* [Bug libgcj/40860] [4.4/4.5 regression] regressions in libjava testsuite on arm-linux
2009-07-26 10:51 [Bug libgcj/40860] New: [4.4/4.5 regression] regressions in libjava testsuite on arm-linux debian-gcc at lists dot debian dot org
@ 2009-07-29 23:07 ` jsm28 at gcc dot gnu dot org
2009-10-15 12:54 ` jakub at gcc dot gnu dot org
` (43 subsequent siblings)
44 siblings, 0 replies; 46+ messages in thread
From: jsm28 at gcc dot gnu dot org @ 2009-07-29 23:07 UTC (permalink / raw)
To: java-prs
--
jsm28 at gcc dot gnu dot org changed:
What |Removed |Added
----------------------------------------------------------------------------
Priority|P3 |P4
Target Milestone|--- |4.4.2
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=40860
^ permalink raw reply [flat|nested] 46+ messages in thread* [Bug libgcj/40860] [4.4/4.5 regression] regressions in libjava testsuite on arm-linux
2009-07-26 10:51 [Bug libgcj/40860] New: [4.4/4.5 regression] regressions in libjava testsuite on arm-linux debian-gcc at lists dot debian dot org
2009-07-29 23:07 ` [Bug libgcj/40860] " jsm28 at gcc dot gnu dot org
@ 2009-10-15 12:54 ` jakub at gcc dot gnu dot org
2010-01-21 13:16 ` jakub at gcc dot gnu dot org
` (42 subsequent siblings)
44 siblings, 0 replies; 46+ messages in thread
From: jakub at gcc dot gnu dot org @ 2009-10-15 12:54 UTC (permalink / raw)
To: java-prs
--
jakub at gcc dot gnu dot org changed:
What |Removed |Added
----------------------------------------------------------------------------
Target Milestone|4.4.2 |4.4.3
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=40860
^ permalink raw reply [flat|nested] 46+ messages in thread* [Bug libgcj/40860] [4.4/4.5 regression] regressions in libjava testsuite on arm-linux
2009-07-26 10:51 [Bug libgcj/40860] New: [4.4/4.5 regression] regressions in libjava testsuite on arm-linux debian-gcc at lists dot debian dot org
2009-07-29 23:07 ` [Bug libgcj/40860] " jsm28 at gcc dot gnu dot org
2009-10-15 12:54 ` jakub at gcc dot gnu dot org
@ 2010-01-21 13:16 ` jakub at gcc dot gnu dot org
2010-01-24 16:04 ` mikpe at it dot uu dot se
` (41 subsequent siblings)
44 siblings, 0 replies; 46+ messages in thread
From: jakub at gcc dot gnu dot org @ 2010-01-21 13:16 UTC (permalink / raw)
To: java-prs
--
jakub at gcc dot gnu dot org changed:
What |Removed |Added
----------------------------------------------------------------------------
Target Milestone|4.4.3 |4.4.4
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=40860
^ permalink raw reply [flat|nested] 46+ messages in thread* [Bug libgcj/40860] [4.4/4.5 regression] regressions in libjava testsuite on arm-linux
2009-07-26 10:51 [Bug libgcj/40860] New: [4.4/4.5 regression] regressions in libjava testsuite on arm-linux debian-gcc at lists dot debian dot org
` (2 preceding siblings ...)
2010-01-21 13:16 ` jakub at gcc dot gnu dot org
@ 2010-01-24 16:04 ` mikpe at it dot uu dot se
2010-01-25 9:33 ` mikpe at it dot uu dot se
` (40 subsequent siblings)
44 siblings, 0 replies; 46+ messages in thread
From: mikpe at it dot uu dot se @ 2010-01-24 16:04 UTC (permalink / raw)
To: java-prs
------- Comment #1 from mikpe at it dot uu dot se 2010-01-24 16:04 -------
I'm now seeing failures in StackTrace2 and Throw_3 on arm-linux-gnueabi with
gcc-4.3 branch which I didn't use to see. gcc-4.4 branch doesn't fail for me on
these two, but both 4.4 and 4.3 fail (as always) on Throw_2.
--
mikpe at it dot uu dot se changed:
What |Removed |Added
----------------------------------------------------------------------------
CC| |mikpe at it dot uu dot se
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=40860
^ permalink raw reply [flat|nested] 46+ messages in thread* [Bug libgcj/40860] [4.4/4.5 regression] regressions in libjava testsuite on arm-linux
2009-07-26 10:51 [Bug libgcj/40860] New: [4.4/4.5 regression] regressions in libjava testsuite on arm-linux debian-gcc at lists dot debian dot org
` (3 preceding siblings ...)
2010-01-24 16:04 ` mikpe at it dot uu dot se
@ 2010-01-25 9:33 ` mikpe at it dot uu dot se
2010-02-03 14:51 ` mikpe at it dot uu dot se
` (39 subsequent siblings)
44 siblings, 0 replies; 46+ messages in thread
From: mikpe at it dot uu dot se @ 2010-01-25 9:33 UTC (permalink / raw)
To: java-prs
------- Comment #2 from mikpe at it dot uu dot se 2010-01-25 09:33 -------
I had very recently updated binutils from 2.19.1 to 2.20 + branch updates +
some fixes from trunk. Going back to binutils 2.19.1 while changing nothing
else eliminated the StackTrace2 and Throw_3 failures from gcc-4.3.
My last gcc-4.4 build and test suite run was probably still with
binutils-2.19.1, which might explain why I didn't see those errors there.
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=40860
^ permalink raw reply [flat|nested] 46+ messages in thread* [Bug libgcj/40860] [4.4/4.5 regression] regressions in libjava testsuite on arm-linux
2009-07-26 10:51 [Bug libgcj/40860] New: [4.4/4.5 regression] regressions in libjava testsuite on arm-linux debian-gcc at lists dot debian dot org
` (4 preceding siblings ...)
2010-01-25 9:33 ` mikpe at it dot uu dot se
@ 2010-02-03 14:51 ` mikpe at it dot uu dot se
2010-02-05 15:40 ` mikpe at it dot uu dot se
` (38 subsequent siblings)
44 siblings, 0 replies; 46+ messages in thread
From: mikpe at it dot uu dot se @ 2010-02-03 14:51 UTC (permalink / raw)
To: java-prs
------- Comment #3 from mikpe at it dot uu dot se 2010-02-03 14:51 -------
gcc-4.4.3 with binutils-2.20 fails StackTrace2 and Throw_3, reverting to
binutils-2.19.1 eliminated those two failures.
Bootstrapping with binutils-2.19.1 but running the test suite with
binutils-2.20 did not trigger the failures. Rebuilding just libjava with
binutils-2.20 also did not trigger the failures.
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=40860
^ permalink raw reply [flat|nested] 46+ messages in thread* [Bug libgcj/40860] [4.4/4.5 regression] regressions in libjava testsuite on arm-linux
2009-07-26 10:51 [Bug libgcj/40860] New: [4.4/4.5 regression] regressions in libjava testsuite on arm-linux debian-gcc at lists dot debian dot org
` (5 preceding siblings ...)
2010-02-03 14:51 ` mikpe at it dot uu dot se
@ 2010-02-05 15:40 ` mikpe at it dot uu dot se
2010-02-06 15:36 ` mikpe at it dot uu dot se
` (37 subsequent siblings)
44 siblings, 0 replies; 46+ messages in thread
From: mikpe at it dot uu dot se @ 2010-02-05 15:40 UTC (permalink / raw)
To: java-prs
------- Comment #4 from mikpe at it dot uu dot se 2010-02-05 15:40 -------
Further experiments show that the failures are triggered by using the
binutils-2.20 version of ld when bootstrapping.
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=40860
^ permalink raw reply [flat|nested] 46+ messages in thread* [Bug libgcj/40860] [4.4/4.5 regression] regressions in libjava testsuite on arm-linux
2009-07-26 10:51 [Bug libgcj/40860] New: [4.4/4.5 regression] regressions in libjava testsuite on arm-linux debian-gcc at lists dot debian dot org
` (6 preceding siblings ...)
2010-02-05 15:40 ` mikpe at it dot uu dot se
@ 2010-02-06 15:36 ` mikpe at it dot uu dot se
2010-02-13 20:49 ` mikpe at it dot uu dot se
` (36 subsequent siblings)
44 siblings, 0 replies; 46+ messages in thread
From: mikpe at it dot uu dot se @ 2010-02-06 15:36 UTC (permalink / raw)
To: java-prs
------- Comment #5 from mikpe at it dot uu dot se 2010-02-06 15:36 -------
Same failures when bootstrapped with binutils-2.20.51.20100202. So there's no
fix on binutils trunk to be identified and backported. I'll try to identify the
point where the regression was introduced instead.
