View Full Version : [VU+ Ultimo4K] 5.3.030 crashing. Fixed in 5.3 032
Nasty looking crash after couch flashing 5.3.030 and restoring settings from 5.3.029
The box boots up ok until the programme info bar appears and then crashes before live tv is shown.
The GUI restarts and the box is ok, but if I reboot it crashes again with the same sequence. Same crash starting from deep standby.
ET10K couch flashed ok.
Tried re-flashing, no better.
< 30.6001> 15:08:41.5856 [Avahi] timeout elapsed
< 30.6003> 15:08:41.5858 [Avahi] client state: 2
< 30.6005> 15:08:41.5859 [Avahi] avahi_timeout_new
< 30.6022> 15:08:41.5877 [Avahi] avahi_timeout_free
< 30.6025> 15:08:41.5880 [Avahi] avahi_timeout_new
< 30.6040> 15:08:41.5895 [Avahi] avahi_timeout_free
< 30.6043> 15:08:41.5898 [Avahi] avahi_timeout_new
< 30.6056> 15:08:41.5910 [Avahi] avahi_timeout_free
< 30.6062> 15:08:41.5917 [Avahi] avahi_timeout_new
< 30.6079> 15:08:41.5934 [Avahi] avahi_timeout_free
< 30.6081> 15:08:41.5935 [Avahi] avahi_timeout_new
< 30.6088> 15:08:41.5943 [Avahi] avahi_timeout_free
< 30.6089> 15:08:41.5943 [Avahi] avahi_timeout_update
< 30.6089> 15:08:41.5944 [Avahi] Registered vuultimo4k (_e2stream._tcp) on vuultimo4k:8001
< 30.6090> 15:08:41.5945 [Avahi] avahi_timeout_new
< 30.6114> 15:08:41.5969 [Avahi] avahi_timeout_free
< 30.6116> 15:08:41.5971 [Avahi] avahi_timeout_new
< 30.6127> 15:08:41.5982 [Avahi] avahi_timeout_free
< 30.6135> 15:08:41.5990 Backtrace:
< 30.6137> 15:08:41.5992 /usr/bin/enigma2(_Z17handleFatalSignaliP9siginfo_tPv) [0x1853A4]
< 30.6138> 15:08:41.5993 /lib/libc.so.6(__default_rt_sa_restorer) [0xB6CACE20]
< 30.6139> 15:08:41.5993 /lib/libc.so.6(gsignal) [0xB6CABFB4]
< 30.6139> 15:08:41.5994 /lib/libc.so.6(abort) [0xB6C9A200]
< 30.6139> 15:08:41.5994 -------FATAL SIGNAL (6)
Same problem with no settings restored.
Exits the wizard ok, but a reboot crashes.
Need to repeat to get hold of the crash log, not stored on hdd by default.
Kind of indicates hardware error, its really early in initialisation - but maybe if the settings backup was from 029, re- flash and don't restore settings as 029 had some incorrect information in SystemInfo.
.... you beat me to it!
Go back to 028/029 and see what happens ....... there are only 3 changes between 029 & 030
Post #2, continued.
Couch flashed 5.3.029 + settings restore, all fine.
Crash and debug logs now attached.... maybe a clue?
dbus[1703]: arguments to dbus_message_iter_append_basic() were incorrect, assertion "_dbus_check_is_valid_utf8 (*string_p)" failed in file ../../dbus-1.12.10/dbus/dbus-message.c line 2754.
This is normally a bug in some application using the D-Bus library.
D-Bus not built with -rdynamic so unable to print a backtrace
Post #2, continued.
Couch flashed 5.3.029 + settings restore, all fine.
Crash and debug logs now attached.... maybe a clue?
dbus[1703]: arguments to dbus_message_iter_append_basic() were incorrect, assertion "_dbus_check_is_valid_utf8 (*string_p)" failed in file ../../dbus-1.12.10/dbus/dbus-message.c line 2754.
This is normally a bug in some application using the D-Bus library.
D-Bus not built with -rdynamic so unable to print a backtrace
Cannot see anything in OE-A that could cause this - (its a pity you don't have a multiboot box!)
