still SBUS qlogicpti no-go

Florian Lohoff (flo@rfc822.org)
Wed, 1 Sep 1999 00:11:49 +0200


Hi,
i still have the same problems with the qlogic Wide SCSI Controllers.

I meanwhile did some days of testing and i can really nail it down
to the Driver for the qlogic controller (qlogicpti.c) or the
Special Revision of Controllers i have (Tried 5 alltogether in
3 machines with a total of >30 SCSI Drives with at least 10 different
Wide SCSI Cables, and 6 different Sun Multipacks)

I have gotten back ~50 Qlogic SBUS Single Ended Wide SCSI Controllers
in ~50 Sparcserver 5 machines each equipped with a Sun Multipack with
4x4 GB Wide SCSI Drives. These machines have done Web Proxieng for
at least 3 Years so i suppose they worked as they are in front of me.

I have NO problems at all plugging a Multipack into my Symbios
Logic PCI Controller (Tekram) in my Linux/i386 box.

When i plug it into the SparcServer 5 (Same Cables, Multipack, Cable)
i am able to start rebuilding the Raid 5 but at >70% the machine starts
to have SCSI Errors for ALL devices on the wide channel.

Aug 31 22:59:49 spill kernel: scsi : aborting command due to timeout :
pid 240699, scsi3, channel 0, id 9, lun 0
Read (10) 00 00 38 b9 a0 00 00 80 00
Aug 31 22:59:49 spill kernel: qlogicpti : Aborting cmd for tgt[9] lun[0]
Aug 31 22:59:49 spill kernel: qlogicpti : scsi abort failure: 4005
Aug 31 22:59:49 spill kernel: SCSI host 3 abort (pid 240699) timed out -
resetting
Aug 31 22:59:49 spill kernel: SCSI bus is being reset for host 3 channel 0.
Aug 31 22:59:49 spill kernel: qlogicpti : Resetting SCSI bus!

The Qlogic controllers are ISP1000 based, Revision 01F, Build in 1993.

Is there anyone seeing similar problems ?
Is there a newer sbus qlogic driver anywhere ? (There is no diff
between 2.2.12 which i run and 2.3.15)

I am trying to do a Raid5 (0.90) on 6 Drives, block size 32K. No problem on
i386 with Symbios, but no-go on Sparc.

Flo

-- 
Florian Lohoff		flo@rfc822.org		      	+49-5241-470566
  ...  The failure can be random; however, when it does occur, it is
  catastrophic and is repeatable  ...             Cisco Field Notice

- 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/