• 1
  • 99
  • 100
  • 101(current)
  • 102
  • 103
  • 342
Intel VAAPI howto with Leia v18 nightly based on Ubuntu 18.04 server
Okay...

* same issue with startup color space as Isengard
* videos all start 4-5 seconds into the video, instead of at 0
* resuming a video tends to result in the picture playing in fast-motion for several seconds

And the biggie... I can definitely trigger problems. The system doesn't crash like it does with Isengard, but instead I get a totally garbled lurching picture with the UI occasionally coming on and flashing "buffering". Once started, this happens on all videos until I reboot. https://www.youtube.com/watch?v=ONfn4ouW3Vw

Going to enable logging and capture that.
Reply
And... I just managed to lock it up identically to Isengard.

http://paste.ubuntu.com/12637082/

It took some time to happen (that's a 2mb logfile), but it was having problems for a while before it crashed.
Reply
(2015-10-02, 10:04)Sunflux Wrote: Okay...

* same issue with startup color space as Isengard

Not an issue we cover here - as this is a kernel hack - a hack we all know that is just a hack

Quote:* videos all start 4-5 seconds into the video, instead of at 0
Thanks for this report.

Quote:* resuming a video tends to result in the picture playing in fast-motion for several seconds

Also thanks for this.

Quote:And the biggie... I can definitely trigger problems. The system doesn't crash like it does with Isengard, but instead I get a totally garbled lurching picture with the UI occasionally coming on and flashing "buffering". Once started, this happens on all videos until I reboot. https://www.youtube.com/watch?v=ONfn4ouW3Vw

Going to enable logging and capture that.

Thanks for the below Debug Log.
First decide what functions / features you expect from a system. Then decide for the hardware. Don't waste your money on crap.
Reply
Your log looks damn broken :-) ... Sadly I cannot outrule any SAMBA issues ...
First decide what functions / features you expect from a system. Then decide for the hardware. Don't waste your money on crap.
Reply
Remember it doesn't happen if VAAPI render is disabled...
Reply
Just in case you were wondering, here's a stress test log with VAAPI render disabled. Certainly not problem-free, but the system didn't crash and video was still playing fine when I rebooted.

http://paste.ubuntu.com/12637308/
Reply
I was just able to order a Beebox in Germany and my friend will bring it to my place In Asia in 3 weeks :-)
Feeling excited to start testing.
Been following this thread and read all 100 pages :-)

Thank you for all the work
Reply
(2015-10-01, 21:07)BigL-New Wrote: File which works:

DISPLAY=:0
OUTPUT=`xrandr -q | sed '/ connected/!d;s/ .*//;q'`
xrandr --output $OUTPUT --set "Broadcast RGB" "Full"
xrandr --output $OUTPUT --set "Broadcast RGB" "Video 16:235 pass-through"
/usr/bin/kodi --standalone

File with problems:

DISPLAY=:0
OUTPUT=`xrandr -q | sed '/ connected/!d;s/ .*//;q'`
xrandr --output $OUTPUT --set "Broadcast RGB" "Full"
xrandr --output $OUTPUT --set "Broadcast RGB" "Video 16:235 pass-through"
xrandr --output $OUTPUT --off
sleep 1
xrandr --output $OUTPUT --auto
/usr/bin/kodi --standalone

But please remember:

1. Problematic file works - Kodi runs. Only display is black. And after running "xrandr --output $OUTPUT --auto" on terminal Kodi shows.
2. For some reason it doesn't work even if i run Kodi in the background ("/usr/bin/kodi --standalone &") and add "xrandr --output $OUTPUT --auto" as next line. Screen is still black. To show Kodi i need to run it from terminal (through ssh).
Fritsch do you know why above file doesn't work as expected? It's really annoing, i've tried few things but still no luck.

