I would think XS_SELTIMEOUT is XS_SELTIMEOUT (or whatever Jason has
renamed it to) no matter how you get there, whether it be from the plain
250ms nominal selection timeout or via a watchdog that comes along 2ms
later and notices this wedged condition.

That is, if I've understood you correctly....


On Sun, 31 Oct 1999, Ignatios Souvatzis wrote:

> I'm trying to debug the Amiga sbic.
> 
> afaict, if a data phase was expected in a command, sbic fails to tell the
> upper layer that a scsi bus timeout occured (on select, for a target that
> is not there).
> 
> What should sbic.c do to tell scsipi about this condition?
> 
> 	-is
>