linux-kernel - Re: [PATCH RFC] remove custom Michael MIC implementation

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <20170403090359.GK3755@eros>
Date:   Mon, 3 Apr 2017 19:03:59 +1000
From:   "Tobin C. Harding" <me@...in.cc>
To:     Kalle Valo <kvalo@...eaurora.org>
Cc:     Wolfram Sang <wsa@...-dreams.de>, linux-kernel@...r.kernel.org,
        Greg Kroah-Hartman <gregkh@...uxfoundation.org>,
        devel@...uxdriverproject.org, Tycho Andersen <tycho@...ho.ws>,
        linux-wireless@...r.kernel.org
Subject: Re: [PATCH RFC] remove custom Michael MIC implementation

On Mon, Apr 03, 2017 at 08:19:40AM +0300, Kalle Valo wrote:
> + linux-wireless
> 
> "Tobin C. Harding" <me@...in.cc> writes:
> 
> > On Fri, Mar 31, 2017 at 09:58:51AM +0200, Wolfram Sang wrote:
> >> 
> >> > The code is untested, I have hardware in the mail.
> >> 
> >> Cool!
> >
> > The card I have is a Spectec FCC ID: S2Y-WLAN-11B-G which I believe is
> > a SDW-823 and should use the ks7010 driver. I am going to attempt to
> > get it running on a Raspberry Pi B+. I ordered the wrong size break
> > out board originally so waiting on the new one now.
> >
> >> 
> >> > If any one is interested and has any comments I would really like to
> >> > hear them. I am open to all suggestions (even down to trivial coding
> >> > style issues).
> >> 
> >> I'll just repeat that the key move to get this driver out of staging is
> >> to get away from the WEXT interface to CFG80211. Otherwise no chance
> >> that wireless maintainers will even look at it. This is a huge change
> >> but once it is done, features like Michael MIC come with it for free
> >> (from what I recall, I am not a wireless expert myself).
> >
> > That would explain why I could not find more than the Orinoco driver
> > using the Michael MIC module directly.
> 
> I think cfg80211 and mac80211 got mixed up. mac80211 (the full IEEE
> 802.11 stack for "softmac" devices) provides Michael MIC implementation,
> but cfg80211 (for "hardmac" devices) does not.

Cool, thanks for clarifying. Hilariously I was just sitting down
trying to figure out what was up after a day spent trying to merge
ideas from ath6kl (fullmac) and ks7010. I finally spent some time
reading the cw1200 driver (softmac) for further inspiration.

> >> Without the CFG80211 conversion, replacing the Michael custom
> >> implementation with the in-kernel one makes the driver a tad better and
> >> is good exercise. However, it will sadly not help to get the driver out
> >> of staging.
> >
> > I'll drop it then. Could you please tell me, is there any thing else
> > more I need to do to let LKML know that this RFC is dropped? Or is
> > this reply enough. I don't want to use any ones time unnecessarily.
> >
> >> 
> >> But if you want a clean WEXT driver first, this is a step in the right
> >> direction.
> >
> > Let's go for a CFG80211 driver and get out of staging :) So next step
> > is I guess study the ath6kl driver, learn how CFG80211 is done and
> > implement that interface in ks7010? Oh, and test that it works.
> 
> Please keep linux-wireless list in loop so that people on that list can
> help.

How newbie friendly is the linux-wireless list please? I am having
trouble separating the data path code from the control path. I think I
will spend a few more days on it though before asking any questions.

Except one: do you know off the top of your head of a canonical
implementation of a softmac wi-fi driver.

thanks,
Tobin.