Subject: Re: possible bug in sd.c
To: None <thorpej@nas.nasa.gov>
From: Mike Hibler <mike@cs.utah.edu>
List: port-hp300
Date: 03/22/1995 12:23:03
> To: port-hp300@NetBSD.ORG
> Subject: possible bug in sd.c
> From: Jason Thorpe <thorpej@nas.nasa.gov>
> Date: Wed, 22 Mar 1995 10:11:50 -0800
> 
> Folks --
> 
> I've been having some trouble adding a Quantum 730S to my hp380 which I 
> believe to be a bug in sd.c.  Basically, the disk refuses to 
> autoconfigure upon reboot *unless* I power-cycle the machine and hit 
> <space> during the memory test to force the roms to probe for 
> bootblocks.  The roms frob the disk (the light blinks) as it works down 
> the chain (id6 -> id0) finds the bootblock on id0 and loads it, etc.
> 
A couple of possibilities:

The drive isn't spun up.  Some drives don't automatically spin up when you
power them on (though there is usually a jumper to enable this).  There is
a "start unit" SCSI command to do this.  This isn't done in the 300 driver.

Also, a lot of drives won't do anything until you clear any pending
"unit conditions" (by requesting the sense data).  This may be done in the
300 driver.

Is it a command or data transfer that is hanging/failing?  I suppose the
drive could be defaulting to sync mode (though this would be counter to
the spec I think) but that would only affect data transfers.