Subject: Re: xen: boot hangs after detection of fwohci
To: Greg Troxel <gdt@ir.bbn.com>
From: Thor Lancelot Simon <tls@rek.tjls.com>
List: port-i386
Date: 02/24/2006 09:56:25
On Fri, Feb 24, 2006 at 09:50:42AM -0500, Greg Troxel wrote:
> Berndt Josef Wulf <wulf@ping.net.au> writes:
> 
> > Now that it is running, the other thing I noticed was that the
> > system appears to spend a awful long time accessing the harddisk
> > making the whole system very sluggish.
> 
> This seems odd; my NetBSD xen systems have very good disk performance,
> 70 MB/s raw read rate from disk (SATA), 57 MB/s read from ffs on the
> disk, and 54 MB/s read from the 'raw' device in a domU (/dev/xbd0d)
> when that device is backed by a file in dom0.

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?)

-- 
  Thor Lancelot Simon	                                     tls@rek.tjls.com

  "We cannot usually in social life pursue a single value or a single moral
   aim, untroubled by the need to compromise with others."      - H.L.A. Hart