Jason Thorpe (thorpej@nas.nasa.gov) wrote:

: poorly) those reasons ... In any case, it's clear that a lot of though 
: and engineering has to be put into the bounce-buffer and, on a larger 
: scale, the machine-independent ISA solution, as well.  We should try 
: really hard to do it right the first time (well, having already failed 

Hm, I think I fully understand why a trivial bounce buffer scheme for
isa/i386 wont work on Alphas. Ok, accepted. Can't "we", for a suitable 
definition of "we", try to get a trivial bounce buffer scheme inside

#if defined(I386) && defined(STUPID_I386_BOUNCE_BUFFERS)

#endif

into those few drivers, which might profit from it? After all, all Amiga
drivers for boards on the Zorro bus, which were developed for the Z2
bus only, have the same problem, and work around quietly ... only thing
you need is a 24bit-dma-able memory allocator.

: that, how about the second or third time :-).  Being someone who uses 
: Alphas and x86 boxes (and an IBM RT and an Apollo, which just might run 

Ah :-) Should I get the diagnose diskette out of the closet and try to find 
out whats wrong with our 6150-135? (Hm, from the noise it makes, its the disk 
which doesn't manage to start up anymore).

: NetBSD one day :-), this is fairly important to me both as a user and as 
: a member of core.

Regards,
	Ignatios Souvatzis

(who was very happy when netbsd-1.0, unlike 0.9withlotsofpatches, was at
least able to detect the 24bit address problem on the machine he uses as
dial in point, instead of just randomly crashing)