Re: [regression] Bug 216753 - 6e 6 ghz bands are disabled since 5.16 on intel ax211

[Greenman, Gregory]

Hi Dave,

On Tue, 2023-01-03 at 11:37 -0600, Dave Chiluk wrote:
> On Thu, Dec 1, 2022 at 5:33 AM Coelho, Luciano <luciano.coelho@xxxxxxxxx> wrote:
> > 
> > On Thu, 2022-12-01 at 11:14 +0100, Thorsten Leemhuis wrote:
> > > Hi, this is your Linux kernel regression tracker.
> > > 
> > > Luca, I noticed a regression report in bugzilla where I'd like your
> > > advice on. To quote https://bugzilla.kernel.org/show_bug.cgi?id=216753
> > 
> > Hi Thorsten wearing-the-regression-hat, 🙂
> > 
> > I'm not the maintainer of iwlwifi anymore, so I'm adding the new
> > maintainer here, Gregory Greenman.
> > 
> > Gregory, can you take a look?
> > 
> 
> @Gregory Greenman as I'm sure this got buried over the holidays, can
> you take a look at this and advise?  This is definitely a regression,
> but I don't think a lot of people are noticing it or don't yet have
> 6ghz access points.  I can write up a patch removing the offending
> commit (698b166ed), or I can add an iwlwifi option to ignore the 6e
> ACPI bit.  Which would you prefer?
> 
> Dell has been of little help which I pretty much expected.
> 
I'll try to explain, the problem here is not technical. After some
internal checks, it appears that we (wifi driver) aren't allowed to
decide if 6E should be enabled or not. Because of the legal restrictions,
OEM should make this decision and enable/disable 6E in the BIOS. This
commit only gets the value from the BIOS and configures the firmware
accordingly. So, unfortunately, legal restriction is the reason we cannot
revert/overwrite 6E enablement...

> @Luciano, as you were the author of the original change, and I'm not
> familiar enough with ACPI, is the below code reading the enable bits
> from the BIOS ACPI table or is this somehow coming out of the network
> card through some UEFI extensions?  I'm trying to figure out which of
> Dell or Intel need to update their firmware?  I think some Lenovo's
> have similar problems, so I suspect it's a BIOS ACPI table problem.
> 
>  ret = iwl_acpi_get_dsm_u32(mvm->fwrt.dev, 0,
>                                    DSM_FUNC_ENABLE_6E,
>                                    &iwl_guid, &value);

It comes from the BIOS ACPI table.

> 
> Thanks,
> Dave.