I have the same kind of issue on a Thinkpad T61.

I don't really know how/when it started, this T61 is not my main laptop
anymore but this morning I had to do stuff on it and noticed network
connection wasn't working while the network cable was plugged correctly.
I modprobe-r'ed e1000e, then modprobed it again, and eth0 didn't
reappeared. I tried to reboot, just in case, and saw this behavior:

* high load (1-2) when idle
* kworker threads activity (40-60 in spikes)
* unresponsiveness of the system

I've tried to boot with pcie_ports=compat but it didn't do anything.

Unloading e1000e driver fix the cpu usage, but again after a new
modprobe the interface doesn't reappear (but the cpu usage either).

What really puzzles me is that I can't really link that to a kernel
upgrade. I can't say if it happened after the latest kernel 3.0.0-4
upgrade, but even reverting to 2.6.32 doesnt help (in this case it's the
'events' thread which takes huge cpu usage.

I didn't upgrade the bios recently either, so I really don't know what
could have lead to that.

Sorry for the undetailed bug report, if you have any idea how to debug
that a bit more, please say so
--
Yves-Alexis

I have the same kind of issue on a Thinkpad T61.

I don't really know how/when it started, this T61 is not my main laptop
anymore but this morning I had to do stuff on it and noticed network
connection wasn't working while the network cable was plugged correctly.
I modprobe-r'ed e1000e, then modprobed it again, and eth0 didn't
reappeared. I tried to reboot, just in case, and saw this behavior:

* high load (1-2) when idle
* kworker threads activity (40-60 in spikes)
* unresponsiveness of the system

I've tried to boot with pcie_ports=compat but it didn't do anything.

Unloading e1000e driver fix the cpu usage, but again after a new
modprobe the interface doesn't reappear (but the cpu usage either).

What really puzzles me is that I can't really link that to a kernel
upgrade. I can't say if it happened after the latest kernel 3.0.0-4
upgrade, but even reverting to 2.6.32 doesnt help (in this case it's the
'events' thread which takes huge cpu usage.

I didn't upgrade the bios recently either, so I really don't know what
could have lead to that.

Sorry for the undetailed bug report, if you have any idea how to debug
that a bit more, please say so
--
Yves-Alexis