Hello,

> >Hmm, that's something for the original author - compare do_pending_intr()
> >and ext_intr(), one of them explicitly sets PSL_EE, the other one masks it
> >out - so do_pending_intr() only intends to turn off interrupts at some
> >point, but ext_intr() explixitly turns them on. Looks intentional.
> 
> Looks like he's making sure it is on in ext_intr() around the KERNEL_LOCK,
> as well as in do_pending_int(), and making sure it is on no matter what
> when exiting ext_intr(). When leaving do_pending_int(), the state is
> restored, but I find the use of "emsr" and "dmsr" to imply that he expects
> PSL_EE to be on when getting here. This is not the case when
> do_pending_int() is called from Idle() via spllower().

Maybe you're right - something weird just happened. Playing something in xmms while fsck'ing  the USB disk didn't disrupt audio playback with xmms on a remote display, but on a local display it did. I get lengthy periods where apparently only clock interrupts get processed, anywhere from 0.5-5 seconds, not always directly following a lost interrupt. 

have fun
Michael