> That seems to be a common problem with recent kernels.  They appear to
   	> lose clock interrupts.  Maybe other interrupts as well.
   
   	Is it known why it happens?
   
   Initial investigation shows some correllation between lost
   interrupts and exiting the scheduler.  However, there's
   quite some difference between that and actually isolating
   the root cause.
   
   Charles posted some of his findings recently.  I think
   that was in tech-kern.


hmm.  this reminds me of lennart's post about tearing down the
processes address space taking a Long Time, on tech-kern(?)
recently...


.mrg.