On Mon, 29 Oct 2007 21:22:28 +0100
Juan RP <juan@xtrarom.org> wrote:

> 
> Hi,
> 
> a few people prefered to make coretemp(4) a real autoconf(9) driver
> rather than an option, I've heard complaints about options being
> unhelpful for LKMs for example.
> 
> Currently we do have "options INTEL_CORETEMP" and registers a driver
> per struct cpu_info pointer via identcpu.
> 
> I've implemented "coretemp* at cpu?". Do we want to use this approach
> rather than an option?
> 
> IMHO it's more logical to me to attach coretemp(4) at cpu, because there's
> a sensor per cpu.

Quentin Garnier suggested to use a new interface attribute to the cpu
device, and I added a cpudevbus attrib to it.

I changed the code in x86/cpu.c to use config_found_ia() as suggested and
now coretemp must be specified as:

"coretemp* at cpudevbus?"

http://www.netbsd.org/~xtraeme/coretemp-driver2.diff

Quentin, is this ok for you? more suggestions?

-- 
Juan Romero Pardines	- The NetBSD Project
http://plog.xtrarom.org	- NetBSD/pkgsrc news in Spanish