>On Fri, 12 Jun 1998 06:08:39 +0900 (JST) 
> Noriyuki Soda <soda@sra.co.jp> wrote:
>
> > About bus_dmamap_callback_t, I think that reducing bounce memory size 
> > is certainly good thing. But it is pain if we must use callback in all 
> > cases.  (Think about hardware_scb in ahc driver.)
> > I think NetBSD's interface is more convenient than FreeBSD's.
> > Isn't there other way to reduce bounce memory ?
>
>Yes, lots of ways.  But you don't have to change the interface to
>bus_dma to do it.  The bounce buffer size issue is largely an issue
>of the implementation that drivers never see.  To be quite honest,
>I implemented it the way I did because bounce buffers were SORELY
>needed, and didn't have the time to implement them differently underneath.

Can you explain how you can remove the need to allocate worse case scenario
bounce buffer space without adding some way of handing deferred requests?
I'm not contending the fact that the NetBSD implementation could be smarter
about copying less data when a bounce has to occur which is easy to do
within the constraints of the current API.

>Jason R. Thorpe                                       thorpej@nas.nasa.gov
>NASA Ames Research Center                            Home: +1 408 866 1912
>NAS: M/S 258-5                                       Work: +1 650 604 0935
>Moffett Field, CA 94035                             Pager: +1 650 428 6939

--
Justin