At 21:39 Uhr +0200 27.6.2001, Manuel Bouyer wrote:
>On Wed, Jun 27, 2001 at 07:58:12PM +0200, Hauke Fath wrote:

>> Seems the NetBSD esp driver has a problem with at least _some_ host adapter
>> hardware...
>
>On SCSI errors solaris downgrades to a lower bus frequency.
>Maybe we should do the same ...

Does it log the fact somewhere?

While downgrading bus speed after SCSI errors may be a good idea, I am not
sure it is happening here:

[hauke@pizza] ~ > w
 10:44pm  up 26 day(s),  1:11,  1 user,  load average: 0.00, 0.01, 0.01
User     tty           login@  idle   JCPU   PCPU  what
hauke    pts/0        10:43pm                      w

[Bonnie ss10 sunos 5.7 disks 1&2 Quantum Atlas III]

              -------Sequential Output-------- ---Sequential Input-- --Random--
              -Per Char- --Block--- -Rewrite-- -Per Char- --Block--- --Seeks---
Machine    MB K/sec %CPU K/sec %CPU K/sec %CPU K/sec %CPU K/sec %CPU  /sec %CPU
ss10_d1   500  1645 94.2  5658 83.9  2788 62.8  1659 95.4  8851 82.5 151.5 20.9
ss10_d2   500  1648 94.5  5690 84.5  2868 64.8  1663 95.6  9057 84.4 162.6 20.6

-- and no SCSI error messages in the syslog. 9 MB/sec pretty much maxes out
a fast/narrow SCSI bus.

	hauke

--
"It's never straight up and down"     (DEVO)