Re: [PATCH v1 3/4] platform/x86: xo15-ebook: Register ACPI notify handler directly

[Rafael J. Wysocki]

On Mon, May 11, 2026 at 6:44 PM Ilpo Järvinen
<ilpo.jarvinen@xxxxxxxxxxxxxxx> wrote:
>
> On Mon, 11 May 2026, Rafael J. Wysocki wrote:
>
> > On Mon, May 11, 2026 at 6:17 PM Ilpo Järvinen
> > <ilpo.jarvinen@xxxxxxxxxxxxxxx> wrote:
> > >
> > > On Mon, 11 May 2026, Rafael J. Wysocki wrote:
> > >
> > > > On Mon, May 11, 2026 at 3:59 PM Ilpo Järvinen
> > > > <ilpo.jarvinen@xxxxxxxxxxxxxxx> wrote:
> > > > >
> > > > > On Fri, 8 May 2026, Rafael J. Wysocki wrote:
> > > > >
> > > > > > From: "Rafael J. Wysocki" <rafael.j.wysocki@xxxxxxxxx>
> > > > > >
> > > > > > To facilitate subsequent conversion of the driver to a platform one,
> > > > > > make it install an ACPI notify handler directly instead of using
> > > > > > a .notify() callback in struct acpi_driver.
> > > > > >
> > > > > > No intentional functional impact.
> > > > > >
> > > > > > Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@xxxxxxxxx>
> > > > > > ---
> > > > > >  drivers/platform/x86/xo15-ebook.c | 19 ++++++++++++++-----
> > > > > >  1 file changed, 14 insertions(+), 5 deletions(-)
> > > > > >
> > > > > > diff --git a/drivers/platform/x86/xo15-ebook.c b/drivers/platform/x86/xo15-ebook.c
> > > > > > index 616f4bb3461a..8af1b9078db8 100644
> > > > > > --- a/drivers/platform/x86/xo15-ebook.c
> > > > > > +++ b/drivers/platform/x86/xo15-ebook.c
> > > > > > @@ -57,16 +57,15 @@ static int ebook_send_state(struct acpi_device *device)
> > > > > >       return 0;
> > > > > >  }
> > > > > >
> > > > > > -static void ebook_switch_notify(struct acpi_device *device, u32 event)
> > > > > > +static void ebook_switch_notify(acpi_handle handle, u32 event, void *data)
> > > > > >  {
> > > > > >       switch (event) {
> > > > > >       case ACPI_FIXED_HARDWARE_EVENT:
> > > > > >       case XO15_EBOOK_NOTIFY_STATUS:
> > > > > > -             ebook_send_state(device);
> > > > > > +             ebook_send_state(data);
> > > > > >               break;
> > > > > >       default:
> > > > > > -             acpi_handle_debug(device->handle,
> > > > > > -                               "Unsupported event [0x%x]\n", event);
> > > > > > +             acpi_handle_debug(handle, "Unsupported event [0x%x]\n", event);
> > > > > >               break;
> > > > > >       }
> > > > > >  }
> > > > > > @@ -123,6 +122,11 @@ static int ebook_switch_add(struct acpi_device *device)
> > > > > >       if (error)
> > > > > >               goto err_free_input;
> > > > > >
> > > > > > +     error = acpi_dev_install_notify_handler(device, ACPI_DEVICE_NOTIFY,
> > > > > > +                                             ebook_switch_notify, device);
> > > > > > +     if (error)
> > > > > > +             goto err_unregister_input;
> > > > > > +
> > > > > >       ebook_send_state(device);
> > > > > >
> > > > > >       if (device->wakeup.flags.valid) {
> > > > > > @@ -139,6 +143,10 @@ static int ebook_switch_add(struct acpi_device *device)
> > > > > >  err_free_button:
> > > > > >       kfree(button);
> > > > > >       return error;
> > > > > > +
> > > > > > +err_unregister_input:
> > > > > > +     input_unregister_device(input);
> > > > > > +     goto err_free_button;
> > > > >
> > > > > The end result would be much simpler if there would be patch 5 to convert
> > > > > to devm_*().
> > > >
> > > > Well, I can add one.
> > > >
> > > > Do you want me to resend an update of the series for this?
> > >
> > > Ah, right. I took these now to review-ilpo-next.
> > >
> > > (I was initially thinking of making a little bit different suggestion
> > > which would have required changes to these patches but then realized it
> > > was better to just add the devm_*() conversion after these).
> >
> > So I guess I'll prepare a follow-up patch for the devm_ conversion.
> >
> > Note though that there's no devm_ counterpart of
> > acpi_dev_install_notify_handler() ATM, so I may need to add one while
> > at it.
>
> While adding it will be able to remove acpi_dev_remove_notify_handler()
> call from ebook_switch_remove(), there's no such call in rollbacks for
> ebook_switch_probe() so it wouldn't be strictly required.

Right.

> But it certainly wouldn't hurt to have one for other drivers.

Sure, I definitely have a plan to add one.