On Tue, 22 Mar 2005 00:00:36 +0100, RUMI Szabolcs <rumi_ml@rtfm.hu> wrote:
>Hello,
>
>I have to apologize to all people who wanted to help me, or me to
>help them on PowerNow-related things but the laptop I was testing
>with didn't belong to me and it was taken away so that I do not have
>any testbed right now.
>
>The PowerNow patch from Martin Vegiard, including my modifications,
>is very far from perfect, I'd only call it a very preliminary support. If =
it
>does work on a certain laptop at all then it seems to work stable,
>I don't think it is really dangerous to use it, but if you'd ask me
>officially, I'd advise against using it on production systems.

I've been using my PowerNow patch daily, since NetBSD-2.0G.
I have used it with 2.0 and -current for quite some time.
Numerous people have confirmed the same with me.

Please note, that it is probably not the best idea to test PowerNow on
broken hardware.

>It would be a lot better if someone who owns one or more PowerNow
>laptops or Cool'n'Quiet (=3DPowerNow for Opteron) systems could rewrite
>the whole thing from scratch. I appended some information I gathered
>to PR#26239 hoping that it might help with this.

For starters, PowerNow works with AMD K7 cpus (arch =3D=3D i386),
Cool'n'Quiet works with 64 bit AMD Athlon 64 bit cpus (arch =3D=3D amd64)
and those _two_ technologies don't work exactly the same way.

There is no reason to rewrite the whole thing from scratch :
- no one has encountered problems. (with non-broken bioses)
- no one has encountered instability/crashes.

My code is clean and conforms to AMD's specifications.=20
It has been committed to two serious BSD-type operating systems already.

For a list of known laptops (amd-k7) with broken bioses and an explanation =
:
http://www.codemonkey.org.uk/projects/cpufreq/powernow-k7.shtml

The latest version of the PowerNow diff can be found at :
http://www.deadnetwork.com/patches/powernow.patch

Thanks for your time
Martin V=E9giard