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=40860
^ permalink raw reply [flat|nested] 46+ messages in thread* [Bug libgcj/40860] [4.4/4.5 regression] regressions in libjava testsuite on arm-linux
2009-07-26 10:51 [Bug libgcj/40860] New: [4.4/4.5 regression] regressions in libjava testsuite on arm-linux debian-gcc at lists dot debian dot org
` (7 preceding siblings ...)
2010-02-06 15:36 ` mikpe at it dot uu dot se
@ 2010-02-13 20:49 ` mikpe at it dot uu dot se
2010-02-15 15:32 ` doko at ubuntu dot com
` (35 subsequent siblings)
44 siblings, 0 replies; 46+ messages in thread
From: mikpe at it dot uu dot se @ 2010-02-13 20:49 UTC (permalink / raw)
To: java-prs
------- Comment #6 from mikpe at it dot uu dot se 2010-02-13 20:49 -------
A bisection through the binutils weekly and daily snapshots has identified
090505 as the last good snapshot and 090506 as the first bad one.
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=40860
^ permalink raw reply [flat|nested] 46+ messages in thread* [Bug libgcj/40860] [4.4/4.5 regression] regressions in libjava testsuite on arm-linux
2009-07-26 10:51 [Bug libgcj/40860] New: [4.4/4.5 regression] regressions in libjava testsuite on arm-linux debian-gcc at lists dot debian dot org
` (8 preceding siblings ...)
2010-02-13 20:49 ` mikpe at it dot uu dot se
@ 2010-02-15 15:32 ` doko at ubuntu dot com
2010-02-15 22:26 ` mikpe at it dot uu dot se
` (34 subsequent siblings)
44 siblings, 0 replies; 46+ messages in thread
From: doko at ubuntu dot com @ 2010-02-15 15:32 UTC (permalink / raw)
To: java-prs
------- Comment #7 from doko at ubuntu dot com 2010-02-15 15:32 -------
this seems to be the patch for the ARM unwind table linker processing? if yes,
see http://sourceware.org/bugzilla/show_bug.cgi?id=10695 for some comments.
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=40860
^ permalink raw reply [flat|nested] 46+ messages in thread* [Bug libgcj/40860] [4.4/4.5 regression] regressions in libjava testsuite on arm-linux
2009-07-26 10:51 [Bug libgcj/40860] New: [4.4/4.5 regression] regressions in libjava testsuite on arm-linux debian-gcc at lists dot debian dot org
` (9 preceding siblings ...)
2010-02-15 15:32 ` doko at ubuntu dot com
@ 2010-02-15 22:26 ` mikpe at it dot uu dot se
2010-02-16 16:34 ` doko at ubuntu dot com
` (33 subsequent siblings)
44 siblings, 0 replies; 46+ messages in thread
From: mikpe at it dot uu dot se @ 2010-02-15 22:26 UTC (permalink / raw)
To: java-prs
------- Comment #8 from mikpe at it dot uu dot se 2010-02-15 22:26 -------
I've identified <http://sourceware.org/ml/binutils-cvs/2009-05/msg00021.html>
as the source of this regression.
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=40860
^ permalink raw reply [flat|nested] 46+ messages in thread* [Bug libgcj/40860] [4.4/4.5 regression] regressions in libjava testsuite on arm-linux
2009-07-26 10:51 [Bug libgcj/40860] New: [4.4/4.5 regression] regressions in libjava testsuite on arm-linux debian-gcc at lists dot debian dot org
` (10 preceding siblings ...)
2010-02-15 22:26 ` mikpe at it dot uu dot se
@ 2010-02-16 16:34 ` doko at ubuntu dot com
2010-02-19 23:32 ` mikpe at it dot uu dot se
` (32 subsequent siblings)
44 siblings, 0 replies; 46+ messages in thread
From: doko at ubuntu dot com @ 2010-02-16 16:34 UTC (permalink / raw)
To: java-prs
------- Comment #9 from doko at ubuntu dot com 2010-02-16 16:34 -------
confirmed with binutils trunk 20100216, checking with gold from the trunk.
--
doko at ubuntu dot com changed:
What |Removed |Added
----------------------------------------------------------------------------
CC| |jsm28 at gcc dot gnu dot org
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=40860
^ permalink raw reply [flat|nested] 46+ messages in thread* [Bug libgcj/40860] [4.4/4.5 regression] regressions in libjava testsuite on arm-linux
2009-07-26 10:51 [Bug libgcj/40860] New: [4.4/4.5 regression] regressions in libjava testsuite on arm-linux debian-gcc at lists dot debian dot org
` (11 preceding siblings ...)
2010-02-16 16:34 ` doko at ubuntu dot com
@ 2010-02-19 23:32 ` mikpe at it dot uu dot se
2010-02-22 21:49 ` mikpe at it dot uu dot se
` (31 subsequent siblings)
44 siblings, 0 replies; 46+ messages in thread
From: mikpe at it dot uu dot se @ 2010-02-19 23:32 UTC (permalink / raw)
To: java-prs
------- Comment #10 from mikpe at it dot uu dot se 2010-02-19 23:32 -------
The binutils patch that broke libjava came from CodeSourcery, so I decided to
test their latest G++ lite for arm-linux-gnueabi (2009q3-67). That compiler is
based on gcc-4.4.1, and it does not have the libjava regressions with current
binutils that we see with FSF gcc.
Looking at the source diffs and changelogs, I've identified four distinct
unwind-related changes in their gcc, two of which are ARM-specific. I'll test
them on FSF 4.4 next.
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=40860
^ permalink raw reply [flat|nested] 46+ messages in thread* [Bug libgcj/40860] [4.4/4.5 regression] regressions in libjava testsuite on arm-linux
2009-07-26 10:51 [Bug libgcj/40860] New: [4.4/4.5 regression] regressions in libjava testsuite on arm-linux debian-gcc at lists dot debian dot org
` (12 preceding siblings ...)
2010-02-19 23:32 ` mikpe at it dot uu dot se
@ 2010-02-22 21:49 ` mikpe at it dot uu dot se
2010-02-28 10:07 ` aph at gcc dot gnu dot org
` (30 subsequent siblings)
44 siblings, 0 replies; 46+ messages in thread
From: mikpe at it dot uu dot se @ 2010-02-22 21:49 UTC (permalink / raw)
To: java-prs
------- Comment #11 from mikpe at it dot uu dot se 2010-02-22 21:49 -------
A recent 4.4 with the four unwind-related fixes I identified in CodeSourcery
G++ lite 2009q3-67 applied still regresses in libjava as before with recent
binutils.
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=40860
^ permalink raw reply [flat|nested] 46+ messages in thread* [Bug libgcj/40860] [4.4/4.5 regression] regressions in libjava testsuite on arm-linux
2009-07-26 10:51 [Bug libgcj/40860] New: [4.4/4.5 regression] regressions in libjava testsuite on arm-linux debian-gcc at lists dot debian dot org
` (13 preceding siblings ...)
2010-02-22 21:49 ` mikpe at it dot uu dot se
@ 2010-02-28 10:07 ` aph at gcc dot gnu dot org
2010-03-04 10:17 ` mikpe at it dot uu dot se
` (29 subsequent siblings)
44 siblings, 0 replies; 46+ messages in thread
From: aph at gcc dot gnu dot org @ 2010-02-28 10:07 UTC (permalink / raw)
To: java-prs
------- Comment #12 from aph at gcc dot gnu dot org 2010-02-28 10:07 -------
I can't duplicate this problem with gcc trunk and binutils 2.20-0ubuntu2.
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=40860
^ permalink raw reply [flat|nested] 46+ messages in thread* [Bug libgcj/40860] [4.4/4.5 regression] regressions in libjava testsuite on arm-linux
2009-07-26 10:51 [Bug libgcj/40860] New: [4.4/4.5 regression] regressions in libjava testsuite on arm-linux debian-gcc at lists dot debian dot org
` (14 preceding siblings ...)
2010-02-28 10:07 ` aph at gcc dot gnu dot org
@ 2010-03-04 10:17 ` mikpe at it dot uu dot se
2010-03-15 9:09 ` mikpe at it dot uu dot se
` (28 subsequent siblings)
44 siblings, 0 replies; 46+ messages in thread
From: mikpe at it dot uu dot se @ 2010-03-04 10:17 UTC (permalink / raw)
To: java-prs
------- Comment #13 from mikpe at it dot uu dot se 2010-03-04 10:16 -------
With binutils-2.20.51.20100223, gcc-4.5-20090327 regresses libjava as described
here but gcc-4.5-20100225 does not.
