>Jonathan Stone said:

>I take a certain interest in news postings that describe the bounce
>buffer support in FreeBSD and Linux as performance killers.

	You gain more performace from 16M -> 32M than you would ever loose by
implementing bounce-buffers; surely you realize that.

>>	*Why is this still broken?*

>The hardware is broken, not the OS.

	The '42 series cards are very good cards. They're very popular cards.
A lot of people have them. Even Adaptec's own drivers for the card do bounce
buffers, and almost every other OS that supports the '42 do bounce buffering.

	Also you must realize that not everyone has the BUS or MOTHERBOARD to
support a VLB or PCI card that can address the full range. And, not everyone
can afford to upgrade the memory and the controller at the same time .. 

	What "The Problem" is is that there _exist_ patches for NetBSD that
allow bounce buffering. BUT .. they're not available!

	So it may not be the cleanest situation, who cares? Make it 
if'def'out-able and put an "options  BROKEN_1542_BOUNCE" in the kernel..

	Then perhaps people who *have* this card don't have to *scrap* it..

Cheers,
Phil
--
Phillip F Knaack               flipk@iastate.edu
Database Programmer, NCREMP    Student Development Group
ISU Extension                  Project Vincent, Iowa State University