[PATCH v2] Revert "dma-buf/sync-file: Avoid enable fence signaling if poll(.timeout=0)"

Daniel Vetter daniel at ffwll.ch
Fri Nov 18 09:37:49 UTC 2016


On Fri, Nov 18, 2016 at 08:56:11AM +0000, Chris Wilson wrote:
> On Fri, Nov 18, 2016 at 05:26:43PM +0900, Gustavo Padovan wrote:
> > From: Gustavo Padovan <gustavo.padovan at collabora.co.uk>
> > 
> > This reverts commit ecebca79f6976ddaddfd054d699272515869ea28.
> > 
> > Do not enable fence callback on poll() when using fence_array causes the
> > fence_array to not signal.
> > 
> > For now we will revert the change and enable signaling everytime time
> > poll is called with timeout=0 as well.
> > 
> > Cc: Chris Wilson <chris at chris-wilson.co.uk>
> > Signed-off-by: Gustavo Padovan <gustavo.padovan at collabora.co.uk>
> 
> Acked-by: Chris Wilson <chris at chris-wilson.co.uk>

Applied to drm-misc.

> I have some patches to use a bit on fence_array->flags to indicate where
> we can use this shortcut. I'm hoping someone has a better idea.

Fundamentally this is a property of how the fence is implemented. A flag
indicating that this fence will eventually signal, even when
enable_signalling isn't called seems like a perfectly ok approach.

We might want to spell out the different expectations clearly though, i.e.
if you block on a fence you must still call ->enable_signalling for prompt
signalling (and maybe it's needed to handle blocking waits or whatever).
-Daniel
-- 
Daniel Vetter
Software Engineer, Intel Corporation
http://blog.ffwll.ch


More information about the dri-devel mailing list