Script 'mail_helper' called by obssrc
Hello community,

here is the log from the commit of package frr for openSUSE:Factory checked in 
at 2024-04-14 11:55:04
++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
Comparing /work/SRC/openSUSE:Factory/frr (Old)
 and      /work/SRC/openSUSE:Factory/.frr.new.26366 (New)
++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++

Package is "frr"

Sun Apr 14 11:55:04 2024 rev:30 rq:1167542 version:8.4

Changes:
--------
--- /work/SRC/openSUSE:Factory/frr/frr.changes  2024-02-08 19:03:28.949955164 
+0100
+++ /work/SRC/openSUSE:Factory/.frr.new.26366/frr.changes       2024-04-14 
12:24:50.934062837 +0200
@@ -1,0 +2,8 @@
+Wed Apr 10 18:59:00 UTC 2024 - Clemens Famulla-Conrad <cfamullacon...@suse.com>
+
+- add
+  0019-bgpd-fix-error-handling-when-receiving-BGP-Prefix-SID-attribute.patch:
+  * Apply upstream fix on error handling when receiving BGP Prefix
+    SID attribute (bsc#1222518,CVE-2024-31948,gh#FRRouting/frr#15628)
+
+-------------------------------------------------------------------

New:
----
  0019-bgpd-fix-error-handling-when-receiving-BGP-Prefix-SID-attribute.patch

BETA DEBUG BEGIN:
  New:- add
  0019-bgpd-fix-error-handling-when-receiving-BGP-Prefix-SID-attribute.patch:
  * Apply upstream fix on error handling when receiving BGP Prefix
BETA DEBUG END:

++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++

Other differences:
------------------
++++++ frr.spec ++++++
--- /var/tmp/diff_new_pack.oIMuL5/_old  2024-04-14 12:24:51.626088038 +0200
+++ /var/tmp/diff_new_pack.oIMuL5/_new  2024-04-14 12:24:51.630088183 +0200
@@ -57,6 +57,7 @@
 Patch16:        0016-bgpd-Ignore-handling-NLRIs-if-we-received-MP_UNREACH.patch
 Patch17:        0017-bgpd-Fix-use-beyond-end-of-stream-of-labeled-unicast.patch
 Patch18:        0018-bgpd-Flowspec-overflow-issue.patch
+Patch19:        
0019-bgpd-fix-error-handling-when-receiving-BGP-Prefix-SID-attribute.patch
 BuildRequires:  autoconf
 BuildRequires:  automake
 BuildRequires:  bison >= 2.7

++++++ 
0019-bgpd-fix-error-handling-when-receiving-BGP-Prefix-SID-attribute.patch 
++++++
>From 51679e4504546584d98673b76ed8e12a8bc74fe0 Mon Sep 17 00:00:00 2001
From: Donatas Abraitis <dona...@opensourcerouting.org>
Date: Wed, 27 Mar 2024 18:42:56 +0200
Subject: [PATCH 1/2] bgpd: Fix error handling when receiving BGP Prefix SID
 attribute
References: bsc#1222518 CVE-2024-31948 gh#FRRouting/frr#15628


Without this patch, we always set the BGP Prefix SID attribute flag without
checking if it's malformed or not. RFC8669 says that this attribute MUST be 
discarded.

Also, this fixes the bgpd crash when a malformed Prefix SID attribute is 
received,
with malformed transitive flags and/or TLVs.

Reported-by: Iggy Frankovic <iggyf...@amazon.com>
Signed-off-by: Donatas Abraitis <dona...@opensourcerouting.org>
(cherry picked from commit ba6a8f1a31e1a88df2de69ea46068e8bd9b97138)
---
 bgpd/bgp_attr.c | 5 +++--
 1 file changed, 3 insertions(+), 2 deletions(-)

diff --git a/bgpd/bgp_attr.c b/bgpd/bgp_attr.c
index 7144c4bfa73d..2e2845b8fa7e 100644
--- a/bgpd/bgp_attr.c
+++ b/bgpd/bgp_attr.c
@@ -1400,6 +1400,7 @@ bgp_attr_malformed(struct bgp_attr_parser_args *args, 
uint8_t subcode,
        case BGP_ATTR_AS4_AGGREGATOR:
        case BGP_ATTR_AGGREGATOR:
        case BGP_ATTR_ATOMIC_AGGREGATE:
+       case BGP_ATTR_PREFIX_SID:
                return BGP_ATTR_PARSE_PROCEED;
 
