[Top] [All Lists]

damaged drive, kernel panics, yet again...

To: lord@xxxxxxx
Subject: damaged drive, kernel panics, yet again...
From: sooo lame <kszysiu@xxxxxxxxxxxx>
Date: Sun, 28 Jan 2001 03:43:25 +0100
Cc: linux-xfs@xxxxxxxxxxx
Sender: owner-linux-xfs@xxxxxxxxxxx
User-agent: Mutt/1.2.5i

i had no time till today to do something about that
damaged(?) drive... here are my results

i set a little PC - AMD K6-2 266 / i430TX (PIIX4E) 
with linux for diagnostic purposes


(that's what i changed)

i compiled CVS snapshot of xfs tree
   27.01.2001 21:52 CET
   page_buf.c 1.49

and i mounted damaged drive (Samsung SV4084D) onto /temp
then i executed 

$ find /temp | xargs -l1 ~/script > ~/list

script looks:
--- script ---
echo "$*"
echo \"$*\" | xargs -l1 cat > /dev/null 2>&1

and tail'ed -f ~/list on other virtual console...

after few files i got:
hdb: dma_intr: status=0x51  { DriveReady SeekComplete Error }
hdb: dma_intr: error=0x40 { UncorrectableError }, LBAsect=2046479, 
end_request: I/O error, dev 03:41 (hdb), sector 2046416
end_pg_buffer_io_async not uptodate 0 page 0xc0017ad8
kernel BUG at page_buf_io.c:923!
invalid operand: 0000
CPU:    0
EIP:    0010:[<c01636a8>]
EFLAGS: 00010096
eax: 00000021   ebx: c0017ad8   ecx: 00000001   edx: 00000001
esi: c0cf4bc0   edi: 00000008   ebp: 00000000   esp: c02bbea4
ds: 0018   es: 0018   ss: 0018
Process swapper (pid: 0, stackpage=c02bb000)
Stack: c0269f5b c026a0dd 0000039b c0cf4bc0 c03328c0 00000008 00000000 c01eba04
       c0cf4bc0 00000000 c03328c0 00000046 c031a2e0 00000040 00000008 c0202704
       c03328c0 00000000 c031124c c02b4c60 c0311164 00000008 c020948d 00000000
Call Trace: [<c01eba04>] [<c0205704>] [<c020948d>] [<c02062ab>] [<c020a204>] 
[<c020a928>] [<c0207027>]
       [<c020a160>] [<c010a13d>] [<c010a2a7>] [<c0107160>] [<c0108e70>] 
[<c0107160>] [<c0107183>] [<c01071da>]
       [<c0105000>] [<c0100191>]

Code: 0f 0b 83 c4 0c eb 19 90 66 81 7e 08 00 10 75 09 c7 43 44 ff
Kernel panic: Aiee, killing interrupt handler!
In interrupt handler - not syncing

i see the way i can save files on non-damaged area (by lookin' at the list
and 'smart' excluding files that didn't get on the list...)

the problems are file related (not inode-) 
(you asked about it)

will XFS ever be able to deal with such thing?

(or maybe it's not XFS fault at all)

best regards
- Krzysztof

yet again, sorry for my terrible english
(don't be suggested by .co.uk MX ... it's only MX... ;-)

<Prev in Thread] Current Thread [Next in Thread>
  • damaged drive, kernel panics, yet again..., sooo lame <=