On 07/08/14 10:43, Bin Cheng wrote:
Hi,
Case pr61772.c scans specific string in assembly file, and it is run for
many different option combinations. When it's tested against different lto
option combinations on targets like ARM, the generated pr61772.s may only
contain lto object content, rathe
On Thu, Aug 7, 2014 at 8:06 PM, Richard Biener
wrote:
> On Thu, Aug 7, 2014 at 11:46 AM, Bin Cheng wrote:
>> Hi,
>> As analyzed in PR62032, this patch fixes the latent lto bug by switching
>> arguments of lto_define_builtins, otherwise vsnprintf-chk.c would fail on
>> arm/aarch64 with lto options
On Fri, Aug 08, 2014 at 10:11:25AM +0400, Yury Gribov wrote:
> commit b90c4abf0b804c60b548f94fe59118d029363140
> Author: Yury Gribov
> Date: Wed Aug 6 11:05:39 2014 +0400
>
> Added fnspec to internal functions.
>
> 2014-08-08 Yury Gribov
>
> * gimple.c (gimple_call_fn
On Fri, Aug 08, 2014 at 10:11:30AM +0400, Yury Gribov wrote:
> commit 22a79d39656d3d600e97c7ae599b86b6d25efe00
> Author: Yury Gribov
> Date: Wed Aug 6 11:06:24 2014 +0400
>
> Move inlining of Asan memory checks to sanopt pass.
> Change asan-instrumentation-with-call-threshold to more cl
On Fri, Aug 8, 2014 at 4:05 PM, Kyrill Tkachov wrote:
>
> On 07/08/14 10:43, Bin Cheng wrote:
>>
>> Hi,
>> Case pr61772.c scans specific string in assembly file, and it is run for
>> many different option combinations. When it's tested against different
>> lto
>> option combinations on targets li
Hi,
On 08/07/2014 07:44 PM, Jason Merrill wrote:
On 08/07/2014 12:41 PM, Paolo Carlini wrote:
Ok. Probably at some point I should review which kind of diagnostics
the various implementations emit: for example *for enum27.C*
currently clang talks about *narrowing*, the kind of of diagnostics
w
On Thu, Aug 7, 2014 at 4:58 PM, Kyrill Tkachov wrote:
> Hi all,
>
> This test in the torture suite fails on -flto because it scans for a
> sentinel string inside an asm block that it wants to make sure is not
> optimised away. However, with -flto gcc uses slim LTO objects which contain
> a binary
On Fri, Aug 8, 2014 at 12:42 AM, Prathamesh Kulkarni
wrote:
> reject operator to be used as variable in for.
> eg:
> (for mult in plus minus)
>
>
> * genmatch.c (get_operator): New function.
> (e_operation::e_operation): Adjust to call get_operator.
> (parse_for): Add call to get_
On Fri, Aug 8, 2014 at 1:59 AM, Yi Yang wrote:
> Remove a redundant assignment "*predictor = PRED_BUILTIN_EXPECT;",
> since six lines later *predictor is assigned again.
Ok.
Thanks,
Richard.
> --
>
> 2014-08-07 Yi Yang
>
> gcc:
> * predict.c (expr_expected_value_1): Remove the redundant
On Fri, Aug 8, 2014 at 10:05 AM, Bin.Cheng wrote:
> On Thu, Aug 7, 2014 at 8:06 PM, Richard Biener
> wrote:
>> On Thu, Aug 7, 2014 at 11:46 AM, Bin Cheng wrote:
>>> Hi,
>>> As analyzed in PR62032, this patch fixes the latent lto bug by switching
>>> arguments of lto_define_builtins, otherwise vs
On Fri, Aug 8, 2014 at 4:54 PM, Richard Biener
wrote:
> On Fri, Aug 8, 2014 at 10:05 AM, Bin.Cheng wrote:
>> On Thu, Aug 7, 2014 at 8:06 PM, Richard Biener
>> wrote:
>>> On Thu, Aug 7, 2014 at 11:46 AM, Bin Cheng wrote:
Hi,
As analyzed in PR62032, this patch fixes the latent lto bug b
Put check for EOF in parse_for and parse_if,
else it results in segmentation fault.
eg:
(if (0)
Not sure why it segfaults.
