Re: [PATCH v7 6/6] Documentation: ACPI: Document allow-low-power-probe _DSD property

[Sakari Ailus]

Hi Rafael,

Thank you for the review.

On Wed, Sep 02, 2020 at 05:57:01PM +0200, Rafael J. Wysocki wrote:
> On Tue, Sep 1, 2020 at 11:03 PM Sakari Ailus
> <sakari.ailus@xxxxxxxxxxxxxxx> wrote:
> >
> > Document the probe-low-power _DSD property and how it is used with I²C
> > drivers.
> 
> I would reorder the series to make this go right after the [1/6] or
> maybe even fold it into that patch.
> 
> The point is that the changes in [1/6] clearly depend on the property
> defined here.

Ack.

> 
> > Signed-off-by: Sakari Ailus <sakari.ailus@xxxxxxxxxxxxxxx>
> > ---
> >  .../acpi/dsd/allow-low-power-probe.rst        | 28 +++++++++++++++++++
> >  Documentation/firmware-guide/acpi/index.rst   |  1 +
> >  2 files changed, 29 insertions(+)
> >  create mode 100644 Documentation/firmware-guide/acpi/dsd/allow-low-power-probe.rst
> >
> > diff --git a/Documentation/firmware-guide/acpi/dsd/allow-low-power-probe.rst b/Documentation/firmware-guide/acpi/dsd/allow-low-power-probe.rst
> > new file mode 100644
> > index 0000000000000..6fcc89162b898
> > --- /dev/null
> > +++ b/Documentation/firmware-guide/acpi/dsd/allow-low-power-probe.rst
> > @@ -0,0 +1,28 @@
> > +.. SPDX-License-Identifier: GPL-2.0
> > +
> > +======================================
> > +Probing I²C devices in low power state
> > +======================================
> > +
> > +Introduction
> > +============
> > +
> > +In some cases it may be preferred to leave certain devices powered off for the
> > +entire system bootup if powering on these devices has adverse side effects,
> > +beyond just powering on the said device. Linux recognizes the _DSD property
> > +"allow-low-power-probe" that can be used for this purpose.
> 
> It would be good to refer to the document defining the generic _DSD
> properties mechanism and the GUID used for that from here.

I'll add a reference to the device properties UUID for _DSD spec.

> 
> The meaning of  "_DSD property" may not be entirely clear to the
> reader as it stands.
> 
> And maybe call the property "i2c-allow-low-power-probe" or similar, to
> indicate that it is specific to i2c.

The bus determines that already, and it's only defined for I²C here. Should
we need this in the future for e.g. I3C, there would be no need to think of
renaming it, just changing the documentation and implementation.

I don't have a strong opinion on that though. 

> 
> > +
> > +How it works
> > +============
> > +
> > +The property "allow-low-power-probe" boolean property may be used to tell Linux
> 
> "boolean device property" ?

Oops. The other "property" was supposed to be removed by now. But yes, I
agree with the suggestion.

> 
> > +that the I²C framework should instruct the kernel ACPI framework to leave the
> > +device in the low power state. If the driver indicates its support for this by
> > +setting the I2C_DRV_FL_ALLOW_LOW_POWER_PROBE flag in struct i2c_driver.flags
> > +field and the "allow-low-power-probe" property is present, the device will not
> > +be powered on for probe.
> > +
> > +The downside is that as the device is not powered on, even if there's a problem
> > +with the device, the driver likely probes just fine but the first user will
> > +find out the device doesn't work, instead of a failure at probe time. This
> > +feature should thus be used sparingly.
> 
> It would be good to give an ASL example with this property provided.

I'll add that for v8.

> 
> > diff --git a/Documentation/firmware-guide/acpi/index.rst b/Documentation/firmware-guide/acpi/index.rst
> > index ad3b5afdae77e..0070fcde9e669 100644
> > --- a/Documentation/firmware-guide/acpi/index.rst
> > +++ b/Documentation/firmware-guide/acpi/index.rst
> > @@ -11,6 +11,7 @@ ACPI Support
> >     dsd/graph
> >     dsd/data-node-references
> >     dsd/leds
> > +   dsd/allow-low-power-probe
> >     enumeration
> >     osi
> >     method-customizing
> > --

-- 
Kind regards,

Sakari Ailus