•   
  • 1
  • 33
  • 34
  • 35(current)
  • 36
  • 37
  • 156
  •   
  Thread Closed
OpenELEC Testbuilds for RaspberryPi Part 3 (Kodi 14.0)
(2014-06-22, 23:00)MilhouseVH Wrote:
(2014-06-22, 22:39)freaktm Wrote: Until tonight, now XBMC won't start after a reboot.

I get this error (or something similar, the text shows for a very short duration!) when Openelec is loading:

Code:
failed to setup user config dir
Check systemctl status userconfig.service

This would suggest your /storage filesystem is no longer readable/writeable. Reformat the /storage partition and restore a backup, or if you don't have a backup then you'll have to go through the setup process again.

It is very much readable. It's located on an USB stick - is there some way for me to diagnose it from SSH, or do I need to do that from another PC? I've tried running fsck, but even after unmounting /storage, it's still busy.

Edit: I see, that fsck on boot was introduced in #0607 - is there a way for me to manually add this single feature right now, or should I just try to do the update from my PC, to the latest build?
Image
(2014-06-21, 12:44)popcornmix Wrote:
(2014-06-21, 00:51)pootler Wrote: You have to laugh though, I am now back with Millhouse helix build ( the latest ), and Live tv today had serious audio/video drift!

Do you only get audio drift with live TV?
Ideally I would like you to record a file that contains the audio drift.
Then the behaviour would be reproducible, and I could download the file and investigate it.

Have you ever had this issue with a recording from live TV?


OK - I waited for the current build ( helix ) to drift which it did after about an hour this time.
The drift was very noticeable.
Quickly recorded the channel for ten minutes, and then played it back.
Surprise, surprise, when playing back the recorded clip, IT WAS IN PERFECT SYNC!

Does that help?

Pootler
(2014-06-23, 14:56)pootler Wrote: OK - I waited for the current build ( helix ) to drift which it did after about an hour this time.
The drift was very noticeable.
Quickly recorded the channel for ten minutes, and then played it back.
Surprise, surprise, when playing back the recorded clip, IT WAS IN PERFECT SYNC!

Does that help?

My understanding is that a "corrupt" packet puts the audio out of sync.
Recording after the sync has gone wrong is too late (it's just like stopping and restarting live tv - it will be back in sync).
The recording needs to start when video is in sync, and catch the point where sync gets broken.
Hopefully the file will then repeatably go out of sync at the same point each time it is played and can be debugged.
How do I run fsck on boot?
(2014-06-24, 04:11)JustinBeaird Wrote: How do I run fsck on boot?

In the latest builds it will run automatically. You'll see nothing unless you have a serious filesystem error. Minor errors will be fixed automatically, sometimes requiring a reboot, but when a more serious error is detected you'll be dropped into a debug shell where you can try to fix the errors yourself.
Texture Cache Maintenance Utility: Preload your texture cache for optimal UI performance. Remotely manage media libraries. Purge unused artwork to free up space. Find missing media. Configurable QA check to highlight metadata issues. Aid in diagnosis of library and cache related problems.
(2014-06-23, 09:02)freaktm Wrote: Edit: I see, that fsck on boot was introduced in #0607 - is there a way for me to manually add this single feature right now, or should I just try to do the update from my PC, to the latest build?

If you want to debug the boot process then add "debug break=all" to the end of your line in cmdline.txt and use a keyboard connected to the Pi to work through the boot steps until /storage is mounted to see what shape it's in.

Or you can add "textmode" to your line in cmdline.txt and that will give you a text console in OpenELEC where you can investigate /storage.

Alternatively, fsck the USB drive in a PC running Linux, or the Pi while running Raspbian.

Falling all of that then yes, manually update your SD card with the latest build.
Texture Cache Maintenance Utility: Preload your texture cache for optimal UI performance. Remotely manage media libraries. Purge unused artwork to free up space. Find missing media. Configurable QA check to highlight metadata issues. Aid in diagnosis of library and cache related problems.
After a lot of debugging, I found that the USB drive had locked up completely in read-only mode.
Trying a lot of different low-level USB tools, which should be able to fix the issue (resetting firmware, stuff like that), with no luck, I gave up and copied the contents to a new USB key. So all is good again.
Next time I reboot, I'll be a little more cautious with just pulling the plug on the Pi Smile
Image
Should Airplay-Video be working?
I definitely enabled Airplay in the settings... but I can't seem to find this functionality
being offered in the options of my Android Tab, LG Gpad 8.3 (e.g. with Youtube).

Any clues?
Edit: Not even working from my iphone 5.
What was the last build when Airplay worked successfully? How is it in official 4.0.5? Are you following the Airplay thread - the YouTube app seems to be broken, and a number of other users are reporting problems with video on various platforms.
Texture Cache Maintenance Utility: Preload your texture cache for optimal UI performance. Remotely manage media libraries. Purge unused artwork to free up space. Find missing media. Configurable QA check to highlight metadata issues. Aid in diagnosis of library and cache related problems.
(2014-06-20, 01:46)popcornmix Wrote: There is a hack you can use if you are having problems. Use dvdplayer to navigate the menus and start the feature.
Stop, then context menu, and "play using omxplayer". It will offer to resume from where you were and you can play the feature with the generally better behaved omxplayer.

