Pentium-M - not recognized?

[Nate Lawson]

Colin Percival wrote:
> Eric Anderson wrote:
> 
>>Eric Anderson wrote:
>>
>>>Eric Anderson wrote:
>>>
>>>>I have a Dell Latitude D610 laptop, and my kernel says:
>>>>
>>>>CPU claims to support Enhanced Speedstep, but is not recognized.
>>>>Please update driver or contact the maintainer.
>>>>cpu_vendor = GenuineIntel msr = 6120e2606000e26, bus_clk = 64
>>>>
>>>>What does that mean to me?  How can I fix it?
> 
> 
> If the BIOS knows about enhanced speedstep and ACPI (which it should)
> then you can probably get access to the different settings via the
> acpi_perf.

Yes, I have a patch that does this locally.

>>>I hate when I reply to myself, but.. After doing some more research, I
>>>see that I have a Pentium M 750, but in cpufreq/est.c, I notice there
>>>isn't an entry for it.
>>>...
>>
>>I think a section for the 750 needs to
>>be added, but I'm not sure what to add exactly, and I'm now afraid I'll
>>light my computer on fire if I try.
> 
> 
> That's exactly the problem I have.  In earlier Pentium M processors, Intel
> helpfully published a complete listing of the frequency/voltage pairs,
> which I encoded into est.c.  With the 133MHz Pentium M processors, Intel
> decided not to publish all this; instead, they've just published the lowest
> and highest pairs.
> 
> I am told that the data I need is available in the "Prescott, Nocona, and
> Potomac Processor BIOS Writer's Guide", but I can't get my hands on that
> since I haven't signed an NDA, and I've been too busy lately to work out
> how one goes about the process of obtaining and signing an Intel NDA.  (I
> have no objection in principle to NDAs, as long as I'm able to publish the
> resulting code.)

Yep.  I finally was able to borrow a friend's T42 for the weekend and 
can test/finish my patch.  I'll commit it once it's working.

-- 
Nate