wpa_supplicant Windows port, again (WPA2)

[Bryan Kadzban]

Jouni Malinen wrote:
> Unfortunately, it looks like increasing polling frequency is not
> really going to help much. Either there is some extra latency in
> select() on Windows or WinPcap takes some time to get the frames. I
> have the code for using faster polling in CVS, but this did not seem
> to change 4-Way Handshake timing that much. It may be better for some
> cases, e.g., if two frames can be received with one poll, but other
> than that, the results were just marginally better than with the
> previous version.

Well, 0.4.5 did fix the problem I was having.  It successfully completes
the 4-way handshake now, every time I've tried it.

That's without turning debug off, too.  However, on system startup, I
had to turn it off, because on the machine I tested it on, there seemed
to be a race between wpa_supplicant and DHCP, or something.  Turning
debug off helped a lot (but didn't quite totally fix the problem;
netlogon still starts a little too early).  But I'll be looking into
that a lot more tomorrow; I think it has something to do with the DHCP
service.

Anyway, the one thing I would like to know for sure is, after the
CTRL-EVENT-CONNECTED string gets sent to all connected frontends, should
the machine be able to pass network traffic?  (I.e., am I watching for
the right string from wpa_supplicant?)  If not, is there something else
that I should be watching for that signals when the card is capable of
using the network?

Thanks!
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 256 bytes
Desc: OpenPGP digital signature
Url : http://lists.shmoo.com/pipermail/hostap/attachments/20050926/40afd600/attachment.pgp