[linux-lvm] Lvremove crashes in kcopyd line 145

["Kottaridis, Chris" <chris kottaridis windriver com>]

I am running into a problem where I get an error when doing an lvremove after creating a snapshot. I am running 2.6.10:

=======================================================

kernel BUG at drivers/md/kcopyd.c:145!
invalid operand: 0000 [#1]
SMP
Modules linked in: ipmi_watchdog vmc vMCstub ipmibus ipmi_serial ipmi_devintf
ipmi_msghandler pmemfs
CPU:    2
EIP:    0060:[<c038af34>]    Not tainted VLI
EFLAGS: 00010283   (2.6.10-pne)
EIP is at client_free_pages+0x38/0x42
eax: 00000100   ebx: f6afcd00   ecx: 00000000   edx: 00000000
esi: f88b2080   edi: 0000002c   ebp: f6829ee8   esp: f6829ee0
ds: 007b   es: 007b   ss: 0068
Process lvremove (pid: 8622, threadinfo=f6828000 task=f70e4c70)
Stack: f6afcd00 f6afcd00 f6829ef8 c038b979 f6afcd00 f7014180 f6829f10 c038c339
       f6afcd00 f6f42ca0 f88b2080 f6f7ab80 f6829f2c c0386099 f88b2080 00000000
       f6a4a480 f8866000 00000004 f6829f40 c0387f3b f6f7ab80 00000000 c052b620
       f6829f54 c038874f f6a4a480 f6828000 00000000 f6829f7c c0389cc7 f8866000
       00004000 f6829fa0 c03886ec f8866000 c134fd04 f6d7ea80 c0389bee f6829fbc
       c017cf94 f75f2c6c f6d7ea80 c134fd04 080de8d0 ffffffe7 f70e4c70 f70e4de8
Call Trace:
 [<c01034fb>] show_stack+0x80/0x96
 [<c010368e>] show_registers+0x15d/0x1d6
 [<c01039ec>] die+0x106/0x194
 [<c010400f>] do_trap+0xdc/0x13e
 [<c01042ed>] do_invalid_op+0xb5/0xbf
 [<c01031a3>] error_code+0x2b/0x30
 [<c038b979>] kcopyd_client_destroy+0x1d/0x37
 [<c038c339>] snapshot_dtr+0x70/0x7e
 [<c0386099>] table_destroy+0x62/0xcc
 [<c0387f3b>] __hash_remove+0x5b/0x84
 [<c038874f>] dev_remove+0x63/0xc5
 [<c0389cc7>] ctl_ioctl+0xd9/0x137
 [<c017cf94>] sys_ioctl+0x207/0x267
 [<c010263c>] no_dpa_vsyscall_enter+0x8/0x1b
Code: 39 43 14 75 26 8b 43 0c 89 04 24 e8 61 ff ff ff c7 43 0c 00 00 00 00 c7
43 10 00 00 00 00 c7 43 14 00 00 00 00 83 c4 04 5b 5d c3 <0f> 0b 91 00 2f f1
49 c0 eb d0 55 89 e5 83 ec 18 c7 44 24 14 00
 <6>Kcore timestamp : 1165881319.900042
Kcore HighResolution timestamp : B43B8A9DA6
Oops: 0000 [#2]
SMP
=========================================

 

It seems to be running into the BUG_ON in this code:

 

static void client_free_pages(struct kcopyd_client *mc)

{

      BUG_ON(kc->nr_free_pages != kc->nr_pages);

      drop_pages(kc->pages);

      kc->pages = NULL;

      kc->nr_free_pages = kc->nr_pages = 0;

}

 

 

Interestingly enough if I put like a 10 second sleep after the original snapshot creation then it seems OK, but if I sleep 5 seconds or less then it breaks as above.

 

Any ideas on what is going on would be appreciated.

 

Thanks

 

Chris Kottaridis

Senior Engineer

Wind River Systems

719-522-9786