Subject: Re: CVS commit: src/sys/dev/ic
To: Sam Leffler <sam@errno.com>
From: enami tsugutomo <enami@but-b.or.jp>
List: source-changes
Date: 10/17/2003 08:07:31 
> Actually this was just a bug, not a bus dma api difference.

Also, why null packet test is done after bus_dmamap_load (not before)?
If i read the source correctly, at least NetBSD/i386's bus_dmamap_load
function returns with bf_nseg == dm_nsegs == 1 when data length == 0
(though, bus_dmamap_load is inconsistent since dm_segs[0] isn't
initialized).

enami.