Login at Kodi Home

quintesse · 2012-11-17, 19:50

Upgraded XBMCbuntu Eden to Frodo Beta 1 (adding PPA and doing apt-get dist-upgrade).
Everything works fine (except for album art problem reported by others).
The only problem I'm having (which worked before) is streaming from a UPnP source, XBMC will crash immediately... or maybe not crash because the crash log doesn't show anything interesting, it might just exit because there's no stack trace in the log:

############### STACK TRACE #################
############# END STACK TRACE ###############

The only lines that show in the og at the moment of the crash are:

18:15:01 T:2896165744 NOTICE: Thread CGUIDialogCache start, auto delete: true
18:15:02 T:3078407952 WARNING: CAEUtil::GuessChLayout - This method should really never be used, please fix the code that called this
18:15:02 T:2896165744 NOTICE: Thread CFileCache start, auto delete: false
18:15:02 T:3078407952 WARNING: CAEUtil::GuessChLayout - This method should really never be used, please fix the code that called this

I can of course post the full log if so desired.

un1versal · (This post was last modified: 2012-11-17, 19:59 by un1versal.)

Hello

First pastebining a full debug log is not optional Rolleyes

its sort of mandatory as well as full crashlog (they are two separate logs) and ppl need to see all of it
Second those errors are AE (audioEngine) related Nothing to do with uPnP

uNI

quintesse · (This post was last modified: 2012-11-17, 20:22 by quintesse.)

Ok, first of all it seems installing "gdb" to get the crash log wasn't enough, for some reason I had to reboot the system first because in this new log there *is* a stack trace:

http://pastebin.ca/2252466

And the debug log: http://paste.ubuntu.com/1365615/
And the old debug log: http://paste.ubuntu.com/1365640/

quintesse · 2012-11-17, 20:17

And I'm not saying it has anything to do with UPnP but it's the correct description of the moment when the application crashes: accessing a UPnP stream. All other sources I have (using SMB) work fine for pictures, music and video.

un1versal · (This post was last modified: 2012-11-17, 20:22 by un1versal.)

yes its AE look at https://github.com/xbmc/xbmc/pull/1779

See if you can compile with that, and if it fixes it and report

you can also add streamsilece to your advancedsettings.xml

http://wiki.xbmc.org/index.php?title=AudioEngine

Something funky is going on with your AE for sure.

Code:
19:12:25 T:3078289168  NOTICE:         m_channels        : FL,FR,BL,BR,FC,LFE,SL,SR,UNKNOWN1,UNKNOWN2,UNKNOWN3,UNKNOWN4,UNKNOWN5,UNKNOWN6,UNKNOWN7,UNKNOWN8

along with the AE-utils error idk suspicious.

uNi

un1versal · (This post was last modified: 2012-11-17, 20:41 by un1versal.)

FYI your AE doesnt even initialize

for e.g.

Code:
43:30 T:3047798592   DEBUG: CSoftAE::InternalOpenSink - ALSA Initialized:

43:30 T:3047798592   DEBUG:   Output Device : HDA NVidia

43:30 T:3047798592   DEBUG:   Sample Rate   : 44100

43:30 T:3047798592   DEBUG:   Sample Format : AE_FMT_S32NE

43:30 T:3047798592   DEBUG:   Channel Count : 2

43:30 T:3047798592   DEBUG:   Channel Layout: FL,FR

43:30 T:3047798592   DEBUG:   Frames        : 512

43:30 T:3047798592   DEBUG:   Frame Samples : 1024

43:30 T:3047798592   DEBUG:   Frame Size    : 8

43:30 T:3047798592   DEBUG: CSoftAE::InternalOpenSink - Using speaker layout: 2.0

43:30 T:3047798592   DEBUG: CSoftAE::InternalOpenSink - Internal Buffer Size: 4096

is missing from your log. (that was taken from mine as a example)

You could set only 2.0 speaker to see if this changes but your AE is funky, like a pair of old smelly socks

uNi

quintesse · 2012-11-17, 21:32

It does, you might have been looking in the old log where debugging was turned on halfway through, but in the "new" log it shows:

