Re: [Libreboot-dev] c3 and c4 states for macbooks

[Albin]

Hi

Den Tue, 17 May 2016 18:06:38 +0200
skrev Re: [Libreboot-dev] c3 and c4 states for macbooks:

> Hi
> 
> This is not tested. But since it is supposed to be cpu thing and not a
> mb thing I believe this to be no problem.
> 
> How bricked was it? brick like no grub or hangs in or while booting
> GNU/linux?

I'm pretty sure it was the former.  I remember recording the failed boot
using GNU screen but I can't find the log file at the moment.

> If it hangs in gnu/linux then it possible to disable cstates all
> together while booting by adding "processor.max_cstate=0" in kernel
> arguments and flash back your working image.
> It would then also be interesting to know which c-state causes
> problems. try to disable just c4 by adding "processor.max_cstate=3"
> in the kernel argument.

Unfortunately I cannot test this.  However, I've attatched a diff of
the changes that Leah made for the experimental ROM that I
mentioned.  (You can also find it in the git archive: 
docs/resources/misc/donotusethis_macbook_acpi.diff)

After I could confirm that it didn't work, phcoder found c states 1 and
2.  If they are different from the values in the attached diff (I
haven't checked myself), then I guess it's not as simple as a "cpu
thing".

I'd be happy to test a new ROM on the macbook, but I don't feel
like trying the exact same thing that bricked my computer the last time,
not unless someone could help me with a fallback image that loads a
ROM that is confirmed to be working if the test ROM should fail.


Best regards

Albin

> 
> Kind regards
> 
> Albin <address@hidden> writes:
> 
> > This is awesome news, Arthur, but has it been tested?  Do you have a
> > macbook2,1 yourself?
> >
> > I've been waiting a long time for this.  Two years ago I tried the
> > first libreboot version for the macbook2,1 which had no c states at
> > all and got 85--90 degrees hot.  Then I tried Leah's experimental
> > ROM which was created, I believe, like you've done it -- by copying
> > the c-state values from the x60.  This didn't work, however, and my
> > computer got bricked. I've since unbricked it but I'm hesitant to
> > try it again, unless somebody could prepare an image with a working
> > fallback mechanism.
> >
> > Cheers!
> >
> > Albin
> >
> >
> >
> > Den 2016-05-17 kl. 11:47, skrev Arthur Heymans:  
> >> Hi
> >> 
> >> Now the macbook have up to c2 states. This patch just makes the
> >> c-states entries identical to the x60/t60.
> >> 
> >> This should just work as it cpu-specific and those targets have
> >> cpu's that support those low power states
> >> 
> >> 
> >> 
> >> 
> >> This would make the macbooks run cooler and have a longer battery
> >> on not cpu intensive work.
> >>   
> >  
> <#secure method=pgpmime mode=sign>
>

donotusethis_macbook_acpi.diff
Description: Binary data