BTW kernel with hack doesn't have latest intel video driver - a lot of messages "*ERROR* timed out waiting for Punit DDR DVFS request" in dmesg. Maybe it's important for file above and also for all of the tests? (at least on BSW)
Reply
I think I've found problems - here is full log from sysylog while starting Kodi:
Code:
Oct  2 18:01:26 bee systemd[1]: Started kodi-standalone using xinit.
Oct  2 18:01:26 bee systemd[1]: Starting kodi-standalone using xinit...
Oct  2 18:01:26 bee systemd[1]: Created slice user-1001.slice.
Oct  2 18:01:26 bee systemd[1]: Starting user-1001.slice.
Oct  2 18:01:26 bee systemd[1]: Starting User Manager for UID 1001...
Oct  2 18:01:26 bee systemd[1]: Started Session c16 of user kodi.
Oct  2 18:01:26 bee systemd[1]: Starting Session c16 of user kodi.
Oct  2 18:01:26 bee systemd[8238]: Reached target Timers.
Oct  2 18:01:26 bee systemd[8238]: Starting Timers.
Oct  2 18:01:26 bee systemd[8238]: Reached target Sockets.
Oct  2 18:01:26 bee systemd[8238]: Starting Sockets.
Oct  2 18:01:26 bee systemd[8238]: Reached target Paths.
Oct  2 18:01:26 bee systemd[8238]: Starting Paths.
Oct  2 18:01:26 bee systemd[8238]: Reached target Basic System.
Oct  2 18:01:26 bee systemd[8238]: Starting Basic System.
Oct  2 18:01:26 bee systemd[8238]: Reached target Default.
Oct  2 18:01:26 bee systemd[8238]: Startup finished in 24ms.
Oct  2 18:01:26 bee systemd[8238]: Starting Default.
Oct  2 18:01:26 bee systemd[1]: Started User Manager for UID 1001.
Oct  2 18:01:26 bee xinit[8237]: X.Org X Server 1.17.1
Oct  2 18:01:26 bee xinit[8237]: Release Date: 2015-02-10
Oct  2 18:01:26 bee xinit[8237]: X Protocol Version 11, Revision 0
Oct  2 18:01:26 bee xinit[8237]: Build Operating System: Linux 3.19.0-28-generic x86_64 Ubuntu
Oct  2 18:01:26 bee xinit[8237]: Current Operating System: Linux test 4.1.9-lauri3+ #2 SMP Thu Oct 1 08:36:41 CEST 2015 x86_64
Oct  2 18:01:26 bee xinit[8237]: Kernel command line: BOOT_IMAGE=/vmlinuz-4.1.9-lauri3+ root=/dev/mapper/test--vg-root ro
Oct  2 18:01:26 bee xinit[8237]: Build Date: 11 September 2015  10:30:58AM
Oct  2 18:01:26 bee xinit[8237]: xorg-server 2:1.17.1-0ubuntu3.1 (For technical support please see http://www.ubuntu.com/support)
Oct  2 18:01:26 bee xinit[8237]: Current version of pixman: 0.32.6
Oct  2 18:01:26 bee xinit[8237]: Before reporting problems, check http://wiki.x.org
Oct  2 18:01:26 bee xinit[8237]: to make sure that you have the latest version.
Oct  2 18:01:26 bee xinit[8237]: Markers: (--) probed, (**) from config file, (==) default setting,
Oct  2 18:01:26 bee xinit[8237]: (++) from command line, (!!) notice, (II) informational,
Oct  2 18:01:26 bee xinit[8237]: (WW) warning, (EE) error, (NI) not implemented, (??) unknown.
Oct  2 18:01:26 bee xinit[8237]: (==) Log file: "/var/log/Xorg.0.log", Time: Fri Oct  2 18:01:26 2015
Oct  2 18:01:26 bee xinit[8237]: (==) Using system config directory "/usr/share/X11/xorg.conf.d"
Oct  2 18:01:26 bee acpid: client 8117[0:0] has disconnected
Oct  2 18:01:26 bee acpid: client connected from 8244[0:0]
Oct  2 18:01:26 bee acpid: 1 client rule loaded
Oct  2 18:01:26 bee xinit[8237]: The XKEYBOARD keymap compiler (xkbcomp) reports:
Oct  2 18:01:26 bee xinit[8237]: > Warning:          Type "ONE_LEVEL" has 1 levels, but <RALT> has 2 symbols
Oct  2 18:01:26 bee xinit[8237]: >                   Ignoring extra symbols
Oct  2 18:01:26 bee xinit[8237]: Errors from xkbcomp are not fatal to the X server
Oct  2 18:01:26 bee xinit[8237]: Openbox-Message: Unable to find a valid menu file "/var/lib/openbox/debian-menu.xml"
Oct  2 18:01:26 bee kernel: [84255.045935] [drm:valleyview_update_wm [i915]] *ERROR* timed out waiting for Punit DDR DVFS request
Oct  2 18:01:27 bee kernel: [84255.259833] [drm:valleyview_update_wm [i915]] *ERROR* timed out waiting for Punit DDR DVFS request
Oct  2 18:01:27 bee kernel: [84255.370882] [drm:valleyview_update_wm [i915]] *ERROR* timed out waiting for Punit DDR DVFS request
Oct  2 18:01:27 bee lircd-0.9.0[745]: accepted new client on /run/lirc/lircd
Oct  2 18:01:27 bee xinit[8237]: /usr/lib/kodi/kodi-xrandr: Output HDMI1 is not disconnected but has no modes
Oct  2 18:01:37 bee xinit[8237]: xrandr: Output HDMI1 is not disconnected but has no modes
Oct  2 18:01:37 bee xinit[8237]: xrandr: cannot find preferred mode

