Skip to content

Commit

Permalink
[202211][FRR][CVE] Add FRR patches to fix CVEs: CVE-2022-43681 CVE-20…
Browse files Browse the repository at this point in the history
…22-40318… (sonic-net#15263)

… CVE-2022-40302

Add patches from PRs
FRRouting/frr#12043
FRRouting/frr#12247

#### Why I did it
To fix CVEs GHSA-x7mf-v6gh-vm4g GHSA-9rqq-99cf-35g5 GHSA-j7hm-p94x-q9pw

##### Work item tracking
- Microsoft ADO **(number only)**: 23268946

#### How I did it
Added patches from the FRR fix PRs
  • Loading branch information
DavidZagury authored Jun 19, 2023
1 parent f593323 commit 72c5562
Show file tree
Hide file tree
Showing 3 changed files with 179 additions and 0 deletions.
Original file line number Diff line number Diff line change
@@ -0,0 +1,130 @@
From 3e46b43e3788f0f87bae56a86b54d412b4710286 Mon Sep 17 00:00:00 2001
From: Donald Sharp <sharpd@nvidia.com>
Date: Fri, 30 Sep 2022 08:51:45 -0400
Subject: [PATCH 1/2] bgpd: Ensure FRR has enough data to read 2 bytes in
peek_for_as4_capability

In peek_for_as4_capability the code is checking that the
stream has at least 2 bytes to read ( the opt_type and the
opt_length ). However if BGP_OPEN_EXT_OPT_PARAMS_CAPABLE(peer)
is configured then FRR is reading 3 bytes. Which is not good
since the packet could be badly formated. Ensure that
FRR has the appropriate data length to read the data.

Signed-off-by: Donald Sharp <sharpd@nvidia.com>
---
bgpd/bgp_open.c | 27 +++++++++++++++++++++------
1 file changed, 21 insertions(+), 6 deletions(-)

diff --git a/bgpd/bgp_open.c b/bgpd/bgp_open.c
index 7248f034a5a..a760a7ca013 100644
--- a/bgpd/bgp_open.c
+++ b/bgpd/bgp_open.c
@@ -1185,15 +1185,30 @@ as_t peek_for_as4_capability(struct peer *peer, uint16_t length)
uint8_t opt_type;
uint16_t opt_length;

- /* Check the length. */
- if (stream_get_getp(s) + 2 > end)
+ /* Ensure we can read the option type */
+ if (stream_get_getp(s) + 1 > end)
goto end;

- /* Fetch option type and length. */
+ /* Fetch the option type */
opt_type = stream_getc(s);
- opt_length = BGP_OPEN_EXT_OPT_PARAMS_CAPABLE(peer)
- ? stream_getw(s)
- : stream_getc(s);
+
+ /*
+ * Check the length and fetch the opt_length
+ * If the peer is BGP_OPEN_EXT_OPT_PARAMS_CAPABLE(peer)
+ * then we do a getw which is 2 bytes. So we need to
+ * ensure that we can read that as well
+ */
+ if (BGP_OPEN_EXT_OPT_PARAMS_CAPABLE(peer)) {
+ if (stream_get_getp(s) + 2 > end)
+ goto end;
+
+ opt_length = stream_getw(s);
+ } else {
+ if (stream_get_getp(s) + 1 > end)
+ goto end;
+
+ opt_length = stream_getc(s);
+ }

/* Option length check. */
if (stream_get_getp(s) + opt_length > end)

From 1117baca3c592877a4d8a13ed6a1d9bd83977487 Mon Sep 17 00:00:00 2001
From: Donald Sharp <sharpd@nvidia.com>
Date: Fri, 30 Sep 2022 08:57:43 -0400
Subject: [PATCH 2/2] bgpd: Ensure FRR has enough data to read 2 bytes in
bgp_open_option_parse

In bgp_open_option_parse the code is checking that the
stream has at least 2 bytes to read ( the opt_type and
the opt_length). However if BGP_OPEN_EXT_OPT_PARAMS_CAPABLE(peer)
is configured then FRR is reading 3 bytes. Which is not good
since the packet could be badly formateed. Ensure that
FRR has the appropriate data length to read the data.