I got following back-trace (in gdb):
#0 _cpp_lex_direct (pfile=pfile@entry=0x66f560) at ../../src/libcpp/lex.c:2183
#1 0x0041d50c in _cpp_lex_token (pfile=pfile
On 08/08/14 09:24, Bin.Cheng wrote:
On Fri, Aug 8, 2014 at 4:05 PM, Kyrill Tkachov wrote:
On 07/08/14 10:43, Bin Cheng wrote:
Hi,
Case pr61772.c scans specific string in assembly file, and it is run for
many different option combinations. When it's tested against different
lto
option combina
... I'm attaching some code accompanying my blabber. Note - in fact I
noticed this some days ago - that probably the idea of passing complain
to check_narrowing is good anyway, otherwise we have a function emitting
unconditional warnings called by functions which have a tsubst_flags_t.
Paolo.
> OK for mainline.
Thanks Uros. Committed to revision 213572
I would like to backport to 4.9 branch too.
Is it OK?
- Ganesh
Hello Manuel,
Manuel López-Ibáñez writes:
> 2014-08-03 Manuel López-Ibáñez
>
> PR fortran/44054
> c-family/
> * c-format.c: Handle Fortran flags.
> * diagnostic.c (build_message_string): Make it extern.
> * diagnostic.h (build_message_string): Make it extern.
Small nit; the d
Hi all,
I have been working on Asan global optimization pass lately. The goal is
to remove redundant Asan checks from sanitized code. This should
hopefully reduce Asan's speed/size overhead (which is currently ~2x).
The patch is not yet ready for trunk (e.g. I haven't done bootstrap,
etc. but
On Fri, Aug 08, 2014 at 02:26:25PM +0400, Yury Gribov wrote:
> I have been working on Asan global optimization pass lately. The goal is to
> remove redundant Asan checks from sanitized code. This should hopefully
> reduce Asan's speed/size overhead (which is currently ~2x). The patch is not
> yet r
On 08/08/14 01:30, Janis Johnson wrote:
> Running an arm-none-eabi test that adds "-mthumb" for a multilib that
> uses "-march=armv5te -mfloat-abi=hard" works fine if the test doesn't
> have any calls, but if it does then the effective-target checks for
> arm_thumb1_ok and arm_thumb2_ok pass but th
On 08/08/14 01:31, Janis Johnson wrote:
> Five tests in gcc.target/arm add -mthumb without first checking that the
> option is compatible with the current multilib flags. This patch fixes
> that by skipping each test if neither arm_thumb1_ok or arm_thumb2_ok is
> true. Tested by running gcc.targe
On Fri, Aug 8, 2014 at 2:37 PM, Jakub Jelinek wrote:
> On Fri, Aug 08, 2014 at 02:26:25PM +0400, Yury Gribov wrote:
>> I have been working on Asan global optimization pass lately. The goal is to
>> remove redundant Asan checks from sanitized code. This should hopefully
>> reduce Asan's speed/size
On 07/08/14 09:46, Kyrill Tkachov wrote:
Hi all,
As Bin pointed out, setting -march=armv5 explicitly on this test causing a FAIL
when testing with /-mthumb.
The bug fixed by the commit that added this test occured only on pre-SMP
architectures, i.e. ARMv5 and earlier.
I think it's best to just
On Fri, Aug 8, 2014 at 12:14 PM, Gopalasubramanian, Ganesh
wrote:
>> OK for mainline.
>
> Thanks Uros. Committed to revision 213572
> I would like to backport to 4.9 branch too.
> Is it OK?
Yes, looks safe for backport.
Thanks,
Uros.
Bootstrapped on x86_64-unknown-linux-gnu.
Richard.
2014-08-08 Richard Biener
Merge from trunk r213455 through r213753.
Steven,
this patch fixes:
- PR62004 (the if-conversion pass part, the tail-merge part is still todo), and
- PR62030.
In both cases, a valid program with a dead type-unsafe access is transformed by
the if-conversion pass into an invalid program with a live type-unsafe access.
The transformatio
On Fri, 8 Aug 2014, Tom de Vries wrote:
> Steven,
>
> this patch fixes:
> - PR62004 (the if-conversion pass part, the tail-merge part is still todo),
> and
> - PR62030.
>
> In both cases, a valid program with a dead type-unsafe access is transformed
> by the if-conversion pass into an invalid pr
On 30 Jul 10:50, Uros Bizjak wrote:
> (Please take care not to break bootstrap after every part, so the
> source will always be bisectable).
Sure, that is the plan.
--
Thanks, K
This merges making lto_input_block a class, getting rid of
LTO_INIT_INPUT_BLOCK. It also simplifies lto headers by
removing unused and write-only fields and making them
a inheritance hierarchy.
LTO bootstrap running on x86_64-unknown-linux-gnu.
Richard.
