2012-07-24, 23:48
Hi all,
I would like your help for a bug i always encountered with XBMC on my installation, here are the symptoms:
- after a few random hours of idle ( minimum 2-3 hours, sometimes up to 8-10 hours) XBMC freezes and nothing responds, no mouse, no keyboard
- i still can SSH to my host, i can see 1 core CPU fully busy by xbmc.bin process
- Impossible to kill xbmc process even with kill -9, trying to restart lightdm doesn't change anything
- Only a reboot recovers the system
I could find with the hour shown by xbmc (as it freezes the hour keeps the same), that it was related to a message in /var/log/kern.log beginning by:
[fglrx] ASIC hang happened
I searched on many forums and posts related to ATI driver and XBMC, i could not find something reliable...
XBMC log does not show anything special, same messages as usual when the bug occurs.
I am actualy running Xubuntu 12.04, kernel 3.2.26, xbmc-xvba-eden (recently updated) with catalyst 12.4, i had the same issue with previous XBMC eden pre-compiled version, Ubuntu kernels and so on.
The only workaround i found for now was to write a small shell watchdog that watches kern.log for the "ASIC hang happened" pattern and initiate system reboot.
I could also try to get more stability with 12.6 catalyst (i don't believe it ^^) but this version is very buggy, HD radeon 6310 are seen as unsupported hardware and you have to apply a workaround to delete unsupported watermarck screen, also aticonfig command can't be used...
Has anyone encountered the same issue ?
Please find the extracted kern.log : http://pastebin.com/8KLg14bu
Thanks for your help!
I would like your help for a bug i always encountered with XBMC on my installation, here are the symptoms:
- after a few random hours of idle ( minimum 2-3 hours, sometimes up to 8-10 hours) XBMC freezes and nothing responds, no mouse, no keyboard
- i still can SSH to my host, i can see 1 core CPU fully busy by xbmc.bin process
- Impossible to kill xbmc process even with kill -9, trying to restart lightdm doesn't change anything
- Only a reboot recovers the system
I could find with the hour shown by xbmc (as it freezes the hour keeps the same), that it was related to a message in /var/log/kern.log beginning by:
[fglrx] ASIC hang happened
I searched on many forums and posts related to ATI driver and XBMC, i could not find something reliable...
XBMC log does not show anything special, same messages as usual when the bug occurs.
I am actualy running Xubuntu 12.04, kernel 3.2.26, xbmc-xvba-eden (recently updated) with catalyst 12.4, i had the same issue with previous XBMC eden pre-compiled version, Ubuntu kernels and so on.
The only workaround i found for now was to write a small shell watchdog that watches kern.log for the "ASIC hang happened" pattern and initiate system reboot.
I could also try to get more stability with 12.6 catalyst (i don't believe it ^^) but this version is very buggy, HD radeon 6310 are seen as unsupported hardware and you have to apply a workaround to delete unsupported watermarck screen, also aticonfig command can't be used...
Has anyone encountered the same issue ?
Please find the extracted kern.log : http://pastebin.com/8KLg14bu
Thanks for your help!