Hi,
On Fri, May 16, 2014 at 07:25:59PM +0200, Jan Hubicka wrote:
> Hi,
> this patch adds code to remove write only static variables. While analyzing
> effectivity of LTO on firefox, I noticed that surprisingly large part of
> binary's data segment is occupied by these. Fixed thus.
> (this is qui
On 05/19/14 15:20, John David Anglin wrote:
The problem compiling c-common.c is very hard to debug. These routines
are called intensively and the ICE occurs after several million calls.
It takes a couple of hours of running under gdb to reach the failing
call to reset_insn_used_flags just counti
My CROSSING_JUMP_P patch didn't handle insns that were copied by
emit_copy_of_insn_after, which previously would copy REG_CROSSING_JUMP
notes by virtue of them being != REG_LABEL_OPERAND.
This patch adds the missing copy. Tested with profiledbootstrap on
x86_64-linux-gnu. Committed as obvious.
Jeff Law writes:
> On 05/20/14 02:16, Richard Sandiford wrote:
>> get_attr_enabled was showing up high in a -O0 compile of fold-const.ii.
>> At the moment, extract_insn calls this function for every alternative
>> on each extraction, which can be expensive for instructions like
>> moves that have
On Tue, 20 May 2014, Jan Hubicka wrote:
> Hi,
> as disucssed some time ago, our assumption that every symbol of shared
> library can
> be interposed at runtime is expensive and prevents a lot of useful
> optimizations,
> including inlining or IPA propagation.
>
> While this is useful feature, i
On 05/19/2014 11:27 PM, Maxim Kuvyrkov wrote:
Changes to ia64 and rs6000 are mostly mechanical to update hook return values,
but port maintainers may wish to review those.
I'm not a maintainer, but the rs6000 changes look good to me.
-Pat
On Tue, 20 May 2014, Marek Polacek wrote:
> * is missing tests for long doubles/-mlong-double-128,
Also missing tests for float - as far as I can see, only double is tested.
Ideally all of float, double, long double, __float128 (where supported),
__float80 (where supported) would be tested (th
> Yes, but why is that better than providing them directly? The latter
> would seem to work better with non-C-family languages like Ada.
That's correct, enumeration types don't have base types in Ada, i.e. they
are their own base types. But if the attributes cannot be expressed in DWARF,
then
Hi Richard,
Apologies for appearing to be pushy on this. I appreciate all of this work
is done in your spare time, I doubt I could manage that. I missed the
important part which was to ask if you had thoughts on the user visible
parts of the patch but you have been through the code now anyway...
Svante Signell, le Fri 16 May 2014 10:03:05 +0200, a écrit :
> is used in gcc-4.9-4.9.0/src/libgo/go/net/fd_unix.go:
> func dupCloseOnExec(fd int) (newfd int, err error) {
> if atomic.LoadInt32(&tryDupCloexec) == 1 && syscall.F_DUPFD_CLOEXEC!=0 {
> r0, _, e1 := syscall.Syscall(syscall.SYS_FCNTL, ui
On Tue, May 20, 2014 at 5:46 PM, Pat Haugen wrote:
> On 05/19/2014 11:27 PM, Maxim Kuvyrkov wrote:
>>
>> Changes to ia64 and rs6000 are mostly mechanical to update hook return
>> values, but port maintainers may wish to review those.
>
> I'm not a maintainer, but the rs6000 changes look good to me
> From: gcc-patches-ow...@gcc.gnu.org [mailto:gcc-patches-
>
> I'll send the new patch as soon as all the tests are done.
Here you are. I also simplified the tests a bit having the reference as a
function
parameter instead of a local one.
Updated ChangeLogs:
*** gcc/ChangeLog ***
2014-05-20
One of the consequences of the (now-fixed) bug in PR60179 is that Nios
II code using target pragmas to specify custom instructions failed to
generate those instructions with -flto. We came up with this test case
for it, but there didn't seem to be an existing hook to scan LTO output.
Is this
Bug # 61144 is a regression in 4.9.0 that breaks building of musl libc
due to aggressive and semantically-incorrect constant folding of weak
aliases. The attached patch seems to fix the issue. A weak alias
should never be a candidate for constant folding because it may always
be replaced by a stron
On May 14, 2014, at 4:13 AM, Yury Gribov wrote:
> Hi,
>
> Asan and Tsan allow sanitized applications to tweak runtime behavior via API
> defined in headers in libsanitizer/include/sanitizer. This patch adds
> installation code for these headers and a small test.
>
> Bootstrapped and regtested
On 21 May 2014 00:54, Jeff Law wrote:
> On 05/20/14 01:11, Zhenqiang Chen wrote:
>>
>> Hi,
>>
>> The patch fix ICE issue triggered by shrink-wrapping enhancement.
>>
>> Bootstrap and no make check regression on X86-64.
>>
>> OK for trunk?
>>
>> Thanks!
>> -Zhenqiang
>>
>>
>> 2014-05-20 Zhenqiang
On Tue, 20 May 2014, Richard Earnshaw wrote:
> >>> If it's not broken in 4.8.2 but broken on the branch head then it's OK
> >>> for the branch.
> >>
> >> I thought I'd double-check with you that it is fine to push this change
> >> to trunk first. OK to apply?
> >
> > Of course it should go to
>
> Unfortunately, this commit has caused the following ICE for me when
> LTO building 471.omnetpp from SPEC 2006 or Firefox (but not libxul,
> something that gets built earlier):
>
> lto1: internal compiler error: in gimple_get_virt_method_for_vtable,
> at gimple-fold.c:3276
> 0x7437a3 gimple_ge
> On Tue, 20 May 2014, Jan Hubicka wrote:
>
> > Hi,
> > as disucssed some time ago, our assumption that every symbol of shared
> > library can
> > be interposed at runtime is expensive and prevents a lot of useful
> > optimizations,
> > including inlining or IPA propagation.
> >
> > While this
> The new test fails on x86_64 Ubuntu 12.04 native build:
Yes, this should be fixed once kcc's patches get into trunk
(https://gcc.gnu.org/ml/gcc-patches/2014-05/msg00756.html). Current
asan_interface.h is indeed C++-only.
-Y
Hi,
on AIX GCC 4.9 as well as mainline dies on callgraph corruption problem. THis
is caused
by ipa-inline walking node that is removed by ipa-inline-transform (because it
is a local
alias that become unreachable).
This patch solves it by adding a hook to inform user of inline_call when this
hap
Hi,
devirt-11.C is somewhat fragile testcase that originally was testing a specific
series of events
in inliner that are no longer happening anyway. It tests for a specific number
of devirutalization that
depends on several aspects. It seems more practical to relax this check,
because it is kin
Hi,
can_refer_decl_in_current_unit_p is a black magic that carefuly papers around
a mine field created by combination of C++ ABI and ELF visibility attributes.
As such it is not pretty and grown up into quite a maze of rules.
Well, I think it is time to clean it up: for example it does twice very
101 - 123 of 123 matches
Mail list logo
