[Bug 218686] Fail to set energy_performance_preference of amd processor on asus ga403uv

[bugzilla-daemon@kernel.org]

https://bugzilla.kernel.org/show_bug.cgi?id=218686

--- Comment #78 from Chema (chemasanchezgarabito@gmail.com) ---
(In reply to xiaojian.du from comment #77)
> (In reply to Chema from comment #75)
> 
> > Thanks for checking
> > 
> > It seems that way, asus has made it work in windows but not in the standard
> > way needed by linux. But either way the amd pstate driver is working, but
> > the epp change is not working.
> > 
> > What could be the easiest to try to solve it from our part, modifying the
> > kernel   in some way? If windows is working it should be a workaround to
> > make it work maybe disable some checks or something like this
> > https://github.com/archlinux/linux/blob/
> > b1bc554e009e3aeed7e4cfd2e717c7a34a98c683/arch/x86/kernel/acpi/cppc.c#L14 or
> > this
> > https://github.com/archlinux/linux/blob/
> > defaf1a2113a22b00dfa1abc0fd2014820eaf065/drivers/acpi/cppc_acpi.c#L1491...?
> > 
> > Cause ASUS has not made any comment about it, I dont know from your part it
> > they have internally said something to you.
> > 
> > Thanks
> 
> A simple tricky hack can't solve this problem.
> From the perspective of amd-pstate driver, it is the most reliable solution
> to check the CPUID register.
> Any other workaround to enable CPPC feature forcely and load epp mode in
> Linux OS is ugly and unacceptable.
> I recommand to wait for further actions from Asus.
> 
> Anyway, acpi-cpu driver still can be the second option for now.

Thanks

But have you any initial idea of how a workaround could work? Obviously for
developing it my self locally and so people could have some solution while asus
does something, if ever does, considering they do not support linux

Yes, currently I use acpi-cpu, but powersave governor makes the laptop laggy
and schedutil that should be a balance makes it hot and drains battery faster.

So currently for my self linux in this laptop is not much a good option, that
why I was wondering if a local workaround would be possible.

Thanks.

-- 
You may reply to this email to add a comment.

You are receiving this mail because:
You are the assignee for the bug.