RE: speedstep-centrino: ENODEV

[Pallipadi, Venkatesh]

>-----Original Message-----
>From: Sune Mølgaard [mailto:sune@xxxxxxxxxxxx] 
>Sent: Thursday, October 19, 2006 11:19 AM
>To: Pallipadi, Venkatesh
>Cc: Jiri Slaby; Linux kernel mailing list; linux-acpi@xxxxxxxxxxxxxxx
>Subject: Re: speedstep-centrino: ENODEV
>
>Thanks for taking an interest, if I haven't said so before.
>
>Now, I was about to write that I couldn't open /dev/mem, but luckily 
>remembered to try sudo'ing acpidump :-$
>
>Enclosed is the dump, created while running 2.6.18.1.
>

OK. Things seem to be fine with the BIOS. 

You said it was working with 2.6.15. Do you remember whether kernel was using acpi-cpufreq or speedstep-centrino?

One change that has happened in this region is that If your BIOS supports both speedstep-centrino and acpi-cpufreq, with 2.6.15 any one of those drivers would have worked. But now with 2.6.18, acpi-cpufreq will not work in the above case and you have to use speedstep_centrino. This was done because, speedstep-centrino has more features than acpi-cpufreq and also doing this helped to elimiate issues with lot of systems with kernel trying to do multiple ACPI PDC writes when BIOS doesn't expect it to. 

In short:
(1) If you were using acpi-cpufreq in 2.6.15, there is a high chance that it wont work with 2.6.18 and you should be able to use speedstep-centrino in its place. Make sure you have properly configured speedstep-centrino (You should select X86_SPEEDSTEP_CENTRINO_ACPI along with X86_SPEEDSTEP_CENTRINO).

(2) If you were using speedstep-centrino in 2.6.15 and now it doesn't work with 2.6.18, then we have a new regression here and we need to root cause it further by enabling cpufreq.debug and getting more debug messages to see where it is failing....

Thanks,
Venki
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/