Re: [PATCH] amd76x_pm: C3 powersaving for AMD K7

[Tony Lindgren <tony@atomide.com>]

* Erik Slagter <erik@slagter.name> [060202 01:50]:
> On Thu, 2006-02-02 at 01:50 +0000, Alan Cox wrote:
> > On Mer, 2006-02-01 at 20:35 -0500, Brown, Len wrote:
> > > This endeavor is full of risk, and I would be extremely careful
> > > about enabling features that the BIOS explicitly disabled --
> > > unless the hardware manufacturer publicly publishes
> > > support for the feature, or the errata that you're working around.

Well the same code could also be used to disable C states on buggy
hardware. On my Fujitsu Lifebook p2120 C states are enabled, but don't
work. But we don't have currently anything in place to fix them or
disable them. And enabling C states on amd76x could be a cmdline option.
 
> > Folks had code that supported AMD76x by banging the hardware directly.
> > On some AMD76x systems it caused corruption. Nobody AFAIK ever figured
> > out if it was an errata (nothing obvious in the docs/errata list) or a
> > bug in the code doing the banging on the chip or some other bit of
> > hardware on the mainboard that needed extra handling.
> 
> Looks like it would be beneficial if someone, preferably with some
> authority ;-), would try to get Tyan and/or AMD to reveal this
> information.
> 
> I have read the docs for the AMD768 chipset numerous times, but I least
> I (okay okay, I am not the expert here ;-)) cannot find anything that
> would prevent this from working correctly.
> 
> AFAIK Tyan has been more or less cooperative with linux developers in
> the past?! I'd be more than happy to ask them, but I don't think they
> will take me very serious

The C states on amd76x are quite usable. The only problem I had on my
Tyan s2460 was the some BIOS bug puts the machine into sleep at first
when they're enabled :) This does not happen on other systems.

Tony