        /* Core attributes, particularly ones which may influence route
@@ -3146,8 +3147,6 @@ enum bgp_attr_parse_ret bgp_attr_prefix_sid(struct 
bgp_attr_parser_args *args)
        struct attr *const attr = args->attr;
        enum bgp_attr_parse_ret ret;
 
-       attr->flag |= ATTR_FLAG_BIT(BGP_ATTR_PREFIX_SID);
-
        uint8_t type;
        uint16_t length;
        size_t headersz = sizeof(type) + sizeof(length);
@@ -3197,6 +3196,8 @@ enum bgp_attr_parse_ret bgp_attr_prefix_sid(struct 
bgp_attr_parser_args *args)
                }
        }
 
+       SET_FLAG(attr->flag, ATTR_FLAG_BIT(BGP_ATTR_PREFIX_SID));
+
        return BGP_ATTR_PARSE_PROCEED;
 }
 

>From 9240abccb564043c85180916b77cad5b194a49c9 Mon Sep 17 00:00:00 2001
From: Donatas Abraitis <dona...@opensourcerouting.org>
Date: Wed, 27 Mar 2024 19:08:38 +0200
Subject: [PATCH 2/2] bgpd: Prevent from one more CVE triggering this place
References: bsc#1222518 CVE-2024-31948 gh#FRRouting/frr#15628
Upstream: submitted

If we receive an attribute that is handled by bgp_attr_malformed(), use
treat-as-withdraw behavior for unknown (or missing to add - if new) attributes.

Signed-off-by: Donatas Abraitis <dona...@opensourcerouting.org>
(cherry picked from commit babb23b74855e23c987a63f8256d24e28c044d07)
---
 bgpd/bgp_attr.c | 33 ++++++++++++++++++++++-----------
 1 file changed, 22 insertions(+), 11 deletions(-)

diff --git a/bgpd/bgp_attr.c b/bgpd/bgp_attr.c
index 2e2845b8fa7e..7570598a3d7f 100644
--- a/bgpd/bgp_attr.c
+++ b/bgpd/bgp_attr.c
@@ -1391,6 +1391,15 @@ bgp_attr_malformed(struct bgp_attr_parser_args *args, 
uint8_t subcode,
                        (args->startp - STREAM_DATA(BGP_INPUT(peer)))
                                + args->total);
 
+       /* Partial optional attributes that are malformed should not cause
+        * the whole session to be reset. Instead treat it as a withdrawal
+        * of the routes, if possible.
+        */
+       if (CHECK_FLAG(flags, BGP_ATTR_FLAG_TRANS) &&
+           CHECK_FLAG(flags, BGP_ATTR_FLAG_OPTIONAL) &&
+           CHECK_FLAG(flags, BGP_ATTR_FLAG_PARTIAL))
+               return BGP_ATTR_PARSE_WITHDRAW;
+
        switch (args->type) {
        /* where an attribute is relatively inconsequential, e.g. it does not
         * affect route selection, and can be safely ignored, then any such
@@ -1425,19 +1434,21 @@ bgp_attr_malformed(struct bgp_attr_parser_args *args, 
uint8_t subcode,
                bgp_notify_send_with_data(peer, BGP_NOTIFY_UPDATE_ERR, subcode,
                                          notify_datap, length);
                return BGP_ATTR_PARSE_ERROR;
+       default:
+               /* Unknown attributes, that are handled by this function
+                * should be treated as withdraw, to prevent one more CVE
+                * from being introduced.
+                * RFC 7606 says:
+                * The "treat-as-withdraw" approach is generally preferred
+                * and the "session reset" approach is discouraged.
+                */
+               flog_err(EC_BGP_ATTR_FLAG,
+                        "%s(%u) attribute received, while it is not known how 
to handle it, treating as withdraw",
+                        lookup_msg(attr_str, args->type, NULL), args->type);
+               break;
        }
 
-       /* Partial optional attributes that are malformed should not cause
-        * the whole session to be reset. Instead treat it as a withdrawal
-        * of the routes, if possible.
-        */
-       if (CHECK_FLAG(flags, BGP_ATTR_FLAG_TRANS)
-           && CHECK_FLAG(flags, BGP_ATTR_FLAG_OPTIONAL)
-           && CHECK_FLAG(flags, BGP_ATTR_FLAG_PARTIAL))
-               return BGP_ATTR_PARSE_WITHDRAW;
-
-       /* default to reset */
-       return BGP_ATTR_PARSE_ERROR_NOTIFYPLS;
+       return BGP_ATTR_PARSE_WITHDRAW;
 }
 
 /* Find out what is wrong with the path attribute flag bits and log the error.

Reply via email to