Tom Gundersen [off-list ref] writes:
What I suggest to do is:
Leave NET_NAME_UNKNOWN in alloc_etherdev() (and any similar
constructs), but in the cases where we know that the final name set
before netdev_register() is some template, I'd still set this to
NET_NAME_ENUM as we do now to make it clear which cases have been
reviewed. So the patchset does not change much, but it would eliminate
any risk of false positives.
Does that sound reasonable to you?
Yes, sure. Although I'm not at all convinced about the value. But I
don't need to be, as that is not for me to decide :-)
Bjørn