Subject: uvm_pageidlezero()
To: None <tech-kern@netbsd.org>
From: Michael <macallan18@earthlink.net>
List: tech-kern
Date: 06/20/2005 16:37:35
--Signature_Mon__20_Jun_2005_16_37_35_-0400_LudySB25rvLLhsmR
Content-Type: text/plain; charset=US-ASCII
Content-Transfer-Encoding: quoted-printable

Hello,

I've been hacking the SPARCbook again, this time I figured out how to
put the CPU to sleep to cool down / save power. So I added another
function pointer to struct cpu_info and hacked locore.s to call it when
idle.
The problem - the idle loop contains code to call uvm_pageidlezero()
under certain circumstances, but we can have only one thing - a sleeping
CPU or one that zeroes idle pages, not both. Since macppc does a similar
thing on some CPUs ( notably G3s ) I had a look at this code and found
that apparently no powerpc port ever calls uvm_pageidlezero. Arm and
m68k don't use it either.

So - is it really necessary to run uvm_pageidlezero in the idle loop?
Should I call it when CPU sleep mode isn't available or disabled?=20

have fun
Michael

--Signature_Mon__20_Jun_2005_16_37_35_-0400_LudySB25rvLLhsmR
Content-Type: application/pgp-signature

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.0 (NetBSD)

iQEVAwUBQrcpEcpnzkX8Yg2nAQKyOQf/a08M3asJ7/GRybdBZ4bDX6ZpHVtA17Jz
p/1/j4gcDFM9rQ0ckwr4nQDXx84VdJ6H7P+m34BhnrvSqdgeuPVIeebi0MWAEfdu
gPQSMbCRohZV1AV3TTC4KqgvWKbvW5CYXZ8AMDNIbtoE/OVFoGlBWAMQ2pR0NCfu
ut2ukevVq42vlauqAUEFdBYjIynYrPbnZeM7B12dspo7rEzjD6Zzq6ayv+cDP3zH
pxcIUqdEbGcq8eT3bA7LXTv8xckpcjnfshNK/IG83ZBkWVUsQ47CMffcdKvxkxt3
IAoZk7S/ehmQ4Q/tRCtVk8G4xkhA2ZbdaIWNquSGNhgz4Sw5a5UbLA==
=gvfL
-----END PGP SIGNATURE-----

--Signature_Mon__20_Jun_2005_16_37_35_-0400_LudySB25rvLLhsmR--