2014-08-08 Richard Biener
*
On Fri, Aug 8, 2014 at 11:50 AM, Prathamesh Kulkarni
wrote:
> Put check for EOF in parse_for and parse_if,
> else it results in segmentation fault.
>
> eg:
> (if (0)
>
>
> Not sure why it segfaults.
Hmm, I think this is a bug in libcpp which seems to
crash when you do cpp_peek_token (r, 0); cp
On Tue, 05 Aug 2014 21:34:00 +0200, Tom Tromey wrote:
> > "Jeff" == Jeff Law writes:
> Jeff> Does this deserve a mention in the news file?
>
> I suppose so, I will get someone here to write it.
Attached (based on Tom's PATCH 0/5 mail).
Jan
Index: htdocs/gcc-5/changes.html
=
Hello,
This patch introduces `-mavx512bw' compiler switch.
Bootstrapped.
gcc/
* common/config/i386/i386-common.c
(OPTION_MASK_ISA_AVX512BW_SET) : Define.
(OPTION_MASK_ISA_AVX512BW_UNSET): Ditto.
(OPTION_MASK_ISA_AVX512VL_UNSET) : Ditto.
(ix86_handle_option):
Hello,
This patch introduces `-mavx512vl' option to GCC
Bootstrapped.
gcc/
* common/config/i386/i386-common.c
(OPTION_MASK_ISA_AVX512VL_SET): Define.
(OPTION_MASK_ISA_AVX512F_UNSET): Update.
(ix86_handle_option): Handle OPT_mavx512vl.
* config/i386/cpuid.h
This patch revisits the normalization and checking of constraints. In
particular, it ensures that associated constraints are always
normalized so that checking constraints will never instantiate a
concept declaration.
This also removes the constraint check from fn_type_unification as per
discussio
> I think this is fine. On the other side, I think the test case itself
> is unnecessarily large. I would assume the majority of the code could
> be deleted while still triggering the bug. Could you give it a shot.
I've attached an improved version of the patch.
> Is there anything else you still
On 08/08/2014 15:11, Roman Gareev wrote:
I think this is fine. On the other side, I think the test case itself
>is unnecessarily large. I would assume the majority of the code could
>be deleted while still triggering the bug. Could you give it a shot.
I've attached an improved version of the pat
On 08/08/2014 06:03 AM, Paolo Carlini wrote:
Note - in fact I
noticed this some days ago - that probably the idea of passing complain
to check_narrowing is good anyway, otherwise we have a function emitting
unconditional warnings called by functions which have a tsubst_flags_t.
Makes sense.
Bu
On 08-08-14 13:36, Richard Biener wrote:
On Fri, 8 Aug 2014, Tom de Vries wrote:
Steven,
this patch fixes:
- PR62004 (the if-conversion pass part, the tail-merge part is still todo),
and
- PR62030.
In both cases, a valid program with a dead type-unsafe access is transformed
by the if-conversi
On Tue, Aug 5, 2014 at 10:31 AM, Zhenqiang Chen
wrote:
> Hi,
>
> For some large constants, ARM will split them during expanding, which
> makes impossible to hoist them out the loop or shared by different
> references (refer the test case in the patch).
>
> The patch keeps some constants in registe
Recent changes to the MIPS binutils sources have made it necessary for
GCC to pass the -msoft-float to the assembler if it wants an object
file marked as soft-float. This patch makes GCC pass any -mhard-float,
or -msoft-float flags that were used on the compile line on to the
assembler so the exec
This libgcc patch is the second of two patches needed to build GCC soft-float
multilibs when using the latest binutils. It skips assembling any of the code
in mips16.S when in soft-float mode because the code is not used when in
soft-float mode and because doing so generates errors during assembly
Hi,
On 08/08/2014 03:55 PM, Jason Merrill wrote:
On 08/08/2014 06:03 AM, Paolo Carlini wrote:
Note - in fact I
noticed this some days ago - that probably the idea of passing complain
to check_narrowing is good anyway, otherwise we have a function emitting
unconditional warnings called by functi
This patch to the Go frontend fixes a compiler crash if a compound
statement winds up doing nothing at all. The test case is
http://codereview.appspot.com/124020044 and will go into the master Go
testsuite. Bootstrapped and ran Go testsuite on
x86_64-unknown-linux-gnu. Committed to mainline.
Ia
Thank you for the comment, I've updated the patch.
OK for trunk and 4.9 branch?
2014-08-08 Guozhi Wei
* config/rs6000/rs6000.md (*movdi_internal64): Add a new constraint.
