--- Comment #16 from slyfox at inbox dot ru 2009-08-29 10:26 ---
(In reply to comment #15)
> Fixed everywhere.
>
Just tested on my sample (got this only revision from 4_4 branch and applied to
gentoo gcc sources).
All works as expected.
Thanks!
--
http://gcc.gnu.org/bugzilla/show
--- Comment #15 from ubizjak at gmail dot com 2009-08-23 13:04 ---
Fixed everywhere.
--
ubizjak at gmail dot com changed:
What|Removed |Added
Status|ASSIGNED
--- Comment #14 from uros at gcc dot gnu dot org 2009-08-23 13:03 ---
Subject: Bug 40718
Author: uros
Date: Sun Aug 23 13:03:39 2009
New Revision: 151033
URL: http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=151033
Log:
PR target/40718
* config/i386/i386.c (*call_pop_1)
--- Comment #13 from uros at gcc dot gnu dot org 2009-08-23 12:38 ---
Subject: Bug 40718
Author: uros
Date: Sun Aug 23 12:37:53 2009
New Revision: 151030
URL: http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=151030
Log:
PR target/40718
* config/i386/i386.c (*call_pop_1)
--- Comment #12 from uros at gcc dot gnu dot org 2009-08-23 12:14 ---
Subject: Bug 40718
Author: uros
Date: Sun Aug 23 12:14:26 2009
New Revision: 151029
URL: http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=151029
Log:
PR target/40718
* config/i386/i386.c (*call_pop_1)
--- Comment #11 from ubizjak at gmail dot com 2009-08-23 12:08 ---
Patched gcc compiles testcase from comment 8 to:
_ZN5Stuff8do_stuffEmm:
.LFB4:
.cfi_startproc
.cfi_personality 0x0,__gxx_personality_v0
pushl %ebp
.cfi_def_cfa_offset 8
movl%
--- Comment #10 from ubizjak at gmail dot com 2009-08-23 11:57 ---
Created an attachment (id=18416)
--> (http://gcc.gnu.org/bugzilla/attachment.cgi?id=18416&action=view)
Additional patch
Additional patch to fix call_value_pop_1.
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=40718
--- Comment #9 from ubizjak at gmail dot com 2009-08-23 11:41 ---
(In reply to comment #8)
> This patch fixes for me Dmitry's sample, but does not fix mine. Still
> SIGSEGVs.
> I've managed to
> place whole testcase in one file:
Ah, the same cure should be applied to "*call_value_pop_
--- Comment #8 from slyfox at inbox dot ru 2009-08-23 11:12 ---
(In reply to comment #7)
> Subject: Bug 40718
>
> Author: uros
> Date: Sun Aug 23 09:46:00 2009
> New Revision: 151028
>
> URL: http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=151028
> Log:
> PR target/40718
>
--- Comment #7 from uros at gcc dot gnu dot org 2009-08-23 09:46 ---
Subject: Bug 40718
Author: uros
Date: Sun Aug 23 09:46:00 2009
New Revision: 151028
URL: http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=151028
Log:
PR target/40718
* config/i386/i386.c (*call_pop_1):
--- Comment #6 from ubizjak at gmail dot com 2009-08-23 09:08 ---
Created an attachment (id=18415)
--> (http://gcc.gnu.org/bugzilla/attachment.cgi?id=18415&action=view)
Updated patch
This patch is bootstrapped and regression tested.
--
ubizjak at gmail dot com changed:
--- Comment #5 from slyfox at inbox dot ru 2009-08-22 18:38 ---
(In reply to comment #3)
> Created an attachment (id=18413)
--> (http://gcc.gnu.org/bugzilla/attachment.cgi?id=18413&action=view) [edit]
> patch to fix the failure
>
> Sibcalls of any kind should be done through call-clobb
--- Comment #4 from ubizjak at gmail dot com 2009-08-22 13:51 ---
Patched gcc:
bar:
pushl %ebp
movl%esp, %ebp
pushl %ebx
subl$20, %esp
movl8(%ebp), %ebx
movl%ebx, (%esp)
callfoo
subl$4, %esp
--- Comment #3 from ubizjak at gmail dot com 2009-08-22 13:43 ---
Created an attachment (id=18413)
--> (http://gcc.gnu.org/bugzilla/attachment.cgi?id=18413&action=view)
patch to fix the failure
Sibcalls of any kind should be done through call-clobbered regs only.
This patch fixes the
--- Comment #2 from ubizjak at gmail dot com 2009-08-22 13:38 ---
Have a patch.
--
ubizjak at gmail dot com changed:
What|Removed |Added
AssignedTo|unassigned at
15 matches
Mail list logo
