netdev
[Top] [All Lists]

Re: [RFC] Wireless extensions rethink

To: "Feldman, Scott" <scott.feldman@xxxxxxxxx>
Subject: Re: [RFC] Wireless extensions rethink
From: Gertjan van Wingerde <gwingerde@xxxxxxx>
Date: Tue, 15 Jun 2004 18:39:26 +0200
Cc: netdev@xxxxxxxxxxx
In-reply-to: <C6F5CF431189FA4CBAEC9E7DD5441E0103AF626C@xxxxxxxxxxxxxxxxxxxxxxxxxxxx>
References: <C6F5CF431189FA4CBAEC9E7DD5441E0103AF626C@xxxxxxxxxxxxxxxxxxxxxxxxxxxx>
Sender: netdev-bounce@xxxxxxxxxxx
User-agent: Mozilla Thunderbird 0.6 (X11/20040526)
Scott,

Sounds like a good idea. I'll start refactoring my work towards that approach. 
Please bear
with me for a couple of days and I'll post a draft patch for this.

--- Gertjan.


Feldman, Scott wrote:
I was thinking along the same lines, however I was taking the ethtool interface as the starting point (using a single ioctl for all wireless operations). The private handlers would just have to be converted to plain ioctls handled by the driver itself.

The attached patch can be used as a starting point for this.
It is not complete (not by far), but it shows the basic structure.
I've called the structure wlantool_ops, again using the example set by ethtool.

Comments?


What if we just use the ethtool ioctl that's already defined, and extend
ethtool with a wireless option:

ethtool -w DEVNAME \
        [ nwid N|off|on} ] \
        [ freq x.xx ] \
        [ mode ad-hoc|managed|master|repeater|... ] \
        [ sens N ] \
        [ ... ]

Each one of the sub-options to -w would have it's own ETHTOOL_[G|S]W...
command as well as a type-safe ethtool_op.

Running ethtool DEVNAME dumps ETHTOOL_GW... :

Wireless settings for eth0:
        nwid: AB34
        freq: 2.422G
        mode: managed
        sens: -80

Good/bad idea?

-scott



<Prev in Thread] Current Thread [Next in Thread>