Subject: problems with NCR PCI SCSI adaptor panics
To: None <current-users@NetBSD.ORG>
From: Charlie Root <email@example.com>
Date: 10/23/1995 12:18:05
I have been having a problem with my system for a long while now, but
it is intermittent and hard to track, so I have not said much. I also
hoped the latest fixes to the ncr.c code would fix the problem, but
I have an NCR PCI SCSI card with 2.0G Fast!SCSI-II drive attached. It
mostly runs fine, but on rare, seemingly random occasions (always on
weekends or after hours of course!), it panics and spews error messages
across the screen so fast that I can't read them properly.
However, it happened again this morning and I managed to get a little
more of the message. When I try a soft reboot, the machine hangs as it
ticks off its memory (you know the start where it goes 6097675+4765176+...
seems to be hanging with a permanent disk light as it gets to the start of
the second number - is this the swap space??)
The only hope I have is switching the machine off and on again, which fixes
the reboot problem.
I am running an Oct 18 build of the 1.1_ALPHA kernel with ncr support.
The messages I see when the panic occurs seem to be:
COMMAND FAILED (4.28)
assert 'cp' failed 'dev/pci/ncr.c' line 5494
over and over.
The last line points to an assertion in pci.c, which is inside a case statement in ncr_int_sir() and seems to be checking through the ncr's ccb list for the
device and missing.
Any ideas would be much appreciated, as this is most annoying for our users.