I tried that (after hacking in a mapping for the context menu using Lircmap.xml), but there was no "Play using..." menu. Furthermore, selecting the Resume option simply navigated into the top-level directory, showing the AUDIO_TS/VIDEO_TS structure, not actually playing anything. But that gave me an idea.

A brief switch to omxplayer in advancedsettings.xml shows that omxplayer does indeed play the video much much better - if you can actually get past the menu system into the feature. No stutter, and I'm sure it looks higher quality even ignoring that. Trickplay is still a bit flaky, with unregulated speed and both forward/backwards skips, but again a bit better than dvdplayer I think. It does still suffer from the close-to-end-of-stream blackouts. (Near the end of a video sequence, the picture suddenly goes black for a second or so, then recovers and displays the final little bit of the video just before the end. Audio continues perfectly during this. You notice it mainly on the copyright/studio splashes before the menu, but it does affect the main feature too - if you bother to watch right to the end of the credits. My wild guess is that the end-of-stream notification is overtaking buffered video data, both in the GPU pipe and xbmc-side, and causing the GPU pipe to flush, then when the right timestamp arrives the remaining xbmc-side buffer which wasn't flushed gets passed in, decoded and played. I've no idea whether I've guessed the decode architecture right there though...)

So, switching back to dvdplayer, I played, passed the menu, stopped, navigated down through the top-level directory, into VIDEO_TS, and hit context menu on VIDEO_TS.IFO itself, where there *was* a Play Using... option. First attempt had no resume option though and just started from the beginning and into the menu again. Second attempt I skipped forwards in dvdplayer before stopping, then I did get a resume option, then skip back to 00:00:00 once omxplayer had started.

What a faff! But the resulting playback is good enough that I can probably inflict this on guests this week!

(More so given that the real DVD player is dying a death and now skips all over the place itself...)

(2014-06-20, 01:46)popcornmix Wrote: Sure, that shouldn't be necessary and I'd like to improve dvdplayer.
What is always useful is samples of files that clearly misbehave. Upload a sample, explain exactly what you press and how it misbehaves and I'll investigate.
I don't tend to watch dvd images. I have a few for testing but they seem to behave pretty well.

I'll try and put a minimal set of files to demonstrate the menu sticking I mentioned earlier - that I've not seen mentioned anywhere else yet. I've just turned on debug logging and during the pause the following gets output: http://xbmclogs.com/show.php?id=234382 . At :24 I played the DVD, at :37 selected the play item from the DVD menu, at :45 the sound cut out, at :56 the screen went blank, at :07 the feature started playing. 30 seconds from click to result.

The dvdplayer stuttering is well attested to, and I get the impression others are hearing the same audio "veeping" I am too, so presumably any old DVD would do if you can rip it to a folder structure. I can probably sort something out if you still need a demo piece though.
(2014-06-25, 02:17)jsxbmc Wrote: The dvdplayer stuttering is well attested to, and I get the impression others are hearing the same audio "veeping" I am too, so presumably any old DVD would do if you can rip it to a folder structure. I can probably sort something out if you still need a demo piece though.

Would be useful. I've got a couple of dvd images that do seem to play fine.
It's always best to try to reproduce exactly the issue that is being reported, and having the same sample file gives the best chance of that happening.
Post-mortem from last night:

I followed the procedure I gave above. You do have to skip forwards in dvdplayer before the Resume option becomes available. It's not clear where the cutoff is, but under a minute appears not to be long enough, 7-8 minutes is, so in practice a chapter skip forwards in dvdplayer, then chapter skip back in omxplayer to return to the beginning works.

The film then played fine until about 12 minutes in. Then the video froze. 10-20 seconds later the audio stopped. Bringing up the OSD showed the current play time as 26 minutes in (which it definitely wasn't.) This happened on the two different films we watched, with pretty much the same timings on both.

So I resorted to just using omxplayer from the beginning, having to drive the menu system "blind", then both films played fine all the way through. It's not that menus don't work *at all* in omxplayer, and least not usually. Though it does sometimes hang on the copyright/studio splashes before the menu appears and be unable to proceed, that is highly disk dependent. It's just that it either doesn't display the menu video at all, or corrupts it badly, and doesn't display button highlights, so you can't actually see where you're navigating/what option your selecting. But if you already know the menu layout it's usually just about possible to get through it. (And it works in dvdplayer so you can at least use that to work out the right sequence of buttons to press.)

