Hi,
this is a variant of patch I commited.
It takes care to load the constructor to memory in sem_variable::equals
and donot touch it earlier. I also made sem_variable::parse to skip volatile
and reigster variables.
Bootstrapped/regtested x86_64-linux, lto-bootstrapped with -fmerge-all-constants
On Sat, Feb 28, 2015 at 5:42 PM, H.J. Lu wrote:
> Ue copy relocation in PIE improves performance. But copy relocation
> can't be used to access protected symbols defined in shared libaries
> and linker in binutils 2.26 enforces doesn't allow it. GCC doesn't
> know if an external definition is pr
On Sun, Mar 01, 2015 at 07:33:14AM -0800, H.J. Lu wrote:
> +mcopyreloc-in-pie
I'm not calling for a change in the name of the option (*), but
technically it isn't completely correct to call your optimisation
"copy reloc in pie". What you are really doing is using a linker
generated variable (in .
As I mention in the PR...
What's happening here is that the ipa_polymorphic_call_context
constructor is calling walk_ssa_copies on a PHI node that has no
arguments. This happens because finalize_jump_threads eventually
removes some PHI arguments as it's redirecting some edges, leaving a PHI
On Mon, Mar 2, 2015 at 1:07 AM, Joseph Myers wrote:
>
> On Sat, 28 Feb 2015, Max Filippov wrote:
>
> > This allows a target to support both windowed and non-windowed ABI.
> >
> > 2015-02-28 Max Filippov
> >
> > gcc/
> > * builtins.c (expand_builtin_return_addr): Add
> > RETURN_ADDR_
Jan Hubicka writes:
>
> The hash itself is quite simple (borrowing incremental hash for constants
> adding
> very simple match for other stuff + logic to skip things that may match even
> if
> they are syntactticaly different). The hash can be strenghtened significantly,
> but I suppose we may d
Hi there,
If target mode isn't specified via either gcc configuration option
--with-mode or command line, this patch intends to improve gcc driver to
automatically add option -mthumb for thumb-only target. Tested with gcc
regression test for various arm targets, no regression. Is it OK?
BR,
Terry
Hi,
my comments to your patch was not complete. ipa-icf is quite confused about
how to get the constructor. This is how it works:
In during in_lto_p DECL_INITIAL may be set to error_mark_node and once you
really want to have the constructor there is varpool_node->get_constuctor
method. Because in
> > >and thus we will hapilly merge types with different TYPE_ALIGN.
> > >Should func_checker::types_compatible be extended to compare these?
> > >Clearly TYPE_ALIGN matters for vectorizer and other plaes...
> >
> > But it matters on MEM_REFs only, and you can't use canonical types for that.
> >
> >and thus we will hapilly merge types with different TYPE_ALIGN.
> >Should func_checker::types_compatible be extended to compare these?
> >Clearly TYPE_ALIGN matters for vectorizer and other plaes...
>
> But it matters on MEM_REFs only, and you can't use canonical types for that.
> And we don't
This patch changes the coarray/coindexed check to match what the error
message already claimed.
OK for the trunk?
Tobias
diff --git a/gcc/fortran/check.c b/gcc/fortran/check.c
index 3be4fb1..cdb5ff1 100644
--- a/gcc/fortran/check.c
+++ b/gcc/fortran/check.c
@@ -1022,7 +1022,7 @@ gfc_check_atomi
The following patch is supposed to fix the problem related to
initializing an array of aggregates with members, which have
user-defined default constructors.
Bootstrapped and regtested on x86_64-unknown-linux-gnu.
diff --git a/gcc/cp/init.c b/gcc/cp/init.c
index 0274663..0f53146 100644
--- a/gcc/
On Sat, 28 Feb 2015, Max Filippov wrote:
> This allows a target to support both windowed and non-windowed ABI.
>
> 2015-02-28 Max Filippov
>
> gcc/
> * builtins.c (expand_builtin_return_addr): Add
> RETURN_ADDR_IN_PREVIOUS_FRAME to 'if' condition.
If you change the semantics of a
Bootstrap and regression tested on x86_64.
A ubsan-bootstrap for C and C++ works too (although there
are other unrelated runtime errors).
I tried a bootstrap with -fsanitize=bounds-strict and
there are indeed many additional runtime errors due the
struct hack with last array element of size 1
Hi,
getting more familiar with ipa-icf I start to understand why it is a compile
time hog for firefox. I assumed it spends most of time streaming in bodies and
comparing them to see they are equivalent (after all about 10% of all symbols
turns out to be and each of that needs to be read to memory)
Dear Mikael,
That looks to be OK for 4.8, 4.9 and 5.0.