Try this: save these 3 modules: /usr/lib/enigma2/python/Screens/InfobarGenerics.pyo and separately /usr/lib/enigma2/python/RecordTimer.pyo and /usr/lib/enigma2/python/Components/EpgListGrid.pyo
Then reflash and see if problem re -occurs. If it does 1st copy over teh InforbarGenerics.pyo (rename the existing one) and reboot. ANy difference??
If not copy over the other 2(having renamed the old ones) and reboot .
OK - I'll have to restore settings to get networking up quickly, but it shouldn't affect the outcome.
BrianTheTechieSnail
18-05-20, 16:52
Same here.
Upgrade from 5.3.029 to 5.3.030 and after reboot device crashes and recovers (automatically restarts GUI maybe, not sure).
Reboot again - same crash and recover after boot.
restore image of 5.3.029 - back to normal.
I didn't investigate yet whether anything wasn't working after crash and recover.
Cannot see anything in OE-A that could cause this - (its a pity you don't have a multiboot box!)
Try this: save these 3 modules: /usr/lib/enigma2/python/Screens/InfobarGenerics.pyo and separately /usr/lib/enigma2/python/RecordTimer.pyo and /usr/lib/enigma2/python/Components/EpgListGrid.pyo
Then reflash and see if problem re -occurs. If it does 1st copy over teh InforbarGenerics.pyo (rename the existing one) and reboot. ANy difference??
If not copy over the other 2(having renamed the old ones) and reboot .Changing all 3 has made no difference, still crashes the same.
Have to say I am bemused - so few changes between the 2 releases 029/030
Have to say I am bemused - so few changes between the 2 releases 029/030
Fortunately, I'm not the only one (with the same problem).
I use openmultiboot plugin on my uno4k when testing images.
Added vix 5.3.030 to openmultiboot with no restore no crash then restored settings and still fine.
So switched to my flash image that had vix 5.2.029 and did a software update to 5.3.030 and no issues with crashes.
Possibly box issues? Any specific driver updates between images?
Possibly box issues? Any specific driver updates between images? Very unlikely, drivers are over a year old, and I'm assuming @BrianTheTechHead's box is a zgemma, so no.
Solo2 did online update from 29 to 30 and no issues noticed yet.
I've had a Fatal signal before but it's been a very, very long time (touch wood) and wasn't, at least for me, easily creatable again (so more of a one-off).
Solo2 did online update from 29 to 30 and no issues noticed yet.
My ET10K was ok as well, couch flashed like the Ultimo4K, which crashes (just the once, the GUI restart doesn't see it again, which might be a clue?)
This crash happens ever time.
Just did an online update from 5.3.029 to 5.3.030 and..........
It still crashes.
Also the tried and tested "switch off at the mains for 5 minutes" didn't help.
On the offchance this is something to do with me, can you try replacing /usr/lib/enigma2/python/RecordTimer.py with this
https://raw.githubusercontent.com/SimonCapewell/enigma2/62b879a227d765b0edb9182ffecbe16e255d16dd/RecordTimer.py
On the offchance this is something to do with me, can you try replacing /usr/lib/enigma2/python/RecordTimer.py with this
https://raw.githubusercontent.com/SimonCapewell/enigma2/62b879a227d765b0edb9182ffecbe16e255d16dd/RecordTimer.py
OK will take 5 minutes or so.
On the offchance this is something to do with me, can you try replacing /usr/lib/enigma2/python/RecordTimer.py with this
https://raw.githubusercontent.com/SimonCapewell/enigma2/62b879a227d765b0edb9182ffecbe16e255d16dd/RecordTimer.py
I wish I could say it was, but it still crashes.
Is this commit a pure coincidence, or......
https://github.com/OpenPLi/enigma2/commit/e1923b7c0d0304fe76218a13bf2d968fc2960bcd
Nasty looking crash after couch flashing 5.3.030 and restoring settings from 5.3.029
The box boots up ok until the programme info bar appears and then crashes before live tv is shown.
The GUI restarts and the box is ok, but if I reboot it crashes again with the same sequence. Same crash starting from deep standby.
ET10K couch flashed ok.
Tried re-flashing, no better.
ET10K
Software update from 29 to 30 - no crashes
Couch flash from 29 to 30 with settings and plugin restore from 29 - no crashes (My own manual 29 backup and not one produced as a result upgrade/flash)
Is this commit a pure coincidence, or......
https://github.com/OpenPLi/enigma2/commit/e1923b7c0d0304fe76218a13bf2d968fc2960bcd
No we don‘t have it on ViX.... but it kind of fits - but why does 019 work?