Starting bisection on trunk (with 24+ hour build/test cycles, that's going to
take a while).
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=40860
^ permalink raw reply [flat|nested] 46+ messages in thread* [Bug libgcj/40860] [4.4/4.5 regression] regressions in libjava testsuite on arm-linux
2009-07-26 10:51 [Bug libgcj/40860] New: [4.4/4.5 regression] regressions in libjava testsuite on arm-linux debian-gcc at lists dot debian dot org
` (15 preceding siblings ...)
2010-03-04 10:17 ` mikpe at it dot uu dot se
@ 2010-03-15 9:09 ` mikpe at it dot uu dot se
2010-03-15 9:16 ` rearnsha at gcc dot gnu dot org
` (27 subsequent siblings)
44 siblings, 0 replies; 46+ messages in thread
From: mikpe at it dot uu dot se @ 2010-03-15 9:09 UTC (permalink / raw)
To: java-prs
------- Comment #14 from mikpe at it dot uu dot se 2010-03-15 09:09 -------
The bug was fixed for 4.5 by r148072:
2009-06-02 Richard Earnshaw <rearnsha@arm.com>
* arm.c (arm_get_frame_offsets): Prefer using r3 for padding a
push/pop multiple to 8-byte alignment.
That change applies cleanly to 4.4 and fixes the bug there as well, although
I've only done a smallish non-bootstrap build and libjava-only test with it.
I'll do a full bootstrap and regression test next.
This change is included in CodeSourcery's 4.4-based G++ lite, which explains
why their compiler works with the newer binutils.
Looking at the patch submission
http://gcc.gnu.org/ml/gcc-patches/2009-06/msg00116.html
I see no indication that this should have any effect on exception handling or
stack unwinding, but obviously it does. I'm adding Richard E. to the cc: list.
--
mikpe at it dot uu dot se changed:
What |Removed |Added
----------------------------------------------------------------------------
CC| |rearnsha at arm dot com
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=40860
^ permalink raw reply [flat|nested] 46+ messages in thread* [Bug libgcj/40860] [4.4/4.5 regression] regressions in libjava testsuite on arm-linux
2009-07-26 10:51 [Bug libgcj/40860] New: [4.4/4.5 regression] regressions in libjava testsuite on arm-linux debian-gcc at lists dot debian dot org
` (16 preceding siblings ...)
2010-03-15 9:09 ` mikpe at it dot uu dot se
@ 2010-03-15 9:16 ` rearnsha at gcc dot gnu dot org
2010-03-16 13:42 ` doko at ubuntu dot com
` (26 subsequent siblings)
44 siblings, 0 replies; 46+ messages in thread
From: rearnsha at gcc dot gnu dot org @ 2010-03-15 9:16 UTC (permalink / raw)
To: java-prs
------- Comment #15 from rearnsha at gcc dot gnu dot org 2010-03-15 09:16 -------
(In reply to comment #14)
> The bug was fixed for 4.5 by r148072:
>
> 2009-06-02 Richard Earnshaw <rearnsha@arm.com>
>
> * arm.c (arm_get_frame_offsets): Prefer using r3 for padding a
> push/pop multiple to 8-byte alignment.
>
I doubt that's really fixed the bug. More likely it's caused it to go latent
by changing the unwind information in some subtle way.
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=40860
^ permalink raw reply [flat|nested] 46+ messages in thread* [Bug libgcj/40860] [4.4/4.5 regression] regressions in libjava testsuite on arm-linux
2009-07-26 10:51 [Bug libgcj/40860] New: [4.4/4.5 regression] regressions in libjava testsuite on arm-linux debian-gcc at lists dot debian dot org
` (17 preceding siblings ...)
2010-03-15 9:16 ` rearnsha at gcc dot gnu dot org
@ 2010-03-16 13:42 ` doko at ubuntu dot com
2010-03-16 17:29 ` mikpe at it dot uu dot se
` (25 subsequent siblings)
44 siblings, 0 replies; 46+ messages in thread
From: doko at ubuntu dot com @ 2010-03-16 13:42 UTC (permalink / raw)
To: java-prs
------- Comment #16 from doko at ubuntu dot com 2010-03-16 13:42 -------
this change doesn't make any change to the libjava test results (configured
--with-mode=thumb).
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=40860
^ permalink raw reply [flat|nested] 46+ messages in thread* [Bug libgcj/40860] [4.4/4.5 regression] regressions in libjava testsuite on arm-linux
2009-07-26 10:51 [Bug libgcj/40860] New: [4.4/4.5 regression] regressions in libjava testsuite on arm-linux debian-gcc at lists dot debian dot org
` (18 preceding siblings ...)
2010-03-16 13:42 ` doko at ubuntu dot com
@ 2010-03-16 17:29 ` mikpe at it dot uu dot se
2010-03-16 23:30 ` doko at ubuntu dot com
` (24 subsequent siblings)
44 siblings, 0 replies; 46+ messages in thread
From: mikpe at it dot uu dot se @ 2010-03-16 17:29 UTC (permalink / raw)
To: java-prs
------- Comment #17 from mikpe at it dot uu dot se 2010-03-16 17:29 -------
I decided to attack things from the binutils side. The problematic binutils
change rolls two distinct changes into one (add cantunwind table entries, merge
adjacent equivalent table entries). Disabling the "merge" part eliminated the
libjava regressions in my ARM mode build:
--- binutils-2.20.51/bfd/elf32-arm.c.~1~
+++ binutils-2.20.51/bfd/elf32-arm.c
@@ -9301,10 +9306,12 @@ elf32_arm_fix_exidx_coverage (asection *
if (elide)
{
+#if 0
add_unwind_table_edit (&unwind_edit_head, &unwind_edit_tail,
DELETE_EXIDX_ENTRY, NULL, j / 8);
deleted_exidx_bytes += 8;
+#endif
}
last_unwind_type = unwind_type;
I'll keep experimenting.
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=40860
^ permalink raw reply [flat|nested] 46+ messages in thread* [Bug libgcj/40860] [4.4/4.5 regression] regressions in libjava testsuite on arm-linux
2009-07-26 10:51 [Bug libgcj/40860] New: [4.4/4.5 regression] regressions in libjava testsuite on arm-linux debian-gcc at lists dot debian dot org
` (19 preceding siblings ...)
2010-03-16 17:29 ` mikpe at it dot uu dot se
@ 2010-03-16 23:30 ` doko at ubuntu dot com
2010-03-16 23:39 ` mikpe at it dot uu dot se
` (23 subsequent siblings)
44 siblings, 0 replies; 46+ messages in thread
From: doko at ubuntu dot com @ 2010-03-16 23:30 UTC (permalink / raw)
To: java-prs
------- Comment #18 from doko at ubuntu dot com 2010-03-16 23:30 -------
rebuilding binutils 2.20.1 with the patch from comment #17 shows one
regression:
Running /home/doko/binutils/binutils-2.20.1/ld/testsuite/ld-arm/arm-elf.exp ...
FAIL: ld-arm/unwind-4
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=40860
^ permalink raw reply [flat|nested] 46+ messages in thread* [Bug libgcj/40860] [4.4/4.5 regression] regressions in libjava testsuite on arm-linux
2009-07-26 10:51 [Bug libgcj/40860] New: [4.4/4.5 regression] regressions in libjava testsuite on arm-linux debian-gcc at lists dot debian dot org
` (20 preceding siblings ...)
2010-03-16 23:30 ` doko at ubuntu dot com
@ 2010-03-16 23:39 ` mikpe at it dot uu dot se
2010-03-17 10:51 ` doko at ubuntu dot com
` (22 subsequent siblings)
44 siblings, 0 replies; 46+ messages in thread
From: mikpe at it dot uu dot se @ 2010-03-16 23:39 UTC (permalink / raw)
To: java-prs
------- Comment #19 from mikpe at it dot uu dot se 2010-03-16 23:39 -------
(In reply to comment #18)
> rebuilding binutils 2.20.1 with the patch from comment #17 shows one
> regression:
>
> Running /home/doko/binutils/binutils-2.20.1/ld/testsuite/ld-arm/arm-elf.exp ...
> FAIL: ld-arm/unwind-4
I guess I should've run the binutils testsuite first. (Looking..) Ah, but that
failure is Ok, as it's checking for the merging that the patch disabled.
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=40860
^ permalink raw reply [flat|nested] 46+ messages in thread* [Bug libgcj/40860] [4.4/4.5 regression] regressions in libjava testsuite on arm-linux
2009-07-26 10:51 [Bug libgcj/40860] New: [4.4/4.5 regression] regressions in libjava testsuite on arm-linux debian-gcc at lists dot debian dot org
` (21 preceding siblings ...)
2010-03-16 23:39 ` mikpe at it dot uu dot se
@ 2010-03-17 10:51 ` doko at ubuntu dot com
2010-03-17 21:13 ` mikpe at it dot uu dot se
` (21 subsequent siblings)
44 siblings, 0 replies; 46+ messages in thread
From: doko at ubuntu dot com @ 2010-03-17 10:51 UTC (permalink / raw)
To: java-prs
------- Comment #20 from doko at ubuntu dot com 2010-03-17 10:51 -------
no change in the libjava testsuite when built with these binutils
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=40860
^ permalink raw reply [flat|nested] 46+ messages in thread* [Bug libgcj/40860] [4.4/4.5 regression] regressions in libjava testsuite on arm-linux
2009-07-26 10:51 [Bug libgcj/40860] New: [4.4/4.5 regression] regressions in libjava testsuite on arm-linux debian-gcc at lists dot debian dot org
` (22 preceding siblings ...)
2010-03-17 10:51 ` doko at ubuntu dot com
@ 2010-03-17 21:13 ` mikpe at it dot uu dot se
2010-03-17 21:23 ` mikpe at it dot uu dot se
` (20 subsequent siblings)
44 siblings, 0 replies; 46+ messages in thread
From: mikpe at it dot uu dot se @ 2010-03-17 21:13 UTC (permalink / raw)
To: java-prs
------- Comment #21 from mikpe at it dot uu dot se 2010-03-17 21:13 -------
(In reply to comment #20)
> no change in the libjava testsuite when built with these binutils
But that's still thumb not arm like in comment #16? All my results are from
plain arm (armv5tel) builds.
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=40860
^ permalink raw reply [flat|nested] 46+ messages in thread* [Bug libgcj/40860] [4.4/4.5 regression] regressions in libjava testsuite on arm-linux
2009-07-26 10:51 [Bug libgcj/40860] New: [4.4/4.5 regression] regressions in libjava testsuite on arm-linux debian-gcc at lists dot debian dot org
` (23 preceding siblings ...)
2010-03-17 21:13 ` mikpe at it dot uu dot se
@ 2010-03-17 21:23 ` mikpe at it dot uu dot se
2010-03-19 23:20 ` mikpe at it dot uu dot se
` (19 subsequent siblings)
44 siblings, 0 replies; 46+ messages in thread
From: mikpe at it dot uu dot se @ 2010-03-17 21:23 UTC (permalink / raw)
To: java-prs
------- Comment #22 from mikpe at it dot uu dot se 2010-03-17 21:23 -------
I did another binutils experiment. I reverted my patch to disable general
merging of table entries, and instead disabled generating new and merging
cantunwind entries. With that binutils libjava regressed just like with vanilla
post-2.19.1 binutils. The general merging bit seems to be the problem.
--- binutils-2.20.51/bfd/elf32-arm.c.~1~
+++ binutils-2.20.51/bfd/elf32-arm.c
@@ -9148,6 +9148,7 @@ adjust_exidx_size(asection *exidx_sec, i
bfd_set_section_size (out_sec->owner, out_sec, out_sec->size +adjust);
}
+#if 0
/* Insert an EXIDX_CANTUNWIND marker at the end of a section. */
static void
insert_cantunwind_after(asection *text_sec, asection *exidx_sec)
@@ -9162,6 +9163,7 @@ insert_cantunwind_after(asection *text_s
adjust_exidx_size(exidx_sec, 8);
}
+#endif
/* Scan .ARM.exidx tables, and create a list describing edits which should be
made to those tables, such that:
@@ -9248,7 +9250,9 @@ elf32_arm_fix_exidx_coverage (asection *
if (sec->size == 0)
continue;
+#if 0
insert_cantunwind_after(last_text_sec, last_exidx_sec);
+#endif
last_unwind_type = 0;
continue;
}
@@ -9282,8 +9286,10 @@ elf32_arm_fix_exidx_coverage (asection *
/* An EXIDX_CANTUNWIND entry. */
if (second_word == 1)
{
+#if 0
if (last_unwind_type == 0)
elide = 1;
+#endif
unwind_type = 0;
}
/* Inlined unwinding data. Merge if equal to previous. */
@@ -9326,8 +9332,10 @@ elf32_arm_fix_exidx_coverage (asection *
}
/* Add terminating CANTUNWIND entry. */
+#if 0
if (last_exidx_sec && last_unwind_type != 0)
insert_cantunwind_after(last_text_sec, last_exidx_sec);
+#endif
return TRUE;
}
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=40860
^ permalink raw reply [flat|nested] 46+ messages in thread* [Bug libgcj/40860] [4.4/4.5 regression] regressions in libjava testsuite on arm-linux
2009-07-26 10:51 [Bug libgcj/40860] New: [4.4/4.5 regression] regressions in libjava testsuite on arm-linux debian-gcc at lists dot debian dot org
` (24 preceding siblings ...)
2010-03-17 21:23 ` mikpe at it dot uu dot se
@ 2010-03-19 23:20 ` mikpe at it dot uu dot se
2010-03-20 18:54 ` ramana at gcc dot gnu dot org
` (18 subsequent siblings)
44 siblings, 0 replies; 46+ messages in thread
From: mikpe at it dot uu dot se @ 2010-03-19 23:20 UTC (permalink / raw)
To: java-prs
------- Comment #23 from mikpe at it dot uu dot se 2010-03-19 23:20 -------
While working on another debugging patch I noticed something that I think might
explain the libjava regressions, especially the stack trace ones.
The binutils change breaks _Unwind_GetRegionStart().
Example: We have three functions, f1() starts at address 100 and is 20 bytes,
f2() starts at address 120 and is 10 bytes, and f3() starts at address 130.
f1() and f2() got the same inlined unwind table data U1, while f3() got some
different unwind table data U3.
A pre-binutils-2.20 .ARM.exidx table for this program would be
[0] = { 100, U1 }, [1] = { 120, U1 }, [2] = { 130, U3 }, ...
binutils-2.20 and later merge entry [1] with entry [0] since they are adjacent
and have the same unwind table data, producing a final .ARM.exidx table:
[0] = { 100, U1 }, [1] = { 130, U3 }, ...
So an exception in or unwind through f2() will now return { 100, U1 } rather
than { 120, U1 }.
The first thing that happens with the found table entry is that the function
start address is stashed away (unwind-arm.c:get_eit_entry(), the assignment to
ucbp->pr_cache.fnstart). Then __gnu_unwind_pr_common() uses fnstart while
interpreting the unwind table data; I haven't studied that code but since
libstdc++ test cases don't regress I assume the incorrect fnstart doesn't
matter for inlined unwind table data.
However, pr-support.c:_Unwind_GetRegionStart() also uses fnstart, so users of
_Unwind_GetRegionStart() will see f1() not f2() for exceptions in or unwinds
through f2(). In particular, libjava/stacktrace.cc looks like it will construct
bogus stack traces (see its uses of 'start_ip').
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=40860
^ permalink raw reply [flat|nested] 46+ messages in thread* [Bug libgcj/40860] [4.4/4.5 regression] regressions in libjava testsuite on arm-linux
2009-07-26 10:51 [Bug libgcj/40860] New: [4.4/4.5 regression] regressions in libjava testsuite on arm-linux debian-gcc at lists dot debian dot org
` (25 preceding siblings ...)
2010-03-19 23:20 ` mikpe at it dot uu dot se
@ 2010-03-20 18:54 ` ramana at gcc dot gnu dot org
2010-03-20 22:10 ` mikpe at it dot uu dot se
` (17 subsequent siblings)
44 siblings, 0 replies; 46+ messages in thread
From: ramana at gcc dot gnu dot org @ 2010-03-20 18:54 UTC (permalink / raw)
To: java-prs
------- Comment #24 from ramana at gcc dot gnu dot org 2010-03-20 18:54 -------
This rings a bell and the comment in
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=29206#c8
is probably related.
If so, this is essentially a dup of PR29206 and I'm not sure what more we can
do with this.
cheers
Ramana
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=40860
^ permalink raw reply [flat|nested] 46+ messages in thread* [Bug libgcj/40860] [4.4/4.5 regression] regressions in libjava testsuite on arm-linux
2009-07-26 10:51 [Bug libgcj/40860] New: [4.4/4.5 regression] regressions in libjava testsuite on arm-linux debian-gcc at lists dot debian dot org
` (26 preceding siblings ...)
2010-03-20 18:54 ` ramana at gcc dot gnu dot org
@ 2010-03-20 22:10 ` mikpe at it dot uu dot se
2010-03-20 22:36 ` mikpe at it dot uu dot se
` (16 subsequent siblings)
44 siblings, 0 replies; 46+ messages in thread
From: mikpe at it dot uu dot se @ 2010-03-20 22:10 UTC (permalink / raw)
To: java-prs
------- Comment #25 from mikpe at it dot uu dot se 2010-03-20 22:10 -------
(In reply to comment #24)
> This rings a bell and the comment in
> http://gcc.gnu.org/bugzilla/show_bug.cgi?id=29206#c8
> is probably related.
I don't think this is at all the same as PR29206. PR29206 is very old now, so I
don't think the one comment about EABI is accurate now, and people have been
using EABI java builds succesfully. The test case given (gcj-dbtool -n) works
for me. If anything, I suspect that maybe the ARM unwinder loops at end of
stack bug is at play; for java at least there's a patch in Debian to work
around that (I use it in my toolchain too).
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=40860
^ permalink raw reply [flat|nested] 46+ messages in thread* [Bug libgcj/40860] [4.4/4.5 regression] regressions in libjava testsuite on arm-linux
2009-07-26 10:51 [Bug libgcj/40860] New: [4.4/4.5 regression] regressions in libjava testsuite on arm-linux debian-gcc at lists dot debian dot org
` (27 preceding siblings ...)
2010-03-20 22:10 ` mikpe at it dot uu dot se
@ 2010-03-20 22:36 ` mikpe at it dot uu dot se
2010-03-22 23:48 ` rearnsha at gcc dot gnu dot org
` (15 subsequent siblings)
44 siblings, 0 replies; 46+ messages in thread
From: mikpe at it dot uu dot se @ 2010-03-20 22:36 UTC (permalink / raw)
To: java-prs
------- Comment #26 from mikpe at it dot uu dot se 2010-03-20 22:36 -------
Created an attachment (id=20147)
--> (http://gcc.gnu.org/bugzilla/attachment.cgi?id=20147&action=view)
test program to take a stack trace using _Unwind_ API
I'm attaching a test program which sets up a predictable call chain, and then
uses _Unwind_Backtrace() and _Unwind_GetRegionStart() to get the entry points
of the functions in the call chain into an array. It prints the corresponding
function names and verifies that the function entry points are correct.
With gcc-4.4.3 and binutils-2.19.1 on armv5tel-unknown-linux-gnueabi this works
fine:
> gcc -funwind-tables -fno-unit-at-a-time -O -o stacktrace-2.19.1 stacktrace.c
> ./stacktrace-2.19.1
bar
f2
main
With the same gcc binary but binutils-20100316 the program fails:
> PATH=/tmp/binutils-2.20.51-install/bin:$PATH gcc -funwind-tables -fno-unit-at-a-time -O stacktrace.c -o stacktrace-2.20.51
> ./stacktrace-2.20.51
bar
f1
f1
Abort
The reason for the failure is that ld merged the .ARM.exidx entries for f2 and
main with the one for f1, which breaks _Unwind_GetRegionStart(). First we see
that f1, f2, and main are adjacent in memory:
> nm -v stacktrace-2.20.51
...
00008588 T tracefn
000085bc T bar
0000865c T f1
0000866c T f2
0000867c T main
00008698 T __libc_csu_fini
...
With binutils-2.19.1 they each get their own .ARM.exidx entry:
> /tmp/binutils-2.20.51-install/bin/readelf --unwind stacktrace-2.19.1
Unwind table index '.ARM.exidx' at offset 0x778 contains 7 entries:
0x845c <_start>: 0x1 [cantunwind]
0x8508 <printf_fn>: 0x80a8b0b0
Compact model 0
0xa8 pop {r4r14}
0xb0 finish
0xb0 finish
0x8588 <tracefn>: 0x80aab0b0
Compact model 0
0xaa pop {r4, r5, r6r14}
0xb0 finish
0xb0 finish
0x85bc <bar>: 0x8003a8b0
Compact model 0
0x03 vsp = vsp + 16
0xa8 pop {r4r14}
0xb0 finish
0x865c <f1>: 0x80a8b0b0
Compact model 0
0xa8 pop {r4r14}
0xb0 finish
0xb0 finish
0x866c <f2>: 0x80a8b0b0
Compact model 0
0xa8 pop {r4r14}
0xb0 finish
0xb0 finish
0x867c <main>: 0x80a8b0b0
Compact model 0
0xa8 pop {r4r14}
0xb0 finish
0xb0 finish
But with binutils-20100316 f2 and main don't have their own entries, so PCs in
them are reported as belonging to f1 instead:
> /tmp/binutils-2.20.51-install/bin/readelf --unwind stacktrace-2.20.51
Unwind table index '.ARM.exidx' at offset 0x778 contains 6 entries:
0x845c <_start>: 0x1 [cantunwind]
0x8508 <printf_fn>: 0x80a8b0b0
Compact model 0
0xa8 pop {r4r14}
0xb0 finish
0xb0 finish
0x8588 <tracefn>: 0x80aab0b0
Compact model 0
0xaa pop {r4, r5, r6r14}
0xb0 finish
0xb0 finish
0x85bc <bar>: 0x8003a8b0
Compact model 0
0x03 vsp = vsp + 16
0xa8 pop {r4r14}
0xb0 finish
0x865c <f1>: 0x80a8b0b0
Compact model 0
0xa8 pop {r4r14}
0xb0 finish
0xb0 finish
0x8698 <__libc_csu_fini>: 0x1 [cantunwind]
This test program also works on i686-linux, sparc64-linux -m32/-m64, and
powerpc64-linux -m32.
I understand the benefits of shrinking .ARM.exidx tables, but it does break
parts of gcc's <unwind.h> implementation (and thus libjava), so should be
opt-in via an explicit option and not done by default.
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=40860
^ permalink raw reply [flat|nested] 46+ messages in thread* [Bug libgcj/40860] [4.4/4.5 regression] regressions in libjava testsuite on arm-linux
2009-07-26 10:51 [Bug libgcj/40860] New: [4.4/4.5 regression] regressions in libjava testsuite on arm-linux debian-gcc at lists dot debian dot org
` (28 preceding siblings ...)
2010-03-20 22:36 ` mikpe at it dot uu dot se
@ 2010-03-22 23:48 ` rearnsha at gcc dot gnu dot org
2010-03-30 13:21 ` mikpe at it dot uu dot se
` (14 subsequent siblings)
44 siblings, 0 replies; 46+ messages in thread
From: rearnsha at gcc dot gnu dot org @ 2010-03-22 23:48 UTC (permalink / raw)
To: java-prs
------- Comment #27 from rearnsha at gcc dot gnu dot org 2010-03-22 23:48 -------
The ARM ABI permits merging of unwind entries, so this should never default to
opt-in across the entire tool-chain. It might be that when GCC links java
programs that it should pass an option to the linker to disable the
optimization, but that's a different story.
It may be that the java unwinder will need to just do more work to figure out
the real call sequence. I'm not sufficiently familiar with the code to know
the answer here.
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=40860
^ permalink raw reply [flat|nested] 46+ messages in thread* [Bug libgcj/40860] [4.4/4.5 regression] regressions in libjava testsuite on arm-linux
2009-07-26 10:51 [Bug libgcj/40860] New: [4.4/4.5 regression] regressions in libjava testsuite on arm-linux debian-gcc at lists dot debian dot org
` (29 preceding siblings ...)
2010-03-22 23:48 ` rearnsha at gcc dot gnu dot org
@ 2010-03-30 13:21 ` mikpe at it dot uu dot se
2010-03-30 14:04 ` pbrook at gcc dot gnu dot org
` (13 subsequent siblings)
44 siblings, 0 replies; 46+ messages in thread
From: mikpe at it dot uu dot se @ 2010-03-30 13:21 UTC (permalink / raw)
To: java-prs
------- Comment #28 from mikpe at it dot uu dot se 2010-03-30 13:21 -------
I've looked at the amount of .ARM.exidx entry merging being done and its
consequences for the various unwinders in gcc. Currently only table entries
with immediate (inlined) data are merged, and for that all of gcc except for
libjava seems to be Ok. However, gcc can still leak bogus unwind data via
_Unwind_GetRegionStart, so I'm proposing a patch like the following:
--- gcc-4.4.3/gcc/config/arm/unwind-arm.c.~1~
+++ gcc-4.4.3/gcc/config/arm/unwind-arm.c
@@ -621,7 +621,6 @@ get_eit_entry (_Unwind_Control_Block *uc
UCB_PR_ADDR (ucbp) = 0;
return _URC_FAILURE;
}
- ucbp->pr_cache.fnstart = selfrel_offset31 (&eitp->fnoffset);
/* Can this frame be unwound at all? */
if (eitp->content == EXIDX_CANTUNWIND)
@@ -637,6 +636,15 @@ get_eit_entry (_Unwind_Control_Block *uc
/* It is immediate data. */
ucbp->pr_cache.ehtp = (_Unwind_EHT_Header *)&eitp->content;
ucbp->pr_cache.additional = 1;
+ /* Adjacent EIT entries with identical immediate data may be merged,
+ making fnoffset/fnstart inaccurate. The ARM unwinder doesn't need
+ fnstart for immediate EIT data. Other PRs than ARM's often use
+ fnstart to derive the locations of landing pads, but such PRs cannot
+ use immediate data in EIT entries, so are not affected by this issue.
+ However, code constructing stack traces may see stack frames for
+ functions with immediate data EIT entries. Clear fnstart to ensure
+ _Unwind_GetRegionStart doesn't return wrong data in this case. */
+ ucbp->pr_cache.fnstart = 0;
}
else
{
@@ -645,6 +653,7 @@ get_eit_entry (_Unwind_Control_Block *uc
ucbp->pr_cache.ehtp =
(_Unwind_EHT_Header *) selfrel_offset31 (&eitp->content);
ucbp->pr_cache.additional = 0;
+ ucbp->pr_cache.fnstart = selfrel_offset31 (&eitp->fnoffset);
}
/* Discover the personality routine address. */
This caused no regressions for c/c++/objc/obj-c++, but libjava got two more
(ExtraClassLoader and InvokeInterface).
The problem with libjava appears to be its stacktrace.cc module. It uses
_Unwind_GetRegionStart to realign any interior PC to its function start PC,
then it uses that to look up method and class in a hash table keyed by method
start PC. With the .ARM.exidx merging, _Unwind_GetRegionStart can return the
PC for a different method, possibly also in a different class, which totally
breaks this. With my patch above libjava's stacktrace.cc can detect this case
and switch to a linear search instead. I'll try to implement that soonish.
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=40860
^ permalink raw reply [flat|nested] 46+ messages in thread* [Bug libgcj/40860] [4.4/4.5 regression] regressions in libjava testsuite on arm-linux
2009-07-26 10:51 [Bug libgcj/40860] New: [4.4/4.5 regression] regressions in libjava testsuite on arm-linux debian-gcc at lists dot debian dot org
` (30 preceding siblings ...)
2010-03-30 13:21 ` mikpe at it dot uu dot se
@ 2010-03-30 14:04 ` pbrook at gcc dot gnu dot org
2010-03-30 15:09 ` mikpe at it dot uu dot se
` (12 subsequent siblings)
44 siblings, 0 replies; 46+ messages in thread
From: pbrook at gcc dot gnu dot org @ 2010-03-30 14:04 UTC (permalink / raw)
To: java-prs
------- Comment #29 from pbrook at gcc dot gnu dot org 2010-03-30 14:04 -------
Wouldn't it be better to just remove _Unwind_GetRegionStart?
This function is not part of the ARM EABI, and removing it would expose any
(already broken) users at compile time.
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=40860
^ permalink raw reply [flat|nested] 46+ messages in thread* [Bug libgcj/40860] [4.4/4.5 regression] regressions in libjava testsuite on arm-linux
2009-07-26 10:51 [Bug libgcj/40860] New: [4.4/4.5 regression] regressions in libjava testsuite on arm-linux debian-gcc at lists dot debian dot org
` (31 preceding siblings ...)
2010-03-30 14:04 ` pbrook at gcc dot gnu dot org
@ 2010-03-30 15:09 ` mikpe at it dot uu dot se
2010-03-31 8:47 ` rearnsha at gcc dot gnu dot org
` (11 subsequent siblings)
44 siblings, 0 replies; 46+ messages in thread
From: mikpe at it dot uu dot se @ 2010-03-30 15:09 UTC (permalink / raw)
To: java-prs
------- Comment #30 from mikpe at it dot uu dot se 2010-03-30 15:09 -------
(In reply to comment #29)
> Wouldn't it be better to just remove _Unwind_GetRegionStart?
> This function is not part of the ARM EABI, and removing it would expose any
> (already broken) users at compile time.
No.
First it'd break most of gcc since the c, c++, and objc unwinders use it. And
they generally use it to provide a base address when interpreting LSDA and
computing landing pad addresses.
Second, all _Unwind_GetRegionStart does is give r/o access to the fnstart value
in ARM's UCB. But ARM's own unwinder uses fnstart in __gnu_unwind_pr_common,
so if fnstart is broken then so it ARM's unwinder.
ARM's unwinder is in the same boat as the c/c++/objc ones. It works because
.ARM.exidx merging is limited to immediate table data, but the code using
fnstart (by luck or design) only runs when the table contains non-immediate
data, and in those cases fnstart is accurate.
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=40860
^ permalink raw reply [flat|nested] 46+ messages in thread* [Bug libgcj/40860] [4.4/4.5 regression] regressions in libjava testsuite on arm-linux
2009-07-26 10:51 [Bug libgcj/40860] New: [4.4/4.5 regression] regressions in libjava testsuite on arm-linux debian-gcc at lists dot debian dot org
` (32 preceding siblings ...)
2010-03-30 15:09 ` mikpe at it dot uu dot se
@ 2010-03-31 8:47 ` rearnsha at gcc dot gnu dot org
2010-03-31 21:44 ` mikpe at it dot uu dot se
` (10 subsequent siblings)
44 siblings, 0 replies; 46+ messages in thread
From: rearnsha at gcc dot gnu dot org @ 2010-03-31 8:47 UTC (permalink / raw)
To: java-prs
------- Comment #31 from rearnsha at gcc dot gnu dot org 2010-03-31 08:47 -------
(In reply to comment #30)
> (In reply to comment #29)
> > Wouldn't it be better to just remove _Unwind_GetRegionStart?
> > This function is not part of the ARM EABI, and removing it would expose any
> > (already broken) users at compile time.
>
> No.
>
> First it'd break most of gcc since the c, c++, and objc unwinders use it. And
> they generally use it to provide a base address when interpreting LSDA and
> computing landing pad addresses.
>
> Second, all _Unwind_GetRegionStart does is give r/o access to the fnstart value
> in ARM's UCB. But ARM's own unwinder uses fnstart in __gnu_unwind_pr_common,
> so if fnstart is broken then so it ARM's unwinder.
>
> ARM's unwinder is in the same boat as the c/c++/objc ones. It works because
> .ARM.exidx merging is limited to immediate table data, but the code using
> fnstart (by luck or design) only runs when the table contains non-immediate
> data, and in those cases fnstart is accurate.
>
There appears to be a mistaken presumption running through this thread that
there is a 1<->1 mapping between unwind blocks and source language functions.
This is not the case, and any code written with such a presumption is just
wrong code.
Just because such behaviour may work on other machines does not make the
presumption correct or the ARM unwinding code wrong.
1) Compilers may inline functions. If they do so, then unwind blocks will get
merged.
2) Compilers may create extra frames (though currently rarely). If they do,
then a function may have more than one frame.
3) They may do some combination of the above.
4) The ARM frame-unwinding annotations are designed to unwind C++ exceptions.
If they don't work outside that specification that does not make them wrong;
they simply weren't designed for the other (mis-)uses to which some people are
trying to put them.
Source language code is run on an abstract machine. When it's mapped onto real
hardware the reality can be very different. You can't rely on the two
corresponding directly beyond the semantics of the language.
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=40860
^ permalink raw reply [flat|nested] 46+ messages in thread* [Bug libgcj/40860] [4.4/4.5 regression] regressions in libjava testsuite on arm-linux
2009-07-26 10:51 [Bug libgcj/40860] New: [4.4/4.5 regression] regressions in libjava testsuite on arm-linux debian-gcc at lists dot debian dot org
` (33 preceding siblings ...)
2010-03-31 8:47 ` rearnsha at gcc dot gnu dot org
@ 2010-03-31 21:44 ` mikpe at it dot uu dot se
2010-04-08 12:14 ` [Bug libgcj/40860] [4.4/4.5/4.6 " mikpe at it dot uu dot se
` (9 subsequent siblings)
44 siblings, 0 replies; 46+ messages in thread
From: mikpe at it dot uu dot se @ 2010-03-31 21:44 UTC (permalink / raw)
To: java-prs
------- Comment #32 from mikpe at it dot uu dot se 2010-03-31 21:43 -------
(In reply to comment #31)
> There appears to be a mistaken presumption running through this thread that
> there is a 1<->1 mapping between unwind blocks and source language functions.
> This is not the case, and any code written with such a presumption is just
> wrong code.
Just to clarify: I'm not defending the parts of gcc that depend on the fact
that gcc's ARM EABI backend has implemented _Unwind_GetRegionStart. I'm just
trying to explain that those dependencies currently do exist.
> 4) The ARM frame-unwinding annotations are designed to unwind C++ exceptions.
> If they don't work outside that specification that does not make them wrong;
> they simply weren't designed for the other (mis-)uses to which some people are
> trying to put them.
The extreme logical conclusion is that gcc's ARM backend shouldn't even pretend
to implement _Unwind_GetRegionStart, as Paul suggested. Doing that in an
arm-linux-gnueabi cross-compiler causes compile-time breakage for most
languages including C++.
A less extreme conclusion is to make _Unwind_GetRegionStart always return 0.
That solves the compile-time dependencies but not the functional ones (however
technically invalid they are). Ignoring libjava's stacktrace module the
functional dependencies are all the same: LSDA blobs attached to exception
table entries are in DWARF format and contain deltas for landing pads relative
to the enclosing function entry. GCC assumes that _Unwind_GetRegionStart
works, so always outputs a dummy @LPStart in the LSDA
(output_function_exception_table in gcc/except.c); at runtime the PRs' LSDA
parsers check if @LPStart is encoded as a dummy or a proper value, if it's a
dummy they set it to the return value of _Unwind_GetRegionStart. Even the
libstdc++ unwinder does this and thus depends on _Unwind_GetRegionStart on ARM
EABI.
So to eliminate the functional dependency on _Unwind_GetRegionStart we have to
instruct output_function_exception_table to output a proper @LPStart with each
LSDA blob not a dummy one. I'll look into that next.
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=40860
^ permalink raw reply [flat|nested] 46+ messages in thread* [Bug libgcj/40860] [4.4/4.5/4.6 regression] regressions in libjava testsuite on arm-linux
2009-07-26 10:51 [Bug libgcj/40860] New: [4.4/4.5 regression] regressions in libjava testsuite on arm-linux debian-gcc at lists dot debian dot org
` (34 preceding siblings ...)
2010-03-31 21:44 ` mikpe at it dot uu dot se
@ 2010-04-08 12:14 ` mikpe at it dot uu dot se
2010-04-12 19:03 ` mikpe at it dot uu dot se
` (8 subsequent siblings)
44 siblings, 0 replies; 46+ messages in thread
From: mikpe at it dot uu dot se @ 2010-04-08 12:14 UTC (permalink / raw)
To: java-prs
------- Comment #33 from mikpe at it dot uu dot se 2010-04-08 12:14 -------
Created an attachment (id=20333)
--> (http://gcc.gnu.org/bugzilla/attachment.cgi?id=20333&action=view)
eliminate use of _Unwind_GetRegionStart on ARM, part 1
Here's a preliminary patch to eliminate the functional dependency on
_Unwind_GetRegionStart on ARM. There are two key changes:
output_function_exception_table outputs a proper @LPStart pointer in each LSDA
blob, and every PR's parse_lsda_header sets @Start to @LPstart if @Start is
NULL.
Without these updates making _Unwind_GetRegionStart return NULL on ARM causes
massive regressions in the c++/objc/obj-c++ test suites, and some in the c and
libffi test suites. With these updates those regressions are eliminated.
There are however two new regressions:
1. There are several orders of magnitude more alignment exceptions during the
test suite run, probably in the read_encoded_value calls that now trigger.
2. For initial testing I ran this on i686 with its _Unwind_GetRegionStart made
to return NULL too, and while it worked the g++.old-deja/g++.law/weak.C test
case failed with a linkage error. It looks like gcc emitted exception table
data for a discarded piece of code. On my ARM box the test suite considered
that test case UNSUPPORTED so never ran it.
The first regression can probably be fixed by avoiding to define unneeded
@LPStart pointers and making read_encoded_value mis-alignment safe.
The second regression might be fixable by making the @LPStart symbol references
weak.
More fixes will be needed for libjava's stacktrace mess though.
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=40860
^ permalink raw reply [flat|nested] 46+ messages in thread* [Bug libgcj/40860] [4.4/4.5/4.6 regression] regressions in libjava testsuite on arm-linux
2009-07-26 10:51 [Bug libgcj/40860] New: [4.4/4.5 regression] regressions in libjava testsuite on arm-linux debian-gcc at lists dot debian dot org
` (35 preceding siblings ...)
2010-04-08 12:14 ` [Bug libgcj/40860] [4.4/4.5/4.6 " mikpe at it dot uu dot se
@ 2010-04-12 19:03 ` mikpe at it dot uu dot se
2010-04-13 16:36 ` aph at gcc dot gnu dot org
` (7 subsequent siblings)
44 siblings, 0 replies; 46+ messages in thread
From: mikpe at it dot uu dot se @ 2010-04-12 19:03 UTC (permalink / raw)
To: java-prs
------- Comment #34 from mikpe at it dot uu dot se 2010-04-12 19:03 -------
Created an attachment (id=20371)
--> (http://gcc.gnu.org/bugzilla/attachment.cgi?id=20371&action=view)
eliminate use of _Unwind_GetRegionStart on ARM, take 2
The previous patch to eliminate _Unwind_GetRegionStart on ARM suffered from
major misalignment exceptions. The function pointers in the @LPStart fields
weren't resolved by the linker but apparently required runtime fixup by glibc.
Since they are always misaligned (the LSDA blob starts aligned with a format
byte followed by the @LPStart pointer), each @LPStart value resulted in two
alignment exceptions. Ouch.
The new version of the patch makes the @LPStart pointers PC-relative which
eliminates the runtime fixups and thus the alignment exceptions.
--
mikpe at it dot uu dot se changed:
What |Removed |Added
----------------------------------------------------------------------------
Attachment #20333|0 |1
is obsolete| |
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=40860
^ permalink raw reply [flat|nested] 46+ messages in thread* [Bug libgcj/40860] [4.4/4.5/4.6 regression] regressions in libjava testsuite on arm-linux
2009-07-26 10:51 [Bug libgcj/40860] New: [4.4/4.5 regression] regressions in libjava testsuite on arm-linux debian-gcc at lists dot debian dot org
` (36 preceding siblings ...)
2010-04-12 19:03 ` mikpe at it dot uu dot se
@ 2010-04-13 16:36 ` aph at gcc dot gnu dot org
2010-04-13 16:52 ` mikpe at it dot uu dot se
` (6 subsequent siblings)
44 siblings, 0 replies; 46+ messages in thread
From: aph at gcc dot gnu dot org @ 2010-04-13 16:36 UTC (permalink / raw)
To: java-prs
------- Comment #35 from aph at gcc dot gnu dot org 2010-04-13 16:36 -------
I've been trying this on gcc trunk, and it doesn't have the problem.
It seems that merging is not done.
I get
...
0x8684 <f2>: @0x8808
Compact model 1
0xb1 0x08 pop {r3}
0x84 0x00 pop {r14}
0xb0 finish
0xb0 finish
...
0x86b0 <f1>: @0x8820
Compact model 1
0xb1 0x08 pop {r3}
0x84 0x00 pop {r14}
0xb0 finish
0xb0 finish
Is it maybe the case that "Compact model 1" unwinder data isn't yet being
merged?
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=40860
^ permalink raw reply [flat|nested] 46+ messages in thread* [Bug libgcj/40860] [4.4/4.5/4.6 regression] regressions in libjava testsuite on arm-linux
2009-07-26 10:51 [Bug libgcj/40860] New: [4.4/4.5 regression] regressions in libjava testsuite on arm-linux debian-gcc at lists dot debian dot org
` (37 preceding siblings ...)
2010-04-13 16:36 ` aph at gcc dot gnu dot org
@ 2010-04-13 16:52 ` mikpe at it dot uu dot se
2010-04-13 17:02 ` aph at redhat dot com
` (5 subsequent siblings)
44 siblings, 0 replies; 46+ messages in thread
From: mikpe at it dot uu dot se @ 2010-04-13 16:52 UTC (permalink / raw)
To: java-prs
------- Comment #36 from mikpe at it dot uu dot se 2010-04-13 16:51 -------
(In reply to comment #35)
> I've been trying this on gcc trunk, and it doesn't have the problem.
