From: Brandon S. Allbery (bsa@kf8nh.wariat.org)
Date: 08/24/93


From: bsa@kf8nh.wariat.org (Brandon S. Allbery)
Subject: Re: SCSI Performance (Yet Again)
Date: Tue, 24 Aug 1993 22:21:55 GMT

In article <1993Aug24.220727.8357@kf8nh.wariat.org> bsa@kf8nh.wariat.org (Brandon S. Allbery) writes:
>In article <110135@hydra.gatech.EDU> gt8134b@prism.gatech.EDU (Howlin' Bob) writes:
>>mean he won't let them in. Everyone's busy. Would you write them?
>>It shouldn't be too hard.
>
>I keep threatening to implement raw devices --- eventually I will find the
>time to do so.

I should add: it *is* somewhat difficult. You see, one of the other
performance increases in raw devices is that they do I/O directly from the
user's buffer --- thereby saving a copy to kernel space and possible sleep on
available memory.

On *my* system this won't cause a problem; I only have 12MB. But on a 32MB
ISA machine the driver had better check the physical address of the user's
buffer and do a copy if it's over 16MB, or things will go BOOM! Not to
mention the 64K boundary problem... I think SCO copies the data anyway to
avoid this. (Hm. Maybe it's not useful after all --- on this braindamaged
architecture, at least. Grrr.)

++Brandon

-- 
Brandon S. Allbery         kf8nh@kf8nh.ampr.org          bsa@kf8nh.wariat.org
"MSDOS didn't get as bad as it is overnight -- it took over ten years
of careful development."  ---dmeggins@aix1.uottawa.ca