Re: [PATCH V5 2/2] virtio: i2c: Allow zero-length transactions

[Cornelia Huck <cohuck@redhat.com>]

On Wed, Oct 13 2021, Viresh Kumar <viresh.kumar@linaro.org> wrote:

> On 13-10-21, 11:26, Cornelia Huck wrote:
>> On Tue, Oct 12 2021, Viresh Kumar <viresh.kumar@linaro.org> wrote:
>> > @@ -124,13 +140,23 @@ \subsubsection{Device Operation: Operation Status}\label{sec:Device Types / I2C
>> >  
>> >  \drivernormative{\subsubsection}{Device Operation}{Device Types / I2C Adapter Device / Device Operation}
>> >  
>> > +A driver SHOULD implement the VIRTIO_I2C_F_ZERO_LENGTH_REQUEST feature.
>> 
>> Make this MUST, or maybe "MUST accept"? The central point of the feature
>> flag (at least for me) was to be able to change the semantics and format
>> without things breaking silently.
>
> Right.
>
>> I don't think we need to keep old
>> driver revisions compliant?
>
> No, we don't.

Thinking some more, we'll also probably want the driver to reject any
device that does not offer the feature. Just keep those old device/driver
revisions to themselves :)