2.6.27-11 + EeePC 1000H = unstable
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
linux (Ubuntu) |
Expired
|
Undecided
|
Unassigned |
Bug Description
Hi,
I'm running 8.10 on an EeePC 1000H witth a custom OCZ core V2 120GB SSD disk. Every thing worked well util I upgraded to version 2.6.27-11. I had the following symptoms :
* Black screen after resuming from suspend to memory and total system freeze.
* Unresponsiveness of X applications after resuming a screensaver session before total freeze of the system.
* The last weirdness was the most interesting. I never had any interesting kernel messages with dmesg but I had an application that was impossible to kill even with the -9 signal. "ps aux" reported that it was in "D" state as shown below :
root 10114 0.0 0.0 2248 832 ? D 13:40 0:00 /sbin/dhclient -d -sf /usr/lib/
According to the docs, "D" state means that the process is waiting for some IO to complete. It cannot be killed. When I tried to logout and restart a new X session, the system completely froze.
I built a custom 2.6.28.4 kernel with kernelcheck and I had the same problems. So I reverted to the old 2.6.27-9 which remains rock stable.
Seems that there is some regression starting from 2.6.27-11.
Cheers,
Jacques
Hi,
I have some news about this issue. Finally it seems not be linked with the kernel but with the intel Xorg driver...
Situation that seems to trigger the freeze :
* run this script with something connected to the VGA connector :
xrandr --output LVDS --mode 800x600
xrandr --output VGA --mode 800x600
* disconnect the VGA plug
* wait for the screensaver to turn off the LCD blacklight
-> _sometimes_ the computer will never wake up again. It is a hard crash, i.e. the kernel is frozen (e.g. if you unplug a USB optical mouse and if you plug it again, it will not light up, so there is no more kernel USB initialization activity).
This bug has a very rare occurrence rate so I waited long enough to have some relevant additional informations. Hope it can help.
Jacques