BrianTheTechieSnail
18-05-20, 21:30
I'm assuming @BrianTheTechHead's box is a zgemma
Correct, Zgemma H7S.
ET10K
Software update from 29 to 30 - no crashes
Couch flash from 29 to 30 with settings and plugin restore from 29 - no crashes (My own manual 29 backup and not one produced as a result upgrade/flash)
I know, just like mine, mentioned in post #1.
Any Ultimo4k owners (or H7S's) out there now running 5.3.030?
I've just software updated a H7C from 5.3.027 to 5.3.030.
abu baniaz
19-05-20, 00:09
Is this commit a pure coincidence, or......
https://github.com/OpenPLi/enigma2/commit/e1923b7c0d0304fe76218a13bf2d968fc2960bcd
That commit was added to fix this crash reported by IMS, which seems to be the same as the one being reported.
Backtrace:
/usr/bin/enigma2(_Z17handleFatalSignaliP9siginfo_tPv) [0xA4DDC]
/lib/libc.so.6(__default_rt_sa_restorer) [0xB6C4FC60]
/lib/libc.so.6(gsignal) [0xB6C4E94C]
/lib/libc.so.6(abort) [0xB6C4FDC4]
-------FATAL SIGNAL
Reason was
This issue appears with the new "speedup" from netifaces.
Parsing the output from ip/ifconfig commands didn't show that issue.
Why it appears now? I have no idea, but the flow is now the same as before, now we are starting some services earlier / faster.
If creating an std::string and passing the const c_str to enigma fixes the issue that is enough for me.
thecaretaker
19-05-20, 04:56
Same problem here. Here is my log...
60228
That commit was added to fix this crash reported by IMS, which seems to be the same as the one being reported.
Backtrace:
/usr/bin/enigma2(_Z17handleFatalSignaliP9siginfo_tPv) [0xA4DDC]
/lib/libc.so.6(__default_rt_sa_restorer) [0xB6C4FC60]
/lib/libc.so.6(gsignal) [0xB6C4E94C]
/lib/libc.so.6(abort) [0xB6C4FDC4]
-------FATAL SIGNAL
Reason was
This issue appears with the new "speedup" from netifaces.
Parsing the output from ip/ifconfig commands didn't show that issue.
Why it appears now? I have no idea, but the flow is now the same as before, now we are starting some services earlier / faster.
If creating an std::string and passing the const c_str to enigma fixes the issue that is enough for me.So you reckon this commit may well fix my (and other posters) crashes?
EDIT:
The PLi discussion mentioned this....
Before we never had the chance to see that error, simply because we where starting OWIF toooooooooo late.
Now OWIF starts tooooooooo fast
so I deleted OWIF and the box no longer crashes.
so I deleted OWIF and the box no longer crashes.
This is NOT a solution, I was just testing out a theory.:)
This is NOT a solution, I was just testing out a theory.:)
Well the OpenPli change has been added to ViX so next build ..............................
Well the OpenPli change has been added to ViX so next build ..............................
Good to hear, I enjoy the PLi forum, probably for the wrong reasons :p , but it's also a good memory test, and that commit somehow rang the proverbial bell.
6023160232
Hope these may be helpful
I am having issues with box booting then restarting a few times before I can even change the channel after a few restarts it works ok , I am using 0.29
Updating to 5.3.030 caused crashing with log when starting but continued starting and seemed to work alright with settings from 5.3.029. Updating showed a warning at first and so it is. Crash but goes on starting alright.
I flashed back to 5.3.029 with settings from 5.3.029, no crashing. I keep that, so far.
Will there be any remedy in the future updates?
abu baniaz
20-05-20, 06:22
Of course there will be a fix. We have just run a Dev image to test the "fix". Once that is confirmed fine, the 5.3 031 Release image will be run. Not all boxes have this issue.
Sorry for the inconvenience.
I noticed my Ultimo4K did the same when I updated to 30, I've not rebooted since so haven't noticed any further GUI restarts. If you still need logs, I can try a reboot but can't do it until later in the Day.
There are plenty of logs submitted, all exactly the same, so I wouldn't bother unless you think your crash is different.
