[2959] in linux-scsi channel archive
Re: BusLogic problems...
daemon@ATHENA.MIT.EDU (Leonard N. Zubkoff)
Thu Dec 18 02:25:54 1997
Date: Wed, 17 Dec 1997 23:21:16 -0800
From: "Leonard N. Zubkoff" <lnz@dandelion.com>
To: tobotras@jet.msk.su
CC: linux-scsi@vger.rutgers.edu
In-reply-to: <lc0en3bb2tl.fsf@goliath.jet.msk.su> (message from Boris Tobotras
on 18 Dec 1997 09:08:54 +0300)
Date: 18 Dec 1997 09:08:54 +0300
From: Boris Tobotras <tobotras@jet.msk.su>
>>>>> "Leonard" == Leonard N Zubkoff writes:
Leonard> The problem is most likely due to the Quantum Stratus disk drives.
Leonard> Even with the latest firmware (see
Leonard> http://www.dandelion.com/Linux/Quantum.html), they still seem to
Leonard> be buggy. If you have the BusLogic driver compiled into your
Leonard> kernel, you can boot with the command line "BusLogic=0,7" to limit
Leonard> the queue depth to 7, which should make this problem either go
Leonard> away completely or become much less frequent. The next release of
Leonard> the BusLogic driver will support setting the queue depth on a
Leonard> per-target basis as well as allow passing options to the driver
Leonard> from insmod.
Leonard> Another possibility is that there have been a few reports of
Leonard> problems with the 5.07B BT-958 firmware. You might try using
Leonard> 5.06J if there are any remaining problems after limiting the queue
Leonard> depth.
Leonard> I'm beginning to think we should just blacklist the Quantum
Leonard> Tempest and Stratus drives.
What about Atlas? I had repeatable system lock with"queue full"
sympthom, with aha2940uw and Altas-II. I had to disable tagged queueing
complety to survive. That was at the times of 2.0.30. With .31, tagged
queuing seems not to cause the troubles, given the fact I didn't dare to
enable SCB paging. But all of that doesn't look promising for me.
Unfortunately, Quantum seems to be able to get their tagged queuing
implementation correct only after the drive is essentially obsolete. There
were many problems with the Atlas I L912 firmware and it took forever for them
to officially release L915, even though people were having serious problems.
The Atlas II LXY4 firmware has a bug where the drive can essentially become
catatonic after a Queue Full and won't resume processing outstanding commands
until an external event occurs to wake it up. I demonstrated this to Quantum
back in June but they still have not officially released a fix; however, the
LYK8 firmware that appears in the TESTCODE directory on their FTP server seems
to fix this problem. Use at your own risk, obviously.
The other problem is that the AIC7XXX driver did not handle Queue Full
gracefully, so even when the Atlas II didn't become catatonic the driver itself
would have problems. I believe Doug Ledford's workarounds for this made it
into 2.0.33.
Leonard