My autostart file is as above (the one with problems) - i've changed only sleep to 10 to better see which command has problems.
Reply
(2015-10-02, 11:06)fritsch Wrote:
(2015-10-02, 10:04)Sunflux Wrote: * videos all start 4-5 seconds into the video, instead of at 0
Thanks for this report.

also noticed something similar to this the other day but hadn't had time to come report it. i believe it is related to the change introduce a week or two ago where playback doesn't start until the refresh rate has changed. it noticed when going back through the settings the pause for X seconds (which i believe was renamed) had been reverted to off. previously on my system i had a 5 second delay. i put a 5 second delay back in and when playing a file the refresh rate switch occurred and the ui stayed displayed for a few seconds. after several adjustment i was not able to get perfect playback. either i will miss a couple seconds of video and have to back up or i will see the ui before play back start. i believe i settled on switching it back to off for the time being. i can update to the latest (i think i'm on the 9/26 build) and do some more testing and provide logs later if it will help.

edit: updated to the 10/2 passthrough build and this is working correctly for me. i also didn't have the colorspace issue others have reported.
Reply
(2015-10-02, 18:24)BigL-New Wrote: I think I've found problems - here is full log from sysylog while starting Kodi:
Code:
Oct  2 18:01:26 bee systemd[1]: Started kodi-standalone using xinit.
Oct  2 18:01:26 bee systemd[1]: Starting kodi-standalone using xinit...
Oct  2 18:01:26 bee systemd[1]: Created slice user-1001.slice.
Oct  2 18:01:26 bee systemd[1]: Starting user-1001.slice.
Oct  2 18:01:26 bee systemd[1]: Starting User Manager for UID 1001...
Oct  2 18:01:26 bee systemd[1]: Started Session c16 of user kodi.
Oct  2 18:01:26 bee systemd[1]: Starting Session c16 of user kodi.
Oct  2 18:01:26 bee systemd[8238]: Reached target Timers.
Oct  2 18:01:26 bee systemd[8238]: Starting Timers.
Oct  2 18:01:26 bee systemd[8238]: Reached target Sockets.
Oct  2 18:01:26 bee systemd[8238]: Starting Sockets.
Oct  2 18:01:26 bee systemd[8238]: Reached target Paths.
Oct  2 18:01:26 bee systemd[8238]: Starting Paths.
Oct  2 18:01:26 bee systemd[8238]: Reached target Basic System.
Oct  2 18:01:26 bee systemd[8238]: Starting Basic System.
Oct  2 18:01:26 bee systemd[8238]: Reached target Default.
Oct  2 18:01:26 bee systemd[8238]: Startup finished in 24ms.
Oct  2 18:01:26 bee systemd[8238]: Starting Default.
Oct  2 18:01:26 bee systemd[1]: Started User Manager for UID 1001.
Oct  2 18:01:26 bee xinit[8237]: X.Org X Server 1.17.1
Oct  2 18:01:26 bee xinit[8237]: Release Date: 2015-02-10
Oct  2 18:01:26 bee xinit[8237]: X Protocol Version 11, Revision 0
