On Mon, 2008-11-24 at 09:29 -0500, James Carlson wrote:
> Sebastien Roy writes:
> > That would only make sense to me if there was a documented way of
> > knowing what secobj name was being used by NWAM for any given WiFi link.
> > Otherwise, it makes it difficult to diagnose problems through the CLI by
> > doing:
> > 
> > dladm connect-wifi -e <essid> -k <secobj-name> <link>
> > 
> > ... after NWAM has automatically created <secobj-name> through the GUI.
> 
> I think that's 'just' a debugging issue.  Normal users should never
> have a need to do anything like that, and when they do, I'd be ok with
> saying something like this:
> 
>   NWAM stores the security objects using hashes.  If you only have one
>   key stored, then "dladm show-secobj" will identify the one needed.
>   Otherwise, you can recreate the hash used in NWAM by doing this:
> 
>       echo 'ESSID name' 'BSSID string' | digest -a sha1
> 
>   For example:
> 
>       % echo 'sunwifi' '0:18:b0:7d:34:c1' | digest -a sha1
>       09a75098435200505fb5de8e59a3a4edfe544781
>       % pfexec dladm connect-wifi -k 
> nwam-09a75098435200505fb5de8e59a3a4edfe544781
> 
> 
> Well, *something* like that.  Another problem here is that the secobj
> names are of limited length, and dladm(1M) doesn't tell you what the
> limits are.  You have to guess.

Instead of having to crunch the hashing algorithm, we could instead
include the secobj name in the /etc/nwam/known_wifi_nets file (or
equivalent for Phase 1).  The nwamd daemon could also grab the name out
of there instead after having generated the key.

-Seb


_______________________________________________
networking-discuss mailing list
[email protected]

Reply via email to