Re: oops on resume

[Dominik Brodowski <linux@dominikbrodowski.de>]

On Thu, Aug 05, 2004 at 09:29:41PM +0100, Dave Jones wrote:
> On Thu, Aug 05, 2004 at 04:19:30PM -0400, Brian J. Murrell wrote:
> 
>  > Yeah, me too.  Any (one have any) idea how to calculate fsb for that
>  > sucker?
> 
> Hunting for docs on developer.intel.com right now. So far I've turned
> up nothing useful. Can you edit pentium4_get_frequency() in arch/i386/kernel/cpu/cpufreq/speedstep-lib.c
> so that before the switch (fsb_code) {   add a
> 		printk ("FSB code = %d\n", fsb_code);

_and_ enable SPEEDSTEP_DEBUG in both speedstep-ich.c and speedstep-lib.c

> Hopefully, Intel are predictable and this is going to be '4'.
> 
>  > I may as well throw that code in there too to see if it
>  > alleviates the problem.
> 
> reminds me, moving speedstep's dprintk over to the same style
> as longhaul/powernow-k7 should happen sometime, then it's just a
> module parameter away instead of a recompile.

... though I'd prefer if it'd be done at the cpufreq-core level, i.e.
cpufreq_dprintk(); with a sysfs-exported ( /sys/parameters/cpufreq/debug )
boot-time parameter cpufreq.debug=1

	Dominik