fdomain + CD-ROM access -> kernel panic

Dean Martin Townsley (townsley@physics.ucsb.edu)
Sat, 28 Aug 1999 18:26:07 -0700


I am using
Linux version 2.2.12 (gcc version egcs-2.91.66 19990314/Linux (egcs-1.1.2 release))
on an RH 6.0 system (this has also occured with 2.2.10 and 2.2.11) with a
AHA2920 (fdomain) scsi adaptor, a VIA motherboard, AMD K6-3D, and an Acer
CRW-6206A cd-r/w drive.

I can panic the system by copying a few tens of Megs of files from the CD
drive (through ide-cd) to the scsi disk. The system panics after the
message:

kernel: scsi : aborting command due to timeout : pid 2 114, scsi0, channel 0, id 6, lun 0 Write (6) 13 6a 12 80 00

and the last screen contains the messages (typed in from the screen)

FIFO status = 0x08
Int. Condition = 0x00
Configuration 1 = 0x87
Configuration 2 = 0x03
scsi0 channel 0 : resetting for second half of retries.
SCSI bus is being reset for host 0 channel 0.
scsi: <fdomain> SCSI bus Reset
Future Domain 16-bit SCSI Driver Version 5.50
scsi0: <fdomain> No BIOS; using scsi id 7
scsi0: <fdomain> TMC-36C70 (PCI bus) chip at 0xec00 irq 11
unknown (34), target = 6 cmnd = 0x0a pieces = 64 size = 65536
sent_command = 0, have_data_in = 0, timeout = 0
in_interrupt_flag = 0
IMR = 0x02c0, IRR = 0x0080, ISR = 0x0000
SCSI Status = 0x11
TMC Status = 0x40 (interrupt)
Interrupt Status = 0x00
FIFO status = 0x08
Int. Condition = 0x00
Configuration 1 = 0x87
Configuration 2 = 0x03
Kernel panic: scsi: <fdomain> fdomain_16x0_queue() NOT REENTRANT!
In swapper task - not syncing

There was more before this, but it scrolled off and there's no scrollback
once the kernel panics. Also note these are from different crashes as I had
to get the system to log to the network to catch the syslog message hopefully
they are still consistent.

Miscellaneous info:

Both the SCSI and the CD work ok individually. The system boots and I can
use it normally, and I can use the CD normally as long as there is only a
very small amount of disk activity. I can burn a cd (through ide-scsi ->
scsi-cd) as long as the source is coming in over the network and I don't do
anything that generates significant disk activity. Notice I say significant,
I can copy maybe a few hundred k to a 1 Meg file across cd -> scsi drive
without killing the system.

I tried the fix proposed in the archives by Alan Cox in "fdomain.o w/ RH6.0"
changing in drivers/scsi/fdomain.h:
can_queue: 1, \
to
can_queue: 0, \
and it did not help.

I originally started using this drive through ide-scsi, with the same
problem, and have switched to ide-cd in hopes of isolating the bug more
closely. Also I do not have another CD-ROM drive to try in this computer so
I cannot eliminate the drive as the source of the problem. Also, copies
from an (old, i.e. slow by modern standards) 240MB IDE drive on the system to
the SCSI drive seem to work fine.

I hope this is a well-enough formatted report. Also I am not subscribed to
linux-kernel so a cc: would be appreciated.
Thanks to all,
-Dean Townsley
<townsley@physics.ucsb.edu>

-
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/