Login at Kodi Home

darzur · 2014-09-17, 07:40

(2014-09-16, 20:10)mikeb93 Wrote: My TV still turns itself on after sending the pi to suspend mode. Heres a log:
i hope theres something in there. I turned off my TV at around 17.20, came home 3 hours later finding my TV turned on and the skin says "working..." so there must have been something that wakes the pi and turns on the TV.

Noticed that too, but in my case it's KODI that restarts itself. Since maybe two updates KODI (only KODI, not entire OpenElec) restarts randomly. Then, if it was suspended, it brings TV out of standby.

mikeb93 · 2014-09-17, 08:09

(2014-09-17, 07:40)darzur Wrote:
(2014-09-16, 20:10)mikeb93 Wrote: My TV still turns itself on after sending the pi to suspend mode. Heres a log:
i hope theres something in there. I turned off my TV at around 17.20, came home 3 hours later finding my TV turned on and the skin says "working..." so there must have been something that wakes the pi and turns on the TV.

Noticed that too, but in my case it's KODI that restarts itself. Since maybe two updates KODI (only KODI, not entire OpenElec) restarts randomly. Then, if it was suspended, it brings TV out of standby.

I've had this problem for months now... This is really annoying..

FAMMAR · (This post was last modified: 2014-09-17, 08:20 by FAMMAR.)

(2014-09-17, 08:09)mikeb93 Wrote:
(2014-09-17, 07:40)darzur Wrote:
(2014-09-16, 20:10)mikeb93 Wrote: My TV still turns itself on after sending the pi to suspend mode. Heres a log:
i hope theres something in there. I turned off my TV at around 17.20, came home 3 hours later finding my TV turned on and the skin says "working..." so there must have been something that wakes the pi and turns on the TV.

Noticed that too, but in my case it's KODI that restarts itself. Since maybe two updates KODI (only KODI, not entire OpenElec) restarts randomly. Then, if it was suspended, it brings TV out of standby.

I've had this problem for months now... This is really annoying..

Could it be CEC settings in XBMC?

Go to System > System > Input Devices > Peripherals > CEC Adapter >Disable/Ignore When TV is turned off and put also the 3e and 4e Line on TV

Milhouse · (This post was last modified: 2014-10-28, 02:24 by Milhouse.)

(2014-09-17, 07:40)darzur Wrote: Noticed that too, but in my case it's KODI that restarts itself. Since maybe two updates KODI (only KODI, not entire OpenElec) restarts randomly. Then, if it was suspended, it brings TV out of standby.

(2014-09-17, 08:09)mikeb93 Wrote: I've had this problem for months now... This is really annoying..

For those with unexplained crashes/restarts, I've uploaded a recent debug build based on #0915 (although the build code is #0916, meh).

*DEBUG* download link (245MB)

To test:

You'll need a 512MB Pi with a FAT partition of at least 300MB. Install the debug build as normal, copying the tar to your .update/Update folder and reboot.
Reduce gpu_mem to 128 or as low as you can live with while still able to reproduce the problem. It may be possible for gdb to load all the symbols with higher gpu_mem and a dose of swap (128MB+) but expect it to be quite slow.
Shut down and then restart Kodi with gdb:

Code:
systemctl stop kodi gdb -iex "thread apply all bt" -ex run --args /usr/lib/kodi/kodi.bin --standalone -fs --lircdev /run/lirc/lircd
When kodi.bin eventually crashes, control will return to gdb. Now is the time to print the backtrace. In gdb, type:

Code:
bt

Here's an example gdb session: http://pastebin.com/0MtvQUeY - the "bt" command is entered at line 133.

Copy and paste your gdb session to pastebin.com or xbmclogs.com.

Once a backtrace is available it should give some clues (maybe even all the clues) to understand what is happening and enable a developer to work on a fix.

Without a backtrace it's very unlikely there will be a fix for random crashes.

Edit: Fix commands due to rebranding

darzur · 2014-09-17, 16:53

OK, installed it but it's completely unusable with gdb active. I'll leave it overnight and see what will happen.

Milhouse · 2014-09-17, 17:05

Unusable in what sense - stuck loading symbols? Does xbmc even start (blue bubbles, menu etc.)? As long as you have enough free RAM to begin with, I find running under gdb very little different to running a non-debug build. I have a 512MB Pi with 128MB swap and have no problems running gdb when setting gpu_mem=128.

darzur · (This post was last modified: 2014-09-17, 17:18 by darzur.)

It's painfully slow, but it works. Starting Youtube addon took 5 minutes. My Pi is 512mb too, gpu_mem=128 and 128 swap.

