Login at Kodi Home

rbrohman · 2011-08-28, 19:22

Hi,

I've got an xbmc live box, running on a sandy bridge Core i3, using on board HDMI output to the tv. I set this up using the guide found at http://forum.xbmc.org/showthread.php?tid=96669

Everything seems to be running well, except that I'm getting seemingly random crashes to a red screen that sort of looks like corrupted video or something. Here's a screenshot:

When this occurs, the box no longer accepts any input. If I have an ssh session open when this happens, the connection is lost as well. Only way out is to hard reboot using the physical power button on the box.

It seems to happen randomly. Sometimes I can watch 3 or 4 episodes without an issue, sometimes I only get 5 minutes between crashes. It has also happened while just idling on the home screen.

Here's an extract from the debug log file, though I'm not sure that it shows anything regarding the crash:

Code:
44:57 T:2624506736   DEBUG: CDVDPlayer::HandleMessages - player started 1

44:58 T:3044759904    INFO: CheckIdle - Closing session to http://xoap.weather.com (easy=0xa790d6f0, multi=0xadb1f328)

44:58 T:2574150512   DEBUG: CDVDPlayerAudio:: Discontinuity - was:1026754.088000, should be:1004925.728461, error:-21828.359539

44:59 T:3044759904    INFO: CheckIdle - Closing session to http://www.google.com (easy=0xa7900018, multi=(nil))

45:01 T:2582543216   DEBUG: CPullupCorrection: detected pattern of length 1: 33366.67, frameduration: 33366.666667

45:04 T:3044759904   DEBUG: LIRC: Update - NEW at 2269450:000000037ff07bef 00 VolUp mceusb (VolUp)

45:04 T:3044759904   DEBUG: OnKey: leftctrl (d0) pressed, action is VolumeUp

45:04 T:3044759904   DEBUG: ------ Window Init (DialogVolumeBar.xml) ------

45:04 T:3044759904   DEBUG: Alloc resources: 0.13ms (0.00 ms skin load)

45:05 T:3044759904   DEBUG: LIRC: Update - NEW at 2269684:000000037ff07bef 00 VolUp mceusb (VolUp)

45:05 T:3044759904   DEBUG: OnKey: leftctrl (d0) pressed, action is VolumeUp

45:05 T:3044759904   DEBUG: LIRC: Update - NEW at 2270051:000000037ff07bef 00 VolUp mceusb (VolUp)

45:05 T:3044759904   DEBUG: OnKey: leftctrl (d0) pressed, action is VolumeUp

45:05 T:3044759904   DEBUG: LIRC: Update - NEW at 2270418:000000037ff07bef 00 VolUp mceusb (VolUp)

45:05 T:3044759904   DEBUG: OnKey: leftctrl (d0) pressed, action is VolumeUp

45:06 T:3044759904   DEBUG: LIRC: Update - NEW at 2270719:000000037ff07bef 00 VolUp mceusb (VolUp)

45:06 T:3044759904   DEBUG: OnKey: leftctrl (d0) pressed, action is VolumeUp

45:06 T:3044759904   DEBUG: LIRC: Update - NEW at 2271453:000000037ff07bef 00 VolUp mceusb (VolUp)

45:06 T:3044759904   DEBUG: OnKey: leftctrl (d0) pressed, action is VolumeUp

45:08 T:3044759904   DEBUG: ------ Window Deinit (DialogVolumeBar.xml) ------

45:16 T:2905598832   DEBUG: ### [Qlock] - Delaying 240 secs

45:24 T:2926357360   DEBUG: Thread Jobworker 2926357360 terminating (autodelete)

45:24 T:2934750064   DEBUG: Thread Jobworker 2934750064 terminating (autodelete)

45:24 T:2863635312   DEBUG: Thread Jobworker 2863635312 terminating (autodelete)

45:24 T:2798492528   DEBUG: Thread Jobworker 2798492528 terminating (autodelete)

45:24 T:3044759904   DEBUG: LIRC: Update - NEW at 2289538:000000037ff07bf1 00 Mute mceusb (Mute)

45:24 T:3044759904   DEBUG: OnKey: launch_app1_pc_icon (c0) pressed, action is Mute

45:24 T:3044759904   DEBUG: ------ Window Init (DialogMuteBug.xml) ------

45:24 T:3044759904   DEBUG: Alloc resources: 0.14ms (0.00 ms skin load)

45:25 T:3044759904   DEBUG: LIRC: Update - NEW at 2290472:000000037ff07bf1 00 Mute mceusb (Mute)

45:25 T:3044759904   DEBUG: OnKey: launch_app1_pc_icon (c0) pressed, action is Mute

45:25 T:3044759904   DEBUG: ------ Window Deinit (DialogMuteBug.xml) ------

45:29 T:3044759904   DEBUG: SECTION:UnloadDll(libcurl-gnutls.so.4)

45:29 T:3044759904   DEBUG: Unloading: libcurl-gnutls.so.4

I've posted the full debug log on pastebin as well, since it wouldn't all fit in this post: http://pastebin.com/V8aj8cV6

Any ideas on what might be causing this, or possible resolutions?

