Login at Kodi Home

**fritsch** · (This post was last modified: 2014-07-09, 09:55 by fritsch.)

I am curious if the mainline implementation of VAAPI is also failing with that sample. Perhaps the VAAPI Infrastructure was never put under such an immense load before.
Edit: I sent an Email to the relevant intel devs, you are on CC (as always)

tret · (This post was last modified: 2014-07-09, 15:28 by tret.)

(2014-07-09, 09:06)Matt Devo Wrote:
(2014-07-09, 05:26)tret Wrote: Note 2: Disabling VAAPI allows playback with no freeze although with the high bitrate video it's really too choppy to watch (for me at least) which kinda makes the Chromebox meh unless a better fix/workaround is found.

tret

you need to enable multi-threaded software decoding under video acceleration settings. It's only available when the hardware options are deselected/disabled.

Can't believe I missed that, thanks Matt. This makes higher bitrate video playback much smoother. Will do until a fix for HW acceleration is available.

(2014-07-09, 09:48)fritsch Wrote: I am curious if the mainline implementation of VAAPI is also failing with that sample. Perhaps the VAAPI Infrastructure was never put under such an immense load before.
Edit: I sent an Email to the relevant intel devs, you are on CC (as always)

Thanks for digging into this one Fritsch. Hopefully with that sample we can get a little traction on this one. Really appreciate all your help.

tret

**fritsch** · 2014-07-09, 15:29

Such a fix does not fall out of heaven ... so post your crash dumps to the above linked bugreport and subscribe.

tret · (This post was last modified: 2014-07-09, 15:37 by tret.)

(2014-07-09, 00:20)Matt Devo Wrote:
(2014-07-08, 23:44)tret Wrote: Would it be possible to get a 1-2 minute sample including the lockup timestamp? I'd like to duplicate your testing and confirm the same results. Also, is there any way to analyze the video to find out what it is at that timestamp that differs from other video?

link: https://dl.dropboxusercontent.com/u/9830...sample.mkv

sample is 30s long, ~97MB. Will reliably freeze at ~14s.

Matt,

How long will this link stay active? It'd be excellent if the devs can access it for further testing, analysis, bug reporting, etc.

Edit: I can host the file on tretflix.com if you prefer or if this link expires.

(2014-07-09, 15:29)fritsch Wrote: Such a fix does not fall out of heaven ... so post your crash dumps to the above linked bugreport and subscribe.

Absolutely, this one correct?

https://bugs.freedesktop.org/show_bug.cgi?id=78960

Also, do you think it would be helpful to link the sample to the bug report as well?

**fritsch** · (This post was last modified: 2014-07-09, 15:40 by fritsch.)

I already mailed the above link to the intel folks. Those guys responsible for vaapi development are communication with FernetMenta and me for a long time now. I also uploaded it to my dropbox where it will stay forever: https://dl.dropboxusercontent.com/u/5572...sample.mkv

So if you post at this bugreport, make sure to:

a) boot with drm.debug=0xe (kernel parameter)
b) save the crash dump (dmesg tells you where its)
c) Post in the bugreport, link that sample, and also upload your crash dump and the output of dmesg > my_dmesg.txt

That way the intel folks can clearly see if this is the same issue or if it is something else with the same symptomes.

tret · 2014-07-09, 15:40

(2014-07-09, 15:39)fritsch Wrote: I already mailed the above link to the intel folks. Those guys responsible for vaapi development are communication with FernetMenta and me for a long time now. I also uploaded it to my dropbox where it will stay forever: https://dl.dropboxusercontent.com/u/5572...sample.mkv

So if you post at this bugreport, make sure to:

a) boot with drm.debug=0xe (kernel parameter)
b) save the crash dump (dmesg tells you where its)
c) Post in the bugreport, link that sample, and also upload your crash dump and the output of dmesg > my_dmesg.txt

That way the intel folks can clearly see if this is the same issue or if it is something else with the same symptomes.

Perfect, I'll work on this.

HotBlack · 2014-07-10, 10:23

"" $""' @@@@@"
@Aa"Z""z"""" ? "z ta a@" #"""Q

Matt Devo · 2014-07-13, 03:46

so are the Haswell NUC folks affected by this as well, or just ChromeBox users? I ask because the i3 ChromeBox I have here has the same CPU/GPU as the Haswell i3 NUC, so both should be affected similarly, no?

tret · 2014-07-13, 04:21

(2014-07-13, 03:46)Matt Devo Wrote: so are the Haswell NUC folks affected by this as well, or just ChromeBox users? I ask because the i3 ChromeBox I have here has the same CPU/GPU as the Haswell i3 NUC, so both should be affected similarly, no?

Would be nice to run the Joe sample on one of those as well and see if we get the same results.

Matt Devo · 2014-07-13, 06:58

(2014-07-13, 04:21)tret Wrote:
(2014-07-13, 03:46)Matt Devo Wrote: so are the Haswell NUC folks affected by this as well, or just ChromeBox users? I ask because the i3 ChromeBox I have here has the same CPU/GPU as the Haswell i3 NUC, so both should be affected similarly, no?

Would be nice to run the Joe sample on one of those as well and see if we get the same results.

I posted in the Haswell NUC thread, hopefully someone will step up and help out

**fritsch** · (This post was last modified: 2014-07-13, 08:07 by fritsch.)

Yes. They Hang exactly the same. Tested i3 version, a celeron 1820T and on an i5 4200u. All hang.

Matt Devo · 2014-07-13, 09:05

(2014-07-13, 08:07)fritsch Wrote: Yes. They Hang exactly the same. Tested i3 version, a celeron 1820T and on an i5 4200u. All hang.

given that, and the number of Haswell NUC owners, I'm surprised I haven't seen more on this particular issue then

**fritsch** · 2014-07-13, 11:36

For now this is the only sample I can reliably reproduce a Hang. But we track at least two issues here. The one with the renderer problem and this real vaapi issue.

Make sure to open a bugreport and link this sample. Also make sure to let them know this difference.

Matt Devo · 2014-07-15, 19:50

(2014-07-13, 11:36)fritsch Wrote: For now this is the only sample I can reliably reproduce a Hang. But we track at least two issues here. The one with the renderer problem and this real vaapi issue.

Make sure to open a bugreport and link this sample. Also make sure to let them know this difference.

not sure I'm equipped to detail the difference between this issue (reproducible) and the the other ( https://bugs.freedesktop.org/show_bug.cgi?id=78960 ) which I have yet to experience

**fritsch** · 2014-07-15, 19:51

Just file the bug with the above sample, provide all the logfiles and see what the intel guys do with it.