Alan DeKok wrote:
Dan Newcombe <[EMAIL PROTECTED]> wrote:
The short of it is I'm trying to get 802.1x with PEAP to be proxied by
freeradius to an ias radius server.
Start simple. Use PAP, and "radtest" to send the packets. If that
makes FreeRADIUS proxy the packets, then go to PEAP. Otherwise,
you're test is just too complicated, and you don't know what's going
wrong.
I used radtest and the request was proxied just fine, and the ias box
gave me a positive response.
This is really an issue with the kernel, I think. If FreeRADIUS
calls the kernel "send packet" function, it should work.
That was my thoughts - was just hoping someone knew a reason why it may
not be doing so. I'm currently digging deep in the source where the
packet is sent and trying to remember my disused socket programming
memories to see just what it is trying to send, hoping to find something
glaringly obvious.
One thing I've noticed is on the non-PEAP packets, the src address of
the packet going to the IAS box is 172.28.240.73, whereas on the PEAP
packets, it is 127.0.0.1, which is causing sendmsg in
udpfromto.c:sendfromto to return an "Invalid Argument" error which says
on sendmsg:
*EINVAL - *The sum of the /iov_len/ values overflows an *ssize_t*.
Can you ping the IAS server from 172.28.240.73? Can you use
"radtest" on 172.28.240.73 to send packets to IAS?
Yes. I can ping and radtest works, and if it's not a PEAP request,
freeradius works as well - very odd.
-
List info/subscribe/unsubscribe? See http://www.freeradius.org/list/users.html
