--On 26 March 2011 10:28:45 -0700 Jeff Lasman <[email protected]> wrote:

Some (not many) of the users of the exim.conf file I maintain have
requested  that we check to make sure the authenticated user sending
email only use a  From address we have on file for them.

They suggest we only send email if the From address is the same as the
domain  registered to the authenticated sender.

I once thought the RFCs said something about that, but I can't find it
anywhere.  I'm probably not googling as well as I should.

No, the RFCs don't say anything about that. In fact, I think they say that it's perfectly acceptable that the From address and the envelope sender differ. However, if they do differ, then the Sender header should be used.

In an ideal world, you'd not be able to do this without permission of the owner of the From address. That's reasonably achievable where you control the domain of both the sender address and the spoofed address, but not otherwise. I guess you could build a web app that gets permission using a mail-back mechansism, in the same way that mailing list subscriptions often work. However, revocation would become a problem. OAuth for SMTP, anyone?

Exim's mechanisms to restrict spoofing are documented in 44.1 of the docs. Essentially, you need to add the line "control = submission" to an ACL at MAIL, RCPT or pre-data. There are some options available: "sender_retain", and "domain=...", "name=..." which can be used to set the sender address domain and the sender's full name.

You should avoid using "sender_retain" for your purposes.

With these measures, you can ensure that the real sender is identifiable from the message envelope, and from a "Sender" header. However, nothing here prevents anyone from putting whatever they like in the From header.

I suppose that you could simply fix up the "From:" header by putting something like
   "headers_remove = From:Sender
    headers_add = From: $h_sender"

on the router for a properly fixed up message see 44.17. This is an untested suggestion. You may have to keep track of fixed up messages, and make a special router for this purpose.

Our exim.conf file is used by webhosting companies, and it's always
allowed  users, once they authenticate, to use any outgoing email address
they want.  I  make use of that myself,sometimes sending email out with
my gmail return  address, and sending from several domains I use but
maintaining only one  outgoing MTA in my desktop client.

What is the current consensus on this issue?  Is it mentioned in any RFC?

Thanks.

Jeff
--
Jeff Lasman, Nobaloney Internet Services
Post Office Box 52200, Riverside, CA  92517
Our blists address used on lists is for list email only
Phone +1 951 643-5345, or see: "http://www.nobaloney.net/contactus.html";



--
Ian Eiloart
IT Services, University of Sussex
01273-873148 x3148
For new support requests, see http://www.sussex.ac.uk/its/help/



--
## List details at http://lists.exim.org/mailman/listinfo/exim-users
## Exim details at http://www.exim.org/
## Please use the Wiki with this list - http://wiki.exim.org/

Reply via email to