Radeon driver version = 1:7.3.0-1ubuntu3.1 Hardware: HP 4310s laptop, ATI 4330 OS: Linux Mint 17, VLC 2.1.2 HDMI works ok most of the time, but suffers from sporadic random drop outs. When a drop out occurs kern.log shows: HDMI: ELD buf size is 0, force 128 HDMI: invalid ELD data byte 0 Also (don't know if relevant) an occasional: hda-intel: IRQ timing workaround is activated for card #0. Suggest a bigger bdl_pos_adj. When a drop out (entire HDMI signal lost - not just audio) occurs VLC continues playing. A mouse click on desktop causes HDMI output to resume. No screen blanking or any kind of suspend mode is active so I can't speculate on the cause of the drop out. I do however think the driver should be able to gracefully cope with it. At least provide an optional flag that causes the driver to ignore the trigger event for the drop out since obviously it's ok most of the time.
This looks like a sound driver issue rather than a GPU driver issue.
(In reply to comment #1) > This looks like a sound driver issue rather than a GPU driver issue. No, because the entire HDMI signal is lost - not just the audio.
Ok, it seems the problems lie in sound/pci/hda/hda_eld.c which seems outside of your domain. I'll try to debug that code. Sorry about false alarm.
(In reply to comment #2) > (In reply to comment #1) > > This looks like a sound driver issue rather than a GPU driver issue. > > No, because the entire HDMI signal is lost - not just the audio. I suspect the monitor reacts badly to the audio stream glitches which causes it to drop the video as well.
(In reply to comment #4) > (In reply to comment #2) > > (In reply to comment #1) > > > This looks like a sound driver issue rather than a GPU driver issue. > > > > No, because the entire HDMI signal is lost - not just the audio. > > I suspect the monitor reacts badly to the audio stream glitches which causes > it to drop the video as well. You mean like overload protection? I thought about that, but the drop outs don't always occur at the same spot in the same audio segment. Having looked at the Intel ELD code I think it's buggy - on a drop out a "workaround" patch kicks in that's not even appropriate for my hardware. I did get a hold of the Intel ELD author, but he no longer maintains it. He did give me a list of developers to try. I'll experiment with this code and if I don't get anywhere will utilize the list.
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.