Re: proposal for bus_dma(9) change

[Thor Lancelot Simon <tls%rek.tjls.com@localhost>]

On Wed, Mar 04, 2009 at 12:41:50AM +0100, Christoph Egger wrote:
> 
> When bus_dmamap_create() fails, then the map pointer
> is undefined. This is IMO a problem, because the
> dma free routines in many drivers assume the map
> pointer is NULL when invalid.

It seems to me that since the map pointer is "undefined" it is a
perfectly compatible and reasonable API change to set it to NULL.

Assuming you're willing to do the work of finding and changing all
the bus_dma implementations, I can't see why anyone should stop you
from doing so.

When you've finished doing that, at that point we could discuss
whether drivers ought to rely on the change you've made -- it seems
to me that at that point the obvious answer would be "yes".

Thor