I am trying to get Xvideo working with shadowfb on the popular Intel 855GM chipset. MPlayer and other similar programs display just a blue rectangle when using Xv output. I can hear the audio track, and no errors appear in stdout. Also, xvinfo tells that an Xv adaptor is present and the overlay seems to be set up according to Xorg.log: $ cat /var/log/Xorg.0.log | grep overlay [ 22.594] (==) intel(0): video overlay key set to 0x83e [ 23.070] (II) intel(0): Set up overlay video Modesetting is enabled in the kernel. kernel 2.6.36.1 xf86-video-intel 2.13.0 xorg-server 1.9.2
Found a solution: Xv works again after changing the screen depth to 24 instead of 16. Previously Xv was running happily on depth 16, though. The problem occurs also on non-shadowfb X sessions with depth 16.
Do you know which component regressed? ie does 2.6.36 + 2.12 play Xv or 2.6.35 + 2.13?
Well, I've written the kms xv code for i8xx and I've never bothered to check it with anything else than 32 bit depth (the default). But now that somebody actually cares I'll see whether I can fix it. First noob question (o.k. just laziness): How do you set 16 bit depth? xorg.conf snippet preferred ...
@Chris: Hard to say, I upgraded from xorg-server 1.6 and xf86-video-intel 2.7 or so. Later versions of the driver were quite unstable on 855 so I did not test them that much. @Daniel: I usually start X from the command line: "startx -- -depth 16". The following snippet in xorg.conf should also do the trick: Section "Screen" DefaultDepth 16 EndSection
> --- Comment #3 from Daniel Vetter <daniel@ffwll.ch> 2010-12-03 14:28:18 PST --- > Well, I've written the kms xv code for i8xx and I've never bothered to check it > with anything else than 32 bit depth (the default). But now that somebody > actually cares I'll see whether I can fix it. First noob question (o.k. just > laziness): How do you set 16 bit depth? xorg.conf snippet preferred ... > Section "Screen" Identifier "foo" DefaultDepth 16 EndSection
I've just tested it on my i855 and it works here. Unfortunately there are a few unfixed i855 overlay bugs that I can't reproduce and have no other clue as to their cause. Maybe-related: #29574 A few question: - is this a regression for you? - anything reported in dmesg or Xorg.log? - can you try a few odd sizes/fullscreen/external (2nd) monitor/different output formats? If it's not broken all the time for 16 bit, this might give a hint.
Created attachment 40829 [details] dmesg
Created attachment 40830 [details] Xorg.log
Created attachment 40831 [details] xorg.conf
Yes, this is a regression, xv on 16 bit was working for me (although on a rather ancient release of the driver + server). It is not a major inconvenience since xv is working fine for 24 bit. A copy dmesg and Xorg.log attached after playing video with MPlayer. No messages appear on the logs during or after playback. The format of video does not seem to affect the result, neither does fullscreen or trying different sizes. I have no access to external display at the moment, but trying different values for xvattr -a XV_PIPE did not have any effect. I tried this with also version 2.12 of the driver and the result was the same - only a blue rectangle is shown on 16 bit mode.
> --- Comment #10 from sh29112911@gmail.com 2010-12-06 07:11:23 PST --- > Yes, this is a regression, xv on 16 bit was working for me (although on a > rather ancient release of the driver + server). It is not a major inconvenience > since xv is working fine for 24 bit. If the old working version was also using kms, could you perhaps try to bisect this regression? And don't have any idea what could be wrong, and it seems to work here ... If this hasn't ever worked for you with kms bisecting is pointless, because it's a totally new codebase.
(In reply to comment #11) > > --- Comment #10 from sh29112911@gmail.com 2010-12-06 07:11:23 PST --- > > If this hasn't ever worked for you with kms bisecting is pointless, > because it's a totally new codebase. This is the case - previously things were working on non-kms setup only.
This issue is affecting a hardware component which is not being actively worked on anymore. Moving the assignee to the dri-devel list as contact, to give this issue a better coverage.
Worksforme with kernel 3.4.0, and xf86-video-intel.git (with sna enabled).
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.