Re: [Mesa3d-dev] [RFC] Event-based buffer validation for DRI2.

[Francisco J. <cur...@ri...>]

Francisco Jerez <cur...@ri...> writes:

> The current buffer validation approach (AKA the DRI2 glViewport hack)
> is both incorrect (because a compliant OpenGL application may opt for
> the identity as viewport transform and work with window coordinates
> directly) and inefficient (some programs have the habit of calling
> glViewport several times per frame (e.g. OpenArena), causing many
> unnecessary roundtrips).
>
> This changeset gives DRI2 the ability to report drawable changes in an
> asynchronous way, but it's a bit intrusive so I expect all sorts of
> complaints to come.
>
> I've tested this with the gallium and classic nouveau drivers, using
> direct and indirect rendering, with single and double buffered apps,
> with older X servers, and it seems to work... I'm open to any
> comments.
>

A revised changeset follows. Among other minor fixes, it avoids the
DRI2InfoRec ABI breakup, so this version will hopefully be somewhat less
intrusive on the X side.

This time I've tried to preserve backwards compatibility with flush v2
(sigh), as intel has recently started to make use of it in a non-trivial
way.

[dri2proto patch v3] Define an event to notify clients about the validity of their buffers.
[xserver patch v3 1/5] Add a PreConfigureWindow hook.
[xserver patch v3 2/5] dri2: No need to blit from front on DRI2GetBuffers if they're just being reused.
[xserver patch v3 3/5] glx: Enforce a 1:1 correspondence between GLX and X11 windows.
[xserver patch v3 4/5] glx/dri2: Notify the driver when its buffers become invalid.
[xserver patch v3 5/5] dri2: Support the DRI2InvalidateBuffers event.
[mesa patch v3 1/3] dri2: Event driven buffer validation.
[mesa patch v3 2/3] dri/nouveau: Use event driven buffer validation.
[mesa patch v3 3/3] st/dri2: Use event-driven buffer validation.