v3.0.x: TLSv1.3 support for EAP-{TLS,TTLS,PEAP} by jimdigriz · Pull Request #3516 · FreeRADIUS/freeradius-server

jimdigriz · 2020-07-06T14:17:46Z

Picking up from where #2449 left off and aligning to draft-ietf-emu-eap-tls13 (TLS) and draft-ietf-emu-tls-eap-types (TTLS/PEAP) we have this patch set.

TLSv1.3 unit tests disabled for now (though they work) as a patched eapol_test from https://gitlab.com/jimdigriz/hostap/-/compare/master...tls13 is required.

alandekok

I'd just drop this commit. The session resumption cache includes TLS attributes, i.e. expiry, cert information, etc. If we don't cache those attributes, then the user could get back online after his certificate has expired. Which is a security violation.

jimdigriz · 2020-07-06T15:28:39Z

I'd just drop this commit. The session resumption cache includes TLS attributes, i.e. expiry, cert information, etc. If we don't cache those attributes, then the user could get back online after his certificate has expired. Which is a security violation.

Which commit? Nothing was reference. Do you mean 'tls: fix session resumption' (8cf675f)?

jimdigriz · 2020-07-06T16:00:12Z

I'd just drop this commit. The session resumption cache includes TLS attributes, i.e. expiry, cert information, etc. If we don't cache those attributes, then the user could get back online after his certificate has expired. Which is a security violation.
I'd just drop this commit. The session resumption cache includes TLS attributes, i.e. expiry, cert information, etc. If we don't cache those attributes, then the user could get back online after his certificate has expired. Which is a security violation.

Which commit? Nothing was reference. Do you mean 'tls: fix session resumption' (8cf675f)?

Removed, and reworked the unit test so to tickle an attribute so that session resumption works.

jimdigriz · 2020-07-06T16:11:17Z

The only outstanding task I know of is to discard any early_data

arr2036 · 2020-07-06T16:30:18Z

This is meant to be a patch series for TLS 1.3. Strip out the commits not related to TLS 1.3.

jimdigriz · 2020-07-06T17:52:26Z

This is meant to be a patch series for TLS 1.3. Strip out the commits not related to TLS 1.3.

Done

jimdigriz · 2020-07-06T18:08:20Z

Don't merge this, I have found some bugs...

draft-ietf-emu-eap-tls13 (tls) and draft-ietf-emu-tls-eap-types (ttls/peap)

jimdigriz · 2020-07-06T19:16:12Z

Okay, added SSL_CTX_set_max_early_data(..., 0) (openssl/openssl@c39e404) to avoid early data as per https://tools.ietf.org/html/draft-ietf-emu-eap-tls13-10#section-2.1.1 and I tickled out the bug I found in EAP-TTLS (label not being set on a path).

Should be ready for consideration/review.

The other patches that were in here I will include in another PR along with a fix for #1756

alandekok · 2020-07-06T22:37:31Z

-	eaptls_gen_eap_key(handler->request->reply, tls_session->ssl,
-			   handler->type);
+	rad_assert(context_tls13[0] == handler->type);	/* expected by eaptls_gen_eap_key */
+	eaptls_gen_eap_key(handler->request->reply, tls_session->ssl, context_tls13, context_tls13_size);


eaptls_gen_key() always dereferences context, which should be handler->type. So it's not clear why this change is necessary. Why not just leave the old code for eaptls_gen_key() as-is?

Resolved by ~~eced45c~~ f97c486

alandekok · 2020-07-06T22:39:15Z

+	/*
+	 *	EAP (hopefully) does not have middlebox deployments
+	 */
+#ifdef SSL_OP_ENABLE_MIDDLEBOX_COMPAT


I'd add a comment about "If set then dummy Change Cipher Spec (CCS) messages are sent in TLSv1.3. This has the effect of making TLSv1.3 look more like TLSv1.2 so that middleboxes that do not understand TLSv1.3 will not drop the connection. This isn't needed for EAP-TLS, so we disable it"

Resolved by becf52a

alandekok · 2020-07-06T22:39:58Z

 	SSL_CTX_set_options(ctx, ctx_options);

+#if OPENSSL_VERSION_NUMBER >= 0x10100000L && OPENSSL_VERSION_NUMBER < 0x10101000L
+	SSL_CTX_set_max_early_data(ctx, 0);


I'd add a comment "TLS 1.3 introduces the concept of early data (also known as zero round trip data or 0-RTT data). Early data allows a client to send data to a server in the first round trip of a connection, without waiting for the TLS handshake to complete if the client has spoken to the same server recently. This doesn't work for EAP, so we disable early data"

Resolved by 3abc316

alandekok · 2020-07-06T22:41:08Z

+	size_t len;

-	prf_size = strlen(prf_label);
+	len = strlen(label);


I'm not sure why variable names were changed, but whatever.

It is no longer a PRF in TLS 1.3 but a HKDF

alandekok · 2020-07-06T22:42:15Z

 	memcpy(p, s->s3->client_random, SSL3_RANDOM_SIZE);
 	p += SSL3_RANDOM_SIZE;
 	memcpy(p, s->s3->server_random, SSL3_RANDOM_SIZE);
+	p += SSL3_RANDOM_SIZE;


I don't think this line is necessary. Compilers may complain about unused assignments

Resolved by 5deea57

alandekok · 2020-07-06T22:54:36Z

+			 *	TLSv1.3 exports a different key depending on the length
+			 *	requested so ask for *exactly* what the spec requires
+			 */
+			eapttls_gen_challenge(ssl, challenge, vp->vp_length + 1);


https://tools.ietf.org/html/rfc5281#section-11.2.4 says 16 octets of challenge, followed by 1 octet of ident.

