em.4: Mention lack of SR-IOV support in BUGS

[ziaee]

Summary

MFC after: 3 days
Reported by: bms

[kbowling]

I don't think a lack of feature is a bug, there are a ton of other unsupported things. The VF code should work (or was close to it and @jmg was working on it at Juniper) so this may add rather than remove confusion.

[ziaee]

What can I do to make this acceptable to you? I feel very strongly that documenting that this is not implemented (if it is truly not implemented, which it seems to be) will save people time and inform purchasing decisions. We document missing hardware features widely in our manuals, and mdoc(7) recommends it (limitations):

CAVEATS
Common misuses and misunderstandings should be explained in this
section.

BUGS
Known bugs, limitations, and work-arounds should be described in this
section.

[kbowling]

Look at one of the data sheets and you will find about equal number of unsupported features as supported features, neither class is documented exhaustively nor would that information overload be particularly helpful.

The fundamental problem is it's not factual, hence the comment to talk to @jmg.

[ziaee]

I got ahold of jmg and he said there is a way to use this. We should document that, but I don't have a desktop atm to test.

[kbowling]

@ziaee the igbv(4) linkage and the new SR-IOV section in head we talked about on IRC is an attempt, feel free to massage it as you see fit