[Public]
> -----Original Message-----
> From: Rafael J. Wysocki <rafael@xxxxxxxxxx>
> Sent: Thursday, June 30, 2022 06:40
> To: Limonciello, Mario <Mario.Limonciello@xxxxxxx>
> Cc: Hans de Goede <hdegoede@xxxxxxxxxx>; Rafael J . Wysocki
> <rjw@xxxxxxxxxxxxx>; linux-acpi <linux-acpi@xxxxxxxxxxxxxxx>;
> regressions@xxxxxxxxxxxxxxx; Greg KH <gregkh@xxxxxxxxxxxxxxxxxxx>;
> rui.zhang@xxxxxxxxx
> Subject: Re: "Revert "ACPI: Pass the same capabilities to the _OSC regardless
> of the query flag"" is causing regressions
>
> On Wed, Jun 29, 2022 at 10:12 PM Limonciello, Mario
> <Mario.Limonciello@xxxxxxx> wrote:
> >
> > [Public]
> >
> > > > -----Original Message-----
> > > > From: Greg KH <gregkh@xxxxxxxxxxxxxxxxxxx>
> > > > Sent: Thursday, June 23, 2022 08:34
> > > > To: Limonciello, Mario <Mario.Limonciello@xxxxxxx>
> > > > Cc: Hans de Goede <hdegoede@xxxxxxxxxx>; Rafael J . Wysocki
> > > > <rjw@xxxxxxxxxxxxx>; linux-acpi <linux-acpi@xxxxxxxxxxxxxxx>;
> > > > regressions@xxxxxxxxxxxxxxx; Thorsten Leemhuis (regressions address)
> > > > <regressions@xxxxxxxxxxxxx>
> > > > Subject: Re: "Revert "ACPI: Pass the same capabilities to the _OSC
> regardless
> > > > of the query flag"" is causing regressions
> > > >
> > > > On Thu, Jun 23, 2022 at 07:24:34AM -0500, Mario Limonciello wrote:
> > > > > On 6/23/22 05:06, Hans de Goede wrote:
> > > > > > Hi Rafael, Mario,
> > > > > >
> > > > > > Commit 2ca8e6285250 ("Revert "ACPI: Pass the same capabilities to
> the
> > > > > > _OSC regardless of the query flag"") is causing the issues fixed
> > > > > > by the reverted commit to show up again, see:
> > > > > >
> > > > > >
> > > >
> https://nam11.safelinks.protection.outlook.com/?url=https%3A%2F%2Fbugz
> > > >
> illa.kernel.org%2Fshow_bug.cgi%3Fid%3D213023&data=05%7C01%7Cm
> > > >
> > >
> ario.limonciello%40amd.com%7C0040716e869d4021ce3208da551d082d%7C3d
> > > >
> > >
> d8961fe4884e608e11a82d994e183d%7C0%7C0%7C637915880426388833%7CU
> > > >
> nknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI
> > > >
> 6Ik1haWwiLCJXVCI6Mn0%3D%7C3000%7C%7C%7C&sdata=KxmNvCfdm
> > > > qvk5gXteXUDXHVTK45yEt%2BUYO4vaBbLXis%3D&reserved=0
> > > > > >
> > > >
> https://nam11.safelinks.protection.outlook.com/?url=https%3A%2F%2Fbugz
> > > >
> illa.redhat.com%2Fshow_bug.cgi%3Fid%3D1963717&data=05%7C01%7C
> > > >
> > >
> mario.limonciello%40amd.com%7C0040716e869d4021ce3208da551d082d%7C
> > > >
> > >
> 3dd8961fe4884e608e11a82d994e183d%7C0%7C0%7C637915880426388833%7
> > > >
> CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJ
> > > >
> BTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000%7C%7C%7C&sdata=pZRqAG3
> > > >
> %2Bg0QevLOGZ8m9PNxcmkmh58soT2dSLg%2B6qWc%3D&reserved=0
> > > > > >
> > > > > > both of which have comments from the reporters that
> > > > > > the error message is back again; and presumably also
> > > > > > that /sys/devices/system/cpu/cpu0/acpi_cppc is missing
> > > > > > again.
> > > > > >
> > > > > > Can you please take a look and see if we can come up with
> > > > > > something which fixes both the re-surfaced issue, as well
> > > > > > as the issue which the revert tries to address ?
> > > > > >
> > > > >
> > > > >
> > > > > Is it possible that c42fa24b44751c62c86e98430ef915c0609a2ab8 didn't
> > > > backport
> > > > > to the stable trees it popped back up again?
> > > > >
> > > >
> > > > It is only in 5.18, no stable releases that I can see.
> > >
> > > I think that should probably come back to stable along with everything
> this
> > > revert went back to, but before doing that perhaps we can get someone
> > > affected to confirm cherry-picking it helps.
> >
> > Users have reported to the Kernel Bugzilla that just picking that patch didn't
> > help. At it's core it "seems" to me the firmware masks CPPC support in the
> _OSC
> > and previously that was actually ignored. Since the final SSDT with the
> PR0._CPC
> > is loaded dynamically based on whether CPPC was supported this runtime
> error
> > happens.
> >
> > As multiple people have been duplicated into that bug let's work through
> some
> > ideas on it there. To start out I've left a suggestion in there for users to
> have a
> > try with.
>
> Honestly, at this point I'm inclined to revert the commit in question.
Isn't that just trading a regression for a regression?