P.S. Or maybe it needed some time to fully start, because now it looks to working almost normally. Only suspend doesn't turn off/on my TV through CEC.

menakite · (This post was last modified: 2014-09-17, 20:44 by menakite.)

Sorry if I hijack this thread - I'm not actually using OpenELEC - but it seems the more appropriate place to discuss issues with the newclock4 branch (also, others can test and confirm if they experience the same behavior).

While debugging an issue I'm experiencing with PR 5237, I've noticed that when switching from a 3D mode (the movie I'm testing is half SbS, if that matters) to monoscopic - either via the OSD or via a remote button mapped to StereoModeToMono - I actually get 2 images on the TV (side-by-side), although I get the notification that the mode was changed (from CStereoscopicsManager::ApplyStereoMode). Switching again to monoscopic fixes it. Sometimes, but very rare, I don't need to toggle it 2 times. Does anybody experience the same? (or: Is this a know issue?)

spannerz · 2014-09-17, 22:51

Is anybody on this thread using any the latest builds with VU+ *and* running more than one profile?

Everything works absolutely fine on a single profile, but the moment I change profiles VU+ PVR throws up the generic "Connection Lost" error and the PVR refuses to make any attempt to connect on the network (I've run TCPDUMP's on the VU+ box to make sure it isn't getting anything)

The debug files don't give away much, but the following few lines seem to be most relevant (note: this isn't a network issue, in fact killing the XBMC process "fixes" this until I change profiles again)

Code:
27:10 535.148682 T:2909750352   ERROR: ADDON: Dll VU+ / Enigma2 Client - Client returned bad status (1) from Create and is not usable

27:10 535.154419 T:2909750352    INFO: Called Add-on status handler for '1' of clientName:VU+ / Enigma2 Client, clientID:pvr.vuplus (same Thread=no)

27:10 535.155029 T:2909750352 WARNING: UpdateAndInitialiseClients - failed to create add-on VU+ / Enigma2 Client, status = 1

2

This is driving me nuts as I can't move from the old builds onto this lot (this is the main TV and I use profiles to keep the kids stuff separate) - I've tried a variety of builds, all with the same issue. I've also tried upgrades and fresh installs.

I've posted in the PVR forum but haven't had much back.... any pointers would be great (there are a few more logs in that post, but they don't give much away)

Cheers

Sz

Milhouse · 2014-09-18, 01:17

@spannerz: Based on the information you posted in your other VU+ thread, I see it is generating an exception during the socket open - could it be a TCP IPv4/IPv6 issue? Who is the author/maintainer of this addon?

Other than that, you need to go back through the builds and find the last working version - that may provide some clue as to what is going on, or what has changed.

Milhouse · 2014-09-18, 06:13

New OpenELEC Helix build: #0917
(Supercedes previous build)

Code:
# uname -a

Linux rpi512 3.16.2 #1 PREEMPT Thu Sep 18 00:28:58 BST 2014 armv6l GNU/Linux

# vcgencmd version

Sep 15 2014 20:47:23 

Copyright (c) 2012 Broadcom

version 750d0f48c8d332ac1782f36915036fa80081f939 (clean) (release)

# lsb_release

OpenELEC (Milhouse) - Version: devel-20140918002746-r19243-g9e8c0f5 [Build #0917]

Based on tip of OpenELEC master (9e8c0f57, changelog) and tip of XBMC master (b6fe3b89, changelog) with the following modifications:

Includes newclock4 patches
Excludes the OpenELEC fernetmenta patches due to conflicts with newclock4
Excludes the OpenELEC linux-01-RPi_support patch in favour of sourcing these and possibly more recent patches directly from kernel branch rpi-3.16.y
Excludes the OpenELEC xbmc-001-newclock4 patch in favour of sourcing these and possibly more recent patches directly from newclock4 branch
Default setting for "Show RSS Feed" changed to disabled
Disabled "Total Duration" in Confluence (see build #0221 for details)
Adapted service.openelec.settings to take advantage of PR5217
Includes latest libnfs master (ab1af695)
Includes latest libcec master (81e38211)
Includes latest xbmc-pvr-addons master (dcda159b)
Includes latest xbmc-addon-xvdr master (a3724213)
Includes libcec double-key suppression
Includes libcec CEC Standby Fix
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 CONFIG_COREDUMP=y to allow creation of coredumps (see here)
Includes OpenELEC linker pull requests: PR:3479, PR:3480 and PR:3481
Includes PR5222: videodb: Add column name on GROUP query otherwise sub query returns no rows
Includes PR5237: [WIP][3D] add support for stereomode switches in video playlist
Includes PR5247: StringUtils: optimize Split(), new overloads for Split() and Tokenize()
Includes PR5307: [UX] Play button/action should not pause
Includes PR5312: dvdplayer: do a seek after having changed subtitle stream
Includes PR5324: Add setting to hide library clean progress bar.

