Before I do what you requested (it will take few hours to recompile
kernel), I'm sending you a new stack trace and a new information.
- The first trace I sent was from my laptop (uniprocesor).
- The second bug trace with debug=y was from my desktop (dual core smp).
(maybe that's the reason some function calls seem impossible?)
- This third trace I'm sending you is from the laptop again, but this
time it's with the AUFS_DEBUG enabled. (see attachnent).
Maybe now you can better find the problem?
Here is the new information I have:
- as you know, I'm doing all of this in Slax, a Live CD which unions the
filesystem from squashfs compressed images.
- I am not able to reproduce the problem if KDE is not running.
- I noticed I can add many branches without any problem, _unless_ I add
the same filesystem image for the second time (explanation below). Then
it is problematic sometimes (but not in all cases).
So, for example, I can add module1.lzm, module2.lzm, ... module100.lzm,
with no problem at all. (Adding is done by mount -o loop to
/mnt/live/memory(tmpfs)/images/module?.lzm/ and then this directory is
added to union by remount,add:1:/.../modules?.lzm/).
But then I do something maybe unexpected :)
I copy module1.lzm filesystem to aaa.lzm and I insert THE COPY (aaa.lzm)
to the union again. Sometimes it's OK but sometimes THIS fails.
(kbuildsycoca segfaults after the remount,add).
in detail:
- module1.lzm is already part of the union
- so first copy module1.lzm to /aaa.lzm (root is in aufs)
- mv /aaa.lzm /tmpfs_mounted_dir/aaa.lzm
- mount -o loop /tmpfs_mounted_dir/aaa.lzm /tmpfs2/some_dir/
- mount -o remount,add:1:/tmpfs2/some_dir/ aufs /
- start kbuildsycoca. It segfaults after a lstat64() function.
Sometimes I have to repeat the steps above several times before I get
the bug.
Maybe it has something to do with inode numbers? which are the same in
all squashfs filesystems which are inserted repeatedly??? Well I have no
idea :) I'm using default xino path, so it should use correct inode
numbering I guess.
Well, please review the log I'm sending this time, and let me know if
you found anything or if I should start the process of rebuilding the
kernel with debug.
Tomas M
[EMAIL PROTECTED] wrote:
Tomas M:
So I enabled debug and here is the result (attached)
:::
kernel BUG at /usr/src/a/aufs/fs/aufs/iinfo.c:64!
:::
Call Trace:
[<f8cec6be>] au_new_inode+0x45e/0x510 [aufs]
[<f8ced1f9>] aufs_permission+0x149/0x260 [aufs]
[<f8cecd1d>] aufs_lookup+0x7d/0xd0 [aufs]
[<c0165660>] do_lookup+0x130/0x170
Thnak you.
It is much easier to debug.
But this Call Trace is strange because of aufs_permission() between
aufs_lookup() and au_new_inode().
- aufs_lookup() never call aufs_permission()
- aufs_permission() never calls au_new_inode()
I guess that you didn't enable CONFIG_FRAME_POINTER, so I want you to
enable these configurations.
- DEBUG_KERNEL
- FRAME_POINTER
- DEBUG_STACKOVERFLOW
And please show me the output of this command (for the current aufs module)
The result may depend on your compiler.
$ objdump -d /your/current/aufs.ko | perl
/your/kernel/src/scripts/checkstack.pl i386
Junjiro Okajima
aufs 20070409
BUG: unable to handle kernel paging request at virtual address 774188d2
printing eip:
f03ee32e
*pde = 00000000
Oops: 0000 [#1]
SMP
Modules linked in: i915 drm snd_seq_dummy snd_seq_oss snd_seq_midi_event
snd_seq snd_seq_device snd_pcm_oss snd_mixer_oss capability commoncap fuse lp
parport_pc parport psmouse ohci1394 hci_usb ieee1394 bluetooth serio_raw
snd_intel8x0 snd_ac97_codec ac97_bus snd_pcm snd_timer b44 bcm43xx pcspkr
ieee80211softmac ieee80211 ieee80211_crypt snd mii iTCO_wdt iTCO_vendor_support
soundcore ata_generic i2c_i801 i2c_core snd_page_alloc intel_agp agpgart sg
evdev yenta_socket rsrc_nonstatic pcmcia pcmcia_core usb_storage uhci_hcd
ohci_hcd ehci_hcd nls_iso8859_1 nls_cp437 aufs squashfs sqlzma unlzma
CPU: 0
EIP: 0060:[<f03ee32e>] Not tainted VLI
EFLAGS: 00010202 (2.6.20 #1)
EIP is at au_h_iptr_i+0x5e/0x70 [aufs]
eax: 774188ae ebx: e284a068 ecx: ffff0001 edx: 0000001b
esi: 00000009 edi: 00000009 ebp: 00000009 esp: d3581d6c
ds: 007b es: 007b ss: 0068
Process kbuildsycoca (pid: 8883, ti=d3580000 task=ef16ea70 task.ti=d3580000)
Stack: e284a068 e28ab194 f03ea6be d3581dcc 00000001 d3581dc8 00000000 40000000
00000000 00000003 00000000 00000009 00000009 00000009 df576d14 00000354
00080000 00000001 00000009 0000000a 00000009 00000000 00000000 00000009
Call Trace:
[<f03ea6be>] au_new_inode+0x45e/0x510 [aufs]
[<f03ead1d>] aufs_lookup+0x7d/0xd0 [aufs]
[<c0165660>] do_lookup+0x130/0x170
[<c0167033>] __link_path_walk+0x313/0xe90
[<c0167bf5>] link_path_walk+0x45/0xc0
[<c0167e43>] do_path_lookup+0x83/0x1e0
[<c016882c>] __user_walk_fd+0x2c/0x50
[<c01617b6>] vfs_lstat_fd+0x16/0x40
[<c016182f>] sys_lstat64+0xf/0x30
[<c0102fe0>] syscall_call+0x7/0xb
[<c0580033>] packet_bind_spkt+0x13/0x60
=======================
Code: ff ff ff 66 39 c6 7e 0c 0f 0b eb fe 0f 0b eb fe 0f 0b eb fe 89 d8 e8 a2
fe ff ff 0f bf d6 8b 40 18 8d 14 52 8b 04 90 85 c0 74 07 <8b> 50 24 85 d2 7e 03
5b 5e c3 0f 0b eb fe 8d 74 26 00 53 89 c3
EIP: [<f03ee32e>] au_h_iptr_i+0x5e/0x70 [aufs] SS:ESP 0068:d3581d6c
-------------------------------------------------------------------------
Take Surveys. Earn Cash. Influence the Future of IT
Join SourceForge.net's Techsay panel and you'll get the chance to share your
opinions on IT & business topics through brief surveys-and earn cash
http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV
