Kodi Community Forum

Full Version: LibreELEC Testbuilds for RaspberryPi (Kodi 17.0)
You're currently viewing a stripped down version of our content. View the full version with proper formatting.
(2016-10-19, 10:12)Milhouse Wrote: [ -> ]Not a FULL debug log, you've snipped useful and potentially relevant information. No idea if what remains is likely to solve the issue.

Sorry, I am not sure I fully understand. I haven't snipped anything at all. I started the "full debug" option via the kodi remote and opened the file and after it started playing, I hit "fast forward". The screen froze, and after a minute or so, I hit "stop". After about 7 or 8 minutes, it finally stopped the movie and went back to the previous screen. The only lines I edited were the name of the movie. I didn't realize that the name was relevant. Do you want me to enable the debug log and restart Kodi and go through the entire sequence? Will that provide more information?
You must have done something else to it. The complete log file header is missing ... but yeah - turn on debugging, restart kodi and reproduce your issue.
(2016-10-19, 10:12)Milhouse Wrote: [ -> ]Not a FULL debug log, you've snipped useful and potentially relevant information. No idea if what remains is likely to solve the issue.

(2016-10-19, 17:45)Memphiz Wrote: [ -> ]You must have done something else to it. The complete log file header is missing ... but yeah - turn on debugging, restart kodi and reproduce your issue.

Ok, I enabled debug log using Advanced Settings and restarted my Pi3 and went through the entire sequence of starting the movie and hitting "fast forward", etc. Here is the log - http://pastebin.com/H6jM1Dyc

Nothing in the log has been edited or snipped. Let me know if this works. Once again, thanks much for looking into this.
milhouse maybe you can help me with my question. i recently tried to run libreelec completly on usb, which the rpi3 allows. but after booting from usb i only have 120mb ram aviable... after switching back to sd card i get 753mb. i have no clue why this happen.
New LibreELEC.tv Krypton build #1019: RPi / RPi2
(Supercedes previous build)

Code:
# uname -a
Linux rpi512 4.8.2 #1 Wed Oct 19 21:05:02 BST 2016 armv6l GNU/Linux

# vcgencmd version
Oct 13 2016 17:21:14
Copyright (c) 2012 Broadcom
version b301dbab300729c7871442ca50f65c669446af20 (clean) (release)

# lsb_release
LibreELEC (Milhouse) - Version: devel-20161019210358-#1019-g6a0a361 [Build #1019]

# Kodi version
Kodi (17.0-BETA4 Git:b4acb43). Platform: Linux ARM 32-bit

Based on tip of LibreELEC.tv master (6a0a3616, changelog) and tip of XBMC master (b4acb432, changelog) with the following modifications: Build Highlights:
  1. VideoPlayer: fix aspect ratio for matroska containers for stereo modes
  2. PR:10731: Change splash db upgrade logic
Build Details:
  1. LibreELEC.tv:
    • linux: backport Avermedia TD110 support to iMX6 4.4 kernel (6a0a3616)
  2. XBMC:
    • Fix deadlock introduced by #10690 (PR:10726, 1 commit, 1 file changed)
    • Fix Music Standard Roles Nodes (PR:10717, 1 commit, 7 files changed)
    • [peripheral API] Fix unitialized variable (f4cb6526)
  3. peripheral.joystick:
    • [cmake] Cleanup and modernize (e5ee9467)
    • Remove unused member variable in CSettings (c89326e3)
  4. newclock5:
    • New commits in this build:
      • VideoPlayer: fix aspect ratio for matroska containers for stereo modes (762a727c)
  5. Additional commits/pull requests/changes not yet merged upstream:
    • Added: [pkg] PR:10731: Change splash db upgrade logic
(2016-10-19, 21:31)mezo Wrote: [ -> ]milhouse maybe you can help me with my question. i recently tried to run libreelec completly on usb, which the rpi3 allows. but after booting from usb i only have 120mb ram aviable... after switching back to sd card i get 753mb. i have no clue why this happen.

The firmware in these builds (specifically, bootcode.bin, I believe) doesn't yet fully support the network and USB boot functionality - this is still considered experimental. If you're using firmware from branch=next then the 120MB issue has already been fixed, but may remain unfixed in the current branch=master firmware. You could try using the bootcode.bin from branch=next, and everything else from these builds, but if you have issues then wait until net/USB boot changes land in master.
(2016-10-20, 00:04)Milhouse Wrote: [ -> ]
(2016-10-19, 21:31)mezo Wrote: [ -> ]milhouse maybe you can help me with my question. i recently tried to run libreelec completly on usb, which the rpi3 allows. but after booting from usb i only have 120mb ram aviable... after switching back to sd card i get 753mb. i have no clue why this happen.

The firmware in these builds (specifically, bootcode.bin, I believe) doesn't yet fully support the network and USB boot functionality - this is still considered experimental. If you're using firmware from branch=next then the 120MB issue has already been fixed, but may remain unfixed in the current branch=master firmware. You could try using the bootcode.bin from branch=next, and everything else from these builds, but if you have issues then wait until net/USB boot changes land in master.

thanks for your replay. i replaced start.elf and bootcode.bin from the next branch with the files from libreelec (would there even a usb possible without that?). or does libreelec somehow overwrite it with the old files on boot?
I was running build 1010 for a few days now and tonight I noticed a weird issue - one of the CPU cores - "CPU 2" was stuck at 100% load. No idea what triggered this or how to reproduce. I ran "top" via ssh, kodi.bin was using ~120% of the cpu.
(2016-10-19, 20:38)the_jaguar Wrote: [ -> ]
(2016-10-19, 10:12)Milhouse Wrote: [ -> ]Not a FULL debug log, you've snipped useful and potentially relevant information. No idea if what remains is likely to solve the issue.

