Thanks for your quick check Michel. I had removed the condition in my local source in ms_present_check_unflip() method and moved the assignment to ms_present_check_flip() method. Below is the snippet.
static Bool ms_present_check_flip(RRCrtcPtr crtc,
WindowPtr window,
PixmapPtr pixmap,
Bool sync_flip,
PresentFlipReason *reason)
WindowPtr window,
PixmapPtr pixmap,
Bool sync_flip,
PresentFlipReason *reason)
{
if (ms->drmmode.sprites_visible > 0)
return FALSE;
Bool ret = ms_present_check_unflip(crtc, window, pixmap, sync_flip, reason);
ms->flip_window = window;
return ret;
return FALSE;
Bool ret = ms_present_check_unflip(crtc, window, pixmap, sync_flip, reason);
ms->flip_window = window;
return ret;
}
Still the problem exists.
Thanks
Uday Kiran
On Mon, May 25, 2020 at 9:49 PM Michel Dänzer <michel@xxxxxxxxxxx> wrote:
On 2020-05-25 4:58 p.m., uday kiran pichika wrote:
> Thanks Michel..
>
> sorry if my description is not clear.
> Yes I have used with glxgears in fullscreen mode and also a simple glxapp
> which actually opens in fullscreen by default.
>
> In both the cases, the issue is same.
>
> Just wanted to know if AMD has done anything different apart from the
> configurations either in MESA or in its Radeon driver ?
The code you can see in xf86-video-amdgpu and Mesa is working for lots
of people, most of them outside of AMD.
I think I've spotted one bug in your changes, in ms_present_check_unflip:
if (ms->drmmode.sprites_visible > 0)
ms->flip_window = window;
ms_present_check_flip has:
if (ms->drmmode.sprites_visible > 0)
return FALSE;
return ms_present_check_unflip(crtc, window, pixmap, sync_flip, reason);
So the condition can never be true in ms_present_check_unflip, and it
never updates ms->flip_window.
--
Earthling Michel Dänzer | https://redhat.com
Libre software enthusiast | Mesa and X developer
_______________________________________________ dri-devel mailing list dri-devel@xxxxxxxxxxxxxxxxxxxxx https://lists.freedesktop.org/mailman/listinfo/dri-devel
