[Intel-gfx] [RFC 0/7+1] Add in-kernel vblank delaying mechanism

[Daniel Vetter]

On Wed, Nov 19, 2014 at 8:47 PM, Paulo Zanoni <przanoni at gmail.com> wrote:
>
>  2. How should the driver interface look like?
>
>    a. All the possibilities are passed through the function call, so the drm.ko
>       code needs to set the struct members itself.
>    b. The caller already sets the struct members instead of passing them as
>       parameters to the function.
>    c. Something different?

We need b because the caller must allocate the structure (the point
where we can return -ENOMEM to userspace might be much after the point
where we need to schedule a vblank callback for e.g. async flips). But
for simple interfaces we need a few things passed directly I think
(since I expect that we'll reuse vblank callback structures similar to
how we reuse timers/work items).

I'll follow up with a detailed review of the new interface exposed to
drivers and what I think it should look like, need to head off now.
-Daniel
-- 
Daniel Vetter
Software Engineer, Intel Corporation
+41 (0) 79 365 57 48 - http://blog.ffwll.ch