Build Highlights:

Dual audio support added:

popcornmix Wrote:Dual audio for pi sink is now pushed for testing.
I've changed the settings, so now there is "HDMI", "Analogue" and ""HDMI and Analogue" as the choices of output device
(this is because we want AE to close and reopen the sink when the setting changes which the "dual audio" toggle doesn't trigger).

Build Details:

XBMC:
- [dvdplayer] Fix for timestamp correction when no audio or no video (PR:5345, 1 commit, 2 files changed)
- ffmpeg: bump to 2.4 (PR:5371, 2 commits, 5 files changed)
newclock4:
- [PiSink] Initial dual audio support

Milhouse · (This post was last modified: 2014-09-19, 01:17 by Milhouse.)

New OpenELEC Helix build: #0918b
(Supercedes previous build)

Code:
# uname -a

Linux rpi512 3.16.2 #1 PREEMPT Thu Sep 18 21:04:39 BST 2014 armv6l GNU/Linux

# vcgencmd version

Sep 15 2014 20:47:23 

Copyright (c) 2012 Broadcom

version 750d0f48c8d332ac1782f36915036fa80081f939 (clean) (release)

# lsb_release

OpenELEC (Milhouse) - Version: devel-20140918232542-r19243-g9e8c0f5 [Build #0918b]

Based on tip of OpenELEC master (9e8c0f57, changelog) and tip of XBMC master (83d9b055, changelog) with the following modifications:

Includes newclock4 patches
Excludes the OpenELEC fernetmenta patches due to conflicts with newclock4
Excludes the OpenELEC linux-01-RPi_support patch in favour of sourcing these and possibly more recent patches directly from kernel branch rpi-3.16.y
Excludes the OpenELEC xbmc-001-newclock4 patch in favour of sourcing these and possibly more recent patches directly from newclock4 branch
Default setting for "Show RSS Feed" changed to disabled
Disabled "Total Duration" in Confluence (see build #0221 for details)
Adapted service.openelec.settings to take advantage of PR5217
Includes latest libnfs master (ab1af695)
Includes latest libcec master (81e38211)
Includes latest xbmc-pvr-addons master (dcda159b)
Includes latest xbmc-addon-xvdr master (a3724213)
Includes libcec double-key suppression
Includes libcec CEC Standby Fix
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 CONFIG_COREDUMP=y to allow creation of coredumps (see here)
Includes OpenELEC linker pull requests: PR:3479, PR:3480 and PR:3481
Includes PR5222: videodb: Add column name on GROUP query otherwise sub query returns no rows
Includes PR5237: [WIP][3D] add support for stereomode switches in video playlist
Includes PR5247: StringUtils: optimize Split(), new overloads for Split() and Tokenize()
Includes PR5307: [UX] Play button/action should not pause
Includes PR5309: [posix] Implement new PosixFile
Includes PR5312: dvdplayer: do a seek after having changed subtitle stream
Includes PR5324: Add setting to hide library clean progress bar.
Includes PR5381: Fix access after free
Includes PR5382: [fix] uninitialized variable in GUIDialogAddonInfo
Includes PR5386: Fix incorrect usage of S_IFDIR flag

Build Highlights:

Added PR5381 which should prevent Python-related crashes
Added PR5309 which may improve file access performance and resource usage
#0918b: Revert "rbp: Use new dispmanx function for vsync callbacks" due to journal spam

Build Details:

XBMC:
- XBMC core Portuguese translation based on English commit 1ee638d8b22a3a5440e57a9384590558925e12ef (PR:1, 1 commit, 1 file changed)
- replace skin.touched with skin.re-touched (PR:5314, 5 commits, 19 files changed)
- [win32][Fix] Adapt loading of ffmpeg dlls after the bump. (PR:5380, 1 commit, 1 file changed)
- ActiveAE: initialize engine stats before first usage, credits to jimcarroll and Valgrind
newclock4:
- [omxplayer] Only enable audio clock master when A/V sync method is set to audio clock
- [rbp] Enable vsync handler all the time
- [mmal] Avoid busy spinning when no work to do
- (#0918b) Revert "rbp: Use new dispmanx function for vsync callbacks"
Custom:
- Added PR5309: [posix] Implement new PosixFile
- Added PR5381: Fix access after free
- Added PR5382: [fix] uninitialized variable in GUIDialogAddonInfo
- Added PR5386: Fix incorrect usage of S_IFDIR flag

Milhouse · (This post was last modified: 2014-09-19, 00:12 by Milhouse.)

