Created attachment 119520 [details] Xorg.0.log crash with --enable-debug=full, build + compton options Since "sna/dri2: Protect against a new possible recursion during TearFree" ( b1015d4 ) I can't play videos with mpv anymore. Xorg.0.log (for some reason the server runs as root again, but dunno since when) shows two lines recurring almost at a 1:1 pattern: [ 3618.247] (--) intel(0): HDMI max TMDS frequency 250000KHz [ 3619.715] (EE) intel(0): [DRI2] DRI2SwapComplete: bad drawable Without compton (and before said commit) they play as intended, but the "max TMDS" line still shows on every mpv start. Also, with compton sometimes videos briefly do play (their video also starts playing when I resize the window) but after some seconds or fast forwarding the display freezes again without mpv printing an error. That and the server crashes almost instantly (the 1080p manages to show the deskotp) only if started with --enable-debug, attached is a log with =full. Haswell 4770, 2x HDMI (1440p+1080p), drm-intel-nightly
Hmm, the dangers of DRI3/Present + DRI2 interop. :|
commit 64dbcdd4de1fe1941196265d054272d7d93e5d59 Author: Chris Wilson <chris@chris-wilson.co.uk> Date: Tue Nov 10 09:56:07 2015 +0000 sna/dri2: Add a specific recursion indicator for TearFree/DRI2 Avoid conflating the DRI3/Present flag for controlling TearFree by adding a separate flag to indicate when we are waiting inside the TearFree shadow handler to avoid recursion in DRI2. Reported-by: Andreas Reis <andreas.reis@gmail.com> Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=92873 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk> commit 47f6c5b211c0458f2bf1e05a736e466bd8b64763 Author: Chris Wilson <chris@chris-wilson.co.uk> Date: Tue Nov 10 09:50:43 2015 +0000 sna: Tweak TearFree interaction with Present flipping When Present takes over we mark the TearFree shadow as disabled. Before Present relinquishes its flip sequence, it copies from its pixmap to the screen triggering the shadow wait, even though the screeen Pixmap is not current bound (and still marked as disabled). Tweak the conditionals to properly skip this wait. References: https://bugs.freedesktop.org/show_bug.cgi?id=92873 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk> should resolve the stuck dri2 mpv and the assertion failure on startup with dri3.
Both issues resolved. Still getting the harmless line intel(0): HDMI max TMDS frequency 250000KHz in Xorg.#.log whenever I start mpv, though.
(In reply to Andreas Reis from comment #3) > Both issues resolved. Still getting the harmless line > > intel(0): HDMI max TMDS frequency 250000KHz > > in Xorg.#.log whenever I start mpv, though. mpv is doing a forced probe of the display configuration (i.e. is forcing a re-read of the hardware state for all attached outputs). That's a bug in mpv.
*** Bug 93446 has been marked as a duplicate of this bug. ***
Use of freedesktop.org services, including Bugzilla, is subject to our Code of Conduct. How we collect and use information is described in our Privacy Policy.