netdev - [PATCH iproute2-next 10/20] bridge: vni: Guard close_vni

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <20231211140732.11475-11-bpoirier@nvidia.com>
Date: Mon, 11 Dec 2023 09:07:22 -0500
From: Benjamin Poirier <bpoirier@...dia.com>
To: netdev@...r.kernel.org
Cc: Petr Machata <petrm@...dia.com>,
	Roopa Prabhu <roopa@...dia.com>
Subject: [PATCH iproute2-next 10/20] bridge: vni: Guard close_vni_port() call

Currently, the call to open_vni_port() within print_vnifilter_rtm() is
written in a way that is safe if there is a RTM_{NEW,DEL,GET}TUNNEL message
without any VXLAN_VNIFILTER_ENTRY attribute. However the close_vni_port()
call is written in a way that assumes there is always at least one
VXLAN_VNIFILTER_ENTRY attribute within every RTM_*TUNNEL message. At this
time, this assumption is correct. However, the code should be consistent in
its assumptions. Choose the safe approach and fix the asymmetry between the
open_vni_port() and close_vni_port() calls by guarding the latter call with
a check.

Reviewed-by: Petr Machata <petrm@...dia.com>
Tested-by: Petr Machata <petrm@...dia.com>
Signed-off-by: Benjamin Poirier <bpoirier@...dia.com>
---
 bridge/vni.c | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)

diff --git a/bridge/vni.c b/bridge/vni.c
index 8a6ac245..ca5d2e43 100644
--- a/bridge/vni.c
+++ b/bridge/vni.c
@@ -341,7 +341,9 @@ int print_vnifilter_rtm(struct nlmsghdr *n, void *arg)

 		print_vni(t, tmsg->ifindex);
 	}
-	close_vni_port();
+
+	if (!first)
+		close_vni_port();

 	print_string(PRINT_FP, NULL, "%s", _SL_);

-- 
2.43.0