On Mon, Sep 15, 2014 at 02:12:25PM +0100, Chris Wilson wrote:
> On Fri, Sep 12, 2014 at 01:48:34PM +0100, Lionel Landwerlin wrote:
> > This is getting bigger than expected. Adding the locking that Chris
> > suggested on IRC.
> >
> > Thanks for taking time to review Chris.
>
> I'm happy with this
On Fri, Sep 12, 2014 at 01:48:34PM +0100, Lionel Landwerlin wrote:
> This is getting bigger than expected. Adding the locking that Chris
> suggested on IRC.
>
> Thanks for taking time to review Chris.
I'm happy with this series,
Reviewed-by: Chris Wilson
-Chris
--
Chris Wilson, Intel Open Sour
This is getting bigger than expected. Adding the locking that Chris
suggested on IRC.
Thanks for taking time to review Chris.
-
Lionel
Hopefully I got this right this time :)
-
Lionel
Following Chris' review, here is an updated patch using drmMMListHead.
I did a quick read of the benchmarks/tests files in igt, as far as I
can see, drm_intel_bufmgr_destroy() is always called before the drm
file descriptor is closed. So it seems this change shouldn't break
anything.
Cheers,
-
L
Hi there,
Here is a small modification I had to make to get buffers shared
between Mesa and LibVA on Chrome OS. This is required to have
refcounting properly between the 2 API otherwise, Mesa might end up
calling exit() when the kernel tells it that one of the buffer object
used in a batch buffer
