Virtualization - Jul 2021 - [PATCH v15] i2c: virtio: add a virtio i2c frontend driver

On Fri, Jul 23, 2021 at 7:44 AM Jie Deng <jie.deng at intel.com> wrote:
> +
> +       ret = virtio_i2c_setup_vqs(vi);
> +       if (ret)
> +               return ret;
> +
> +       vi->adap.owner = THIS_MODULE;
> +       snprintf(vi->adap.name, sizeof(vi->adap.name),
> +                "i2c_virtio at virtio bus %d", vdev->index);
> +       vi->adap.algo = &virtio_algorithm;
> +       vi->adap.quirks = &virtio_i2c_quirks;
> +       vi->adap.dev.parent = &vdev->dev;
> +       i2c_set_adapdata(&vi->adap, vi);
> +
> +       /*
> +        * Setup ACPI node for controlled devices which will be probed
through
> +        * ACPI.
> +        */
> +       ACPI_COMPANION_SET(&vi->adap.dev, ACPI_COMPANION(pdev));
Since there is now a generic way for virtio drivers to link up with OF
device nodes, maybe this should be handled the same way in the
virtio core rather than the driver?
> index 70a8057a..99aa27b 100644
> --- a/include/uapi/linux/virtio_ids.h
> +++ b/include/uapi/linux/virtio_ids.h
> @@ -55,6 +55,7 @@
>  #define VIRTIO_ID_FS                   26 /* virtio filesystem */
>  #define VIRTIO_ID_PMEM                 27 /* virtio pmem */
>  #define VIRTIO_ID_MAC80211_HWSIM       29 /* virtio mac80211-hwsim */
> +#define VIRTIO_ID_I2C_ADAPTER          34 /* virtio i2c adapter */
>  #define VIRTIO_ID_BT                   40 /* virtio bluetooth */
This will now conflict with Viresh's patch that adds all the other IDs.
Not sure if there is anything to be done about that.

       Arnd

On 23-07-21, 11:03, Arnd Bergmann wrote:
> > index 70a8057a..99aa27b 100644
> > --- a/include/uapi/linux/virtio_ids.h
> > +++ b/include/uapi/linux/virtio_ids.h
> > @@ -55,6 +55,7 @@
> >  #define VIRTIO_ID_FS                   26 /* virtio filesystem */
> >  #define VIRTIO_ID_PMEM                 27 /* virtio pmem */
> >  #define VIRTIO_ID_MAC80211_HWSIM       29 /* virtio mac80211-hwsim */
> > +#define VIRTIO_ID_I2C_ADAPTER          34 /* virtio i2c adapter */
> >  #define VIRTIO_ID_BT                   40 /* virtio bluetooth */
> 
> This will now conflict with Viresh's patch that adds all the other IDs.
> Not sure if there is anything to be done about that.
An easier way of avoiding all such conflicts can be:

- Michael applies my first patch (which sync's the device id's from
specs) for
  5.14-rc4. Rest of the patches can go for 5.15.

- And then Wolfram applies this series over rc4 instead of rc1.

Or we can leave the conflict there for Linus to handle.

-- 
viresh

On 2021/7/23 17:03, Arnd Bergmann wrote:
> On Fri, Jul 23, 2021 at 7:44 AM Jie Deng <jie.deng at intel.com>
wrote:
>
>> +
>> +       ret = virtio_i2c_setup_vqs(vi);
>> +       if (ret)
>> +               return ret;
>> +
>> +       vi->adap.owner = THIS_MODULE;
>> +       snprintf(vi->adap.name, sizeof(vi->adap.name),
>> +                "i2c_virtio at virtio bus %d",
vdev->index);
>> +       vi->adap.algo = &virtio_algorithm;
>> +       vi->adap.quirks = &virtio_i2c_quirks;
>> +       vi->adap.dev.parent = &vdev->dev;
>> +       i2c_set_adapdata(&vi->adap, vi);
>> +
>> +       /*
>> +        * Setup ACPI node for controlled devices which will be probed
through
>> +        * ACPI.
>> +        */
>> +       ACPI_COMPANION_SET(&vi->adap.dev, ACPI_COMPANION(pdev));
> Since there is now a generic way for virtio drivers to link up with OF
> device nodes, maybe this should be handled the same way in the
> virtio core rather than the driver?

I'm currently based on the I2C tree. Has that patch been already merged ?

Anyway, I think we can send an additional patch to remove this line once 
that

"generic way" patch is merged.

Regards,

Jie