2.5.27: JFS oops

Hi Dave,

here goes another jfsCommit oops from kernel 2.5.27.

Distribution: Slackware 8.1
Kernel : 2.5.27
JFS utils : 1.0.20

Filesystems were created with jfs-utils 1.0.18 / kernel 2.4.18 by slackware installation.

Unable to handle kernel NULL pointer dereference at virtual address 00000004
*pde = 00000000
Oops: 0002
CPU:    0
EIP:    0010:[<c0184da8>]       Not tainted

Using defaults from ksymoops -t elf32-i386 -a i386
EFLAGS: 00010282
eax: 00000000 ebx: c880a08c ecx: c7ea0800 edx: c7d5f668
esi: c7cee000 edi: 00000000 ebp: c880a080 esp: c7cee000
ds: 0018 es: 0018 ss: 0018

Stack: c7c7b86c 00000030 00001000 00000000 00000000 c7cee000 c7d5f65c
c7cee000
       c7cee000 00000000 c01924bb c7d5f65c c7d62b98 00000000 00038da8
00c2a128
       c7ea0800 c7d5f65c c6008000 00000030 00000000 00000001 00000000
00000030
Call Trace: [<c01924bb>] [<c019875f>] [<c0197fc0>] [<c01161e4>] [c0198c15>]
[<c0198ec4>] [<c0116350>] [<c0116350>] [<c0105000>] [<c0105000>]
[<c01057de>]
        [<c0198d10>]
Code: 89 50 04 8b 4c 24 44 89 5a 04 89 41 0c 8b 45 30 89 55 0c 89

Trace; c01924bb <release_metapage+19b/270>
Trace; c019875f <txFreeMap+2ef/490>
Trace; c0197fc0 <txUpdateMap+100/370>
Trace; c01161e4 <schedule+1d4/300>
Trace; c0198ec4 <jfs_lazycommit+1b4/300>
Trace; c0116350 <default_wake_function+0/40>
Trace; c0116350 <default_wake_function+0/40>
Trace; c0105000 <_stext+0/0>
Trace; c0105000 <_stext+0/0>
Trace; c01057de <kernel_thread+2e/40>
Trace; c0198d10 <jfs_lazycommit+0/300>

Code; c0184da8 <dbUpdatePMap+458/540>
00000000 <_EIP>:
Code; c0184da8 <dbUpdatePMap+458/540> <=====
0: 89 50 04 mov %edx,0x4(%eax) <=====
Code; c0184dab <dbUpdatePMap+45b/540>
3: 8b 4c 24 44 mov 0x44(%esp,1),%ecx
Code; c0184daf <dbUpdatePMap+45f/540>
7: 89 5a 04 mov %ebx,0x4(%edx)
Code; c0184db2 <dbUpdatePMap+462/540>
a: 89 41 0c mov %eax,0xc(%ecx)
Code; c0184db5 <dbUpdatePMap+465/540>
d: 8b 45 30 mov 0x30(%ebp),%eax
Code; c0184db8 <dbUpdatePMap+468/540>
10: 89 55 0c mov %edx,0xc(%ebp)
Code; c0184dbb <dbUpdatePMap+46b/540>
13: 89 00 mov %eax,(%eax)

Can I use jfs from cvs with current kernels (2.5.29/2.4.19-rc3-ac3) to see how latest changes work?

Best regards,
Axel Siebenwirth

I haven't seen this trap before. I'll take a closer look at it, and let you know what I find.

Yes. You should be able to just replace everything under fs/jfs with what's in cvs.

Thanks,
Shaggy

David Kleikamp
IBM Linux Technology Center

Hi,
I am working on a device driver for a device which supports MSI. I am not sure how to attach an interrupt service routine for it. I will appreciate any help in this regard.

Thanks,
Imran.

Are you saying jfs in 2.5.29 jfs is not a problem? I'm looking to see if I should put it on my problem status report.

2.5.29 had a problem that is fixed in Linus's bk tree. I had posted a patch to this list as well. (Simply remove the calls to d_delete from namei.c.)

Axel is seeing traps that I haven't seen elsewhere, so that's another potential problem. We don't know the cause of that one yet.

David Kleikamp
IBM Linux Technology Center

I've added jfs oops to my problem report status page. I hope it will be clear from the context which one is being discusssed.