Comment # 39 on bug 1184817 from Takashi Iwai

(In reply to teo teo from comment #38)
> > The fact that web browser doesn't open
> 
> That's never happened to me. The browser (which is already open and I've
> been using for a while) stops responding (meaning I'm unable to switch
> between tabs, and/or close tabs, and/or scroll, and/or etc.), but when I
> close it and relaunch it it opens immediately, and usually works for a
> while. And the same with other applications.

Well, that's what I meant -- the unstable behavior of the web browser (or
whatever).

> > When the problem happens, and when you unplug the USB Ethernet device (instead of power-off / reset), does the problem persist?
> 
> Yes.

And what if you logout and re-login?  The problem persists?  If it's a kernel
problem, it should persist.  Also it should persist on the another user, too.

In anyway, we need a way to verify the buggy behavior with a more simpler
application than web browser or those heavy & complex beast.  If you can find
some issue with a simple command-line application (just does some network
access or anything), it can be used for analyzing better what's going on.

Once after we find out a simpler test case, we can take a deeper look at the
system level via performance monitoring or tracing.

And another thing is to try solving the IRQ 9 problem.  If you have no BIOS
update available, at least try irqpoll boot option to see whether the error
message is gone or not.