public inbox for netdev@vger.kernel.org
 help / color / mirror / Atom feed
From: Chuck Lever <cel@kernel.org>
To: John Fastabend <john.fastabend@gmail.com>,
	 Jakub Kicinski <kuba@kernel.org>,
	Sabrina Dubroca <sd@queasysnail.net>
Cc: Eric Dumazet <edumazet@google.com>,
	Simon Horman <horms@kernel.org>,  Paolo Abeni <pabeni@redhat.com>,
	netdev@vger.kernel.org,  kernel-tls-handshake@lists.linux.dev,
	Chuck Lever <chuck.lever@oracle.com>,
	 Hannes Reinecke <hare@suse.de>,
	Alistair Francis <alistair.francis@wdc.com>
Subject: [PATCH net-next v9 2/5] tls: Fix dangling skb pointer in tls_sw_read_sock()
Date: Wed, 29 Apr 2026 17:48:09 -0400	[thread overview]
Message-ID: <20260429-tls-read-sock-v9-2-39e71aa7810f@oracle.com> (raw)
In-Reply-To: <20260429-tls-read-sock-v9-0-39e71aa7810f@oracle.com>

From: Chuck Lever <chuck.lever@oracle.com>

Two related defects in the receive loop of tls_sw_read_sock()
share a single fix.

Per ISO/IEC 9899:2011 section 6.2.4p2, a pointer value becomes
indeterminate when the object it points to reaches the end of
its lifetime; Annex J.2 classifies the use of such a value as
undefined behavior. consume_skb(skb) in the fully-consumed path
frees the skb, but the "do { } while (skb)" loop condition then
evaluates that freed pointer. Although the value is never
dereferenced -- the loop either continues and overwrites skb,
or exits -- any future change that adds a dereference between
consume_skb() and the loop condition would produce a silent
use-after-free.

Separately, when read_actor() consumes only part of a record
(used < rxm->full_len) but desc->count is still non-zero, the
existing code updates rxm in place and falls through to the
next loop iteration. The next iteration then unconditionally
overwrites skb without freeing or requeuing the partially
consumed buffer, leaking the skb and silently dropping stream
data.

A read_actor returning fewer bytes than offered is in every
case a backpressure signal; the only correct response is to
requeue and exit. Replace the do/while with an explicit
for(;;), requeue unconditionally on partial consume, and break
on exhausted desc->count after a full consume.

Fixes: 662fbcec32f4 ("net/tls: implement ->read_sock()")
Reviewed-by: Hannes Reinecke <hare@suse.de>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Reviewed-by: Sabrina Dubroca <sd@queasysnail.net>
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
---
 net/tls/tls_sw.c | 15 +++++++--------
 1 file changed, 7 insertions(+), 8 deletions(-)

diff --git a/net/tls/tls_sw.c b/net/tls/tls_sw.c
index 244ac8ed4b01..c58d3b0b0a8a 100644
--- a/net/tls/tls_sw.c
+++ b/net/tls/tls_sw.c
@@ -2366,7 +2366,7 @@ int tls_sw_read_sock(struct sock *sk, read_descriptor_t *desc,
 		goto read_sock_end;
 
 	decrypted = 0;
-	do {
+	for (;;) {
 		if (!skb_queue_empty(&ctx->rx_list)) {
 			skb = __skb_dequeue(&ctx->rx_list);
 			rxm = strp_msg(skb);
@@ -2411,14 +2411,13 @@ int tls_sw_read_sock(struct sock *sk, read_descriptor_t *desc,
 		if (used < rxm->full_len) {
 			rxm->offset += used;
 			rxm->full_len -= used;
-			if (!desc->count)
-				goto read_sock_requeue;
-		} else {
-			consume_skb(skb);
-			if (!desc->count)
-				skb = NULL;
+			goto read_sock_requeue;
 		}
-	} while (skb);
+		consume_skb(skb);
+		skb = NULL;
+		if (!desc->count)
+			break;
+	}
 
 read_sock_end:
 	tls_rx_reader_release(sk, ctx);

-- 
2.53.0


  parent reply	other threads:[~2026-04-29 21:48 UTC|newest]

Thread overview: 16+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2026-04-29 21:48 [PATCH net-next v9 0/5] TLS read_sock performance scalability Chuck Lever
2026-04-29 21:48 ` [PATCH net-next v9 1/5] tls: Abort the connection on decrypt failure Chuck Lever
2026-05-03  1:20   ` Jakub Kicinski
2026-04-29 21:48 ` Chuck Lever [this message]
2026-05-03  1:05   ` [PATCH net-next v9 2/5] tls: Fix dangling skb pointer in tls_sw_read_sock() Jakub Kicinski
2026-04-29 21:48 ` [PATCH net-next v9 3/5] tls: Factor tls_strp_msg_release() from tls_strp_msg_done() Chuck Lever
2026-05-03  1:09   ` Jakub Kicinski
2026-04-29 21:48 ` [PATCH net-next v9 4/5] tls: Suppress spurious saved_data_ready on all receive paths Chuck Lever
2026-05-03  1:19   ` Jakub Kicinski
2026-04-29 21:48 ` [PATCH net-next v9 5/5] tls: Flush backlog before waiting for a new record Chuck Lever
2026-04-29 23:13 ` [PATCH net-next v9 0/5] TLS read_sock performance scalability Jakub Kicinski
2026-04-29 23:15   ` Chuck Lever
2026-05-03  1:04 ` Jakub Kicinski
2026-05-03 19:34   ` Chuck Lever
2026-05-04 13:33     ` Sabrina Dubroca
2026-05-04 15:59       ` Chuck Lever

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20260429-tls-read-sock-v9-2-39e71aa7810f@oracle.com \
    --to=cel@kernel.org \
    --cc=alistair.francis@wdc.com \
    --cc=chuck.lever@oracle.com \
    --cc=edumazet@google.com \
    --cc=hare@suse.de \
    --cc=horms@kernel.org \
    --cc=john.fastabend@gmail.com \
    --cc=kernel-tls-handshake@lists.linux.dev \
    --cc=kuba@kernel.org \
    --cc=netdev@vger.kernel.org \
    --cc=pabeni@redhat.com \
    --cc=sd@queasysnail.net \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox