* Random Oopses on 2.6.0-test1-mm2
@ 2003-07-26 7:18 Apurva Mehta
2003-07-26 8:54 ` Felipe Alfaro Solana
0 siblings, 1 reply; 3+ messages in thread
From: Apurva Mehta @ 2003-07-26 7:18 UTC (permalink / raw)
To: Linux Kernel Mailing List
[-- Attachment #1: Type: text/plain, Size: 982 bytes --]
Hi,
I get random oopses while using 2.6.0-test1-mm2. I have attached the
oops message and also the ksymoops output. I get the same behaviour
with the 08int patch applied also. The OOPs message attached was
obtained on a 2.6.0-test1-mm2-08int.
Basically, the system just freezes while performing random
tasks. Sometimes it happes just after loading xmms and starting to
play a song. Sometimes I can do that and open my browser and surf a
bit before it freezes. However, every session ends in a
freeze. AFAICS, there is no discernible pattern of usage that leads to
a freeze. It may occur while processing the startup scripts or while
starting X or at any moment during an X session.
Once the system is frozen, I can still move the mouse or switch to a
virtual console. But I cannot run any programs. If the freeze occurs
in X, I do not see the oops message. So the oops message I have
attached happened just after I typed `startx` at the console after a
fresh reboot.
- Apurva
[-- Attachment #2: oops.txt --]
[-- Type: text/plain, Size: 1298 bytes --]
Unable to handle kernel NULL pointer dereference at virtual address
00000014
print eip:
*pde=00000000
Oops: 0000 [#1]
PREEMPT
CPU : 0
EIP : 0060 " [<c0197ad8>] Not tainted VLI
EFLAGS : 00010202
EIP is at journal_dirty_metadata+0x38/0x210
eax : c7aac000 ebx : 00000000 ecx : 00000000 edx : cb4c0e20
esi : cadfd180 edi : cbd37bc0 ebp : cb070220 esp : c7aade78
ds : 007b es : 007b ss : 0068
Process X (pid: 1322, threadinfo = c7aac000 task = c7b34d60 )
Stack:
cbd4c660 c7abd148 c0152c6b c015127f cb4c0e20 cb4c0e20 cb070220 00000001
c7abd148 c01896c7 cb070220 cb4c0e20 00001000 c7aadebc 00000001 00000001
00000000 00000030 00000001 00000001 000006c6 00075081 cbd4a3a0 00004000
Call Trace:
[<c0152c6b>] __getblk+0x2b/0x60
[<c015127f>] wake_up_buffer+0xf/0x30
[<c01896c7>] ext3_getblk+0x107/0x2a0
[<c0189893>] ext3_bread+0x33/0xc0
[<c018f707>] ext3_mkdir+0x107/0x2e0
[<c018f600>] ext3_mkdir+0x0/0x2e0
[<c01601aa>] vfs_mkdir+0x6a/0xc0
[<c01602c6>] sys_mkdir+0xc6/0x100
[<c01292bc>] sys_newuname+0x86/0x90
[<c0109359>] sysenter_past_esp+0x52/0x71
Code: 24 2c 8b 75 00 8b 5a 24 f6 45 18 04 8b 3e 0f 84 df 01 00 00 b8 01 00 00 00
85 c0 0f 85 d1 00 00 00 b8 00 e0 ff ff 21 e0 ff 40 14 <8b> 43 14 3b 45 00 0f 84
6c 01 00 b8 00 e0 ff ff 21 e0 ff 40
<6>note: x[1322] exited with preempt_count 1
giving up
[-- Attachment #3: ksymoops.output --]
[-- Type: text/plain, Size: 2367 bytes --]
ksymoops 2.4.9 on i686 2.6.0-test1-mm2. Options used
-v /plan9/linux_kernel_builds/linux-2.6.0-test1/vmlinux (specified)
-k /proc/kallsyms (specified)
-l /sbin/lsmod (specified)
-o /lib/modules/2.6.0-test1-mm2/ (default)
-m /boot/System.map-2.6.0-test1-mm2 (specified)
Warning (read_ksyms): no kernel symbols in ksyms, is /proc/kallsyms a valid ksyms file?
No modules in ksyms, skipping objects
No ksyms, skipping lsmod
Unable to handle kernel NULL pointer dereference at virtual address
00000014
Oops: 0000 [#1]
Stack:
cbd4c660 c7abd148 c0152c6b c015127f cb4c0e20 cb4c0e20 cb070220 00000001
c7abd148 c01896c7 cb070220 cb4c0e20 00001000 c7aadebc 00000001 00000001
00000000 00000030 00000001 00000001 000006c6 00075081 cbd4a3a0 00004000
Call Trace:
[<c0152c6b>] __getblk+0x2b/0x60
[<c015127f>] wake_up_buffer+0xf/0x30
[<c01896c7>] ext3_getblk+0x107/0x2a0
[<c0189893>] ext3_bread+0x33/0xc0
[<c018f707>] ext3_mkdir+0x107/0x2e0
[<c018f600>] ext3_mkdir+0x0/0x2e0
[<c01601aa>] vfs_mkdir+0x6a/0xc0
[<c01602c6>] sys_mkdir+0xc6/0x100
[<c01292bc>] sys_newuname+0x86/0x90
[<c0109359>] sysenter_past_esp+0x52/0x71
Code: 24 2c 8b 75 00 8b 5a 24 f6 45 18 04 8b 3e 0f 84 df 01 00 00 b8 01 00 00 00
Using defaults from ksymoops -t elf32-i386 -a i386
Trace; c0152c6b <__getblk+2b/60>
Trace; c015127f <wake_up_buffer+f/30>
Trace; c01896c7 <ext3_getblk+107/2a0>
Trace; c0189893 <ext3_bread+33/c0>
Trace; c018f707 <ext3_mkdir+107/2e0>
Trace; c018f600 <ext3_mkdir+0/2e0>
Trace; c01601aa <vfs_mkdir+6a/c0>
Trace; c01602c6 <sys_mkdir+c6/100>
Trace; c01292bc <sys_newuname+8c/90>
Trace; c0109359 <sysenter_past_esp+52/71>
Code; 00000000 Before first symbol
00000000 <_EIP>:
Code; 00000000 Before first symbol
0: 24 2c and $0x2c,%al
Code; 00000002 Before first symbol
2: 8b 75 00 mov 0x0(%ebp),%esi
Code; 00000005 Before first symbol
5: 8b 5a 24 mov 0x24(%edx),%ebx
Code; 00000008 Before first symbol
8: f6 45 18 04 testb $0x4,0x18(%ebp)
Code; 0000000c Before first symbol
c: 8b 3e mov (%esi),%edi
Code; 0000000e Before first symbol
e: 0f 84 df 01 00 00 je 1f3 <_EIP+0x1f3>
Code; 00000014 Before first symbol
14: b8 01 00 00 00 mov $0x1,%eax
1 warning issued. Results may not be reliable.
^ permalink raw reply [flat|nested] 3+ messages in thread
* Re: Random Oopses on 2.6.0-test1-mm2
2003-07-26 7:18 Random Oopses on 2.6.0-test1-mm2 Apurva Mehta
@ 2003-07-26 8:54 ` Felipe Alfaro Solana
2003-07-26 15:38 ` Apurva Mehta
0 siblings, 1 reply; 3+ messages in thread
From: Felipe Alfaro Solana @ 2003-07-26 8:54 UTC (permalink / raw)
To: Apurva Mehta; +Cc: Linux Kernel Mailing List
[-- Attachment #1: Type: text/plain, Size: 342 bytes --]
On Sat, 2003-07-26 at 09:18, Apurva Mehta wrote:
> Hi,
>
> I get random oopses while using 2.6.0-test1-mm2. I have attached the
> oops message and also the ksymoops output. I get the same behaviour
> with the 08int patch applied also. The OOPs message attached was
> obtained on a 2.6.0-test1-mm2-08int.
Please, try the following patch...
[-- Attachment #2: ext3-inode.patch --]
[-- Type: text/x-patch, Size: 1115 bytes --]
diff -puN fs/ext3/inode.c~ext3_getblk-race-fix-fix fs/ext3/inode.c
--- 25/fs/ext3/inode.c~ext3_getblk-race-fix-fix 2003-07-19 22:59:50.000000000 -0700
+++ 25-akpm/fs/ext3/inode.c 2003-07-19 23:07:42.000000000 -0700
@@ -936,19 +936,17 @@ struct buffer_head *ext3_getblk(handle_t
ext3_get_block instead, so it's not a
problem. */
lock_buffer(bh);
- if (!buffer_uptodate(bh)) {
- BUFFER_TRACE(bh, "call get_create_access");
- fatal = ext3_journal_get_create_access(handle, bh);
- if (!fatal) {
- memset(bh->b_data, 0,
- inode->i_sb->s_blocksize);
- set_buffer_uptodate(bh);
- }
+ BUFFER_TRACE(bh, "call get_create_access");
+ fatal = ext3_journal_get_create_access(handle, bh);
+ if (!fatal && !buffer_uptodate(bh)) {
+ memset(bh->b_data, 0, inode->i_sb->s_blocksize);
+ set_buffer_uptodate(bh);
}
unlock_buffer(bh);
BUFFER_TRACE(bh, "call ext3_journal_dirty_metadata");
err = ext3_journal_dirty_metadata(handle, bh);
- if (!fatal) fatal = err;
+ if (!fatal)
+ fatal = err;
} else {
BUFFER_TRACE(bh, "not a new buffer");
}
_
^ permalink raw reply [flat|nested] 3+ messages in thread
* Re: Random Oopses on 2.6.0-test1-mm2
2003-07-26 8:54 ` Felipe Alfaro Solana
@ 2003-07-26 15:38 ` Apurva Mehta
0 siblings, 0 replies; 3+ messages in thread
From: Apurva Mehta @ 2003-07-26 15:38 UTC (permalink / raw)
To: Linux Kernel Mailing List
* Felipe Alfaro Solana <felipe_alfaro@linuxmail.org> [26-07-2003 14:44]:
> On Sat, 2003-07-26 at 09:18, Apurva Mehta wrote:
> > Hi,
> >
> > I get random oopses while using 2.6.0-test1-mm2. I have attached the
> > oops message and also the ksymoops output. I get the same behaviour
> > with the 08int patch applied also. The OOPs message attached was
> > obtained on a 2.6.0-test1-mm2-08int.
>
> Please, try the following patch...
Yes, that seems to have fixed the problem. So far, I see a definite
improvement over the vanilla 2.6.0-test1.
But Ingo's patch made the vanilla 2.6.0-test1 awesome under really
heavy load. I basically was compiling 2.6.0-test1-mm2-O8 from scratch
and the system was (almost) as responsive as if there was no load, with
no audio skipping. That was the first real improvement I have seen
over the vanilla 2.4.21... Will keep you'll posted at how
2.6.0-test1-mm2-O8int compares..
- Apurva
^ permalink raw reply [flat|nested] 3+ messages in thread
end of thread, other threads:[~2003-07-26 15:28 UTC | newest]
Thread overview: 3+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2003-07-26 7:18 Random Oopses on 2.6.0-test1-mm2 Apurva Mehta
2003-07-26 8:54 ` Felipe Alfaro Solana
2003-07-26 15:38 ` Apurva Mehta
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).