cp deadlocks on large copy from SCSI zip disk under SMP (2.2.9)

Jason A. Pfeil (pfeil@nu.cs.fsu.edu)
Tue, 15 Jun 1999 21:46:38 -0400 (EDT)


----
Jason A. Pfeil pfeil@cs.fsu.edu
Network Administrator jpfeil@lsi.fsu.edu
CASDL (850)644-8014; fax: (850)644-4952
Learning Systems Institute University Center C-3519
http://idl.fsu.edu Tallahassee, FL 32306-2540

---------- Forwarded message ----------
Date: Mon, 14 Jun 1999 01:24:56 -0400 (EDT)
From: Jason A. Pfeil <pfeil@cs.fsu.edu>
To: linux-kernel@vger.rutgers.edu
Cc: Jason A. Pfeil <pfeil@nu.cs.fsu.edu>
Subject: cp deadlocks on large copy from SCSI zip disk under SMP (2.2.9)

I have been trying to copy a large file ~49MB from a zip disk and each
time, the file enters an uninterruptible sleep on the disk (/dev/sda4).
The source filesystem is vfat and the destination filesystem is ext2. It
copies fine under a nonSMP kernel. The two kernels are exactly the same
except that one is SMP and the other is not.

Here is some information that I have been able to gather:

Entries in /var/log/messages
============================
Jun 14 01:10:11 brunell kernel: end_scsi_request: buffer-list destroyed
Jun 14 01:10:14 brunell last message repeated 3 times

cp --version
============
[root@brunell 2809]# cp --version
cp (GNU fileutils) 4.0

rpm -qf /bin/cp
===============
[root@brunell 2809]# rpm -qf /bin/cp
fileutils-4.0-1

RedHat 6.0 is the Distro.

uname -a
========
[root@brunell 2809]# uname -a
Linux brunell.jaguars.jap.org 2.2.9 #2 SMP Wed Jun 2 21:29:05 EDT 1999
i686 unknown

/proc filesystem stuff from the PID's entry
===========================================
[root@brunell 2809]# cat status
Name: cp
State: D (disk sleep)
Pid: 2809
PPid: 2786
Uid: 0 0 0 0
Gid: 0 0 0 0
Groups: 0 1 2 3 4 6 10
VmSize: 1072 kB
VmLck: 0 kB
VmRSS: 340 kB
VmData: 16 kB
VmStk: 8 kB
VmExe: 28 kB
VmLib: 992 kB
SigPnd: 0000000000000107
SigBlk: 0000000000000000
SigIgn: 0000000000000000
SigCgt: 0000000000000000
CapInh: 00000000fffffeff
CapPrm: 00000000fffffeff
CapEff: 00000000fffffeff

[root@brunell 2809]# cat cpu
cpu 1 21
cpu0 0 10
cpu1 1 11

[root@brunell 2809]# cat cwd
cat: cwd: Is a directory
[root@brunell 2809]# ls -ld cwd
lrwx------ 1 root root 0 Jun 14 01:20 cwd -> /root

[root@brunell 2809]# cat fd
cat: fd: Invalid argument
[root@brunell 2809]# ls -l fd
total 0
lrwx------ 1 root root 64 Jun 14 01:20 0 -> /dev/pts/2
lrwx------ 1 root root 64 Jun 14 01:20 1 -> /dev/pts/2
lrwx------ 1 root root 64 Jun 14 01:20 2 -> /dev/pts/2
lr-x------ 1 root root 64 Jun 14 01:20 3 ->/mnt/zip/R4_5-x86.zip
l-wx------ 1 root root 64 Jun 14 01:20 4 ->/tmp/R4_5-x86.zip

[root@brunell 2809]# cat maps
08048000-0804f000 r-xp 00000000 03:41 26528 /bin/cp
0804f000-08050000 rw-p 00006000 03:41 26528 /bin/cp
40000000-40012000 r-xp 00000000 03:41 32642 /lib/ld-2.1.1.so
40012000-40013000 rw-p 00011000 03:41 32642 /lib/ld-2.1.1.so
40013000-40014000 rwxp 00000000 00:00 0
40019000-400ff000 r-xp 00000000 03:41 32649 /lib/libc-2.1.1.so
400ff000-40104000 rw-p 000e5000 03:41 32649 /lib/libc-2.1.1.so
40104000-40107000 rw-p 00000000 00:00 0
bfffe000-c0000000 rwxp fffff000 00:00 0

[root@brunell 2809]# cat stat
2809 (cp) D 2786 2809 2781 0 -1 0 13 0 1670 0 1 21 0 0 0 0 0 0 11201450
1097728 85 2147483647 134512640 134537664 3221224784 3221219404 1074509108
263 0 0 0 3222405809 0 0 17 0

[root@brunell 2809]# cat statm
85 85 72 7 0 78 13

I can provide any more files/information that you need. The driver being
used is the aha152x driver with the following append line in
/etc/lilo.conf: append="aha152x=0x140,11,7,1"

Thank you for your help.

----
Jason A. Pfeil pfeil@cs.fsu.edu
Network Administrator jpfeil@lsi.fsu.edu
CASDL (850)644-8014; fax: (850)644-4952
Learning Systems Institute University Center C-3519
http://idl.fsu.edu Tallahassee, FL 32306-2540

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/