Re: [PATCH v3 1/1] i2c: add ACPI support for I2C mux ports

["Rafael J. Wysocki" <rjw@rjwysocki.net>]

On Tuesday, October 20, 2015 03:51:11 PM Mika Westerberg wrote:
> On Mon, Oct 19, 2015 at 03:29:00PM -0700, Dustin Byford wrote:
> > Although I2C mux devices are easily enumerated using ACPI (_HID/_CID or
> > device property compatible string match) enumerating I2C client devices
> > connected through a I2C mux device requires a little extra work.
> > 
> > This change implements a method for describing an I2C device hierarchy that
> > includes mux devices by using an ACPI Device() for each mux channel along
> > with an _ADR to set the channel number for the device.  See
> > Documentation/acpi/i2c-muxes.txt for a simple example.
> > 
> > Signed-off-by: Dustin Byford <dustin@cumulusnetworks.com>
> 
> In general this looks good to me.
> 
> > ---
> >  Documentation/acpi/i2c-muxes.txt | 58 ++++++++++++++++++++++++++++++++++++++++
> >  drivers/i2c/i2c-core.c           | 15 +++++++++--
> >  drivers/i2c/i2c-mux.c            |  8 ++++++
> >  include/linux/acpi.h             |  6 +++++
> >  4 files changed, 85 insertions(+), 2 deletions(-)
> >  create mode 100644 Documentation/acpi/i2c-muxes.txt
> > 
> 
> [...]
> 
> > +	/*
> > +	 * By default, associate I2C adapters with their parent device's ACPI
> > +	 * node.
> > +	 */
> > +	if (!has_acpi_companion(dev)) {
> > +		struct acpi_device *adev = ACPI_COMPANION(dev->parent);
> > +
> > +		if (adev)
> > +			ACPI_COMPANION_SET(dev, adev);
> 
> Instead of always doing this in the I2C core, maybe we can make it
> dependent on the host controller driver. For example the I2C designware
> driver already did this for both DT and ACPI:
> 
> 	adap->dev.parent = &pdev->dev;
> 	adap->dev.of_node = pdev->dev.of_node;
> 	ACPI_COMPANION_SET(&adap->dev, ACPI_COMPANION(&pdev->dev));
> 
> Also I would like to ask what Rafael thinks about this since he authored
> b34bb1ee71158d5b ("ACPI / I2C: Use parent's ACPI_HANDLE() in
> acpi_i2c_register_devices()").
> 
> I don't see a problem multiple Linux devices sharing a single ACPI
> companion device like in this patch but I may be forgetting something ;-)

Well, we already have that in the MFD case, but in principle it may be
problematic for things like power management (say you want to put a
child device into D3, so you use _PS3 on its ACPI companion and then
the parent is powere down instead).

At least, devices in that setup should not be attached to the ACPI PM
domain.

Thanks,
Rafael