Re: (RADIUS) Livingston Radius on Alpha/Linux

Tom Strickland Jr (tom@sourcenet.com)
Wed, 22 Apr 1998 07:06:47 -0400 (EDT)

You're right, I was looing at the Unix side, not thinking about the
radius side.

Then again, hey I am clueless....

<TOM>

tom@SourceNet.com

On Tue, 21 Apr 1998, Per Hedeland wrote:

> Tom Strickland Jr wrote:<tom@sourcenet.com>
> >Encrypt is a one way process, so it should take the
> >input string from the user encrypt it using the same salt (first two
> >charcters ) and then encrypt it. After this is done it should then
> >compare the two encryped strings they should batch if the user knows the
> >password.
> >
> >It should not decrypt.
>
> You're confusing the encryption used for the passwd file with the
> encryption used between RADIUS client and server - the latter must be
> decrypted by the server (using the "shared secret"), after which it can
> do (one-way) encryption for the former.
>
> Sorry, don't know what "Tom Tom Tom"'s problem is, but a reasonable
> guess is that it's a 32- vs 64-bit issue - e.g. some part of the MD5
> implementation assumes that its variables are only 32 bits. This part of
> src/md5.h might provide a clue:
>
> #if defined(__alpha) && defined(__osf__)
> typedef unsigned int UINT4;
> #else
> typedef unsigned long int UINT4;
> #endif
>
> - perhaps you just need to modify the #if a bit...
>
> --Per Hedeland
> per@erix.ericsson.se
>
-
To unsubscribe, email 'majordomo@livingston.com' with
'unsubscribe portmaster-radius' in the body of the message.
Searchable list archive: <URL:http://www.livingston.com/Tech/archive/>