Date: Fri, 26 May 2006 06:55:50 -0700 From: Nate Lawson <nate@root.org> To: Andriy Gapon <avg@icyb.net.ua> Cc: freebsd-acpi@freebsd.org, Nate Lawson <web@root.org> Subject: Re: nforce2 cpufreq Message-ID: <447708E6.7010205@root.org> In-Reply-To: <4475C74C.2080204@icyb.net.ua> References: <4475C74C.2080204@icyb.net.ua>
next in thread | previous in thread | raw e-mail | index | archive | help
Andriy Gapon wrote: > I've recently had a sudden urge to investigate power saving / cpu > throttling options for my desktop Athlon XP system. > It seems that the CPU itself does not provide any interfaces for that, > at least neither of acpi_perf/acpi_throttle/cpufreq seem to detect > anything interesting for them. > Or am I mistaken and doing something wrong ? > > Anyway, my MB is based on nForce2 chipset and I found out that Linux has > cpufreq-nforce2 module that works in their cpufreq framework: > http://www.hasw.net/linux/ > > It seems that that module works by using nForce2 PCI interface for > querying and changing FSB frequency. It also seems that the code is > rather simple and obvious in its logic (save for allegedly > reverse-engineered constants). Not sure though how easy it is to port > that to FreeBSD cpufreq framework. > But the question that I really would like to ask is the following: is it > a proper way to do cpufreq stuff by changing FSB frequency ? Would that > approach fit into our framework ? And finally, would it have any > positive temperature/power consumption effects ? > It's really easy to do. Just see the sys/dev/cpufreq/ichss.c file. -- Nate
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?447708E6.7010205>