Skip to contentSkip to conversation

Send auth fail to client on reneg failure

PATCH In review Eric Thorpe 3 participants 4 messages Apr 11, 2019 – Aug 13, 2020

lore
Share & tools
lore
#0 lore raw

PATCH v2 1/2 Send auth fail to client on reneg failure

Hi All,

This patch relies on Arne's "Add send_control_channel_string_dowork variant" patch.

This patch modifies auth so that on a renegotiation the client is informed of a SESSION re-auth failure during a renegotiation if either their auth-token has expired, or they enter a wrong password in the case of auth-nocache for example.

This also addresses my previous patch for supporting a client reason being rejected.

Regards,
Eric

Signature · 6 lines
--
Eric Thorpe
SparkLabs Developer
https://www.sparklabs.com
https://twitter.com/sparklabs
[email protected]...

From bb0fada899a43e48e2ca38358912f92447422f53 Mon Sep 17 00:00:00 2001
From: Eric Thorpe <[email protected]>
Date: Wed, 10 Apr 2019 19:06:16 +1000
Subject: [PATCH] Send auth failed message to client on renegotiation failure
via user-pass or auth-token expiry

---
 src/openvpn/multi.c      |  1 +
 src/openvpn/push.c       | 30 ++++++++++++++++++++++++++++++
 src/openvpn/ssl_common.h |  1 +
 src/openvpn/ssl_verify.c |  7 +++++++
 4 files changed, 39 insertions(+)
src/openvpn/multi.c +1 -0
index 28c3b8867..a883faa79 100644
@@ -2085,6 +2085,7 @@ multi_connection_established(struct multi_context *m, struct multi_instance *mi)
 
         /* set flag so we don't get called again */
         mi->connection_established_flag = true;
+        mi->context.c2.tls_multi->connection_established = true;
 
         /* increment number of current authenticated clients */
         ++m->n_clients;
src/openvpn/push.c +30 -0
index dd5bd4163..327ae0891 100644
@@ -233,6 +233,36 @@ send_auth_failed(struct context *c, const char *client_reason)
     gc_free(&gc);
 }
 
+/*
+ * Send auth failed message from server to client without scheduling.
+ * Main use for queuing a message during renegotiation
+ */
+void
+send_push_reply_auth_failed(struct tls_multi *multi, const char *client_reason)
+{
+    struct gc_arena gc = gc_new();
+    static const char auth_failed[] = "AUTH_FAILED";
+    size_t len;
+
+    len = (client_reason ? strlen(client_reason)+1 : 0) + sizeof(auth_failed);
+    if (len > PUSH_BUNDLE_SIZE)
+    {
+        len = PUSH_BUNDLE_SIZE;
+    }
+
+    {
+        struct buffer buf = alloc_buf_gc(len, &gc);
+        buf_printf(&buf, auth_failed);
+        if (client_reason)
+        {
+            buf_printf(&buf, ",%s", client_reason);
+        }
+        send_control_channel_string_dowork(multi, BSTR(&buf), D_PUSH);
+    }
+
+    gc_free(&gc);
+}
+
 /*
  * Send restart message from server to client.
  */
src/openvpn/ssl_common.h +1 -0
index ac25ffa78..b82a014a0 100644
@@ -547,6 +547,7 @@ struct tls_multi
     time_t auth_token_tstamp; /**< timestamp of the generated token */
     bool auth_token_sent; /**< If server uses --auth-gen-token and
                            *   token has been sent to client */
+    bool connection_established ; /** Notifies future auth calls this is a reneg */
     /*
      * Our session objects.
      */
src/openvpn/ssl_verify.c +7 -0
index c7e595e46..fd83cde85 100644
@@ -1336,6 +1336,7 @@ verify_user_pass(struct user_pass *up, struct tls_multi *multi,
             && (multi->auth_token_tstamp + session->opt->auth_token_lifetime) < now)
         {
             msg(D_HANDSHAKE, "Auth-token for client expired\n");
+            send_push_reply_auth_failed(multi, "SESSION:Auth-token expired");
             wipe_auth_token(multi);
             ks->authenticated = false;
             goto done;
@@ -1458,6 +1459,12 @@ verify_user_pass(struct user_pass *up, struct tls_multi *multi,
     }
     else
     {
+        if (multi->connection_established)
+        {
+            /* Notify the client */
+            send_push_reply_auth_failed(multi, "SESSION:Auth failed");
+
+        }
         msg(D_TLS_ERRORS, "TLS Auth Error: Auth Username/Password verification failed for peer");
     }
 
#1 lore raw

Am 11.04.19 um 04:07 schrieb Eric Thorpe:

Quoted 9 lines

Hi All,

This patch relies on Arne's "Add send_control_channel_string_dowork variant" patch.

This patch modifies auth so that on a renegotiation the client is informed of a SESSION re-auth failure during a renegotiation if either their auth-token has expired, or they enter a wrong password in the case of auth-nocache for example.

Unfortenately this patch modifies the same code path that the auth-token patch set modifies and creates conflicts.

So this patch depends on the earlier patches of the patch set (which happen to be already applied) but breaks the later ones of the patch set

So while the patch looks good enough, I would like to postpone it after the auth-token set has been applied.

Arne

#2 lore raw
In reply to Eric Thorpe

Hi,

Quoted 2 lines · Eric Thorpe

This patch relies on Arne's "Add send_control_channel_string_dowork variant" patch.

So. After we left you out in the cold for over a year (apologies for that), we should revisit your patch set.

I checked with Arne, he says "no, that functionality is not in yet, but still interesting" - and all the *other* auth/plugin/async/SSO related code we had brewing is in.

So - if you're not too frustrated with us yet: can you re-send?

gert

Signature · 6 lines
"If was one thing all people took for granted, was conviction that if you 
 feed honest figures into a computer, honest figures come out. Never doubted 
 it myself till I met a computer with a sense of humor."
                             Robert A. Heinlein, The Moon is a Harsh Mistress

Gert Doering - Munich, Germany                             [email protected]...
#3 lore raw

Have re-based to master and resent.

Cheers,
Eric

---
Eric Thorpe
SparkLabs Developer
https://www.sparklabs.com
https://twitter.com/sparklabs
[email protected]...

Quoted 15 lines · Gert Doering

Hi,

Quoted 2 lines · Eric Thorpe

This patch relies on Arne's "Add send_control_channel_string_dowork variant" patch.

So. After we left you out in the cold for over a year (apologies for that), we should revisit your patch set.

I checked with Arne, he says "no, that functionality is not in yet, but still interesting" - and all the *other* auth/plugin/async/SSO related code we had brewing is in.

So - if you're not too frustrated with us yet: can you re-send?

gert