[coreboot] T530: Cleaning Intel ME currently does not work <--> CPU fan won't start

[Pavel Alyev]

Not at all. ME can control clock output frequency at GPIO 64/65/66/67. 
 From coreboot you can only set these pins to NATIVE mode. So, if you 
EC/SIO take clock input from PCH, without ME they may work incorrectly.

But looks like at t530 this outputs sets in GPIO mode. David, can you 
dump gpios from system with running ME (can be done by 'inteltool -g').


20.07.2017 20:58, Nico Huber пишет:
> Hi David,
>
> On 20.07.2017 18:44, David Hobach wrote:
>> Dear all,
>>
>> just tested two coreboot + SeaBios images on a T530 that were identical
>> except one time with ME and one time without (using the compile option,
>> rev 54db255529ce8afc689ae425c24b7fb1d45654e8).
>>
>> Unfortunately it seems that ME does some CPU fan initialization (ACPI?)
>> that coreboot doesn't, i.e. on the image without ME the CPU fan didn't
>> start anymore = not usable. I tested with Fedora 26 and ubuntu 14.02 LTS
>> up to temperatures of 98 degrees celsius, but it wouldn't start. I
>> didn't test manual fan control.
> the fan should be controlled by the EC which runs independently of the
> ME and the host CPU. There are some settings the host CPU can write into
> the EC RAM (coreboot should do that, the ME doesn't know anything about
> it). It might be that coreboot does something wrong (e.g. trying to ap-
> ply these settings while the EC is not ready yet after the reset) and
> that it only becomes obvious after the EC has been reset (and even then
> it would depend on the EC firmware, so YMMV). Another explanation could
> be that something broke during the flashing (not due to bad data in the
> flash but due to an overall bad state of the hardware). In which way did
> you flash? did you use the internal flasher? an external one? if exter-
> nal, did you connect an external power supply to the flash chip? All
> bets are off, if you did the the latter.
>
> Though unlikely, the following could also be the case: with the cleaned
> ME the system is in a weird state where it can't correctly write the EC
> RAM registers. But I really doubt that if the host behaves well other-
> wise.
>
> In any case, I'd check the coreboot log for errors.
>
> Nico
>