Even for NetBSD running in domain 0, Xen seems to impose quite a large
  disk latency penalty, though throughput is not seriously effected.  Jed
  did some measurements of this a while ago.  I think most of the cost,
  particularly for domU, is due to the extensive remapping of data via
  the MMU, some done at our behest by the hypervisor.  (Jed, does this
  sound right to you?)

I didn't mean to argue that there were no issues.  But performance on
the systems, dom0 and domU, is qualitatively reasonable, and didn't
get as bad as "awful long time accessing the harddisk" and "very
sluggish".

It would be nice to have careful measurements, and see if this is an
intrinsic cost given the protection model, or if it can be reduced.

-- 
        Greg Troxel <gdt@ir.bbn.com>