On Fri, Sep 19, 2003 at 12:43:08PM -0700, Greywolf wrote:
> Thus spake Perry E. Metzger ("PEM> ") sometime Today...
> 
> PEM> However, my main delays in booting are for things like delays during
> PEM> probe of my IDE devices during autoconf, not the amount of time rc
> PEM> takes to run (which is quite short), so I'm not sure it is worth it.
> 
> ...and SCSI devices as well; did we ever do a mass-reset instead of waiting
> for 2 seconds for *each* bus?

current doesn't do 2 seconds for *each* bus any more.
I also have changes to do the IDE resets in parrallel (but Charles just made
my life harder :(

> 
> If I were to wire down everything in the kernel such that there are no
> wildcard probes, would that speed things up at all?  i.e. are there any
> ways around these delays?  Some of them seem somewhat obnoxiously long
> (I'll have to note where the major pauses happen...)

No

-- 
Manuel Bouyer <bouyer@antioche.eu.org>
     NetBSD: 24 ans d'experience feront toujours la difference
--