Re: systemd-rfkill regression on 5.11 and later kernels

[Johannes Berg]

On Thu, 2021-03-18 at 12:16 +0100, Johannes Berg wrote:
> > Yeah, that's a dilemma.  An oft-seen trick is to add more bytes for
> > the future use, e.g. extend to 16 bytes and fill 0 for the remaining.
> 
> Yeah, I guess I could stick a reserved[15] there, it's small enough.

Actually, that doesn't really help anything either.

If today I require that the reserved bytes are sent as 0 by userspace,
then any potential expansion that requires userspace to set it will
break when userspace does it and runs on an old kernel.

If I don't require the reserved bytes to be set to 0 then somebody will
invariably get it wrong and send garbage, and then we again cannot
extend it.

So ... that all seems pointless. I guess I'll send the patch as it is
now.

johannes