Thanks, any help is greatly appreciated.

rbrohman · 2011-08-30, 14:46

Anyone have any ideas on this, or see anything unusual in that log file?

thanks

mac666 · 2011-08-30, 22:58

What videocard are you using?
Which driver?

rbrohman · 2011-08-31, 14:42

I'm using the onboard video of the core i3.

How can I check the specific video drivers in use? I just followed this guide here to get everything setup http://forum.xbmc.org/showthread.php?tid=96669

rbrohman · 2011-09-03, 14:48

In settings > videos > playback, I have tried changing the render method to software, while also disabling the "allow hardware accelerations (vaapi)" option, but it still crashes ob me last night.

Nobody has any ideas on what this could be? Can I try anything else to help isolate the issue?

Thanks

rbrohman · 2011-09-03, 15:27

I'm wondering if this could be the relevant information from the log file, though its not right at the end where I'd presume the crash report to be:

Code:
11:44:57 T:2624506736 INFO: ffmpeg[9C6ECB70]: Input #0, avi, from '/home/xbmc/Downloads/TV/Whale Wars/Season 1/Whale Wars - S01E04 - We Are Hooligans.avi': 

11:44:57 T:2624506736 INFO: ffmpeg[9C6ECB70]: Metadata: 

11:44:57 T:2624506736 INFO: ffmpeg[9C6ECB70]: encoder : VirtualDubMod 1.5.4.1 (build 2178/release) 

11:44:57 T:2624506736 INFO: ffmpeg[9C6ECB70]: IAS1 : English 

11:44:57 T:2624506736 INFO: ffmpeg[9C6ECB70]: Duration: 00:42:43.06, start: 0.000000, bitrate: 1141 kb/s 

11:44:57 T:2624506736 INFO: ffmpeg[9C6ECB70]: Stream #0.0: Video: mpeg4, yuv420p, 624x352 [PAR 1:1 DAR 39:22], 29.97 tbr, 29.97 tbn, 29.98 tbc 

11:44:57 T:2624506736 INFO: ffmpeg[9C6ECB70]: Stream #0.1: Audio: mp3, 48000 Hz, 2 channels, s16, 136 kb/s 

11:44:57 T:2624506736 NOTICE: Opening video stream: 0 source: 256 

11:44:57 T:2624506736 NOTICE: Creating video codec with codec id: 13 

11:44:57 T:2624506736 DEBUG: CDVDFactoryCodec: compiled in hardware support: CrystalHD:no OpenMax:no VDPAU:no VAAPI:yes 

11:44:57 T:2624506736 DEBUG: FactoryCodec - Video: - Opening 

11:44:57 T:2624506736 DEBUG: SECTION:LoadDLL(special://xbmcbin/system/players/dvdplayer/swscale-0-i486-linux.so) 

11:44:57 T:2624506736 DEBUG: Loading: /usr/lib/xbmc/system/players/dvdplayer/swscale-0-i486-linux.so 

11:44:57 T:2624506736 DEBUG: SECTION:LoadDLL(special://xbmcbin/system/players/dvdplayer/avfilter-1-i486-linux.so) 

11:44:57 T:2624506736 DEBUG: Loading: /usr/lib/xbmc/system/players/dvdplayer/avfilter-1-i486-linux.so 

11:44:57 T:2624506736 NOTICE: CDVDVideoCodecFFmpeg::Open() Using codec: MPEG-4 part 2 

11:44:57 T:2624506736 DEBUG: VAAPI - attempting to open codec 13 with profile -99 at level -99 with 1 reference frames 

11:44:57 T:2624506736 DEBUG: VAAPI - initialize version 0.32 

11:44:57 T:2624506736 DEBUG: VAAPI - profile 0 

11:44:57 T:2624506736 DEBUG: VAAPI - profile 1 

11:44:57 T:2624506736 DEBUG: VAAPI - profile 5 

11:44:57 T:2624506736 DEBUG: VAAPI - profile 6 

11:44:57 T:2624506736 DEBUG: VAAPI - profile 7 

11:44:57 T:2624506736 DEBUG: VAAPI - profile 8 

11:44:57 T:2624506736 DEBUG: VAAPI - profile 9 

11:44:57 T:2624506736 DEBUG: VAAPI - profile 10 

11:44:57 T:2624506736 DEBUG: VAAPI - unable to find a suitable profile 

11:44:57 T:2624506736 DEBUG: VAAPI - destroying display 0xac4d568

rbrohman · 2011-09-14, 05:18

This crash issue was really hurting the WAF, so I decided to just buy a video card with the thought that this must be Sandy Bridge related.

