On 8/6/08, Keith Packard <keithp@keithp.com> wrote:
quoted text
> On Wed, 2008-08-06 at 19:32 +0200, Stephane Marchesin wrote:
>
>  > Right, but the current code will basically force the discrete card
>  > drivers to implement backing store for all allocations.
>
>
> Aside from not really forcing discrete card drivers to to anything
>  (they're welcome to use or not use this stuff as they prefer), I believe
>  discrete cards will need backing store to support paging objects to disk
>  and suspend-to-ram operations.

Well there is no real other clean way than backing store...

quoted text
>
>
>  >  Do we want
>  > this ? Also, for cards that can handle page-based memory allocations,
>  > there is no way to make use of this feature, do we want this too ?
>
>
> I don't see how any of the current code directs how future drivers might
>  work; the user-level interface is reasonably abstract and should allow
>  all kinds of internal organizations.
>
>  Instead of complaining that the current code might not support some
>  abstract hardware, please build something that does work and we'll see
>  how to merge that with code for other drivers.
>

Well, this "abstract hardware" is in stores, it's called G80 and R600.
If intel can't do something on its hardware shouldn't prevent others
from using it. And I seriously doubt even the current interfaces would
be fit for this.

Stephane
--
unsubscribe notice
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/