--- Comment #23 from jakub at gcc dot gnu dot org 2008-10-27 22:06 ---
Fixed.
--
jakub at gcc dot gnu dot org changed:
What|Removed |Added
Status|UNCONFIRMED
--- Comment #22 from jakub at gcc dot gnu dot org 2008-10-27 21:38 ---
Subject: Bug 37378
Author: jakub
Date: Mon Oct 27 21:37:06 2008
New Revision: 141392
URL: http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=141392
Log:
PR target/37378
* df-scan.c (df_bb_refs_collect)
--- Comment #21 from jakub at gcc dot gnu dot org 2008-10-24 20:04 ---
Indeed, changing that line didn't cure anything (but that still doesn't mean
that this hunk shouldn't be either fixed if needed, or killed).
Adding all EH_USES regs as explicit uses for all insns that may throw would
--- Comment #20 from zadeck at naturalbridge dot com 2008-10-24 18:44
---
Subject: Re: [4.4 Regression] Revision 139827 causes Divide_X
jakub at gcc dot gnu dot org wrote:
> --- Comment #19 from jakub at gcc dot gnu dot org 2008-10-24 18:09
> ---
> This hunk in df-scan.c con
--- Comment #19 from jakub at gcc dot gnu dot org 2008-10-24 18:09 ---
This hunk in df-scan.c confuses me:
/* These registers are live everywhere. */
if (!reload_completed)
{
#ifdef EH_USES
/* The ia-64, the only machine that uses this, does not define these
unt
--- Comment #18 from jakub at gcc dot gnu dot org 2008-10-24 16:11 ---
FYI, EH_USES (32) is 1, so dataflows knows (or at least should know) this
register is needed.
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=37378
--- Comment #17 from jakub at gcc dot gnu dot org 2008-10-24 15:32 ---
Created an attachment (id=16538)
--> (http://gcc.gnu.org/bugzilla/attachment.cgi?id=16538&action=view)
prims.ii.188r.dce
*.dce dump on Kenny's request.
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=37378
--- Comment #16 from jakub at gcc dot gnu dot org 2008-10-24 15:22 ---
I guess e.g. ia64_expand_prologue could add all those regs that it
/* Even if we're not going to generate an epilogue, we still
need to save the register so that EH works. */
if (! epilogue_p)
emit_insn (gen_pro
--- Comment #15 from jakub at gcc dot gnu dot org 2008-10-24 14:38 ---
BTW, the unwind info says that rp is saved in r34 even for the .L7 and onwards
hunk.
Anyway, in the 196r.compgotos dump mov r34 = b0 insn is still before the
conditional jump to .L7, it is the scheduler that moves it
--- Comment #14 from jakub at gcc dot gnu dot org 2008-10-24 13:53 ---
I think it is _Jv_divI which now has wrong unwind info (well, not only that).
Here is a self-contained source for _Jv_divI:
#pragma GCC java_exceptions
extern "Java"
{
typedef __java_int jint;
typedef __java_long
--
mmitchel at gcc dot gnu dot org changed:
What|Removed |Added
Priority|P3 |P1
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=37378
--- Comment #13 from jh at suse dot cz 2008-09-06 23:18 ---
Subject: Re: [4.4 Regression] Revision 139827 causes Divide_X
> This looks like a target issue anyways.
The patch had effect of turning code paths leading to trap or noreturn
etc. to be optimized for size...
Honza
--
ht
--- Comment #12 from pinskia at gcc dot gnu dot org 2008-09-06 22:30
---
This looks like a target issue anyways.
--
pinskia at gcc dot gnu dot org changed:
What|Removed |Added
---
13 matches
Mail list logo
