[Intel-gfx] [PATCH] drm/i915: Don't clflush purged objects on unbind

[Eric Anholt]

On Wed, 09 Dec 2009 21:18:31 +0000, Chris Wilson <chris at chris-wilson.co.uk> wrote:
> On Wed, 09 Dec 2009 09:27:15 -0800, Eric Anholt <eric at anholt.net> wrote:
> > Yeah, the question was whether there was any danger in handing pages
> > back to the system where there may be read cachelines that are actually
> > not valid.  We couldn't convince ourselves last night that that was
> > definitely safe.  (Would the CPU drop a write of zeroes to a cacheline
> > where the cache said it was already full of zeroes?)
> 
> Ooh, nasty. Do all DMA engines have similar issues with needing to flush
> the CPU cache-lines?
> -ickle

No, other DMA engines are coherent with the CPU cache.  In fact, it
appeared to us during some early debugging that DMA coming out of the
disk was cache-hot in the CPU (not that surprising of a way to do
things, when we thought about it).
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 197 bytes
Desc: not available
URL: <http://lists.freedesktop.org/archives/intel-gfx/attachments/20091215/4ea7ee79/attachment.sig>