* [PATCH] net: af_key: fix refcount leak in pfkey_spdadd()
@ 2026-06-11 16:37 WenTao Liang
2026-06-11 17:49 ` Sabrina Dubroca
0 siblings, 1 reply; 2+ messages in thread
From: WenTao Liang @ 2026-06-11 16:37 UTC (permalink / raw)
To: steffen.klassert, herbert, davem, edumazet, kuba, pabeni
Cc: horms, netdev, linux-kernel, WenTao Liang, stable
In pfkey_spdadd(), an xfrm policy is allocated via xfrm_policy_alloc()
with a refcount of 1. On the success path the policy is eventually freed
by xfrm_pol_put(), which decrements the refcount and calls
xfrm_policy_destroy() only when it reaches zero. However, all error
paths directly call xfrm_policy_destroy() without releasing the initial
reference, leaking the policy object.
Fix the leak by replacing the direct xfrm_policy_destroy() call with
xfrm_pol_put() in the error unwind. Keep the xp->walk.dead assignment
as it is required by xfrm_policy_destroy()'s BUG_ON check.
Cc: stable@vger.kernel.org
Fixes: 64c31b3f7648 ("[XFRM] xfrm_policy_destroy: Rename and relative fixes.")
Signed-off-by: WenTao Liang <vulab@iscas.ac.cn>
---
net/key/af_key.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/net/key/af_key.c b/net/key/af_key.c
index 9cffeef18cd9..013592086f1e 100644
--- a/net/key/af_key.c
+++ b/net/key/af_key.c
@@ -2370,7 +2370,7 @@ static int pfkey_spdadd(struct sock *sk, struct sk_buff *skb, const struct sadb_
out:
xp->walk.dead = 1;
- xfrm_policy_destroy(xp);
+ xfrm_pol_put(xp);
return err;
}
--
2.50.1 (Apple Git-155)
^ permalink raw reply related [flat|nested] 2+ messages in thread
* Re: [PATCH] net: af_key: fix refcount leak in pfkey_spdadd()
2026-06-11 16:37 [PATCH] net: af_key: fix refcount leak in pfkey_spdadd() WenTao Liang
@ 2026-06-11 17:49 ` Sabrina Dubroca
0 siblings, 0 replies; 2+ messages in thread
From: Sabrina Dubroca @ 2026-06-11 17:49 UTC (permalink / raw)
To: WenTao Liang
Cc: steffen.klassert, herbert, davem, edumazet, kuba, pabeni, horms,
netdev, linux-kernel, stable
2026-06-12, 00:37:43 +0800, WenTao Liang wrote:
> In pfkey_spdadd(), an xfrm policy is allocated via xfrm_policy_alloc()
> with a refcount of 1. On the success path the policy is eventually freed
> by xfrm_pol_put(), which decrements the refcount and calls
> xfrm_policy_destroy() only when it reaches zero. However, all error
> paths directly call xfrm_policy_destroy() without releasing the initial
> reference, leaking the policy object.
Uhm... have you looked at what xfrm_policy_destroy and xfrm_pol_put
do?
--
Sabrina
^ permalink raw reply [flat|nested] 2+ messages in thread
end of thread, other threads:[~2026-06-11 17:49 UTC | newest]
Thread overview: 2+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2026-06-11 16:37 [PATCH] net: af_key: fix refcount leak in pfkey_spdadd() WenTao Liang
2026-06-11 17:49 ` Sabrina Dubroca
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox