise hack below. It simply avoids the transform for
sources that c_strlen can compute a length of. That fixes all
strlenopt testcase apart from strlenopt-8.c which does
memcpy (, flag ? a : b); which then still folds
The following makes sure we fold the reads from string constants
created from folding memcpy (foo, string_cst[0], ...) to
eventually create similar code as if that memcpy folding wasn't
done (move-by-pieces).
Bootstrapped and tested on x86_64-unknown-linux-gnu.
Comments?
(yeah, those
On Mon, Jul 7, 2014 at 6:40 PM, Richard Henderson r...@redhat.com wrote:
On 07/03/2014 02:53 AM, Evgeny Stupachenko wrote:
-expand_vec_perm_palignr (struct expand_vec_perm_d *d)
+expand_vec_perm_palignr (struct expand_vec_perm_d *d, int insn_num)
insn_num might as well be bool avx2, since
On Thu, Jul 10, 2014 at 05:12:53PM +0200, Richard Biener wrote:
The following makes sure we fold the reads from string constants
created from folding memcpy (foo, string_cst[0], ...) to
eventually create similar code as if that memcpy folding wasn't
done (move-by-pieces).
Bootstrapped and
Hi,
On Thu, 10 Jul 2014, Richard Biener wrote:
Apart from the out-of-SSA patch you proposed elsewhere a possibility
is to simply never mark undefined SSA names as
SSA_NAME_OCCURS_IN_ABNORMAL_PHI ... (or not mark those
as must-coalesce).
The insight to note is, that undefined SSA names
which means that the precision of the mode is used to set the size
of the type, which very likely means that the size of the mode is
larger. So the size of the mode will be larger than the size of the
type, which is a lie.
For partial int modes, the precision and size are the same, and
OK.
Jason
I've backported this patch from trunk at r212211. Committed to gcc-4_9
at r212434.
https://gcc.gnu.org/ml/gcc-patches/2014-07/msg00076.html
-cary
2014-07-01 Cary Coutant ccout...@google.com
gcc/
* dwarf2out.c (remove_addr_table_entry): Remove unnecessary hash table
OK.
Jason
This fixes the uses-allocator construction in experimental::any and
re-orders the data members so that I can extract the stored value in a
Python printer without having to walk the inheritance hierarchy of a
std::tuple.
The actual pretty printer for any (and string_view and optional) will
follow
On 07/10/14 09:48, Michael Matz wrote:
Hi,
On Thu, 10 Jul 2014, Richard Biener wrote:
Apart from the out-of-SSA patch you proposed elsewhere a possibility
is to simply never mark undefined SSA names as
SSA_NAME_OCCURS_IN_ABNORMAL_PHI ... (or not mark those
as must-coalesce).
The insight to
On 07 Jul 17:03, Bernd Schmidt wrote:
Is libgomp the only problematic one? (Does the accel compiler even
need one?) It seems to be installed in /usr/lib rather than in a
gcc-specific directory, which is a little surprising to me. It may
be necessary to give the accel compiler a different
On 07/10/2014 08:24 PM, Ilya Verbin wrote:
On 07 Jul 17:03, Bernd Schmidt wrote:
Is libgomp the only problematic one? (Does the accel compiler even
need one?) It seems to be installed in /usr/lib rather than in a
gcc-specific directory, which is a little surprising to me. It may
be necessary to
Hi Thomas,
This patch addresses two bugs openacc reduction clause bugs. The first
bug occurred because I didn't anticipate a GIMPLE_BIND stmt to be passed
to process_reduction_data. Turns out, this could happen with the
collapse clause. That's because the variables which were declared inside
the
Hi Thomas,
These patch enables the collapse clause with a value greater than one.
We had partial support for the collapse clause which supported a value
of 1, because the fortran frontend always associated a collapse clause
with each openacc for loop. However in terms of actual collapse support,
My earlier patch for 61659 caused more virtual functions to be
instantiated when -fdevirtualize is on, leading to additional errors
appearing at higher optimization levels. This patch shifts that
instantiation to a new flag, -fuse-all-virtuals, which is on by default,
and adds an explanatory
Hi,
after more than 2 years, I'm finally back to this issue:
https://gcc.gnu.org/ml/gcc-patches/2012-05/msg01442.html
https://gcc.gnu.org/ml/gcc-patches/2012-05/msg01502.html
and the below draft, which passes testing, tries to implement as closely
as possible what Jason suggested in
On 07/10/2014 08:13 AM, Dodji Seketeli wrote:
Jason Merrill ja...@redhat.com writes:
On 07/04/2014 05:13 AM, Dodji Seketeli wrote:
// preprocessed output
# 3 test.cpp 3 4
sys_token
# 3 test.cpp
3
# 3 test.cpp 3 4
sys_token
Yeah. For Built-in tokens that are expanded like that we only do
Hmm, why aren't we already getting the error from
if (convs-check_narrowing)
check_narrowing (totype, expr);
in convert_like_real? Is it that we need to copy LOOKUP_NO_NARROWING
into convflags in build_user_type_conversion_1?
Jason
The fact that extra line directives are inserted around built-in
tokens isn't ideal, but I must concur with Dodji's assessment that
such a fix belongs in a separate patch.
The purpose of this patch is to resolve a discrepancy between
integrated-cpp and non-integrated-cpp. The locations of
On Jul 10, 2014, at 8:00 PM, Kyrill Tkachov kyrylo.tkac...@arm.com wrote:
On 30/06/14 21:39, Jeff Law wrote:
On 06/27/14 02:29, Kyrill Tkachov wrote:
Hi all,
This patch generalises the TARGET_MACRO_FUSION_PAIR_P hook usage to work
on more than just
compares and conditional branches for
If we handle PTRMEM_CST specially, we also need to handle any
CONSTRUCTORs that wrap them.
Tested x86_64-pc-linux-gnu, applying to trunk and 4.9.
commit a2f90f6d52fa4c3d5391b23914c4177c3272817c
Author: Jason Merrill ja...@redhat.com
Date: Thu Jul 10 17:05:28 2014 -0400
PR c++/61661
since displaying the caret was made the default, the -Wstack-usage warning has
a strange-looking final part:
stack-usage-2.c: In function 'foo2':
stack-usage-2.c:16:1: warning: stack usage is 920 bytes [-Wstack-usage=]
}
^
Your patch hides the caret but the location still points to the
Hi,
On 07/10/2014 10:55 PM, Jason Merrill wrote:
Hmm, why aren't we already getting the error from
if (convs-check_narrowing)
check_narrowing (totype, expr);
in convert_like_real? Is it that we need to copy LOOKUP_NO_NARROWING
into convflags in build_user_type_conversion_1?
Ah, ah,
... the below is another, very safe, option for setting
LOOKUP_NO_NARROWING in flags.
Paolo.
Index: call.c
===
--- call.c (revision 212431)
+++ call.c (working copy)
@@ -3586,7 +3586,8 @@ build_user_type_conversion_1
... and of course the problem with all such ideas is that we easily end
up warning twice in all the simple cases which check_narrowing can
already handle. Something like the attached has more chances of passing
the testsuite while not regressing in terms of duplicate warnings (which
are hard
Index: gcc/testsuite/g++.dg/cpp1y/digit-sep-neg.C
===
--- gcc/testsuite/g++.dg/cpp1y/digit-sep-neg.C(revision 212440)
+++ gcc/testsuite/g++.dg/cpp1y/digit-sep-neg.C(working copy)
@@ -7,7 +7,7 @@
i = 1048''576; // {
On Thu, 10 Jul 2014, Jakub Jelinek wrote:
+Wmemset-transposed-args
+C ObjC C++ ObjC++ Var(warn_memset_transposed_args) Warning LangEnabledBy(C
ObjC C++ ObjC++,Wall)
+Warn about suspicious call to memset where the third argument is constant
zero and second is not zero
calls (plural), like
The Go library reflect package used to have a notion of reflectFlags
that were used to indicate an invalid interface value. That has not
been used for a while. This patch drops testing for it from the libgo
runtime. Bootstrapped and ran Go testsuite on
x86_64-unknown-linux-gnu. Committed to
A new version of the master Go library has a new file runtime/iface.goc.
It is not the same as the existing libgo file runtime/iface.goc. This
patch renames the libgo file to go-iface.goc to avoid any future
confusion. Bootstrapped and ran Go testsuite on
x86_64-unknown-linux-gnu. Committed to
I've ported the libstdc++ pretty printers to work with Python 3 as
well as Python 2, and made a couple of other changes that I think are
improvements along the way. The important change here is the final
patch; the others just made things easier at some point along the way.
I started with a
When I try to build test just libstdc++, or to run the testsuite
from trunk against my installed libstdc++, the testsuite tries to pass
-fdiagnostics-color=never to the system GCC, which is too old to
know what that is.
Since I really just want to test a patch for the gdb pretty-printers,
and
We load our pretty-printers explicitly, and we shouldn't need any other
random -gdb.gdb or -gdb.py files from anywhere, so in this patch we turn
that off by running set auto-load no.
Also, run info share so that the list of loaded libraries ends up in
the logs for the GDB tests.
libstdc++-v3/
Loosely based on Mathias Klose' earlier patch
http://patchwork.ozlabs.org/patch/287368/.
Tested with:
$ make check-target-libstdc++-v3 RUNTESTFLAGS='--directory
libstdc++-prettyprinters'
libstdc++-v3/
PR libstdc++/58962
* python/libstdcxx/v6/printers.py: Port to Python 2+3
Loosely based on Matthias Klose's earlier patch
http://patchwork.ozlabs.org/patch/287368/.
This time with his name spelled correctly; sorry about that!
Tested with:
$ make check-target-libstdc++-v3 RUNTESTFLAGS='--directory
libstdc++-prettyprinters'
libstdc++-v3/
PR libstdc++/58962
80 matches
Mail list logo