Oct  2 18:01:26 bee xinit[8237]: Build Operating System: Linux 3.19.0-28-generic x86_64 Ubuntu
Oct  2 18:01:26 bee xinit[8237]: Current Operating System: Linux test 4.1.9-lauri3+ #2 SMP Thu Oct 1 08:36:41 CEST 2015 x86_64
Oct  2 18:01:26 bee xinit[8237]: Kernel command line: BOOT_IMAGE=/vmlinuz-4.1.9-lauri3+ root=/dev/mapper/test--vg-root ro
Oct  2 18:01:26 bee xinit[8237]: Build Date: 11 September 2015  10:30:58AM
Oct  2 18:01:26 bee xinit[8237]: xorg-server 2:1.17.1-0ubuntu3.1 (For technical support please see http://www.ubuntu.com/support)
Oct  2 18:01:26 bee xinit[8237]: Current version of pixman: 0.32.6
Oct  2 18:01:26 bee xinit[8237]: Before reporting problems, check http://wiki.x.org
Oct  2 18:01:26 bee xinit[8237]: to make sure that you have the latest version.
Oct  2 18:01:26 bee xinit[8237]: Markers: (--) probed, (**) from config file, (==) default setting,
Oct  2 18:01:26 bee xinit[8237]: (++) from command line, (!!) notice, (II) informational,
Oct  2 18:01:26 bee xinit[8237]: (WW) warning, (EE) error, (NI) not implemented, (??) unknown.
Oct  2 18:01:26 bee xinit[8237]: (==) Log file: "/var/log/Xorg.0.log", Time: Fri Oct  2 18:01:26 2015
Oct  2 18:01:26 bee xinit[8237]: (==) Using system config directory "/usr/share/X11/xorg.conf.d"
Oct  2 18:01:26 bee acpid: client 8117[0:0] has disconnected
Oct  2 18:01:26 bee acpid: client connected from 8244[0:0]
Oct  2 18:01:26 bee acpid: 1 client rule loaded
Oct  2 18:01:26 bee xinit[8237]: The XKEYBOARD keymap compiler (xkbcomp) reports:
Oct  2 18:01:26 bee xinit[8237]: > Warning:          Type "ONE_LEVEL" has 1 levels, but <RALT> has 2 symbols
Oct  2 18:01:26 bee xinit[8237]: >                   Ignoring extra symbols
Oct  2 18:01:26 bee xinit[8237]: Errors from xkbcomp are not fatal to the X server
Oct  2 18:01:26 bee xinit[8237]: Openbox-Message: Unable to find a valid menu file "/var/lib/openbox/debian-menu.xml"
Oct  2 18:01:26 bee kernel: [84255.045935] [drm:valleyview_update_wm [i915]] *ERROR* timed out waiting for Punit DDR DVFS request
Oct  2 18:01:27 bee kernel: [84255.259833] [drm:valleyview_update_wm [i915]] *ERROR* timed out waiting for Punit DDR DVFS request
Oct  2 18:01:27 bee kernel: [84255.370882] [drm:valleyview_update_wm [i915]] *ERROR* timed out waiting for Punit DDR DVFS request
Oct  2 18:01:27 bee lircd-0.9.0[745]: accepted new client on /run/lirc/lircd
Oct  2 18:01:27 bee xinit[8237]: /usr/lib/kodi/kodi-xrandr: Output HDMI1 is not disconnected but has no modes
Oct  2 18:01:37 bee xinit[8237]: xrandr: Output HDMI1 is not disconnected but has no modes
Oct  2 18:01:37 bee xinit[8237]: xrandr: cannot find preferred mode

My autostart file is as above (the one with problems) - i've changed only sleep to 10 to better see which command has problems.

