> > Yes, I thought briefly about splitting SMBUS_QUICK into read-write
> > variants too. To me the question is if this feature is still used on modern
> > devices or if this is more a relict of the past. I don't know enough
> > about SMBUS to answer that.

Well, note that there are zero-length messages in I2C allowed as well.
Not only in SMBUS. I mainly use the term SMBUS_QUICK because it covers
both cases.


> Checking on ThunderX:
...
> Do all these other numbers make sense (although there are no
> devices)?

It makes sense in a way that it shows SMBUS_QUICK_WRITE is broken :) It
doesn't react to ACK/NACK properly. So, what needs to be done:

1) remove SMBUS_QUICK as you did in this patch 2) move the length check
so it doesn't only check read messages but also write messages. That is
to prevent handling custom setup I2C messages with a length of 0 (which
is legal). I'd suggest to return -EOPNOTSUPP in this case.

Thanks,

   Wolfram