Fix regression introduced on 272446r

[araujo]

Summary

(based on) PR: https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=206478

lagg(4) supports the protocol none, where it disables any traffic without
disabling the lagg(4) interface itself.

Test Plan

ifconfig lagg0 create
ifconfig lagg0 laggproto none

[araujo]

Rework the patch, now the statement looks more clean and
easy to understand.

[bapt]

sys/net/if_lagg.c

1268

why not a single level of indentation?

[araujo]

sys/net/if_lagg.c

1268

Do you mean a single level of statement? If it is about the statement.

For two reasons:

1. I would prefer have the statement for the LAGG_PROTO explicit, much clear to understand, what is the intention there.

2. Theoretically only one statement might be faster, but in the piece of code, the only way to know that is checking what the compiler will build with this statement. However, as this statement is not implemented in lagg_ioctl(), and it is not a fast path. It is OK to choose have more clarity in the code.

[araujo]

sys/net/if_lagg.c

1268

Just rephrase here:

2. Theoretically only one statement might be faster, but in the piece of code, the only way to know that is checking what the compiler will build with this statement. However, as this statement IS implemented in lagg_ioctl(), and it is not a fast path. It is OK to choose have more clarity in the code.

I just fix a misspelling that might make my affirmation confused.

[araujo]

Abandon this revision in favor of: https://reviews.freebsd.org/D5226
That patch seemed much more reasonable and I made the tests it gets close with the same fix for 10-Stable.

Best,