Subject: Re: Need to force SCSI_POLL when cold
To: Gordon W. Ross <email@example.com>
From: Jeremy Cooper <firstname.lastname@example.org>
Date: 03/24/1997 13:43:15
On Mon, 24 Mar 1997, Gordon W. Ross wrote:
> I'm somewhat dissatisfied with the need for every SCSI driver to
> force SCSI_POLL when cold, as in the change shown here:
> + /*
> + * XXX: Hack: During autoconfig, force polling mode.
> + * Needed as long as sdsize() can be called while cold,
> + * otherwise timeouts will never call back (grumble).
> + */
> + if (cold)
> + flags |= SCSI_POLL;
I think the real heart of the problem is the code in
kern_synch.c:tsleep(). It is not behaving as it says it will when
``cold'' is true.