Skip to contentSkip to conversation

tls: Preserve sk_err across recvmsg() when data has been copied

PATCH Applied Chuck Lever 2 participants 2 messages May 13 – May 15

Subsystems (3)
Networking Network drivers Networking TLS
lore
Share & tools
lore

Chuck Lever

author
#0 lore raw

PATCH 1/1 tls: Preserve sk_err across recvmsg() when data has been copied

From: Chuck Lever <[email protected]>

The sk_err check in tls_rx_rec_wait() consumes the error via sock_error(), which clears sk_err atomically. When the caller (tls_sw_recvmsg, tls_sw_splice_read, or tls_sw_read_sock) already has bytes copied to userspace, it returns those bytes and discards the error from this call. sk_err is now zero on the socket, so the next read syscall observes only RCV_SHUTDOWN and reports a clean EOF instead of the actual error (typically -ECONNRESET).

The race is reachable when tls_read_flush_backlog()'s periodic sk_flush_backlog() triggers tcp_reset() in the middle of a multi-record read.

Pass a has_copied flag to tls_rx_rec_wait(). When has_copied is false, consume sk_err via sock_error() as before. When has_copied is true, report the error from READ_ONCE() but leave sk_err set:
the caller returns the byte count and discards the err from this call, and the next read syscall surfaces the preserved sk_err. This mirrors the tcp_recvmsg() preserve-and-surface pattern.

The decrypt-abort path is unaffected: tls_err_abort() raises sk_err to EBADMSG after tls_rx_rec_wait() returns, and nothing on the caller's return path consumes it, so the EBADMSG surfaces on the next read.

tls_sw_splice_read() passes has_copied=false: it processes one record per call, so no bytes have been copied within the function when tls_rx_rec_wait() runs. A reset that arrives between iterations of splice_direct_to_actor() (the sendfile() path) is still consumed by sock_error() in the later call, and the outer loop returns the prior iterations' byte count and drops the error. tcp_splice_read() exhibits the same pattern at the iteration boundary; addressing it belongs at the splice_direct_to_actor() layer and is out of scope here.

Fixes c46b01839f7a ("tls: rx: periodically flush socket backlog")
Suggested-by Jakub Kicinski <[email protected]>
Cc Jakub Kicinski <[email protected]>
Signed-off-by Chuck Lever <[email protected]>
---
 net/tls/tls_sw.c | 20 ++++++++++++++------
 1 file changed, 14 insertions(+), 6 deletions(-)
net/tls/tls_sw.c +14 -6
index 2590e855f6a5..c4cc4e357848 100644
@@ -1356,9 +1356,14 @@ void tls_sw_splice_eof(struct socket *sock)
 	mutex_unlock(&tls_ctx->tx_lock);
 }
 
+/* When has_copied is true the caller has already moved bytes to
+ * userspace. Report sk_err but leave it set so the next read
+ * surfaces it instead of a spurious EOF, otherwise sk_err is
+ * consumed via sock_error().
+ */
 static int
 tls_rx_rec_wait(struct sock *sk, struct sk_psock *psock, bool nonblock,
-		bool released)
+		bool released, bool has_copied)
 {
 	struct tls_context *tls_ctx = tls_get_ctx(sk);
 	struct tls_sw_context_rx *ctx = tls_sw_ctx_rx(tls_ctx);

@@ -1376,8 +1381,11 @@ tls_rx_rec_wait(struct sock *sk, struct sk_psock *psock, bool nonblock,
 		if (!sk_psock_queue_empty(psock))
 			return 0;
 
-		if (sk->sk_err)
+		if (sk->sk_err) {
+			if (has_copied)
+				return -READ_ONCE(sk->sk_err);
 			return sock_error(sk);
+		}
 
 		if (ret < 0)
 			return ret;

@@ -1413,7 +1421,7 @@ tls_rx_rec_wait(struct sock *sk, struct sk_psock *psock, bool nonblock,
 	}
 
 	if (unlikely(!tls_strp_msg_load(&ctx->strp, released)))
-		return tls_rx_rec_wait(sk, psock, nonblock, false);
+		return tls_rx_rec_wait(sk, psock, nonblock, false, has_copied);
 
 	return 1;
 }

@@ -2100,7 +2108,7 @@ int tls_sw_recvmsg(struct sock *sk,
 		int to_decrypt, chunk;
 
 		err = tls_rx_rec_wait(sk, psock, flags & MSG_DONTWAIT,
-				      released);
+				      released, !!(decrypted + copied));
 		if (err <= 0) {
 			if (psock) {
 				chunk = sk_msg_recvmsg(sk, psock, msg, len,

@@ -2287,7 +2295,7 @@ ssize_t tls_sw_splice_read(struct socket *sock,  loff_t *ppos,
 		struct tls_decrypt_arg darg;
 
 		err = tls_rx_rec_wait(sk, NULL, flags & SPLICE_F_NONBLOCK,
-				      true);
+				      true, false);
 		if (err <= 0)
 			goto splice_read_end;
 

@@ -2373,7 +2381,7 @@ int tls_sw_read_sock(struct sock *sk, read_descriptor_t *desc,
 		} else {
 			struct tls_decrypt_arg darg;
 
-			err = tls_rx_rec_wait(sk, NULL, true, released);
+			err = tls_rx_rec_wait(sk, NULL, true, released, !!copied);
 			if (err <= 0)
 				goto read_sock_end;
Signature · 1 line
2.54.0
CI Applied
#1 lore raw
This patch was applied to netdev/net.git (main) by Jakub Kicinski <[email protected]>:

Hello:

This patch was applied to netdev/net.git (main)
by Jakub Kicinski <[email protected]>:

Quoted 11 lines · On Wed, 13 May 2026 08:58:25 -0400 you wrote:

From: Chuck Lever <[email protected]>

The sk_err check in tls_rx_rec_wait() consumes the error via sock_error(), which clears sk_err atomically. When the caller (tls_sw_recvmsg, tls_sw_splice_read, or tls_sw_read_sock) already has bytes copied to userspace, it returns those bytes and discards the error from this call. sk_err is now zero on the socket, so the next read syscall observes only RCV_SHUTDOWN and reports a clean EOF instead of the actual error (typically -ECONNRESET).

[...]

Here is the summary with links:
- [net] tls: Preserve sk_err across recvmsg() when data has been copied

    https://git.kernel.org/netdev/net/c/f508262ae9f2

You are awesome, thank you!

Signature · 2 lines
Deet-doot-dot, I am a bot.
https://korg.docs.kernel.org/patchwork/pwbot.html