Martin Husemann writes:
> > This trace simply shows that it's getting SIGALRM a lot, and that
> > it's calling gettimeofday() in the signal handler. It's hard to
> > tell what the actual cause is of the loop.
> 
> I've seen this with all versions of communicator on Linux too. It can be
> provoked easily with certain java applets.

Mozilla did something similar a few minutes ago.  Started using all of
the cpu, and ktrace showed:
  1024 mozilla-bin EMUL  "netbsd"
  1024 mozilla-bin CALL  poll(0x8135fc0,0x5,0)
  1024 mozilla-bin RET   poll 1
  1024 mozilla-bin CALL  gettimeofday(0xbfbfcdb8,0)
  1024 mozilla-bin RET   gettimeofday 0
  1024 mozilla-bin CALL  ioctl(0xc,FIONREAD,0xbfbfc588)
  1024 mozilla-bin RET   ioctl 0
  1024 mozilla-bin CALL  poll(0x8135fc0,0x5,0)
  1024 mozilla-bin RET   poll 1
  1024 mozilla-bin CALL  gettimeofday(0xbfbfcdb8,0)
  1024 mozilla-bin RET   gettimeofday 0
  1024 mozilla-bin CALL  ioctl(0xc,FIONREAD,0xbfbfc588)
  1024 mozilla-bin RET   ioctl 0
  1024 mozilla-bin CALL  poll(0x8135fc0,0x5,0)
  ...
(repeats continously)

So I guess there are some issues with handling errors or something. :-)

-Andrew