>On Mon, 21 Dec 1998, Andreas Priebe wrote:
>
>> I know this is a real old story, which I hoped is resolved now -
>> especially after reading the relaese notes of current 1.3.x releases.
>> So I went along and built a kernel from current as of 19.12.
>> 
>> But alas...
>> 
>> Already during boot it gave me lots of the wellknown
>> 
>>   asc_get_status cmdreg 11 fifo count 6 (or 7 or 3)
>> or
>>   rz0 Illegal request
>>   rz1 Illegal request
>> or
>>   SENSE
>> 
>> and dumped so much core files that I changed back imediately to my
>> (relatively) stable kernel from -current as of March 98.
>> 
>> Any ideas what I am doing wrong ... ???
>
>  These would appear to be some of the problems documented in PR#5510,
>which is partially fixed in -current and 1.3.3.  I was able to duplicate
>some of the problems that Greg Wohletz reported in that PR, and fix or
>work around them, but Greg was still having problems that I was not
>able to duplicate on my system and was not able to come up with anything
>for Greg to try that would help me understand where the problem was.  My
>guess is that it depends upong the SCSI devices attached to the system,
>and the devices I have don't exhibit that problem.
>
>  Until I can get a device configuration that has the same problem, this
>is probably not likely to get fixed soon.  [Also, when the pmax finally
>gets switched to the MI ncr53c9x driver, the problem may disappear.]
>
>Michael
>

I've been able to eliminate the probalem on my 200s by using an old version
of the driver (1.2F i believe).  I can provide the replacement code to
anyone who is interested.

Using this driver I have one 5000/200 acting as a squid proxy for several
thousand users, and another acting as an NNTP server, both are performing
well under this rather heavy load.

						--Greg