Kodi Community Forum
v18 LibreELEC Testbuilds for x86_64 (Kodi 18.0) - Printable Version

Kodi Community Forum (https://forum.kodi.tv)
+-- Forum: Support (https://forum.kodi.tv/forumdisplay.php?fid=33)
+--- Forum: General Support (https://forum.kodi.tv/forumdisplay.php?fid=111)
+---- Forum: Linux (https://forum.kodi.tv/forumdisplay.php?fid=52)
---- Thread: v18 LibreELEC Testbuilds for x86_64 (Kodi 18.0) (/showthread.php?tid=298462)



RE: LibreELEC Testbuilds for x86_64 (Kodi 18.0) - HomerJau - 2018-12-03

I’ve noticed recent network lagging too. Intel NUCs. Seems to start around a week or so ago.

Edit: Maybe around the time Linux went to .19?


LibreELEC Testbuilds for x86_64 (Kodi 18.0) - Milhouse - 2018-12-03

Post the kernel log using dmesg | pastebinit.

There has been reports of network performance issues in this thread, introduced wth 4.19-rc1, and there is a possible matching bug report but it doesn't look like anyone with the problem in this thread is interested in posting additional details to the bug so I'm not expecting a fix any time soon.


RE: LibreELEC Testbuilds for x86_64 (Kodi 18.0) - jmerrilljr - 2018-12-03

(2018-12-03, 01:56)Milhouse Wrote: Post the kernel log using dmesg | pastebinit.

There has been reports of network performance issues in this thread, introduced wth 4.19-rc1, and there is a possible matching bug report but it doesn't look like anyone with the problem in this thread is interested in posting additional details to the bug so I'm not expecting a fix any time soon.
So, stick with Milhouse build #1022. It works. The last stable LibreELEC builds work too. It looks like the LibreELEC Alpha builds also have this issue although I only saw one bug report.


RE: LibreELEC Testbuilds for x86_64 (Kodi 18.0) - J.L.C. - 2018-12-03

(2018-12-03, 01:56)Milhouse Wrote: Post the kernel log using dmesg | pastebinit.

There has been reports of network performance issues in this thread, introduced wth 4.19-rc1, and there is a possible matching bug report but it doesn't look like anyone with the problem in this thread is interested in posting additional details to the bug so I'm not expecting a fix any time soon.
 I just reflashed LibreELEC-Generic.x86_64-9.0-Milhouse-20181201210245-#1201-g5dc77c9

Here is a link to the kernel log:
http://ix.io/1vay


RE: LibreELEC Testbuilds for x86_64 (Kodi 18.0) - Milhouse - 2018-12-03

(2018-12-03, 21:39)J.L.C. Wrote:
(2018-12-03, 01:56)Milhouse Wrote: Post the kernel log using dmesg | pastebinit.

There has been reports of network performance issues in this thread, introduced wth 4.19-rc1, and there is a possible matching bug report but it doesn't look like anyone with the problem in this thread is interested in posting additional details to the bug so I'm not expecting a fix any time soon.
 I just reflashed LibreELEC-Generic.x86_64-9.0-Milhouse-20181201210245-#1201-g5dc77c9

Here is a link to the kernel log:
http://ix.io/1vay

Yes, you have the same network related issue that is present since 4.19-rc1.

There is a bug for this, but nobody from this thread has posted anything relevant such as a before/after 4.19-rc1 dmesg logs. And no, a Kodi log is not relevant for a kernel issue.


RE: LibreELEC Testbuilds for x86_64 (Kodi 18.0) - Milhouse - 2018-12-04

New LibreELEC.tv Leia build #1203: Generic
(Supercedes previous build)

SHA256 Checksum: a698c8ee5d29229d22035bebbd7a42dcdd3888c352530f9733234659ed7626a7 (Generic)

# uname -a
Linux NUC 4.19.6 #1 SMP Mon Dec 3 21:36:35 GMT 2018 x86_64 GNU/Linux

# lsb_release
LibreELEC (Milhouse): devel-20181203213420-#1203-gbf8d34d [Build #1203]

# Kodi version
(18.0-RC3 Git:1de0d40). Platform: Linux x86 64-bit

Based on tip of LibreELEC.tv master (bf8d34d, changelog) and tip of XBMC master (1de0d40, changelog) with the following modifications: Build Highlights:
  1. bump to v18.0 rc3
  2. fix libcec and Samsung TVs (ping keydon + thanks!)
Build Details:
  1. LibreELEC.tv:
    • iptables: fix build with 3.10.y, remove hack (PR:3144, 1 commit, 2 files changed)
    • libssh: update to libssh-0.8.5 (PR:3138, 1 commit, 1 file changed)
  2. XBMC:
    • bump to v18.0 rc3 (PR:14978, 1 commit, 1 file changed)
  3. libcec:
    • fixed: only prevent TV polls when a Samsung TV is detected instead of suppressing all logical addresses. issue #424 #444 (f759324)



RE: LibreELEC Testbuilds for x86_64 (Kodi 18.0) - keydon - 2018-12-04

(2018-12-04, 00:19)Milhouse Wrote: New LibreELEC.tv Leia build #1203:
[...]Build Details:
  • libcec:
    • fixed: only prevent TV polls when a Samsung TV is detected instead of suppressing all logical addresses. issue #424 #444 (f759324)
 
I can happily confirm, the libcec-fix is working for me.
Thank you for the excellent bugfix-experience.


