SPF provides 2 results. People get confused because they often think there is only one.
The first result is based on the RFC7489.HELO and the second result is based on the RFC7489.MAILFROM. The first result could allow you to close a connection at the RFC5321.Helo stage, while the second result would allow you to close the connection after the RFC5321.Mailfrom. In practice both results are integrated in higher reputation layers... DMARC uses only the second result (and identifiers). As a side note and as Terry points out, Office 365, only uses the second results for SPF. Many implementations of SPF use only the second result. RFC7489.MAILFROM is defined as RFC5321.MailFrom unless it is empty, in that case it is postmaster@<RFC5321.Helo> Hope this helps. On Mon, May 9, 2016 at 8:17 AM, Terry Zink via dmarc-discuss < dmarc-discuss@dmarc.org> wrote: > This is a good point, I'm not sure about how others do it, but in Office > 365 we compare the 5322.From domain against the domain that was used to > authenticate SPF. That's the 5321.MailFrom unless it is <>, in which case > we use the HELO/EHLO domain. That would allow a DMARC pass in the absence > of a DKIM signature. > > -- Terry > > -----Original Message----- > From: dmarc-discuss [mailto:dmarc-discuss-boun...@dmarc.org] On Behalf Of > Sistemisti Posta via dmarc-discuss > Sent: Monday, May 9, 2016 3:38 AM > To: dmarc-discuss@dmarc.org > Subject: [dmarc-discuss] DMARC and null path > > Hello DMARC users, > > because I'm new in DMARC world I'm trying to understand some details > before to start with policy implementation. > > A detail I would understand is related to mails with null path, or null > sender address, typically sent by Delivery Status Notifications. > > It seems that the only way to pass DMARC with null path is to DKIM sign > the mails. Is it true? > > I ask this because RFC7489 says that exists a condition when DMARC > considers the null path: > > "Note that the RFC5321.HELO identity is not typically used in the > context of DMARC (except when required to "fake" an otherwise null > reverse-path)" > > And: > > "DMARC uses the result of SPF authentication of the MAIL FROM identity. > Section 2.4 of [SPF] describes MAIL FROM processing for cases in which > the MAIL command has a null path." > > RFC4408 says accordingly: > > 'When the reverse-path is null, this document defines the "MAIL FROM" > identity to be the mailbox composed of the localpart "postmaster" and > the "HELO" identity (which may or may not have been checked separately > before).' > > So if a mail with null path and HELO domain equal to RFC5322.From passes > the SPF check, why should DMARC fail? > > See at this live example: > > libero.it descriptive text "v=spf1 ip4:212.48.25.128/25 > ip4:212.48.14.160/27 include:srs.bis.na.blackberry.com > include:srs.bis.eu.blackberry.com include:srs.bis.ap.blackberry.com > include:mail.zendesk.com -all" > _dmarc.libero.it descriptive text "v=DMARC1\; p=quarantine\; ... > > If 212.48.14.166 sends a mail with null path, > RFC5322.From=<local>@libero.it and *helo=libero.it*, then SPF thinks to > have a 'MAIL FROM' like "postmas...@libero.it", passing this result to > DMARC for alignment with RFC5322.From. (If it passes the helo domain is > the same) > > The result I see: > ~~~ > 2016-05-09T09:47:06.481095+02:00 postfix/smtpd[14063]: 3r3Dx63PshzFpVy: > client=smtp-32-i6.italiaonline.it[212.48.14.166] > 2016-05-09T09:47:06.636894+02:00 postfix/qmgr[17134]: 3r3Dx63PshzFpVy: > from=<>, size=308079, nrcpt=x (queue active) > 2016-05-09T09:47:06.551037+02:00 opendkim[6782]: 3r3Dx63PshzFpVy: > smtp-32-i6.italiaonline.it [212.48.14.166] not internal > 2016-05-09T09:47:06.551173+02:00 opendkim[6782]: 3r3Dx63PshzFpVy: not > authenticated > 2016-05-09T09:47:06.551960+02:00 opendkim[6782]: 3r3Dx63PshzFpVy: no > signature data > 2016-05-09T09:47:06.594831+02:00 opendmarc[9812]: SPF: 3r3Dx63PshzFpVy: > libero.it pass > 2016-05-09T09:47:06.595936+02:00 opendmarc[9812]: 3r3Dx63PshzFpVy: > libero.it pass > ~~~ > > The mail with null path and no DKIM signs passes DMARC. For me this is a > correct result; isn't it? > In this particular case we could complain that the client doesn't send > an helo equal to his hostname, but this is not DMARC related. > > > I would implement DMARC. For DSN sent to Internet by any authorized MTA > I would declare an SPF record as: > > mta.example.com IN TXT "v=spf1 a -all" > _dmarc.example.com descriptive text "v=DMARC1\; p=reject\;" > > Let suppose the above host sends a DSN with null path, > helo="mta.example.com" and RFC5322.From=postmas...@mta.example.com. > I expect DMARC passes (in relaxed mode) because SPF passes. > > Could you explain me where I'm wrong? > > Thank you very much for your help. > Best Regards > Marco > _______________________________________________ > dmarc-discuss mailing list > dmarc-discuss@dmarc.org > http://www.dmarc.org/mailman/listinfo/dmarc-discuss > > NOTE: Participating in this list means you agree to the DMARC Note Well > terms (http://www.dmarc.org/note_well.html) > > _______________________________________________ > dmarc-discuss mailing list > dmarc-discuss@dmarc.org > http://www.dmarc.org/mailman/listinfo/dmarc-discuss > > NOTE: Participating in this list means you agree to the DMARC Note Well > terms (http://www.dmarc.org/note_well.html) >
_______________________________________________ dmarc-discuss mailing list dmarc-discuss@dmarc.org http://www.dmarc.org/mailman/listinfo/dmarc-discuss NOTE: Participating in this list means you agree to the DMARC Note Well terms (http://www.dmarc.org/note_well.html)