On Wed, Aug 6, 2014 at 7:28 PM, Segher Boessenkool
wrote:
> On Wed, Aug 06, 2014 at 04:48:26PM -0700, Carrot W
On Aug 7, 2014, at 1:38 AM, Kyrill Tkachov wrote:
>
> Thanks for the detailed explanation, the linker errors I was seeing were
> about relocations being truncated.
Ah, those are bugs in your port! You should be able to generate large code and
then relax it into short small code. Large code,
On Aug 7, 2014, at 12:47 AM, Tom de Vries wrote:
> Done. Committed as attached.
Thanks.
Oddly, this case wasn't tested before.
Tested x86_64-pc-linux-gnu, applying to trunk.
commit b5879e54d61597e2255fd6a25affcb24862f2844
Author: Jason Merrill
Date: Fri Aug 8 10:16:43 2014 -0400
* pt.c (lookup_template_class_1): Copy abi_tag.
diff --git a/gcc/cp/pt.c b/gcc/cp/pt.c
index 30
On 08/08/2014 01:04 PM, Paolo Carlini wrote:
-float d = { i };
+float d = { i }; // { dg-error "narrowing" }
No, this case is the entire point of -Wno-narrowing.
I was suggesting that we ignore -Wno-narrowing *for a constant value*,
but i is not a constant-expression.
Jason
On Fri, Aug 08, 2014 at 02:17:04PM -0400, Jason Merrill wrote:
> --- a/gcc/cp/pt.c
> +++ b/gcc/cp/pt.c
> @@ -7813,6 +7813,12 @@ lookup_template_class_1 (tree d1, tree arglist, tree
> in_decl, tree context,
> }
> }
>
> + if (OVERLOAD_TYPE_P (t)
> + && !DECL_ALIAS_TEMPLA
Hi, Vlad
Why does rs6000_emit_move need special support for SDmode with LRA and
not with reload? In other words, why is this a specific fix for
rs6000 instead of a general improvement for reload?
SDmode has some bizarre restrictions on PPC, but one of the historical
advantages of GCC was reload'
On Fri, Aug 8, 2014 at 1:50 PM, Carrot Wei wrote:
> Thank you for the comment, I've updated the patch.
>
> OK for trunk and 4.9 branch?
>
>
> 2014-08-08 Guozhi Wei
>
> * config/rs6000/rs6000.md (*movdi_internal64): Add a new constraint.
>
>
> On Wed, Aug 6, 2014 at 7:28 PM, Segher Boess
Committed to branch dmalcolm/jit:
This adds two public API entrypoints:
gcc_jit_context_new_function_ptr_type
gcc_jit_context_new_call_through_ptr
gcc/jit/
* TODO.rst: Function ptrs are done.
* internal-api.c
(gcc::jit::recording::context::new_function_ptr_type): New
On Thu, Aug 7, 2014 at 2:41 PM, Michael Meissner
wrote:
> I'm starting to look at updating my old address branch with an eye towards
> getting the changes committed in GCC 4.10. The address branch is meant to
> rewrite handling of addresses in the rs6000 backend, to generalize the
> addresses bef
Committed to branch dmalcolm/jit:
test-combination.c attempts to run all tests within one context, but
was missing a couple of them. Add the missing tests.
gcc/testsuite/
* jit.dg/test-combination.c (create_code): Add missing calls to
create_code_quadratic and create_code_reading
On Fri, 2014-08-08 at 08:38 -0700, Steve Ellcey wrote:
> Recent changes to the MIPS binutils sources have made it necessary for
> GCC to pass the -msoft-float to the assembler if it wants an object
> file marked as soft-float. This patch makes GCC pass any -mhard-float,
> or -msoft-float flags tha
Here is another MIPS patch. This removes the fp64 multilib from the
mips-mti-* targets. With the new fpxx we no longer want special fp64
multilibs in the mti targets. Since it doesn't affect any other targets
hopefully there is no objection to checking this in.
Tested on mips-mti-linux-gnu and
This patch removes skeleton type units that were being produced
with -gsplit-dwarf. These sections were originally intended as
targets for .gdb_index entries that needed to point to type units.
Because of the limitations of the .debug_gnu_pubnames/pubtypes
sections with split DWARF, we were not abl
Hi,
On 08/08/2014 08:22 PM, Jason Merrill wrote:
On 08/08/2014 01:04 PM, Paolo Carlini wrote:
-float d = { i };
+float d = { i }; // { dg-error "narrowing" }
No, this case is the entire point of -Wno-narrowing.
