You are not logged in.
Hi all,
With the update to kernel 2.6.30 I found a bug that was quite annoying to me. It is reported upstream and Jesse Barnes (from Intel) already has an idea of where the problem can be. However, I am a bit surprised that I find no other mention to it anywhere, so I thought that maybe other are having it but didn't notice.
The requirements for possibly having the bug are having a G45 Intel chipset (using the integrated graphics) and using kernel >= 2.6.30.
Way of checking is having a terminal opened and typing the command (as root): "udevadm monitor". If at some point during the session you start to get lots of messages repeated you might be having this bug.
This bug can cause udev process to be using a lot of CPU time, depending on how bad it gets stuck sending uevents.
Offline
So what'd be the best temporary solution? Some guys suggested downgrading in another thread, but which packages to downgrade? Which versions are safe?
Thanks for the info
Offline
It is a kernel bug introduced in 2.6.30, so my temporary solution is to simply stay with 2.6.29.
Where did you find that other thread you mentioned? I was looking around for others seeing this problem but never found any reference.
Offline
You could try 2.6.31-rc8-git2 or to wait until 2.6.31 reaches final (should be in a week max)...
Offline
I believe these two threads are about the same problem:
http://bbs.archlinux.org/viewtopic.php?id=78946
http://bbs.archlinux.org/viewtopic.php?id=78962
I downgraded both the kernel and udev, ignoring dependencies. All seems to be ok now, I'll wait for an official fix/update.
Offline
2.6.31 also has the bug, so for now we must stay with 2.6.29.
Strange, in those threads people had problems with udev. Simptoms are similar, but I can't be sure is the same problem. In my case it's certainly a kernel bug (with a small patch/hack, it is possible to make it go away, though it's not a real fix).
Offline
I am also seeing this prob in kernel 2.6.30-5
Is it fixed in 2.6.31??
Regards
Mukul
Offline
I am also seeing this prob in kernel 2.6.30-5
Is it fixed in 2.6.31??
Ahem, did you read the thread? The person right above you said...
2.6.31 also has the bug, so for now we must stay with 2.6.29.
Anyways, no issue here (but using G45M, not G45).
Offline