It’s unlikely to be this though if you say it’s only hot after a crash.
I see a kernel BUG in your dmesg. Just to confirm: do you have the exact same setup and software installed when you ran the February version as you did with the March version?
Exactly the same, when i’ve tried to update from february, install the march from sd, installed smb and transmission from the store, tried the packages from ssh, no problems with the old kernel… Now i’m on february with the kernel suggested from you wnd get crash… I made a pattern and each installation take it the same way!
You can add it to /etc/rc.local before exit 0, then it will run on boot.
Still not sure what the issue is, but we can revert changes slowly from March to February and see if it gives us clues. It may take a couple of days, but we should find it.
I think my problems might be also related to this higher CPU clock since my V4K was really hot to the touch when the freeze occured. On the February image it never gets this hot, even during playback (20-22°C ambient temperature in the room and the V4K is placed on an open surface in an open-front sideboard.
I’ve went back to March release just now and will report if the crashes reappear. Directly after the update I got the sad face crash again, this time I made a video - is this expected behavior?: https://u.nya.is/dfmucc.mp4
Video is sped up during certain times to make it less boring since you probably know the update process well
Shot of error message directly before the sad face (only visible in the video, it was too fast to even see in real life):
Other than that, I will now test streaming content again and see if the crashes reappear.
Sure I can also try that! But I want to try the “regular” march image first. If the freezes reappear with higher clocks, do you still think it’s a problem with the PSU or can that be ruled out?
So far confirmed on my side:
February image: No freezes, V4K slightly warm during playback
March Update: Freezes occur during playback, V4K very warm, borderline hot to the touch during playback (and directly after freeze).
Edit: @sam_nazarko: And about the sad face directly after the update? Is this “expected”? It happened during the first time I updated from Feb to Mar as well (hence I made a vid during the update). I recall you saying in a different thread that this has something to do with UI code and is normal?! Just want to confirm!
Good. Slightly warm is fine, especially if there are no crashes or unexpected behaviour. There are thermal trip-points too to stop things getting too toasty.
Okay. I know for you that you’re only having problems with playback. @G4eva is it the same for you, or are you experiencing heat and crashes when not playing back content? Are you leaving your screen on the Home screen when not using it?
Would be good to know.
Same as above.
What’s interesting here is that it isn’t actually a reduced clock; but somehow setting it in initramfs was losing us about 15-20% performance.
I’ll produce a kernel with more debugging and some ideas and let you know when it’s ready for testing.
Yes – it’s a (low priority) bug with Script Shortcuts that means whenever Kodi is updated, it crashes once. It occurs on all platforms.
During the day the monitor is off because from 7:30 to 19 i’m at work, the only thing it does is download torrent. It still crashes, it happens even during playback (at night, with temperatures about 10 degrees currently) and the morning I wake up with the monitor locked on a scene and an annoying sound coming from the speakers.
P.S. a monitor was almost burned because it remained all night on the same image, it took two days to unlock the pixels
And I’m sorry for today, but with the February version was disconnected from wifi and does not automatically reconnected, so it was not crash but only offline… So I think it’s not a kernel issue, at least I think…
EDIT: new version, new log
Just installed March version and everything with the same method but limiting the clock frequency as @sam_nazarko said… gonna testing it!