Re: [PATCH v2] drm/client: Detect when ACPI lid is closed during initialization

Re: [PATCH v2] drm/client: Detect when ACPI lid is closed during initialization

[Dmitry Baryshkov]

On Thu, 30 May 2024 at 07:41, Limonciello, Mario
<mario.limonciello@amd.com> wrote:
>
>
> >> Also a direct acpi_lid_open() call seems a bit iffy. But I guess if
> >> someone needs this to work on non-ACPI system they get to figure out
> >> how to abstract it better. acpi_lid_open() does seem to return != 0
> >> when ACPI is not supported, so at least it would err on the side
> >> of enabling everything.
> >
> > Thanks. I was going to comment, but you got it first. I think a proper
> > implementation should check for SW_LID input device instead of simply
> > using acpi_lid_open(). This will handle the issue for other,
> > non-ACPI-based laptops.
> >
>
> Can you suggest how this would actually work?  AFAICT the only way to
> discover if input devices support SW_LID would be to iterate all the
> input devices in the kernel and look for whether ->swbit has SW_LID set.
>
> This then turns into a dependency problem of whether any myriad of
> drivers have started to report SW_LID.  It's also a state machine
> problem because other drivers can be unloaded at will.
>
> And then what do you if more than one sets SW_LID?

It might be easier to handle this in the input subsystem. For example
by using a refcount-like variable which handles all the LIDs and
counts if all of them are closed. Or if any of the LIDs is closed.

>
> IOW - a lot of complexity for a non-ACPI system.  Does such a problem
> exist in non-ACPI systems?

There are non-ACPI laptops. For example Chromebooks. Or Lenovo X13s,
Lenovo Yoga C630, Lenovo Flex5G, etc. We are expecting more to come in
the next few months. And I don't see why they won't have the same
problem.


-- 
With best wishes
Dmitry

Re: [PATCH v2] drm/client: Detect when ACPI lid is closed during initialization

[Dmitry Torokhov]

On Thu, May 30, 2024 at 11:07:53AM +0300, Dmitry Baryshkov wrote:
> On Thu, 30 May 2024 at 07:41, Limonciello, Mario
> <mario.limonciello@amd.com> wrote:
> >
> >
> > >> Also a direct acpi_lid_open() call seems a bit iffy. But I guess if
> > >> someone needs this to work on non-ACPI system they get to figure out
> > >> how to abstract it better. acpi_lid_open() does seem to return != 0
> > >> when ACPI is not supported, so at least it would err on the side
> > >> of enabling everything.
> > >
> > > Thanks. I was going to comment, but you got it first. I think a proper
> > > implementation should check for SW_LID input device instead of simply
> > > using acpi_lid_open(). This will handle the issue for other,
> > > non-ACPI-based laptops.
> > >
> >
> > Can you suggest how this would actually work?  AFAICT the only way to
> > discover if input devices support SW_LID would be to iterate all the
> > input devices in the kernel and look for whether ->swbit has SW_LID set.
> >
> > This then turns into a dependency problem of whether any myriad of
> > drivers have started to report SW_LID.  It's also a state machine
> > problem because other drivers can be unloaded at will.
> >
> > And then what do you if more than one sets SW_LID?
> 
> It might be easier to handle this in the input subsystem. For example
> by using a refcount-like variable which handles all the LIDs and
> counts if all of them are closed. Or if any of the LIDs is closed.

Yes, install an input handler matching on EV_SW/SW_LID so you will get
notified when input devices capable of reporting SW_LID appear and
disappear and also when SW_LID event is being generated, and handle as
you wish. Something like

https://chromium.googlesource.com/chromiumos/third_party/kernel/+/40e9f6a991856ee7d504ac1ccd587e435775cfc4%5E%21/#F0

In practice I think it is pretty safe to assume only 1 lid for a
laptop/device.

Thanks.

-- 
Dmitry