Signed-off-by: Donald Sharp <sharpd@nvidia.com>
---
bgpd/bgp_open.c | 35 ++++++++++++++++++++++++++++-------
1 file changed, 28 insertions(+), 7 deletions(-)

diff --git a/bgpd/bgp_open.c b/bgpd/bgp_open.c
index a760a7ca013..d1667fac261 100644
--- a/bgpd/bgp_open.c
+++ b/bgpd/bgp_open.c
@@ -1278,19 +1278,40 @@ int bgp_open_option_parse(struct peer *peer, uint16_t length,
uint8_t opt_type;
uint16_t opt_length;

- /* Must have at least an OPEN option header */
- if (STREAM_READABLE(s) < 2) {
+ /*
+ * Check that we can read the opt_type and fetch it
+ */
+ if (STREAM_READABLE(s) < 1) {
zlog_info("%s Option length error", peer->host);
bgp_notify_send(peer, BGP_NOTIFY_OPEN_ERR,
BGP_NOTIFY_OPEN_MALFORMED_ATTR);
return -1;
}
-
- /* Fetch option type and length. */
opt_type = stream_getc(s);
- opt_length = BGP_OPEN_EXT_OPT_PARAMS_CAPABLE(peer)
- ? stream_getw(s)
- : stream_getc(s);
+
+ /*
+ * Check the length of the stream to ensure that
+ * FRR can properly read the opt_length. Then read it
+ */
+ if (BGP_OPEN_EXT_OPT_PARAMS_CAPABLE(peer)) {
+ if (STREAM_READABLE(s) < 2) {
+ zlog_info("%s Option length error", peer->host);
+ bgp_notify_send(peer, BGP_NOTIFY_OPEN_ERR,
+ BGP_NOTIFY_OPEN_MALFORMED_ATTR);
+ return -1;
+ }
+
+ opt_length = stream_getw(s);
+ } else {
+ if (STREAM_READABLE(s) < 1) {
+ zlog_info("%s Option length error", peer->host);
+ bgp_notify_send(peer, BGP_NOTIFY_OPEN_ERR,
+ BGP_NOTIFY_OPEN_MALFORMED_ATTR);
+ return -1;
+ }
+
+ opt_length = stream_getc(s);
+ }

/* Option length check. */
if (STREAM_READABLE(s) < opt_length) {
Original file line number Diff line number Diff line change
@@ -0,0 +1,47 @@
From 766eec1b7accffe2c04a5c9ebb14e9f487bb9f78 Mon Sep 17 00:00:00 2001
From: Donald Sharp <sharpd@nvidia.com>
Date: Wed, 2 Nov 2022 13:24:48 -0400
Subject: [PATCH] bgpd: Ensure that bgp open message stream has enough data to
read

If a operator receives an invalid packet that is of insufficient size
then it is possible for BGP to assert during reading of the packet
instead of gracefully resetting the connection with the peer.

Signed-off-by: Donald Sharp <sharpd@nvidia.com>
---
bgpd/bgp_packet.c | 19 +++++++++++++++++++
1 file changed, 19 insertions(+)

diff --git a/bgpd/bgp_packet.c b/bgpd/bgp_packet.c
index 769f9613da8..72d6a923175 100644
--- a/bgpd/bgp_packet.c
+++ b/bgpd/bgp_packet.c
@@ -1386,8 +1386,27 @@ static int bgp_open_receive(struct peer *peer, bgp_size_t size)
|| CHECK_FLAG(peer->flags, PEER_FLAG_EXTENDED_OPT_PARAMS)) {
uint8_t opttype;

+ if (STREAM_READABLE(peer->curr) < 1) {
+ flog_err(
+ EC_BGP_PKT_OPEN,
+ "%s: stream does not have enough bytes for extended optional parameters",
+ peer->host);
+ bgp_notify_send(peer, BGP_NOTIFY_OPEN_ERR,
+ BGP_NOTIFY_OPEN_MALFORMED_ATTR);
+ return BGP_Stop;
+ }
+
opttype = stream_getc(peer->curr);
if (opttype == BGP_OPEN_NON_EXT_OPT_TYPE_EXTENDED_LENGTH) {
+ if (STREAM_READABLE(peer->curr) < 2) {
+ flog_err(
+ EC_BGP_PKT_OPEN,
+ "%s: stream does not have enough bytes to read the extended optional parameters optlen",
+ peer->host);
+ bgp_notify_send(peer, BGP_NOTIFY_OPEN_ERR,
+ BGP_NOTIFY_OPEN_MALFORMED_ATTR);
+ return BGP_Stop;
+ }
optlen = stream_getw(peer->curr);
SET_FLAG(peer->sflags,
PEER_STATUS_EXT_OPT_PARAMS_LENGTH);
2 changes: 2 additions & 0 deletions src/sonic-frr/patch/series
Original file line number Diff line number Diff line change
Expand Up @@ -12,3 +12,5 @@ cross-compile-changes.patch
0010-zebra-Note-when-the-netlink-DUMP-command-is-interrup.patch
0011-bgpd-enhanced-capability-is-always-turned-on-for-int.patch
0012-Ensure-ospf_apiclient_lsa_originate-cannot-accidently-write-into-stack.patch
0027-bgpd-Ensure-FRR-has-enough-data-to-read-in-peek_for_as4_capability-and-bgp_open_option_parse.patch
0028-bgpd-Ensure-that-bgp-open-message-stream-has-enough-data-to-read.patch

0 comments on commit 72c5562

Please sign in to comment.