Strange testcase, though... :-)
Thanks for the patch
Paul
On 15 February 2015 at 18:48, Mikael Morin wrote:
> Hello,
>
> I propose a fix for PR60898, where a symbol is freed despite remaining
> reachable in the symbol tree.
> The problem
Hi,
here is a variant of patch I intend to commit after x86_64 bootstrap®test
and LTO-bootstrap. What i changed are the ipa-icf-gimple.c bits (dorpping all
comparsions relying that external references are already matched and used at
same spots). Second change is sem_function::compare_cgraph_refere
Hello, gentle maintainer.
This is a message from the Translation Project robot.
A revised PO file for textual domain 'gcc' has been submitted
by the Swedish team of translators. The file is available at:
http://translationproject.org/latest/gcc/sv.po
(This file, 'gcc-5.1-b20150208.sv.po',
Hi all,
the appended patch removes the soft-float multilib build option from
powerpc64-*-freebsd*.
These libraries are rarely to never used on 64-bit FreeBSD PowerPC
machines. But the build time of these libraries counts.
Is this ok for trunk? If yes, at what stage can I commit, 5.0/5.1?
T
Hi,
This adds a test case for PR 61142.
Tested with
make -k check-gcc RUNTESTFLAGS="sh.exp=pr61142.c --target_board=sh-sim
\{-m2/-ml,-m2/-mb,-m2a/-mb,-m4/-ml,-m4/-mb,-m4a/-ml,-m4a/-mb}"
Committed as r221089.
Cheers,
Oleg
gcc/testsuite/ChangeLog:
PR target/61142
* gcc.target/sh/s
They were introduced by the latest tree structure re-shuffling.
Tested on x86_64-suse-linux, applied on the mainline as obvious.
2015-03-01 Eric Botcazou
* c-ada-spec.c (is_tagged_type): Add guard for DECL_VINDEX.
(dump_template_types): Adjust DECL_TEMPLATE_INSTANTIATIONS and
> >From 7cedbe5e3736b9eb7b2fab1b931a7bc9ed269f37 Mon Sep 17 00:00:00 2001
> From: mliska
> Date: Fri, 27 Feb 2015 22:42:49 +0100
> Subject: [PATCH 2/4] ICF: Do not consider variable aliases for merge
> operation.
>
> gcc/ChangeLog:
>
> 2015-02-28 Martin Liska
> Jan Hubicka
>
>
On 27 February 2015 at 21:49, Jan Hubicka wrote:
>>
>> ../sysdeps/gnu/siglist.c:72:1: internal compiler error: in
>> address_matters_p, at symtab.c:1908
>> versioned_symbol (libc, __new_sys_sigabbrev, sys_sigabbrev, GLIBC_2_3_3);
>> ^
>> 0x66a080 symtab_node::address_matters_p()
>> /scr
On Sat, Feb 28, 2015 at 5:27 PM, Sandra Loosemore
wrote:
> On 02/28/2015 09:42 AM, H.J. Lu wrote:
>>
>> @@ -22704,6 +22705,13 @@ For systems that use the GNU C Library, the
>> default is on.
>> Specify that the assembler should encode SSE instructions with VEX
>> prefix. The option @option{-m
On 02/28/2015 11:41 PM, Jan Hubicka wrote:
>> Hello.
>>
>> Following patch skips all variable aliases as potential merge candidates.
>> Tested on x86_64-linux.
>>
>> Ready for trunk?
>> Thanks,
>> Martin
>
>> >From fbde2e98f98a71105d18cf3e91e8032d0c657139 Mon Sep 17 00:00:00 2001
>> From: mliska
On Sun, 1 Mar 2015, Tobias Burnus wrote:
>> Ping re: https://gcc.gnu.org/ml/gcc-patches/2015-02/msg00371.html
> While it would be nice if Gerald could find the time to review the patch,
> I recall that he once said that maintainers don't need approvals for
> release-notes changes of their part of
David Malcolm wrote:
On Thu, 2015-02-05 at 21:47 -0500, David Malcolm wrote:
This patch adds an -level section about libgccjit between the
sections on languages and targets.
It has various links to the documentation (now that that's on the
website), and to a couple of 3rd party repos (the Pytho
On March 1, 2015 1:09:50 AM CET, Jan Hubicka wrote:
>> >
>> > Half of FAILs are gone, but the rest is correctly merged (alignment
>matches). Thus I would omit ICF
>> > in this testcase. I'm going to install the patch.
>
>Richard, I wonder what happens with TYPE_ALIGN at LTO. It is not part
>of c
28 matches
Mail list logo
