tech-net archive

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index][Old Index]

Re: ifconfig(8) - a tale of two states



Mouse <mouse%Rodents-Montreal.ORG@localhost> writes:

> Is this inconsistent?  I don't know enough about 802.11, but it seems
> plausible, at least, to me that the radio could be on (media active)
> without the interface having an SSID...particularly for passive traffic
> snooping.

I think this is the crux of the issue.

There are multiple things

  PHY/MAC parts are commanded to be up

  PHY/MAC have formed a link with "the network"

With ethernet, it's not really that interesting if we are up but have no
peer, and it's not complicated.   802.11 is much more complicated, in
that IBSS mode is useful to many, even if we are not aware of peers.


Another point is that "link state" only really conceptually applies to a
technology where there is a single peer and there is a notion of having
a PHY connection to it.

In the case of 802.11 IBSS mode, one can have the radio functioning and
configured, where "active" is appropriate.  But the link state is
neither up nor down, as there is no peer to which we should be
connected.

So I think that either

  the link state abstraction needs a new state NOT_APPLICABLE (spell it
  NA if you feel like it) to indicated that for the current
  configuration, it does not make sense to talk about the link being up
  or down.

  somehow have a notion of not having link state in the NA case, but
  that seems much harder for no good reason.

So, 802.11 should have link state NA for IBSS, HOSTAP, and MONITOR.

Certainly for HOSTAP and IBSS it makes sense to have addresses.

Attachment: signature.asc
Description: PGP signature



Home | Main Index | Thread Index | Old Index