Skip to contentSkip to conversation

sctp: hold shkey across socket migration

PATCH In review David Lee 4 participants 6 messages Jul 31 – Sep 23

Subsystems (3)
lore
Share & tools
lore

David Lee

author
#0 lore raw

PATCH 1/1 sctp: hold shkey across socket migration

sctp_sock_migrate() transfers queued DATA skbs from the old socket to the new one. skb_orphan() invokes sctp_wfree() during that transfer and drops the skb-owned shared-key reference.

If userspace has removed that key from the association, this can be the final reference. The following sctp_set_owner_w() then dereferences the freed chunk->shkey while trying to take the new owner reference.

Take a temporary shared-key reference before orphaning the skb and release it after the new owner has taken its reference. This preserves the selected authentication key throughout the ownership transfer.

Fixes 1b1e0bc99474 ("sctp: add refcnt support for sh_key")

Bug found and triaged by OpenAI Security Research and validated by Trail of Bits.

Assisted-by Codex:gpt-5.6-sol gpt-5.5-cyber

---
Trail of Bits has a reproducer for this bug that triggers a KASAN use-after-free and can share if needed.

 net/sctp/socket.c | 12 +++++++++++-
 1 file changed, 11 insertions(+), 1 deletion(-)
net/sctp/socket.c +11 -1
index c7b9e325e..4a08023d5 100644
@@ -147,9 +147,19 @@ static inline void sctp_set_owner_w(struct sctp_chunk *chunk)
 
 static void sctp_clear_owner_w(struct sctp_chunk *chunk)
 {
+	/* Keep the shkey alive until the new owner takes its reference. */
+	if (chunk->shkey)
+		sctp_auth_shkey_hold(chunk->shkey);
 	skb_orphan(chunk->skb);
 }
 
+static void sctp_set_owner_w_migrate(struct sctp_chunk *chunk)
+{
+	sctp_set_owner_w(chunk);
+	if (chunk->shkey)
+		sctp_auth_shkey_release(chunk->shkey);
+}
+
 #define traverse_and_process()	\
 do {				\
 	msg = chunk->msg;	\
@@ -9632,7 +9642,7 @@ static int sctp_sock_migrate(struct sock *oldsk, struct sock *newsk,
 	lock_sock_nested(newsk, SINGLE_DEPTH_NESTING);
 	sctp_for_each_tx_datachunk(assoc, true, sctp_clear_owner_w);
 	sctp_assoc_migrate(assoc, newsk);
-	sctp_for_each_tx_datachunk(assoc, false, sctp_set_owner_w);
+	sctp_for_each_tx_datachunk(assoc, false, sctp_set_owner_w_migrate);
 
 	/* If the association on the newsk is already closed before accept()
 	 * is called, set RCV_SHUTDOWN flag.
Signature · 1 line
2.53.0
maintainer Acked-by on v1 1/1
#1 lore raw
Quoted 13 lines · David Lee

sctp_sock_migrate() transfers queued DATA skbs from the old socket to the new one. skb_orphan() invokes sctp_wfree() during that transfer and drops the skb-owned shared-key reference.

If userspace has removed that key from the association, this can be the final reference. The following sctp_set_owner_w() then dereferences the freed chunk->shkey while trying to take the new owner reference.

Take a temporary shared-key reference before orphaning the skb and release it after the new owner has taken its reference. This preserves the selected authentication key throughout the ownership transfer.

Quoted 6 lines
Fixes 1b1e0bc99474 ("sctp: add refcnt support for sh_key")

Bug found and triaged by OpenAI Security Research and validated by Trail of Bits.

Assisted-by Codex:gpt-5.6-sol gpt-5.5-cyber

The "Fixes: " is not together with others at the end, and the "Signed-off-by:" for yourself is also missed:

Patch looks good to me:

Not sure if you need to repost for fixing the tags, let's see.

Thanks.

maintainer
#3 lore raw
Quoted 5 lines · On Wed, Sep 9, 2026 at 10:08 PM <[email protected]> wrote:

Hi Xin Long,

Is there any progress regarding this patch?

I think you need to post v2 with the tag issues addressed.

Thanks.

#4 lore raw
In reply to David Lee

Hi David and Xin,

My control-chunk migration fix [1] depends on this patch. Xin asked me to wait for it to land so the dependent patch would apply.

Xin, would it be acceptable to send a v3 series carrying this patch as 1/2, without code changes, and my fix as 2/2? I would preserve David's authorship, fix the tag placement noted in your review, and add my sign-off.

David, could you provide the missing Signed-off-by requested in the earlier review?

Thanks,
Jun

[1] https://lore.kernel.org/netdev/[email protected]/

maintainer
#5 lore raw
Quoted 10 lines · Jun Yang

Hi David and Xin,

My control-chunk migration fix [1] depends on this patch. Xin asked me to wait for it to land so the dependent patch would apply.

Xin, would it be acceptable to send a v3 series carrying this patch as 1/2, without code changes, and my fix as 2/2? I would preserve David's authorship, fix the tag placement noted in your review, and add my sign-off.

Sure, sounds good to me.

Thanks.

Quoted 8 lines

David, could you provide the missing Signed-off-by requested in the earlier review?

Thanks,
Jun

[1] https://lore.kernel.org/netdev/[email protected]/