Subject: Re: FreeBSD Bus DMA
To: Justin T. Gibbs <>
From: Jonathan Stone <jonathan@DSG.Stanford.EDU>
List: tech-kern
Date: 06/11/1998 20:37:59
Justin Gibbs writes:

>I tried to point at the potential size savings, but I guess that simply
>doesn't count in your book. 


The size savings for better ISA bounce buffer management and
implementation matter. We've both seen several people explicitly agree
to that.  Nobody has yet said otherwise.  If they are, i want to know
who: I think they're wrong.

But no, the size savings of the order of 0.2% do not count. 
And i do not see now the second affects or enables the first.

If you've attempted to explain that, I missed it.  And I've been
looking for it.  Could you repost it, please?

>I guess I should have also waited another year
>or so until FreeBSD has finished its port to the Alpha and perhaps another
>platform to bother talking about this. 

No, but itd be nice if we didnt have bootless incomatiblities in the
API (from either side). I think you'd agree with that.

>It's not as if I'm trying to toot
>my horn here.  A user asked a question about the differences in the
>implementations and I attempted to answer it.  Sorry to have wasted so much
>of your time.

I dont think that's a fair or accurate summation of events.

I dont know what happened in other fora (private mail or on freebsd
lists) but what you wrote here was:

 }I had hoped to share an identical bus dma interface with
 }NetBSD, but I am unwilling to do this so long as the NetBSD interface
 }sacrifices speed and memory resources for absolutely no gain in

There seem to be adequate grounds for skepticism that this is really
so. Answering the questions that've been posed would go a long way
toward dispelling them.