Bug 80142

Summary: ide-scsi driver causes kernel panic on rewind after tape fills
Product: [Retired] Red Hat Linux Reporter: Kern Sibbald <kern>
Component: kernelAssignee: Pete Zaitcev <zaitcev>
Status: CLOSED WORKSFORME QA Contact: Brian Brock <bbrock>
Severity: high Docs Contact:
Priority: medium    
Version: 4.2   
Target Milestone: ---   
Target Release: ---   
Hardware: All   
OS: Linux   
Fixed In Version: Doc Type: Bug Fix
Doc Text:
Story Points: ---
Clone Of: Environment:
Last Closed: 2003-08-05 04:32:21 UTC Type: ---
Regression: --- Mount Type: ---
Documentation: --- CRM:
Verified Versions: Category: ---
oVirt Team: --- RHEL 7.3 requirements from Atomic Host:
Cloudforms Team: --- Target Upstream Version:

Description Kern Sibbald 2002-12-20 17:01:18 UTC
From Bugzilla Helper:
User-Agent: Mozilla/5.0 Galeon/1.2.5 (X11; Linux i686; U;) Gecko/20020606

Description of problem:
Writing 64K fixed size tape blocks to a Seagate  Model: STT20000A  
tape drive using the ide-scsi driver causes a kernel
panic when the end of tape is reached and a rewind request
is made.

Version-Release number of selected component (if applicable):

How reproducible:

Steps to Reproduce:
1.Run Bacula btape program
2.Enter the "fill" command

Actual Results:  The program reports its progress periodically as follows:
Simple test (single tape) selected.
btape: Wrote label to prelabeled Volume  on device /dev/nst0
btape: btape.c:1121 Wrote block=5000, VolBytes=322,495,678 rate=820.6 KB/s
btape: btape.c:1126 Flush block, write EOF
btape: Error in block.c:229 block.c:228 Expected block-id BB01 or BB02, got ´O½). Bu
btape: btape Error: Re-read last block at EOT failed. ERR=block.c:228 Expected block
btape: btape.c:1377 block.c:228 Expected block-id BB01 or BB02, got ´O½). Buffer dis
btape: btape.c:1379 Block not written: FileIndex=459800 Block=154975 Size=64512
btape: block.c:81 Dump block Block not written 808e9d0: size=64512 BlkNum=154975
               Hdrcksum=71aa652d cksum=71aa652d
btape: block.c:94    Rec: VId=1 VT=1039883876 FI=459798 Strm=contDATA len=26612 p=80
btape: block.c:94    Rec: VId=1 VT=1039883876 FI=459799 Strm=DATA len=32768 p=80969b
btape: btape.c:1172 Done filling tape. Now beginning re-read of tape ...

The kernel panics and rebooting is necessary.

Expected Results:  As above, except that the tape should rewind and
the program should read back what it wrote. 

Additional info:

At the point of the system crash, the program is using
system read(), write() requests and a few simple ioctl() requests
(e.g. rewind). There is absolutely nothing fancy, no
special modes ...

Comment 1 Pete Zaitcev 2003-08-01 18:13:23 UTC
If the system stays up, dmesg should be captured.
Otherwise, a serial console or netdump must be established.

Also, the release field is obviously wrong in the report (4.2).
I need to see /etc/redhat_release and /proc/versions.

Comment 2 Kern Sibbald 2003-08-03 21:09:24 UTC
Unfortunately, I no longer have access to the system that has the
IDE-SCSI drive.