Re: [Ieee80211-devel] Re: [PATCH 23/29] ieee80211 Added ieee80211_radiotap.h

[Mike K. <dr...@ki...>]

> A bunch of information that is put into the radiotap header is put inside=
=20
> struct ieee80211_rx_stats when the driver passes the frame to the ieee802=
11=20
> layer. I think it is a better idea to expand ieee80211_rx_stats to includ=
e=20
> all information necessary for filling in a radiotap header, and let the=
=20
> ieee80211 tack on the header.

Yeah, we talked about this, I was just waiting for things to settle in
the tree before I tackled it.

> I would prefer either that or giving drivers access to the generic monito=
r=20
> mode path so frames can be passed to userspace ASAP.

I'm of 2 minds here...  For a driver to implement radiotap, all it takes
is a struct that the driver fills in that LOOKS like a radiotap dynamic
header to userspace, but is really just a static struct fill.  For the
ieee80211 layer to do the same, it will have to actually dynamically
contstruct the header, per frame, based on what the driver gives us.
It's quicker to do per-driver header construction.

I don't mind implementing both, and including a field that ieee80211
layer can pick to do headers dynamically or not.

Maybe this is too weird an idea, and we should just pick "80211" or
"driver".  I'm willing to write the code for either use case.  I suppose
that rfmon is a fairly special case and won't be as much of a cpu drain
as a wired ethernet NIC or normal mode would be.

> Also, we probably want an ioctl or some way to switch between raw ieee802=
11=20
> frames and radiotap headers.

Sure, *as long as every driver uses the same ioctl* :)  PLEASE PLEASE
PLEASE lets pick a constant ioctl name for this.  I assume we'll want it
as an iwpriv ioctl, thats where the other behavior of this ilk lives.
If we're going to implement an iwpriv control for it, shall we plan
forward and make it an integer not a boolean, so we can have different
headers if we need them?

The ioctl is pretty trivial in any case, but you're right, we should
have it.

-m

--=20
Mike Kershaw/Dragorn <dr...@ki...>
GPG Fingerprint: 3546 89DF 3C9D ED80 3381  A661 D7B2 8822 738B BDB1

"Dear Die-ary,
    Today I stuffed some dolls with dead rats I put in a blender.
    I'm beginning to wonder if, maybe, I really am screwed up."
                   -- Johnny C. (JtHM)