>>>>> On Sat, 18 May 1996 14:11:19 -0700,
	Jason Thorpe <thorpej@nas.nasa.gov> said:

> Sometime during the release cycle for 1.2, we'll have to make a decision 
> whether or not putting in the work-around you suggest is appropriate as 
> a stop-gap, depending on whether or not the underlying problem of SCSI 
> resource allocation has been addressed.

jmarin's problem (AHA-274X/284X and AIC-7850 have few SCBs) cannot be
avoided by resource pre-allocation in ahc_attach()...

# Though, Justin's SCB paging probably will resolve jmarin's problem.

>>>>> On Sat, 18 May 1996 21:26:38 -0700,
	"Justin T. Gibbs" <gibbs@freefall.freebsd.org> said:

> I don't know why NetBSD has problems allocating SCB space on demand.  Is
> malloc(M_NOWAIT) likely to fail if used some time after boot?  There
> is still no protection for calling malloc from an interrupt context as
> well.  This is a separate problem to that of making it safe for a
> resource request to fail from an interrupt context.

Because unlike FreeBSD, NetBSD's sdstart() always use SCSI_NOSLEEP.
--
soda@sra.co.jp		Software Research Associates, Inc., Japan
(Noriyuki Soda)		   software tools and technology group