Re: [PATCH] i2c: designware: Add a new ACPI HID for GOOG5000 I2C controller

[Moritz Fischer]

Hi Andy,

On Fri, Feb 27, 2026 at 08:51:26PM +0200, Andy Shevchenko wrote:
> On Fri, Feb 27, 2026 at 06:23:07PM +0000, Moritz Fischer wrote:
> > Define a new ACPI HID for GOOG5000 as used on Google Axion.
> >
> > This has been validated on Silicon.

> > Cc: Mika Westerberg <mika.westerberg@xxxxxxxxxxxxxxx>
> > Cc: Andy Shevchenko <andriy.shevchenko@xxxxxxxxxxxxxxx>
> > Cc: Jan Dabros <jsd@xxxxxxxxxxxx>

> Please, put Cc list...

> > Signed-off-by: Moritz Fischer <moritzf@xxxxxxxxxx>
> > ---

> ...somewhere here. It will have the same effect on the email. But the  
> positive
> outcome is the noise reduction in the commit message.

> See also this discussion (patch 18 subthread):
> https://lore.kernel.org/lkml/20260123113708.416727-19-bigeasy@xxxxxxxxxxxxx/

Ok, fair. Do you want me to resend or do you want to patch it when
applying?

> >  drivers/i2c/busses/i2c-designware-platdrv.c | 1 +
> >  1 file changed, 1 insertion(+)

> ...

> > +	{ "GOOG5000", 0 },

> The ID seems legit, but we need an ACPI excerpt from DSDT to show how it's
> described (just important pieces of the respective Device() object) and  
> better
> reference to the device / platform in question: Is it already on the  
> market?

I pulled this out of the EDK2 code, I hope this satisfies your request.

    Device (I2C0)
    {
      Name (_HID, "GOOG5000")
      Name (_UID, 0x00)
      Method (_STA) {
        Return (0xf)
      }
      Name (_CRS, ResourceTemplate ()
      {
        QWordMemory (
          ResourceProducer,
          PosDecode,
          MinFixed,
          MaxFixed,
          NonCacheable,
          ReadWrite,
          0x0,
          0xA02048000,
          0xA020480FF,
          0x000000000,
          0x000000100
        )
        Interrupt (ResourceConsumer, Level, ActiveHigh, Exclusive)
	{
          109
        }
      })
      ...
    }

This is being used by SoCs powering Google's Axion product line.

I can't go into more details on timelines etc, but they *are* real
products.

Thanks,
Moritz

> --
> With Best Regards,
> Andy Shevchenko