Problem with the testing kernel:
Quote:Oct 2 18:01:26 bee kernel: [84255.045935] [drm:valleyview_update_wm [i915]] *ERROR* timed out waiting for Punit DDR DVFS request
Oct 2 18:01:27 bee kernel: [84255.259833] [drm:valleyview_update_wm [i915]] *ERROR* timed out waiting for Punit DDR DVFS request
Oct 2 18:01:27 bee kernel: [84255.370882] [drm:valleyview_update_wm [i915]] *ERROR* timed out waiting for Punit DDR DVFS request

You have a braswell and therefore need the other kernel I gave you last time ... which does not have the passthrough patches, but the DVFS fixes ...
First decide what functions / features you expect from a system. Then decide for the hardware. Don't waste your money on crap.
Reply
(2015-10-02, 11:45)Sunflux Wrote: Just in case you were wondering, here's a stress test log with VAAPI render disabled. Certainly not problem-free, but the system didn't crash and video was still playing fine when I rebooted.

http://paste.ubuntu.com/12637308/

It looks good. Fernet found a bug that could have caused the "mem leak". The EGLContext was not destroyed and probably all allocated surfaces would remain unfreed: https://github.com/FernetMenta/xbmc/comm...65734453c7

I will be travelling until Sunday ... so don't expect any new build until then.
First decide what functions / features you expect from a system. Then decide for the hardware. Don't waste your money on crap.
Reply
(2015-10-02, 22:14)fritsch Wrote:
(2015-10-02, 11:45)Sunflux Wrote: Just in case you were wondering, here's a stress test log with VAAPI render disabled. Certainly not problem-free, but the system didn't crash and video was still playing fine when I rebooted.

http://paste.ubuntu.com/12637308/

It looks good. Fernet found a bug that could have caused the "mem leak". The EGLContext was not destroyed and probably all allocated surfaces would remain unfreed: https://github.com/FernetMenta/xbmc/comm...65734453c7

I will be travelling until Sunday ... so don't expect any new build until then.

When I run Kodi in limited 16-235 mode I can easily lock Kodi. The skin doesn't rebuild totally and I have to reboot through JSON. Is that related to the mem leak fix? It doesn't happen when I run Kodi in full configuration.

Log: http://xbmclogs.com/purnggoap

The log wil show I'm running Arctic Zephyr, but it happens with Confluence as well. Depending on the amount of video or sequence of videos, I get garbled video like the YouTube video above or the skin artefacts.

Video: https://youtu.be/2coGdXzwZ9Q

(don't mind the black levels, that's the camera playing tricks Wink)
My XBMC / Kodi history.
XBMC for XBOX -> XBMC for Windows -> XBMC for Android -> Kodi for Android - > Kodi v16 OE (Amlogic) -> Kodi v17 LE (NUC 2x) -> Kodi v18.5 LE (NUC 2x)
Reply
(2015-10-02, 22:14)fritsch Wrote:
(2015-10-02, 11:45)Sunflux Wrote: Just in case you were wondering, here's a stress test log with VAAPI render disabled. Certainly not problem-free, but the system didn't crash and video was still playing fine when I rebooted.

http://paste.ubuntu.com/12637308/

It looks good. Fernet found a bug that could have caused the "mem leak". The EGLContext was not destroyed and probably all allocated surfaces would remain unfreed: https://github.com/FernetMenta/xbmc/comm...65734453c7

I will be travelling until Sunday ... so don't expect any new build until then.

Hey, no problem... just trying to help where I can! I know skipping around like that seems like a ridiculous non-real-world occurrence, but I've just found that forces the issue quickly, especially with that high bitrate file. And if you saw the YouTube garbled video, once that starts, it happens on every single video I play - and shortly after that, the system will die.

The first time the lockup happened, I was 40 minutes into a movie, hit the shortskipback button once, and boom - frozen. Nothing else had occurred before then.
Reply
Hello , I use the Isengard version from 0923. When I watch a BluRay the sound comes slightly delayed so that I have set myself the delay in Kodi +175 . After that it was perfect.WhatI have to put in the advancedsettings.xml ?

< video>
< latency >
<delay > 0 < / delay>
<refresh >
< min > 23 < / min >
< max > 24 < / max >
<delay > 0 < / delay>
< / refresh>
< / latency >
</ video>

Thanks
Reply
  • 1
  • 99
  • 100
  • 101(current)
  • 102
  • 103
  • 342

Logout Mark Read Team Forum Stats Members Help
Intel VAAPI howto with Leia v18 nightly based on Ubuntu 18.04 server18