On Tue, Jan 22, 2002 at 11:53:25PM +0000, Mike Pumford wrote:
> > Hi, 
> > after doing a cvs update yesterday, I built a new kernel and tried
> > what I could get to run.
> > Unfortunately the machine dropped into the debugger immediately
> > as soon as I tried to mount my CDROM like this:
> > 
> > bash-2.05#mount /dev/cd0a /mnt
> > panic: scsipi_done: ASYNC and POLL
> > Stopped in pid 192 (mount) at _cpu_Debugger+0x10: ldmdb r11, {r11,
> > r13, r15}
> > db>
> > 
> This appears to just be some diagnostic code in the MI SCSI layer which 
> doesn't expect a polled SCSI driver. You could try removing the code in 
> scsipi_base.c and see if it will survive anyway. Compiling without 
> DIAGNOSTIC will also remove the check.
> 
> Looks like the only real solution here are some interrupt driven SCSI drivers 
> which will be impossible without actually having a card to develop against.

I stand corrected, you were right. Compiling without DIAGNOSTIC really
helps, must have overseen it last time.

So, the status of Cumana SCSI II is:
It works, even if it seems quite slow, but this has to be tested.
I guess this is a result of the polling mode.

I was able to disklabel a harddisc, newfs and now try to copy my
complete /usr partition onto it (which creates no problems 'til now.)
Doing a cvs update will be a nice test.. ;-)

-- 
Greetings,	
	Jan-Uwe
-------------------------------------------------
PGP-Public-Key on Request.