I was suggesting that we ignore -Wno-narrowing *for a constant value*,
but i is
Pushed to branch dmalcolm/jit:
This adds a new API entrypoint:
gcc_jit_context_new_union_type
gcc/jit/
* libgccjit.h (gcc_jit_context_new_union_type): New entrypoint.
(gcc_jit_lvalue_access_field): Rename first param from "struct_"
to "struct_or_union".
(gcc_jit
On 08/08/2014 04:52 PM, Paolo Carlini wrote:
+ else if (!TREE_CONSTANT (init))
+ {
+ if (complain & tf_warning_or_error)
+ pedwarn (EXPR_LOC_OR_LOC (init, input_location), OPT_Wnarrowing,
+"narrowing conversion of %qE from %qT to %qT inside { }",
+
On 2014-08-08, 2:53 PM, David Edelsohn wrote:
Hi, Vlad
Why does rs6000_emit_move need special support for SDmode with LRA and
not with reload?
rs6000_emit_move has also special SDmode support for *reload too*.
It is not implemented generically in reload as you wrote. Please, look
at the c
I've backported this patch from trunk r213765.
These sections were originally intended as targets for .gdb_index
entries that needed to point to type units. Because of the limitations
of the .debug_gnu_pubnames/pubtypes sections with split DWARF, we were
not able to pass along enough information
Jonathan Wakely writes:
> On 23/07/14 11:58 +0200, Marc Glisse wrote:
> As an aside, we already have divide-by-zero bugs in , it
> would be nice if someone could look at that.
>
> https://gcc.gnu.org/bugzilla/show_bug.cgi?id=60037
Sorry, it took a while to get back to tihs and now the referenced
Hi,
On 08/08/2014 11:18 PM, Jason Merrill wrote:
On 08/08/2014 04:52 PM, Paolo Carlini wrote:
+ else if (!TREE_CONSTANT (init))
+{
+ if (complain & tf_warning_or_error)
+pedwarn (EXPR_LOC_OR_LOC (init, input_location),
OPT_Wnarrowing,
+ "narrowing conversion
.. by^2 the way, I think c++/57460 my have something to do with my
curiosities about non-constant vs SFINAE...
Paolo.
On 08/08/2014 07:26 PM, Paolo Carlini wrote:
Ok (that seems a latent buglet...). By the way, since I wondered briefly
about SFINAE wrt the existing documentation, does what we do in the
*non* constant case really matter for SFINAE?
Yes; in unevaluated context, at least.
Anyway, Ok if testing
The problem is that HJL's patch in 2011 to try and fix his miscompiling of x32
ld.so was in fact incorrect and causes problems with AARCH64 ILP32.
The testcase which is being miscompiled after HJL's patch (which he was also
trying to fix up with his patch on x32):
void f(int *a, int b, long long d)
This patch fixes the original problem that HJL was having with x32 in
PR 47727, it is more constraint than HJL's patch only care about what
is happening inside a CONST; if we allow it for other cases,
the RTL and generated code is incorrect as it does not have the needed
zero extend. This allows
OK? When the second patch is approved?
Thanks,
Andrew Pinski
ChangeLog:
Revert:
2011-08-19 H.J. Lu
PR middle-end/49721
* explow.c (convert_memory_address_addr_space): Also permute the
conversion and addition of constant for zero-extend.
---
gcc/explow
From: Andi Kleen
To use gcc-{ar,ranlib} for boot strap we need to add a -B option
to the tool. Since ar has weird and unusual argument conventions
implement the code by hand instead of using any libraries.
v2: Fix typo
gcc/:
2014-08-04 Andi Kleen
* gcc-ar.c (main): Support -B optio
From: Andi Kleen
Change the bootstrap-lto config file to use slim (non fat) LTO..
Speeds up the LTO bootstrap by ~18% on a 4 core system.
This requires using gcc-ar/ranlib in post stage 1 builds, so these
are passed to all sub builds.
v2: Change existing config file as requested by Honza.
conf
Attached please find the patch and testcase for PR62037.
DEF1 can be a GIMPLE_NOP and gimple_bb will be NULL then. The patch
checks for that.
Bootstrapped on x86_64-suse-linux. OK for trunk? Please commit this
patch if it's OK.
Index: gcc/ChangeLog
===
On 08-08-14 17:17, Tom de Vries wrote:
Maybe instead of a new mem_alias_equal_p simply compare MEM_ATTRs
with mem_attrs_eq_p?
I propose to fix it this way (as attached) on 4.8/4.9/trunk, and maybe do a more
efficient handling on trunk as a follow-up patch.
I'll put this through bootstrap/test
72 matches
Mail list logo