(Actually, before trying the resume method I had a concern about using the Resume function to switch players: how likely is that to work with alternative splicing, such as disks with theatrical and directors cuts, where it adds/removes scenes from the single contiguous video stream as it goes depending on which main menu option you chose. It seems that that could only work if it preserves the entire DVD VM state, and not just the resume position, and can share that between the two different players. I'm pretty sure I've seen dvdplayer mess up the sequencing even without stop/resume, where rewinding past a chapter/splice point then playing doesn't quite backtrack the way you came, or follow the same path though the video second time round.)
New OpenELEC Helix build: #0627
(Supercedes previous build)

Code:
# uname -a
Linux rpi512 3.15.1 #1 PREEMPT Fri Jun 27 00:49:19 BST 2014 armv6l GNU/Linux

# vcgencmd version
Jun 26 2014 21:03:47
Copyright (c) 2012 Broadcom
version 87aee38aa98517e53651d3400bf9261beb20c3e7 (clean) (release)

# lsb_release
OpenELEC_Helix (Milhouse) - Version: devel-20140627011044-r18647-g610ee50

Based on tip of OpenELEC master (610ee50, changelog) and tip of XBMC master (b104278, changelog) with the following modifications:
  • Includes newclock3 patches
  • Excludes the OpenELEC fernetmenta patches due to conflicts with newclock3
  • Excludes the OpenELEC linux-01-RPi_support patch in favour of sourcing these and possibly more recent patches directly from kernel branch rpi-3.15.y
  • Excludes the OpenELEC xbmc-master-newclock3 patch in favour of sourcing these and possibly more recent patches directly from newclock3 branch
  • Default setting for "Show RSS Feed" changed to disabled
  • Disabled "Total Duration" in Confluence (see build #0221 for details)
  • ALSA is enabled and any suitable hardware should be available in XBMC Audio Settings
  • FIQ FSM is enabled by default. See initial post for more details.
  • Includes RPi-specific libCEC commits from mk01/libcec. See build #0527 for details. Also double-key suppression.
  • Increase scan interval of PeripBusCEC from 5000 to 60000, reducing CPU loading by about 2% (1GHz Pi) every 5 seconds (even when CEC is "disabled")
  • Includes libnfs master with directory caching support
Build Highlights:

New firmware. Experimental anaglyph support:
popcornmix Wrote:I've added experimental anaglyph rendering. Not very well tested, but worked a couple of vids. Have you got any red/cyan or green/magenta 3d glasses?

  1. Firmware (Jun 26):
    • effects: Initial code for 3d sbs to anaglyph
      video_decode: Fix nBufferSize setting getting lost. See: #240 (comment)
      video_render: Add command for querying render statistics. See: #287 (comment)
      hvs: Add gencmd to configure only updating display list on odd/even field. See: #292
  2. OpenELEC:
    • cmake: update to cmake-3.0.0, fix installation
    • kmod: update to kmod-18
    • misc-firmware: update to misc-firmware-0.0.11
    • alsa-lib: update to alsa-lib-1.0.28
    • alsa-utils: update to alsa-utils-1.0.28
    • ffmpeg: update to ffmpeg-xbmc-717a0b5
    • llvm: update to llvm-3.4.2
    • libaacs: update to libaacs-0.7.1
    • libva-intel-driver: update to libva-intel-driver-1.3.2
    • mpc: update to mpc-1.0.2
    • cloog: update to cloog-0.18.2
    • samba: update to samba-3.6.24
    • Mesa: update to Mesa-10.2.2
    • service.openelec.settings (Gotham): update to service.openelec.settings-0.3.18
    • service.openelec.settings (Helix): update to service.openelec.settings-0.4.1
    • libbluray: update to libbluray-0.6.0
  3. XBMC:
    • fix: Cosmetic OverlayWatched alignment on select (PR:4939, 1 commit, 1 file changed)
    • Aefixes (PR:4943, 4 commits, 20 files changed)
    • [osx/depends] - add a fake Info.plist to target python26 for allowing codesign... (PR:4926, 2 commits, 2 files changed)
    • [epg] skip -/+ of 12 hours with SkipPrevious/SkipNext (PR:4910, 1 commit, 1 file changed)
    • prompt for master pin in case the addon manager is locked (PR:4927, 1 commit, 3 files changed)
  4. newclock3:
    • [rbp] Resume video output on startup
    • [omxplayer] Experimental support for anaglyph rendering of 3d videos
  5. kernel 3.15.y:
    • vmstat: Workaround for issue where dirty page count goes negative
    • dwc_otg: Fix various issues with root port and transaction errors
    • fiq_fsm: Implement hack for Split Interrupt transactions
Texture Cache Maintenance Utility: Preload your texture cache for optimal UI performance. Remotely manage media libraries. Purge unused artwork to free up space. Find missing media. Configurable QA check to highlight metadata issues. Aid in diagnosis of library and cache related problems.
MilhouseVH: server seems to be down.
  •   
  • 1
  • 33
  • 34
  • 35(current)
  • 36
  • 37
  • 156
  •   
  Thread Closed
 
Thread Rating:
  • 8 Vote(s) - 4.88 Average



Logout Mark Read Team Forum Stats Members Help
OpenELEC Testbuilds for RaspberryPi Part 3 (Kodi 14.0)4.888