I installed a GT430, formated my SSD and started again from scratch. Installed Ubuntu Natty minimal and then compiled the latest xbmc nightly ( XBMC PRE-11.0 GIT:20110910-caab280). I'm now pushing video via hdmi from the GT430, while audio is coming from the optical (TOSLINK) port of my mobo.

No longer getting that crazy red screen - but xbmc is still regularly (and seemingly randomly) crashing on me. When it crashes, I hear a screeching noise coming out of my speakers while xbmc is completely unresponsive, requiring a hard reboot.

I've posted another log file, but again am stumped as to finding the culprit. Is this a complete log file? am I missing something? http://pastebin.com/aNgSSuNG

I'm ready to do whatever is needed to get this resolved, as xbmc is basically unusable right now - I get a crash every 15 minutes it seems. Is it possible that the mobo is toast? Best way to test that theory?

Any and all assistance greatly appreciated. Thanks.

waver123 · 2011-09-14, 05:59

I would run memtest first to check the memory and then run a burn-in program like cpuburn and see if the system has any issue.

rbrohman · 2011-09-14, 06:12

Thanks for the ideas.

I ran memtest just prior to the last Ubuntu install and it came out clean. I'll give cpuburn a shot and see if it shows anything unusual. Any other ideas?

Thanks

**wsnipex** · (This post was last modified: 2011-09-14, 08:52 by wsnipex.)

This is definately xbmc related as it occasionally happened to me also(using ATI and VVAPI). I don't know what resolved it as I was playing around a lot trying to solve the vaapi 24p issue, which I finally did with a couple of patches from FernetMenta.
I guess it won't help you though, since those are afaik only vaapi related.

I can't see anything special in your debug log, you probably should also post the xbmc crash log.

Did you try to compile from source too?

tocinillo · (This post was last modified: 2011-09-14, 13:56 by tocinillo.)

wsnipex Wrote:This is definately xbmc related as it occasionally happened to me also(using ATI and VVAPI). I don't know what resolved it as I was playing around a lot trying to solve the vaapi 24p issue, which I finally did with a couple of patches from FernetMenta.
I guess it won't help you though, since those are afaik only vaapi related.

I can't see anything special in your debug log, you probably should also post the xbmc crash log.

Did you try to compile from source too?

Exactly, the problem is in XBMC (specially with Nvidia), XBMC crashes when "auto change refresh rate" option is enabled. With pixmap fernetmenta branch this problem is fixed: http://forum.xbmc.org/showpost.php?p=818...stcount=18

The problem is that this version is outdated, buy Fernet is working in the latest pre-11 version, but need more time to test...

FernetMenta · 2011-09-14, 18:30

I think the problem you are describing is similar to that I have investigated for vdpau. Decoder resources must not cleared down as long they are referenced by open gl. In case of vdpau this only happens if refresh rate is changing but for VAAPI this can occur even when a playback is stopped.
I have an experimental solution for vdpau but it looks like it's even more complicated for VAAPI since ffmpeg has control over the resources.

rbrohman · 2011-09-14, 20:09

I'm using vdpau rather than vaapi, so I'm definitely interested in testing out your solution. Please let me know how I can help test this.

Thanks

**wsnipex** · 2011-09-15, 10:07

FernetMenta Wrote:I think the problem you are describing is similar to that I have investigated for vdpau. Decoder resources must not cleared down as long they are referenced by open gl. In case of vdpau this only happens if refresh rate is changing but for VAAPI this can occur even when a playback is stopped.
I have an experimental solution for vdpau but it looks like it's even more complicated for VAAPI since ffmpeg has control over the resources.

Did anyone test with the newest ffmpeg? The version used in xbmc is rather old..
Is there a ffmpeg version bump in sight?

evil.oz · 2011-10-07, 07:54

i have the exact same problem on two identical sandybridge machines (i5, intel mobo)...

its totally NON-XMBC related

if i use onboard video i experience graphic corruption and system lock, very similar to your screenshot
if i use addin card (both ati or nvidia), the system simply deadly locks

this happens once every 10 days aprox. it is NOT related with temperature or workload.

os is w7-64, used in a graphics studio (adobe suite mainly). have other w7-64 machines with same apps that do behave properly.

afaik no software (even the worse one) should be able to lock down a machine in this way, so i presume its a hardwre problem.

i am planning to swap out all components to narrow/track down the problem.
in order psu/ram/hdd.

rbrohman, did you get any more clues bout this?