On Thu, Dec 04, 2014 at 08:26:24PM +0300, Yuri Gribov wrote:
> On Thu, Dec 4, 2014 at 8:06 PM, 'Dmitry Vyukov' via address-sanitizer
> wrote:
> > You answered your own question about user space :)
>
> Yeah, I hoped someone would rush to overpersuade me...
While in C unaligned accesses are UB, I
Hi,
this patch fixes implementation of shift in sreal.h that ICEs on 0 value.
Also the comment is copied from shift_right and does not make sense for
sreal.h.
The to_double conversion is useful for debug output: Most of inliner data is
now output as integer that does not make much sense.
Bootstra
> From: Trevor Saunders
>
> Hi,
>
> In this pr we have a section attribute being applied to something in a comdat
> group. Since its ok for things to be in a comdat group and have an implicit
> section it seems it should also be fine to have a section from the attribute.
>
> bootstrapped + reg
So added testcase for this pr (its c++98 only)
So:
ChangeLog testsuite
2014-12-04 Kai Tietz
PR c++/64127
* g++.dg/cpp/pr64127.C: New file.
Tested on x86_64-unknown-linux-gnu.
Ok to apply prior posted patch plus this new testcase?
Regards,
Kai
Index: gcc/gcc/testsuite/g++.dg/c
On Dec 3, 2014, at 1:18 PM, Jeff Law wrote:
> The problem is that waiting for the final spec means effectively GCC doesn't
> support the feature until gcc-6 in 2016 just due to the expected timing of
> the dwarf5 standard and the gcc release schedule. They pretty much couldn't
> line up much w
On Thu, Dec 04, 2014 at 05:34:35PM +0100, Richard Biener wrote:
> On December 4, 2014 4:45:25 PM CET, Marek Polacek wrote:
> >The PR shows a case in which fold introduces undefined behavior in a
> >valid program, because what it does here is
> >-(long int) (ul + ULONG_MAX) - 1 ->
> >~(long int) (u
On December 4, 2014 7:37:32 PM CET, Marek Polacek wrote:
>On Thu, Dec 04, 2014 at 05:34:35PM +0100, Richard Biener wrote:
>> On December 4, 2014 4:45:25 PM CET, Marek Polacek
> wrote:
>> >The PR shows a case in which fold introduces undefined behavior in a
>> >valid program, because what it does h
On Wed, Dec 3, 2014 at 9:19 AM, Andrey Ryabinin wrote:
> On 12/02/2014 08:56 PM, Dmitry Vyukov wrote:
>> Hi,
>>
>> The following patch adds support for instrumentation of globals for
>> Linux kernel (-fsanitize=kernel-address). Kernel only supports
>> constructors with default priority, but the re
On Thu, Dec 04, 2014 at 11:10:15PM +0400, Dmitry Vyukov wrote:
> That would be an option too. I don't know whether it is much better or not.
> Kernel lives without constructors, they are used only by coverage. And
> coverage does not need priorities. So it is only kasan that needs
> priorities. Tha
On December 4, 2014 7:00:12 PM CET, Jan Hubicka wrote:
>Hi,
>this patch fixes implementation of shift in sreal.h that ICEs on 0
>value.
>Also the comment is copied from shift_right and does not make sense for
>sreal.h.
>
>The to_double conversion is useful for debug output: Most of inliner
>data i
will wait for Andrey this time
On Thu, Dec 4, 2014 at 10:12 PM, Jakub Jelinek wrote:
> On Thu, Dec 04, 2014 at 11:10:15PM +0400, Dmitry Vyukov wrote:
>> That would be an option too. I don't know whether it is much better or not.
>> Kernel lives without constructors, they are used only by coverag
On December 4, 2014 6:34:18 PM CET, Jan Hubicka wrote:
>>
>> I think you want get_addr_base_and_unit_offset here. But I really
>wonder
>
>I copied what I found in tree-ssa-alias. The differenc eis that
>get_addr_base_and_unit_offset won't give a range for variable sized
>accesses right?
Yes. Bu
On 04/12/14 11:07, Richard Biener wrote:
On Thu, Dec 4, 2014 at 12:07 PM, Richard Biener
wrote:
On Thu, Dec 4, 2014 at 12:00 PM, Jiong Wang wrote:
which means re-associate the constant imm with the virtual frame pointer.
transform
RA <- fixed_reg + RC
RD <- MEM (RA + const_off
On Thu, Dec 4, 2014 at 5:46 PM, H.J. Lu wrote:
It would probably help reviewers if you pointed to actual path
submission [1], which unfortunately contains the explanation in the
patch itself [2], which further explains that this functionality is
currently only
On 12/04/2014 10:12 AM, Kai Tietz wrote:
else if (cxx_dialect < cxx11
+ && DECL_P (id)
&& !strcmp (IDENTIFIER_POINTER (id), "thread_local"))
This doesn't make any sense: If it's a decl it isn't an identifier. Did
you mean to check for IDENTIFIER_NODE?