(2016-10-19, 17:45)Memphiz Wrote: [ -> ]You must have done something else to it. The complete log file header is missing ... but yeah - turn on debugging, restart kodi and reproduce your issue.

Ok, I enabled debug log using Advanced Settings and restarted my Pi3 and went through the entire sequence of starting the movie and hitting "fast forward", etc. Here is the log - http://pastebin.com/H6jM1Dyc

Nothing in the log has been edited or snipped. Let me know if this works. Once again, thanks much for looking into this.

I just observed another strange thing - if I disable "Decode the stereo stream from 3D files", the movie plays in 2D, but I am able to "fast forward" without any issues. I find this really weird...
(2016-10-20, 00:08)mezo Wrote: [ -> ]thanks for your replay. i replaced start.elf and bootcode.bin from the next branch with the files from libreelec (would there even a usb possible without that?). or does libreelec somehow overwrite it with the old files on boot?

LibreELEC will overwrite the firmware from branch=next each time you upgrade so you'll need to restore the branch=next bootcode.bin and start.elf after every upgrade (these are the two files that allow the RPi3 to boot from USB). However as start.elf includes the GPU firmware you're basically installing non-standard GPU firmware that doesn't include all the updates from branch=master, which is the version we're testing in this thread, so expect problems nobody using the branch=master firmware can reproduce.

As such, please do not be report issues (playback/deinterlace problems, random crashes, GPU hangs etc.) in this thread if your GPU firmware is from branch=next.

Edit: it's possible you need the bootcode.bin/start.elf firmware from branch=next only to program the "program_usb_boot_mode" OTP bit, after which the current firmware provided by LE may allow you to boot from USB, but with only 120MB as that issue is not yet fixed in branch=master. Depending on whether the 120MB issue is in bootcode.bin or start.elf, try replacing only bootcode.bin from branch=next and see if that fixes the memory issue, but avoid using start.elf from branch=next with these test builds - if you do you're on your own... Smile
(2016-10-20, 05:06)smp1 Wrote: [ -> ]I was running build 1010 for a few days now and tonight I noticed a weird issue - one of the CPU cores - "CPU 2" was stuck at 100% load. No idea what triggered this or how to reproduce. I ran "top" via ssh, kodi.bin was using ~120% of the cpu.
These kind of issues are not easy to pin down. If you run "top -Hp $(pidof kodi.bin)" you should see all the threads for the kodi.bin process - if the thread creating the CPU load is called LanguageInvoker then it's an add-on causing the problem (unfortunately can't say which add-on). If it's kodi.bin itself then using strace on the kodi.bin process might yield some clues (strace can be installed from one of the LE add-ons, System Tools I think).
(2016-10-20, 08:43)Milhouse Wrote: [ -> ]
(2016-10-20, 00:08)mezo Wrote: [ -> ]thanks for your replay. i replaced start.elf and bootcode.bin from the next branch with the files from libreelec (would there even a usb possible without that?). or does libreelec somehow overwrite it with the old files on boot?

LibreELEC will overwrite the firmware from branch=next each time you upgrade so you'll need to restore the branch=next bootcode.bin and start.elf after every upgrade (these are the two files that allow the RPi3 to boot from USB). However as start.elf includes the GPU firmware you're basically installing non-standard GPU firmware that doesn't include all the updates from branch=master, which is the version we're testing in this thread, so expect problems nobody using the branch=master firmware can reproduce.

As such, please do not be report issues (playback/deinterlace problems, random crashes, GPU hangs etc.) in this thread if your GPU firmware is from branch=next.

Edit: it's possible you need the bootcode.bin/start.elf firmware from branch=next only to program the "program_usb_boot_mode" OTP bit, after which the current firmware provided by LE may allow you to boot from USB, but with only 120MB as that issue is not yet fixed in branch=master. Depending on whether the 120MB issue is in bootcode.bin or start.elf, try replacing only bootcode.bin from branch=next and see if that fixes the memory issue, but avoid using start.elf from branch=next with these test builds - if you do you're on your own... Smile
i tried it a second time by replacing bootcode.bin/start.elf from branch=next, but somehow i get only 120mb ram after.
(2016-10-20, 12:27)mezo Wrote: [ -> ]i tried it a second time by replacing bootcode.bin/start.elf from branch=next, but somehow i get only 120mb ram after.

Yes, same (after copying start_x.elf from branch=next to /flash/start.elf).

What happens if you only replace bootcode.bin with the version from branch=next, and keep the current LE start.elf? You should have the full memory allocation, what I'm not sure of is whether USB booting will be possible with the start.elf from branch=master.
(2016-10-20, 15:28)Milhouse Wrote: [ -> ]What happens if you only replace bootcode.bin with the version from branch=next, and keep the current LE start.elf? You should have the full memory allocation, what I'm not sure of is whether USB booting will be possible with the start.elf from branch=master.

I believe the answer is no - the "next" branch has significant changes to both bootcode.bin and start.elf.
(2016-10-18, 18:23)popcornmix Wrote: [ -> ]The VC-1 and MPEG-2 files are detected as deinterlaced correctly.
They look fine to me. The file is a deinterlace torture test, so it is intended to be hard to deinterlace.
We produce the same output as, say VLC using yadif x2 (which is considered a very good deinterlacer).
Yadif is something software? The quality corresponds to a average level. Something like hardware "adaptive" on amd. There is "motion-adaptive", which is much better and "vector-adaptive", which is very much better.
VLC never used, only tried. IMO this is the one of most bad players ever Smile
Thank you for the answer.