> It seems that merging is not done.
>
> I get
>
> ...
> 0x8684 <f2>: @0x8808
> Compact model 1
> 0xb1 0x08 pop {r3}
> 0x84 0x00 pop {r14}
> 0xb0 finish
> 0xb0 finish
> ...
> 0x86b0 <f1>: @0x8820
> Compact model 1
> 0xb1 0x08 pop {r3}
> 0x84 0x00 pop {r14}
> 0xb0 finish
> 0xb0 finish
>
> Is it maybe the case that "Compact model 1" unwinder data isn't yet being
> merged?
They have to be adjacent to be merged. Looks like there's something else
between f2 and f1 in your object code. (You are using binutils >= 2.20 right?)
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=40860
^ permalink raw reply [flat|nested] 46+ messages in thread* [Bug libgcj/40860] [4.4/4.5/4.6 regression] regressions in libjava testsuite on arm-linux
2009-07-26 10:51 [Bug libgcj/40860] New: [4.4/4.5 regression] regressions in libjava testsuite on arm-linux debian-gcc at lists dot debian dot org
` (38 preceding siblings ...)
2010-04-13 16:52 ` mikpe at it dot uu dot se
@ 2010-04-13 17:02 ` aph at redhat dot com
2010-04-13 17:25 ` aph at gcc dot gnu dot org
` (4 subsequent siblings)
44 siblings, 0 replies; 46+ messages in thread
From: aph at redhat dot com @ 2010-04-13 17:02 UTC (permalink / raw)
To: java-prs
------- Comment #37 from aph at redhat dot com 2010-04-13 17:02 -------
Subject: Re: [4.4/4.5/4.6 regression] regressions in libjava
testsuite on arm-linux
On 04/13/2010 05:52 PM, mikpe at it dot uu dot se wrote:
>> Is it maybe the case that "Compact model 1" unwinder data isn't yet being
>> merged?
>
> They have to be adjacent to be merged. Looks like there's something else
> between f2 and f1 in your object code. (You are using binutils >= 2.20 right?)
Yes. OK, that explains it: gcc trunk outputs the functions in a completely
different order.
Andrew.
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=40860
^ permalink raw reply [flat|nested] 46+ messages in thread* [Bug libgcj/40860] [4.4/4.5/4.6 regression] regressions in libjava testsuite on arm-linux
2009-07-26 10:51 [Bug libgcj/40860] New: [4.4/4.5 regression] regressions in libjava testsuite on arm-linux debian-gcc at lists dot debian dot org
` (39 preceding siblings ...)
2010-04-13 17:02 ` aph at redhat dot com
@ 2010-04-13 17:25 ` aph at gcc dot gnu dot org
2010-04-21 16:34 ` aph at gcc dot gnu dot org
` (3 subsequent siblings)
44 siblings, 0 replies; 46+ messages in thread
From: aph at gcc dot gnu dot org @ 2010-04-13 17:25 UTC (permalink / raw)
To: java-prs
------- Comment #38 from aph at gcc dot gnu dot org 2010-04-13 17:25 -------
I think I can fairly easily add an option to the linker to suppress table
merging when linking Java libraries, and that will completely solve the
problem, at least from my point of view. To that end, it would not be at all
helpful to make _Unwind_GetRegionStart on ARM return NULL.
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=40860
^ permalink raw reply [flat|nested] 46+ messages in thread* [Bug libgcj/40860] [4.4/4.5/4.6 regression] regressions in libjava testsuite on arm-linux
2009-07-26 10:51 [Bug libgcj/40860] New: [4.4/4.5 regression] regressions in libjava testsuite on arm-linux debian-gcc at lists dot debian dot org
` (40 preceding siblings ...)
2010-04-13 17:25 ` aph at gcc dot gnu dot org
@ 2010-04-21 16:34 ` aph at gcc dot gnu dot org
2010-04-21 17:05 ` aph at gcc dot gnu dot org
` (2 subsequent siblings)
44 siblings, 0 replies; 46+ messages in thread
From: aph at gcc dot gnu dot org @ 2010-04-21 16:34 UTC (permalink / raw)
To: java-prs
------- Comment #39 from aph at gcc dot gnu dot org 2010-04-21 16:34 -------
Subject: Bug 40860
Author: aph
Date: Wed Apr 21 16:34:01 2010
New Revision: 158611
URL: http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=158611
Log:
2010-04-19 Andrew Haley <aph@redhat.com>
PR libgcj/40860
* configure.ac: Handle --no-merge-exidx-entries.
Modified:
trunk/libjava/ChangeLog
trunk/libjava/configure
trunk/libjava/configure.ac
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=40860
^ permalink raw reply [flat|nested] 46+ messages in thread* [Bug libgcj/40860] [4.4/4.5/4.6 regression] regressions in libjava testsuite on arm-linux
2009-07-26 10:51 [Bug libgcj/40860] New: [4.4/4.5 regression] regressions in libjava testsuite on arm-linux debian-gcc at lists dot debian dot org
` (41 preceding siblings ...)
2010-04-21 16:34 ` aph at gcc dot gnu dot org
@ 2010-04-21 17:05 ` aph at gcc dot gnu dot org
2010-04-22 16:07 ` aph at gcc dot gnu dot org
2010-04-22 16:08 ` aph at gcc dot gnu dot org
44 siblings, 0 replies; 46+ messages in thread
From: aph at gcc dot gnu dot org @ 2010-04-21 17:05 UTC (permalink / raw)
To: java-prs
------- Comment #40 from aph at gcc dot gnu dot org 2010-04-21 17:05 -------
Subject: Bug 40860
Author: aph
Date: Wed Apr 21 17:04:42 2010
New Revision: 158614
URL: http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=158614
Log:
2010-04-19 Andrew Haley <aph@redhat.com>
PR libgcj/40860
* configure.ac: Handle --no-merge-exidx-entries.
Modified:
branches/gcc-4_4-branch/libjava/ChangeLog
branches/gcc-4_4-branch/libjava/configure
branches/gcc-4_4-branch/libjava/configure.ac
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=40860
^ permalink raw reply [flat|nested] 46+ messages in thread* [Bug libgcj/40860] [4.4/4.5/4.6 regression] regressions in libjava testsuite on arm-linux
2009-07-26 10:51 [Bug libgcj/40860] New: [4.4/4.5 regression] regressions in libjava testsuite on arm-linux debian-gcc at lists dot debian dot org
` (42 preceding siblings ...)
2010-04-21 17:05 ` aph at gcc dot gnu dot org
@ 2010-04-22 16:07 ` aph at gcc dot gnu dot org
2010-04-22 16:08 ` aph at gcc dot gnu dot org
44 siblings, 0 replies; 46+ messages in thread
From: aph at gcc dot gnu dot org @ 2010-04-22 16:07 UTC (permalink / raw)
To: java-prs
------- Comment #41 from aph at gcc dot gnu dot org 2010-04-22 16:07 -------
Subject: Bug 40860
Author: aph
Date: Thu Apr 22 16:06:39 2010
New Revision: 158648
URL: http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=158648
Log:
2010-04-19 Andrew Haley <aph@redhat.com>
PR libgcj/40860
* configure.ac: Handle --no-merge-exidx-entries.
Modified:
branches/gcc-4_5-branch/libjava/ChangeLog
branches/gcc-4_5-branch/libjava/configure
branches/gcc-4_5-branch/libjava/configure.ac
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=40860
^ permalink raw reply [flat|nested] 46+ messages in thread* [Bug libgcj/40860] [4.4/4.5/4.6 regression] regressions in libjava testsuite on arm-linux
2009-07-26 10:51 [Bug libgcj/40860] New: [4.4/4.5 regression] regressions in libjava testsuite on arm-linux debian-gcc at lists dot debian dot org
` (43 preceding siblings ...)
2010-04-22 16:07 ` aph at gcc dot gnu dot org
@ 2010-04-22 16:08 ` aph at gcc dot gnu dot org
44 siblings, 0 replies; 46+ messages in thread
From: aph at gcc dot gnu dot org @ 2010-04-22 16:08 UTC (permalink / raw)
To: java-prs
------- Comment #42 from aph at gcc dot gnu dot org 2010-04-22 16:08 -------
Fixed.
--
aph at gcc dot gnu dot org changed:
What |Removed |Added
----------------------------------------------------------------------------
Status|UNCONFIRMED |RESOLVED
Resolution| |FIXED
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=40860
^ permalink raw reply [flat|nested] 46+ messages in thread