Jason
Hi,
On 30 Sep 18:53, Ilya Verbin wrote:
> This patch creates 2 vectors with decls: offload_funcs and offload_vars.
> libgomp will use addresses from these arrays to look up offloaded code.
>
> During the compilation they are outputted to:
> * binary __gnu_offload_funcs/vars sections, or using
>
My patch from yesterday fixed this bug.
PR c++/64080
* constexpr.c (cxx_eval_store_expression): Handle non-decl store
targets.
Jason
On Dec 4, 2014, at 7:22 AM, Ilya Enkovich wrote:
>>> --
>>> 2014-12-04 Ilya Enkovich
>>>
>>> PR target/64056
>>> * lib/target-supports.exp (check_effective_target_stpcpy): New.
>
> Thanks for the notice! I see there is also no description for mempcpy target
> check. So add both of
On Thu, Dec 04, 2014 at 10:35:19PM +0300, Ilya Verbin wrote:
> This issue can be resolved by forcing output of such variables.
> Is this fix ok? Should I add a testcase?
Yes, with proper ChangeLog. Yes.
> diff --git a/gcc/varpool.c b/gcc/varpool.c
> index 0526b7f..db28c2a 100644
> --- a/gcc/var
2014-12-04 20:35 GMT+01:00 Jason Merrill :
> On 12/04/2014 10:12 AM, Kai Tietz wrote:
>>
>> else if (cxx_dialect < cxx11
>> + && DECL_P (id)
>> && !strcmp (IDENTIFIER_POINTER (id), "thread_local"))
>
>
> This doesn't make any sense: If it's a decl it isn't an identifi
> On December 4, 2014 7:00:12 PM CET, Jan Hubicka wrote:
> >Hi,
> >this patch fixes implementation of shift in sreal.h that ICEs on 0
> >value.
> >Also the comment is copied from shift_right and does not make sense for
> >sreal.h.
> >
> >The to_double conversion is useful for debug output: Most of
On Thu, Dec 04, 2014 at 02:37:54PM -0500, Jason Merrill wrote:
> My patch from yesterday fixed this bug.
>
> PR c++/64080
> * constexpr.c (cxx_eval_store_expression): Handle non-decl store
> targets.
I don't see such a patch neither on ML nor in git.
Marek
> On December 4, 2014 6:34:18 PM CET, Jan Hubicka wrote:
> >>
> >> I think you want get_addr_base_and_unit_offset here. But I really
> >wonder
> >
> >I copied what I found in tree-ssa-alias. The differenc eis that
> >get_addr_base_and_unit_offset won't give a range for variable sized
> >accesses
Updated version checking the right thing ...
ChangeLog gcc/cp
2014-12-04 Kai Tietz
PR c++/64127
* parser.c (cp_parser_diagnose_invalid_type_name): Check
id for being an identifier before accessing it.
Tested on x86_64-unknown-linux-gnu.
Ok for apply together with testcase?
Rega
On 12/04/2014 03:05 PM, Marek Polacek wrote:
On Thu, Dec 04, 2014 at 02:37:54PM -0500, Jason Merrill wrote:
My patch from yesterday fixed this bug.
PR c++/64080
* constexpr.c (cxx_eval_store_expression): Handle non-decl store
targets.
I don't see such a patch neithe
The problem was that the extended reference temporary never got its
array type completed. Doing it in the extension code didn't work
because by that time the initializer already has incomplete array type.
So let's handle it in grok_reference_init.
Tested x86_64-pc-linux-gnu, applying to trun
cxx_eval_store_expression was too strict; a store to something where we
can't figure out which variable it refers to shouldn't ICE, just make
the expression non-constant.
Tested x86_64-pc-linux-gnu, applying to trunk.
commit 796c5ad03637ba9452639262cef57db83cd60cba
Author: Jason Merrill
Date:
OK, thanks.
Jason
On 12/04/2014 10:57 AM, Kai Tietz wrote:
Well, I spared that as this is an invalid code bug with a lot of
side-errors, which might change over time.
Every patch should have an associated testcase unless it's fixing an
existing test or writing a testcase is somehow not feasible. Many tests
in
On Thu, Dec 04, 2014 at 04:43:34PM +0800, Zhenqiang Chen wrote:
> C code:
>
> if (!--*p)
>
> rtl code:
>
> 6: r91:SI=[r90:SI]
> 7: {r88:SI=r91:SI-0x1;clobber flags:CC;}
> 8: [r90:SI]=r88:SI
> 9: flags:CCZ=cmp(r88:SI,0)
>
> expected output:
>
> 8: {flags:CCZ=cmp([r90:SI]
On Thu, Dec 04, 2014 at 02:49:56PM -0600, Segher Boessenkool wrote:
> On Thu, Dec 04, 2014 at 04:43:34PM +0800, Zhenqiang Chen wrote:
> > C code:
> >
> > if (!--*p)
> >
> > rtl code:
> >
> > 6: r91:SI=[r90:SI]
> > 7: {r88:SI=r91:SI-0x1;clobber flags:CC;}
> > 8: [r90:SI]=r88:SI
>
The auto_ptr<> idea sounds invasive (and made me wonder - why not go the
whole way and use std::string?)