A fix should be in the next release.
Build 5.3.031 is running, most likely be completed tomorrow.
BrianTheTechieSnail
20-05-20, 15:42
I updated to 5.3.031 and it still seems broken to me.
It crashed on first boot after upgrade but not on a second (manually requested) one but it did crash on a third.
Maybe I upgraded too soon?
60237
60238
I updated to 5.3.031 and it still seems broken to me.
It crashed on first boot after upgrade but not on a second (manually requested) one but it did crash on a third.
Maybe I upgraded too soon?
No idea what you have upgraded to! 5.3.031 hasn't even completed building and has not even been uploaded! You must be on a different timezone ahead of us mere mortals :p
BrianTheTechieSnail
20-05-20, 15:56
It seems wishful thinking + short sightedness took over and I just upgraded to 030 again?
Feel free to make me feel like an idiot.
:(
I've just initiated Time Warp and upgraded to build 31 - OK, so I compiled it myself :)
I can report on an Ultimo4K it did not crash the GUI on first start or on a subsequent reboot
BrianTheTechieSnail
20-05-20, 19:58
I think this may be why I thought I has 5.3.031.
60239
I think this may be why I thought I has 5.3.031.
60239
The changlogs get updated when a build starts.
BrianTheTechieSnail
20-05-20, 22:49
The changlogs get updated when a build starts.
Is that just to embarrass people like me or is there some other reason I can't imagine at the moment?
abu baniaz
20-05-20, 23:09
Is that just to embarrass people like me or is there some other reason I can't imagine at the moment?
You made an incorrect assumption, he posted something to clarify what happens so that your understanding as well as other people increases. Most people would have just left it at that, clicked on the thanks button or posted "thanks for that" type of post.
To find out what image you are on.
Menu > Information > About
BrianTheTechieSnail
21-05-20, 00:04
You made an incorrect assumption, he posted something to clarify what happens so that your understanding as well as other people increases. Most people would have just left it at that, clicked on the thanks button or posted "thanks for that" type of post.
To find out what image you are on.
Menu > Information > About
If I' ve gone to Check for update is it not reasonable to expect that I'll be shown information about what I'm about to update to and not a future version that's not available yet?
You MUST surely at least admit that would be a reasonable assumption?
abu baniaz
21-05-20, 00:16
When are you going to supply a fix for this?
BrianTheTechieSnail
21-05-20, 00:27
When are you going to supply a fix for this?
Er. I didn't know I had access to your build scripts.
abu baniaz
21-05-20, 00:42
https://github.com/OpenViX
BrianTheTechieSnail
21-05-20, 01:46
https://github.com/OpenViX
Doesn't seem to be any documentation to speak of.
Several change logs with literally nothing except a list of version numbers in them.
What am I not spotting?
abu baniaz
21-05-20, 01:58
Our files are there. It is a public repository, you have access to it. To help you on your way, the relevant file is here:
https://github.com/OpenViX/enigma2/blob/master/lib/python/Screens/GitCommitInfo.py
Please provide the fix required to solve the issue that you are currently highlighting. We look forward to your solution with great anticipation.
BrianTheTechieSnail
21-05-20, 02:51
Our files are there. It is a public repository, you have access to it. To hep you on your way, the relevant file is here:
https://github.com/OpenViX/enigma2/blob/master/lib/python/Screens/GitCommitInfo.py
Please provide the fix required to solve the issue that you are currently highlighting. We look forward to your solution in great anticipation.
That looks like part of OpenVIX itself.
Not a single comment anywhere in that file.
Does it not get compiled to pyo or something?
Would comments really take too much space?
Maybe it's fixable there but it seems to me that what's needed is for that code to look for version info in a place other than the version info for the sources.
You need to make this piece of code look for release notes on the same server as the binary image files and make the script that uploads the binary images also upload the release notes immediately after the images, or maybe even in the same zip file.
abu baniaz
21-05-20, 03:24
Great analysis and ideas.
There are various threads/posts on forums on how to work with py/pyo files in Enigma2 receivers.
When you do submit your pull request with the fix, please send to the Dev branch.
Just to confirm that the problem is now fixed in the official release of 5.3.032
Thanks.
Powered by vBulletin® Version 4.2.5 Copyright © 2024 vBulletin Solutions Inc. All rights reserved.