RE: LibreELEC Testbuilds for x86_64 (Kodi 18.0) - wheemer - 2018-12-04

please delete.


RE: LibreELEC Testbuilds for x86_64 (Kodi 18.0) - drebbe - 2018-12-04

Milhouse, I know we've been trying your patience on this network problem using a 4.19 kernel, but I'm way out of my comfort/competency zone modifying syslinux.cfg via vi. Frankly, I wouldn't know how to begin to report a linux kernel bug but here is my best attempt to provide, as requested, two complete/untruncated (dmesg | pastebinit with log_buf_len=16M appended to the kernel command line)  kernel logs of 1016 and 1016x .
Perhaps someone more knowledgeable than me can take it from here.

http://ix.io/1vgP - 1016 kernel log

http://ix.io/1vgR - 1016x kernel log

If I've at least managed to do this first step correctly, I'll do my best to do what you ask to further diagnose the issue. 
If I've done it incorrectly, my apologies for wasting more of your time. Regardless, thanks for all you do.


RE: LibreELEC Testbuilds for x86_64 (Kodi 18.0) - Milhouse - 2018-12-05

@drebbe there's really nothing I can do, as it's a kernel bug/regression. I linked the bug in a previous post, someone needs to get involved with the kernel developers and post them their logs etc. as I don't have this issue myself. I don't want to get involved myself for that reason, as they'll generally ask me for information I don't have. If/when a fix is proposed I'm more than able to create a test build that includes it.


RE: LibreELEC Testbuilds for x86_64 (Kodi 18.0) - Milhouse - 2018-12-05

New LibreELEC.tv Leia build #1204: Generic
(Supercedes previous build)

SHA256 Checksum: 8c94c38737c73a58daa85f5a7c158364ca2ee1687829480231b16dee110b1c02 (Generic)

# uname -a
Linux NUC 4.19.6 #1 SMP Tue Dec 4 21:57:45 GMT 2018 x86_64 GNU/Linux

# lsb_release
LibreELEC (Milhouse): devel-20181204215607-#1204-g1f14e69 [Build #1204]

# Kodi version
(18.0-RC3 Git:1de0d40). Platform: Linux x86 64-bit

Based on tip of LibreELEC.tv master (1f14e69, changelog) and tip of XBMC master (1de0d40, changelog) with the following modifications: Build Highlights:
  1. Minors
Build Details:
  1. LibreELEC.tv:
    • fix .config/firmware handling (PR:3141, 2 commits, 3 files changed)
    • libretro-mgba: updated to git 38bad71 (PR:3134, 1 commit, 1 file changed)
  2. Additional commits/pull requests/changes not yet merged upstream:
    • Updated: [env] PR:3130 (perma): linux (RPi/Generic): update to linux-4.19.6



RE: LibreELEC Testbuilds for x86_64 (Kodi 18.0) - drebbe - 2018-12-05

(2018-12-05, 00:29)Milhouse Wrote: @drebbe there's really nothing I can do, as it's a kernel bug/regression. I linked the bug in a previous post, someone needs to get involved with the kernel developers and post them their logs etc. as I don't have this issue myself. I don't want to get involved myself for that reason, as they'll generally ask me for information I don't have. If/when a fix is proposed I'm more than able to create a test build that includes it.
 Ok. Thanks, I get it. Again, sincerely, thanks for all you do.


RE: LibreELEC Testbuilds for x86_64 (Kodi 18.0) - Matt Devo - 2018-12-05

(2018-12-05, 00:29)Milhouse Wrote: @drebbe there's really nothing I can do, as it's a kernel bug/regression. I linked the bug in a previous post, someone needs to get involved with the kernel developers and post them their logs etc. as I don't have this issue myself. I don't want to get involved myself for that reason, as they'll generally ask me for information I don't have. If/when a fix is proposed I'm more than able to create a test build that includes it.
 if you're able to revert this commit, would be an easy way to test:  
https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?id=a99790bf5c7f3d68d8b01e015d3212a98ee7bd57

it looks like there's a few subsequent commits after which may make reversion require some massaging. If it's too much trouble, I'll try to build 8.90.008 here with that reverted and see how it goes


RE: LibreELEC Testbuilds for x86_64 (Kodi 18.0) - Milhouse - 2018-12-05

@matt Devo yes there have been a lot of changes in the area of r8169 which makes it difficult to revert commits from almost 6 months ago. If you have the time, bisecting 4.18.14...4.19-rc1 might be the best approach.


RE: LibreELEC Testbuilds for x86_64 (Kodi 18.0) - meowmoo - 2018-12-05

same Network Issue with 4.19 Kernel on my end.

4.19 http://ix.io/1vim
4.18 http://ix.io/1vil

I posted both on the linked Bug report, hope it helps