19:12:26 T:3078289168 INFO: CSoftAE::InternalOpenSink - sink incompatible, re-starting
19:12:26 T:3078289168 INFO: CAESinkALSA::Initialize - Attempting to open device "default"
19:12:26 T:3078289168 INFO: CAESinkALSA::Initialize - Opened device "default"
19:12:26 T:3078289168 INFO: CAESinkALSA::InitializeHW - Your hardware does not support AE_FMT_FLOAT, trying other formats
19:12:26 T:3078289168 INFO: CAESinkALSA::InitializeHW - Using data format AE_FMT_S32NE
19:12:26 T:3078289168 DEBUG: CAESinkALSA::InitializeHW - Request: periodSize 512, periods 16, bufferSize 8192
19:12:26 T:3078289168 DEBUG: CAESinkALSA::InitializeHW - Got: periodSize 512, periods 16, bufferSize 8192
19:12:26 T:3078289168 DEBUG: CAESinkALSA::InitializeHW - Setting timeout to 186 ms
19:12:26 T:3078289168 DEBUG: CSoftAE::InternalOpenSink - ALSA Initialized:
19:12:26 T:3078289168 DEBUG: Output Device : Default (HDA NVidia HDMI 0)
19:12:26 T:3078289168 DEBUG: Sample Rate : 44100
19:12:26 T:3078289168 DEBUG: Sample Format : AE_FMT_S32NE
19:12:26 T:3078289168 DEBUG: Channel Count : 2
19:12:26 T:3078289168 DEBUG: Channel Layout: FL,FR
19:12:26 T:3078289168 DEBUG: Frames : 512
19:12:26 T:3078289168 DEBUG: Frame Samples : 1024
19:12:26 T:3078289168 DEBUG: Frame Size : 8
19:12:26 T:3078289168 DEBUG: CSoftAE::InternalOpenSink - Using speaker layout: 5.1
19:12:26 T:3078289168 DEBUG: CSoftAE::InternalOpenSink - Internal Buffer Size: 4096
19:12:26 T:3075918704 NOTICE: Thread CSoftAE start, auto delete: false
19:12:26 T:3075918704 INFO: CSoftAE::Run - Thread Started
19:12:26 T:3075918704 INFO: CAESinkALSA - ALSA: pcm_hw.c:587 Sad

snd_pcm_hw_start) SNDRV_PCM_IOCTL_START failed (-32): Broken pipe

un1versal · (This post was last modified: 2012-11-17, 21:39 by un1versal.)

Anyway...

I still think its AE refer to my posts for a possible fix, else, wait for someone else, to provide feedback.

uNi

quintesse · (This post was last modified: 2012-11-17, 21:41 by quintesse.)

BTW sound works perfectly for music and video with digital audio, the receiver correctly detects digital audio and switches to full 5.1 sound. So it's not that things don't work, just when streaming through UPnP (from a source that hasn't changed and used to work before).

Well I'm compiling right now, but it'll still take some time. But anyway thanks for your help!

quintesse · 2012-11-18, 22:31

Ok, just tried the latest version from master, with and without the patch you pointed out before and no difference whatsoever.
The error also does not seem related at all to the problem talked about in the issue, I don't think this is a problem with devices disconnecting or anything or I would have the same problem with other forms of getting the data to the speakers / on the screen.
This part of the crash log:

Program terminated with signal 8, Arithmetic exception.
#0 0x087a07ee in PAPlayer::QueueData(PAPlayer::StreamInfo*) ()

seems to point to something stream related in the PAPlayer, which seems more logical, right?

un1versal · 2012-11-18, 23:08

One of the devs that does upnp should have a gander. I cant ping anyone afaik.

uNi

quintesse · 2012-11-19, 16:28

NB just opened a ticket for this: http://trac.xbmc.org/ticket/13579

As I say in the comments there I figured out that it must be related somehow to the format of the audio stream, when I force the format to MP3 in the source things work just fine.