There may be a problem with build #0918 - my systemd journal is full of the following although xbmc appears to be working OK (idle CPU load is much higher which is what tipped me off):

Code:
Sep 18 22:09:42 rpi512 kernel: vchiq: vchiq_ioctl: cmd VCHIQ_IOC_RELEASE_SERVICE returned error -1 for service UPDH:432

Sep 18 22:09:42 rpi512 kernel: ------------[ cut here ]------------

Sep 18 22:09:42 rpi512 kernel: WARNING: CPU: 0 PID: 439 at drivers/misc/vc04_services/interface/vchiq_arm/vchiq_arm.c:2481 vchiq_release_internal+0xc0/0x268()

Sep 18 22:09:42 rpi512 kernel: Modules linked in: snd_soc_pcm512x_i2c snd_soc_pcm512x snd_soc_bcm2708_i2s snd_soc_wm8804 regmap_mmio regmap_spi regmap_i2c snd_soc_core snd_compress snd_pcm_dmaengine snd_pcm snd_timer snd bcm2708_rng

Sep 18 22:09:42 rpi512 kernel: CPU: 0 PID: 439 Comm: HDispmanx Notif Tainted: G        W     3.16.2 #1

Sep 18 22:09:42 rpi512 kernel: Backtrace:

Sep 18 22:09:42 rpi512 kernel: [<c001167c>] (dump_backtrace) from [<c0011954>] (show_stack+0x18/0x1c)

Sep 18 22:09:42 rpi512 kernel:  r6:c04695cc r5:00000009 r4:00000000 r3:00000000

Sep 18 22:09:42 rpi512 kernel: [<c001193c>] (show_stack) from [<c0642528>] (dump_stack+0x20/0x28)

Sep 18 22:09:42 rpi512 kernel: [<c0642508>] (dump_stack) from [<c001e35c>] (warn_slowpath_common+0x70/0x90)

Sep 18 22:09:42 rpi512 kernel: [<c001e2ec>] (warn_slowpath_common) from [<c001e420>] (warn_slowpath_null+0x24/0x2c)

Sep 18 22:09:42 rpi512 kernel:  r8:cea25548 r7:c0a31dc4 r6:cf2bde18 r5:c09eb21c r4:ce84ce44

Sep 18 22:09:42 rpi512 kernel: [<c001e3fc>] (warn_slowpath_null) from [<c04695cc>] (vchiq_release_internal+0xc0/0x268)

Sep 18 22:09:42 rpi512 kernel: [<c046950c>] (vchiq_release_internal) from [<c046a2d0>] (vchiq_ioctl+0x84c/0x185c)

Sep 18 22:09:42 rpi512 kernel:  r9:ce84ce00 r8:00000008 r7:0000c40c r6:00003002 r5:c09eb21c r4:0000c40d

Sep 18 22:09:42 rpi512 kernel: [<c0469a84>] (vchiq_ioctl) from [<c00c95b0>] (do_vfs_ioctl+0x404/0x5dc)

Sep 18 22:09:42 rpi512 kernel:  r10:00000000 r9:cf2bc000 r8:00000008 r7:ce84d1e0 r6:ce84d1e0 r5:cea51d40

Sep 18 22:09:42 rpi512 kernel:  r4:00003002

Sep 18 22:09:42 rpi512 kernel: [<c00c91ac>] (do_vfs_ioctl) from [<c00c97c4>] (SyS_ioctl+0x3c/0x64)

Sep 18 22:09:42 rpi512 kernel:  r9:cf2bc000 r8:00000008 r7:0000c40d r6:ce84d1e0 r5:00003002 r4:ce84d1e1

Sep 18 22:09:42 rpi512 kernel: [<c00c9788>] (SyS_ioctl) from [<c000e760>] (ret_fast_syscall+0x0/0x30)

Sep 18 22:09:42 rpi512 kernel:  r8:c000e8e4 r7:00000036 r6:00000000 r5:00000000 r4:b655e36c r3:00000062

Sep 18 22:09:42 rpi512 kernel: ---[ end trace 954399c5b5b88742 ]---

Investigating...

Edit: It's a problem caused by the "[rbp] Enable vsync handler all the time" commit. Hopefully I can upload a new build shortly.

Milhouse · 2014-09-19, 01:18

Last release updated with build #0918b.

spannerz · 2014-09-19, 20:52

(2014-09-18, 01:17)Milhouse Wrote: @spannerz: Based on the information you posted in your other VU+ thread, I see it is generating an exception during the socket open - could it be a TCP IPv4/IPv6 issue? Who is the author/maintainer of this addon?

Other than that, you need to go back through the builds and find the last working version - that may provide some clue as to what is going on, or what has changed.

Cheers for the response.... I've already dropped jdembski a PM, hopefully he will turn something around fairly quick.

All the best and keep up the good work

Cheers

Sz