Here's an implementation of the xstrdup_for_dump idea.
To recap the motivation:
cgraph*.c and ipa-*.c use xstrdup on strings when dumping them via
fprintf, leaking all of the duplicated buffe
Ok, Additional change for testcase mentioned in PR.
ChangeLog
2014-12-04 Kai Tietz
PR c++/641000
* g++.dg/template/pr64100.C: New file.
Ok for apply?
Regards,
Kai
Index: gcc/gcc/testsuite/g++.dg/template/pr64100.C
This is the JIT-specific part of the patch for PR jit/64166.
Implement a way to get at dumpfiles from JIT testcases.
Use it from test-sum-of-squares.c to provide a simple selftest of the
dumping functionality, and use it from test-functions.c to add
verification of the fix for PR jit/64020.
gcc/
This is the non-JIT part of the patch for PR jit/64166.
Provide a way for the JIT to lookup a dump_file_info * by switch name,
and to get from there to the filename.
OK for trunk?
gcc/ChangeLog:
PR jit/64166
* dumpfile.c (gcc::dump_manager::get_dump_file_info_by_switch):
assumed -fgnu89-inline until a recent upstream fix;
see http://lists.gnu.org/archive/html/dejagnu/2014-10/msg00011.html
Remove the workaround from jit.exp that used -fgnu89-inline
in favor of a fixincludes to dejagnu.h that applies the upstream fix
to a local copy.
This should make it easier to
On 11/24/2014 03:08 AM, Olivier Hainque wrote:
+ if (init_state->processed_regno[regno])
+return;
I would expect this to go in the loop in
expand_builtin_init_dwarf_reg_sizes, before we look up a span for the
regno. OK with that change.
Jason
OK, thanks.
Jason
> Normally, with -fPIE/-fpie, GCC accesses globals that are extern to the
> module using the GOT. This is two instructions, one to get the address
> of the global from the GOT and the other to get the value. If it turns
> out that the global gets defined in the executable at link-time, it still
>
David Malcolm writes:
> assumed -fgnu89-inline until a recent upstream fix;
> see http://lists.gnu.org/archive/html/dejagnu/2014-10/msg00011.html
>
> Remove the workaround from jit.exp that used -fgnu89-inline
> in favor of a fixincludes to dejagnu.h that applies the upstream fix
> to a local co
On 3 December 2014 at 08:59, Jakub Jelinek wrote:
>
> I think using a default argument for this is fine, though of course
> you need to declare the default argument in the header containing
> the prototype, not in the function definition.
>
> Ok with that change.
This is what I have committed at
> The auto_ptr<> idea sounds invasive (and made me wonder - why not go the
> whole way and use std::string?)
That would work for me too :)
>
> Here's an implementation of the xstrdup_for_dump idea.
>
> To recap the motivation:
>
> cgraph*.c and ipa-*.c use xstrdup on strings when dumping them v
On 4 December 2014 at 11:21, Tobias Burnus
wrote:
> Manuel López-Ibáñez wrote:
>> It is still not clear to me if line_len is the length of the line read
>> or not, is it? If not, is there any way to actually get the length of
>> the line?
>
> Looking at the code in load_line, the line_len in
>
>
On Thu, Dec 4, 2014 at 2:19 PM, Dominique Dhumieres wrote:
>> Normally, with -fPIE/-fpie, GCC accesses globals that are extern to the
>> module using the GOT. This is two instructions, one to get the address
>> of the global from the GOT and the other to get the value. If it turns
>> out that th
On 3 December 2014 at 09:48, Dodji Seketeli wrote:
> Manuel López-Ibáñez writes:
>
>> libcpp uses diagnostic->override_column to give a custom column number
>> to diagnostics. This is taken into account when building the prefix,
>> but it was missing when placing the caret.
>>
>> Before:
>>
>> /h
Ping?
Thanks!
-Zhenqiang
-Original Message-
From: gcc-patches-ow...@gcc.gnu.org [mailto:gcc-patches-ow...@gcc.gnu.org] On
Behalf Of Zhenqiang Chen
Sent: Monday, December 01, 2014 3:14 PM
To: 'H.J. Lu'
Cc: Richard Henderson; GCC Patches
Subject: RE: [PATCH, ifcvt] Fix PR63917
> -Ori
If you generate code with the JIT which references outside symbols there
is currently no way to have a self-contained DSO created. The command
line to invoke the linker is fixed.
The patch below would change that. It builds upon the existing
framework to specify options for the compiler. The li
101 - 147 of 147 matches
Mail list logo
