[Bug 711155] wayland: add wl_drm support to wayland sink

[GStreamer (GNOME Bugzilla)]

https://bugzilla.gnome.org/show_bug.cgi?id=711155

--- Comment #68 from Fabien Dessenne <fabien.dessenne at st.com> ---
(In reply to George Kiagiadakis from comment #67)
> I see a collision here as we are both working on the same thing apparently.
Agree :)

> A few notes on your patch:
> * It still depends on libdrm (I know we have to use drm_fourcc.h, but it
> doesn't justify the dependency for me, as it's just a couple of constants. I
> prefer copying them over).
I am neither satisfied with having such a libdrm dependance, nor with hard
copying the drm_fourcc.h file.
Maybe there is something to do in configure.ac so we can have only a libdrm
header file dependance, not a 'full' dependance. I do not know if this is
possible (AC_CHECK_HEADER?) but it would be probably the best option.


> * It's not based on master (and I can tell it does not apply cleanly on
> master)
Well, I apologize here. To be honest I did not plan to share this patch right
now (as I did not rebase it yet on master) but your patch proposal forced me to
do it now, so we can discuss on the two different options we have now.


> * GST_VIDEO_FORMATS is confusing as a name, since there is another constant
> like that in the gstvideo library
Agree, I need to fix it.


> * Using gst_wl_display_rountrip() is bad. At that time that you call it, the
> event queue is polled by another thread in wldisplay.c, so doing a sync
> roundtrip in another thread while the event loop thread is running in
> parallel is prone to cause races.
Good point. The way you manage it in your proposal seems a better option.


> Finally, I see that you are now using the "memory:dmabuf" caps feature. I
> thought that you were trying to avoid that.
I have tried, tried and tried again. I finally came to this least bad option.


> Do you have reconsidered your
> proposal, or do you agree/disagree with my proposal in the comments above?
> It would be useful to discuss a bit more about it.
Benjamin (who is an office coworker of mine) has a better view on this
discussion, he's going to take over my answer.

Fabien

-- 
You are receiving this mail because:
You are the QA Contact for the bug.
You are the assignee for the bug.