lkml.org 
[lkml]   [1998]   [Feb]   [23]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
Date
From
SubjectRe: Hang in wait_on_inode with SMP 2.1.87
Hi Bill and developer list!

[Cc'ed to linux-kernel]

In article <34EEE5BA.83AF5241@star.net>,
whawes@star.net (Bill Hawes) writes:
> I've attached a patch with some debugging code that may help track things down.
> It enables a magic sysreq option to display a table of the inode state on
> alt-sysrq-i, so if you could get your system to hang again and then dispay the
> inode table, we may get some addditional clues.

Okay, as promised I provoked the bug with a large copy. Someone asked for a
shellscript to reproduce the bug, but it is not reproducable. Sometimes
(okay, really often, indeed) everything works okay. The offending command
was (Kernel Version 2.1.88):
python:/dos#cp -Rv . /mnt

/dos is a vfat mounted Win95, every other partition is ext2

/mnt was a loopback filesystem on /usr with about 140 MB. The loopback was
formatted as ext2, too.

python:/usr#df
Filesystem 1024-blocks Used Available Capacity Mounted on
/dev/sda1 40629 20885 17646 54% /
/dev/sdb5 504628 465722 13256 97% /usr
/dev/sdb6 330402 250517 76527 77% /usr/local
/dev/sdb1 199255 123722 65243 65% /home
/dev/sda3 154456 113296 41160 73% /dos
/dev/sda4 281040 164291 102465 62% /opt


Afer a while the system hung: here is the output from alt+sysreq+i on the
console (written by hand, I hope, everything is correct):

List Inodes Used Hash Link Lock Dirt Pipe Sock Odd? Page
UNUSED 1707 0 0 1698 0 0 0 0 0 0
CLEAN 2209 2209 2126 2209 1 11 11 72 0 7302
08:01 1 1 1 1 1 1 0 0 0 0
00:01 0 0 0 0 0 0 0 0 0 0
08:15 0 0 0 0 0 0 0 0 0 0
08:16 3 3 3 3 0 3 0 0 0 1
08:11 1 1 1 1 0 1 0 0 0 0
08:03 6 6 6 6 0 6 0 0 0 222
07:00 9 9 9 9 0 9 0 0 0 0
Inodes Acc.: 3936, Total pages: 7525

Every disc activity was dead this time, no writes, no reads. So I was not
able to run ps. After switching back to X, the keyboard was nearly dead (no
numlock, but sysreq worked), the mousecursor was the only sign of
"interrupt" working.

Perhaps it is a low-level problem with the aic7xxx driver? A lot of people
seem to have problems with occosional hangs and the Adaptec 2940 controller.

During writing the above output on paper, there was this error message:
EXT2fs-panic (08:11) ext2_read_inode: unable to read inode block - inode
38693 block = 1373770361

I think the block number doesn't make sense. Filesystem sdb1 is home with
200 MB.

Inode Number 38693 is
python:~>find . -inum 38693
../.netscape/bookmarks.html
python:~/.netscape>ll bookmarks.html
-rw------- 1 carsten knwu 6226 Feb 23 10:59 bookmarks.html
Netscape was running at the point of crash, the file permissions and
owner/group are correct.

The Superblock of /dev/sdb1 was corrupt after this one. I had to use a
backup block to correct the error.

I hope this helps.

Regards

Carsten


--
Carsten Gross Internet: carsten@sol.wohnheim.uni-ulm.de
Wohnheim Heilmeyersteige: Sebastian Kneipp Weg 6, 89075 Ulm

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu

\
 
 \ /
  Last update: 2005-03-22 13:41    [W:0.101 / U:0.148 seconds]
©2003-2014 Jasper Spaans. Advertise on this site