--PuGuTyElPB9bOcsM
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
Content-Transfer-Encoding: quoted-printable

On Thu, Sep 29, 2005 at 12:48:40PM -0700, Jason Thorpe wrote:
>=20
> On Sep 29, 2005, at 11:40 AM, Bill Studenmund wrote:
>=20
> >On Thu, Sep 29, 2005 at 07:32:51PM +0900, YAMAMOTO Takashi wrote:
> >>ftp://ftp.netbsd.org/pub/NetBSD/misc/xtraeme/bufq_on_the_fly.diff
> >>
> >>- is there any point to change "sortmethod" on the fly?
> >>  it's characteristics of the device.
> >
> >I thought the answer to your question would be that we have =20
> >different sort
> >methods which are better for different workloads, so that changing =20
> >on the
> >fly would make sense if the workload shifts (or if powerup code =20
> >needs to
> >set up the right scheduler).
>=20
> Queueing method, sure... but the "sort" (cylinder or raw block #) is =20
> really something the device needs to decide.

Ahh. Thank you.

The one reason I could see for changing it would be if the administrator=20
felt the device chose incorrectly. Not sure if it's appropriate to do...

Take care,

Bill

--PuGuTyElPB9bOcsM
Content-Type: application/pgp-signature
Content-Disposition: inline

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.3 (NetBSD)

iD8DBQFDPFwUWz+3JHUci9cRAoofAJ9p/84T1iUBWbNJNd//km6Ekml3WgCeLDaT
JHUklPgwyHfpkylqfUQ7Z7o=
=tmCR
-----END PGP SIGNATURE-----

--PuGuTyElPB9bOcsM--