We might want to check the ident, too...

The comment above this (vintage deKok circa 2003): "We're a little forgiving in that we have loose checks on the length, and we do NOT check the Id (first octet of the response to the challenge). But if the client gets the challenge correct, we're not too worried about the Id."

I suspect this is because this code is in the middle of the VP iterator and may not have seen the challenge yet. Do you want me to break the check out to actually look at the Id?

…code rather than sensibly optimise away FreeRADIUS#3516 (review)

…viour

* eap-tls: tls version 1.3 draft-ietf-emu-eap-tls13 (tls) and draft-ietf-emu-tls-eap-types (ttls/peap) * SQUASH: remove unneeded line, compilers complain about more readable code rather than sensibly optimise away FreeRADIUS#3516 (review) * SQUASH: comment about early data * SQUASH: middleware boxen comment * SQUASH: method of using context is too confusing retain original behaviour

jimdigriz force-pushed the tls13 branch from ae6d0f7 to 7fe495e Compare July 6, 2020 14:33

alandekok requested changes Jul 6, 2020

View reviewed changes

jimdigriz force-pushed the tls13 branch from 7fe495e to bda7c2e Compare July 6, 2020 15:58

jimdigriz force-pushed the tls13 branch 2 times, most recently from c2e9f67 to 195eb03 Compare July 6, 2020 16:18

jimdigriz force-pushed the tls13 branch from 195eb03 to 1b87a97 Compare July 6, 2020 16:54

jimdigriz changed the title ~~TLSv1.3 support for EAP-{TLS,TTLS,PEAP}~~ v3.0.x: TLSv1.3 support for EAP-{TLS,TTLS,PEAP} Jul 6, 2020

jimdigriz force-pushed the tls13 branch 2 times, most recently from 2524e3b to 5a0bce7 Compare July 6, 2020 17:52

jimdigriz force-pushed the tls13 branch from 5a0bce7 to 20c5860 Compare July 6, 2020 18:40

eap-tls: tls version 1.3

adff0fa

draft-ietf-emu-eap-tls13 (tls) and draft-ietf-emu-tls-eap-types (ttls/peap)

jimdigriz force-pushed the tls13 branch from 20c5860 to adff0fa Compare July 6, 2020 19:12

alandekok reviewed Jul 6, 2020

View reviewed changes

jimdigriz added 4 commits July 7, 2020 07:25

SQUASH: remove unneeded line, compilers complain about more readable …

5deea57

…code rather than sensibly optimise away FreeRADIUS#3516 (review)

SQUASH: comment about early data

3abc316

SQUASH: middleware boxen comment

becf52a

SQUASH: method of using context is too confusing retain original beha…

f97c486

…viour

jimdigriz force-pushed the tls13 branch from eced45c to f97c486 Compare July 7, 2020 07:17

alandekok merged commit e8b4d28 into FreeRADIUS:v3.0.x Jul 7, 2020

jimdigriz deleted the tls13 branch July 7, 2020 11:16

This was referenced Jul 7, 2020

Backport some TLS1.3 fixes #2449

Closed

v3.0.x: misc (EAP-)TLS fixes for review #3517

Closed

jimdigriz mentioned this pull request Jul 23, 2020

Fix OpenSSL 1.1.0 builds #3534

Merged

spaceone mentioned this pull request Sep 21, 2023

ttls fails with TLS 1.3 (openssl 1.1.1) #2385

Closed

Uh oh!

Conversation

jimdigriz commented Jul 6, 2020 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

alandekok left a comment

Choose a reason for hiding this comment

Uh oh!

jimdigriz commented Jul 6, 2020

Uh oh!

jimdigriz commented Jul 6, 2020

Uh oh!

jimdigriz commented Jul 6, 2020

Uh oh!

arr2036 commented Jul 6, 2020

Uh oh!

jimdigriz commented Jul 6, 2020

Uh oh!

jimdigriz commented Jul 6, 2020

Uh oh!

jimdigriz commented Jul 6, 2020 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

alandekok Jul 6, 2020 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

jimdigriz Jul 7, 2020 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

alandekok Jul 6, 2020

Choose a reason for hiding this comment

Uh oh!

jimdigriz Jul 7, 2020

Choose a reason for hiding this comment

Uh oh!

alandekok Jul 6, 2020

Choose a reason for hiding this comment

Uh oh!

jimdigriz Jul 7, 2020

Choose a reason for hiding this comment

Uh oh!

alandekok Jul 6, 2020

Choose a reason for hiding this comment

Uh oh!

jimdigriz Jul 7, 2020 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

alandekok Jul 6, 2020

Choose a reason for hiding this comment

Uh oh!

jimdigriz Jul 7, 2020 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

alandekok Jul 6, 2020

Choose a reason for hiding this comment

Uh oh!

jimdigriz Jul 7, 2020 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants

jimdigriz commented Jul 6, 2020 •

edited

Loading

jimdigriz commented Jul 6, 2020 •

edited

Loading

alandekok Jul 6, 2020 •

edited

Loading

jimdigriz Jul 7, 2020 •

edited

Loading

jimdigriz Jul 7, 2020 •

edited

Loading

jimdigriz Jul 7, 2020 •

edited

Loading

jimdigriz Jul 7, 2020 •

edited

Loading