public inbox for netdev@vger.kernel.org
 help / color / mirror / Atom feed
From: Chuck Lever <cel@kernel.org>
To: john.fastabend@gmail.com, kuba@kernel.org, sd@queasysnail.net
Cc: <netdev@vger.kernel.org>, <kernel-tls-handshake@lists.linux.dev>,
	Chuck Lever <chuck.lever@oracle.com>,
	Alistair Francis <alistair.francis@wdc.com>,
	Hannes Reinecke <hare@suse.de>
Subject: [PATCH v1 3/6] tls: Suppress spurious saved_data_ready during read_sock
Date: Thu,  5 Mar 2026 16:13:59 -0500	[thread overview]
Message-ID: <20260305211402.39408-4-cel@kernel.org> (raw)
In-Reply-To: <20260305211402.39408-1-cel@kernel.org>

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

During tls_sw_read_sock(), each record release triggers
tls_strp_msg_done(), which calls back through the strparser into
saved_data_ready(). For a batch of N records, the first N-1
wakeups are pure overhead: the read_sock callback is already
running and will pick up subsequent records on the next iteration.

Remove the per-record wakeup from the record-release path by
introducing tls_rx_rec_release(), which calls
tls_strp_msg_release() instead of tls_strp_msg_done(). Factor
tls_rx_msg_ready() out of tls_strp_read_sock() so that parsing a
record no longer fires the callback directly, and introduce
tls_strp_check_rcv_quiet() for use in tls_rx_rec_wait(), which
parses queued data without notifying. A single
tls_strp_check_rcv() at the read_sock_end exit point fires at
most one notification, and only when data remains after the loop
exits.

Acked-by: Alistair Francis <alistair.francis@wdc.com>
Reviewed-by: Hannes Reinecke <hare@suse.de>
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
---
 net/tls/tls.h      |  1 +
 net/tls/tls_strp.c | 21 +++++++++++++++++++--
 net/tls/tls_sw.c   | 13 +++++++++++--
 3 files changed, 31 insertions(+), 4 deletions(-)

diff --git a/net/tls/tls.h b/net/tls/tls.h
index a97f1acef31d..d58d86e8e43e 100644
--- a/net/tls/tls.h
+++ b/net/tls/tls.h
@@ -193,6 +193,7 @@ int tls_strp_init(struct tls_strparser *strp, struct sock *sk);
 void tls_strp_data_ready(struct tls_strparser *strp);
 
 void tls_strp_check_rcv(struct tls_strparser *strp);
+void tls_strp_check_rcv_quiet(struct tls_strparser *strp);
 void tls_strp_msg_release(struct tls_strparser *strp);
 void tls_strp_msg_done(struct tls_strparser *strp);
 
diff --git a/net/tls/tls_strp.c b/net/tls/tls_strp.c
index a7648ebde162..7b9f5051becb 100644
--- a/net/tls/tls_strp.c
+++ b/net/tls/tls_strp.c
@@ -368,7 +368,6 @@ static int tls_strp_copyin(read_descriptor_t *desc, struct sk_buff *in_skb,
 		desc->count = 0;
 
 		WRITE_ONCE(strp->msg_ready, 1);
-		tls_rx_msg_ready(strp);
 	}
 
 	return ret;
@@ -539,11 +538,27 @@ static int tls_strp_read_sock(struct tls_strparser *strp)
 		return tls_strp_read_copy(strp, false);
 
 	WRITE_ONCE(strp->msg_ready, 1);
-	tls_rx_msg_ready(strp);
 
 	return 0;
 }
 
+/**
+ * tls_strp_check_rcv_quiet - parse without consumer notification
+ * @strp: TLS stream parser instance
+ *
+ * Parse queued data without firing the consumer notification. A subsequent
+ * tls_strp_check_rcv() is required before the socket lock is released;
+ * otherwise queued data stalls until the next tls_strp_data_ready() event.
+ */
+void tls_strp_check_rcv_quiet(struct tls_strparser *strp)
+{
+	if (unlikely(strp->stopped) || strp->msg_ready)
+		return;
+
+	if (tls_strp_read_sock(strp) == -ENOMEM)
+		queue_work(tls_strp_wq, &strp->work);
+}
+
 void tls_strp_check_rcv(struct tls_strparser *strp)
 {
 	if (unlikely(strp->stopped) || strp->msg_ready)
@@ -551,6 +566,8 @@ void tls_strp_check_rcv(struct tls_strparser *strp)
 
 	if (tls_strp_read_sock(strp) == -ENOMEM)
 		queue_work(tls_strp_wq, &strp->work);
+	else if (strp->msg_ready)
+		tls_rx_msg_ready(strp);
 }
 
 /* Lower sock lock held */
diff --git a/net/tls/tls_sw.c b/net/tls/tls_sw.c
index 108d417dcfb7..a5905f4c1ae2 100644
--- a/net/tls/tls_sw.c
+++ b/net/tls/tls_sw.c
@@ -1372,7 +1372,10 @@ tls_rx_rec_wait(struct sock *sk, struct sk_psock *psock, bool nonblock,
 			return ret;
 
 		if (!skb_queue_empty(&sk->sk_receive_queue)) {
-			tls_strp_check_rcv(&ctx->strp);
+			/* tls_strp_check_rcv() is called on the read_sock_end
+			 * path before the socket lock is released.
+			 */
+			tls_strp_check_rcv_quiet(&ctx->strp);
 			if (tls_strp_msg_ready(ctx))
 				break;
 		}
@@ -1853,6 +1856,11 @@ static int tls_record_content_type(struct msghdr *msg, struct tls_msg *tlm,
 	return 1;
 }
 
+static void tls_rx_rec_release(struct tls_sw_context_rx *ctx)
+{
+	tls_strp_msg_release(&ctx->strp);
+}
+
 static void tls_rx_rec_done(struct tls_sw_context_rx *ctx)
 {
 	tls_strp_msg_done(&ctx->strp);
@@ -2383,7 +2391,7 @@ int tls_sw_read_sock(struct sock *sk, read_descriptor_t *desc,
 			tlm = tls_msg(skb);
 			decrypted += rxm->full_len;
 
-			tls_rx_rec_done(ctx);
+			tls_rx_rec_release(ctx);
 		}
 
 		/* read_sock does not support reading control messages */
@@ -2413,6 +2421,7 @@ int tls_sw_read_sock(struct sock *sk, read_descriptor_t *desc,
 	}
 
 read_sock_end:
+	tls_strp_check_rcv(&ctx->strp);
 	tls_rx_reader_release(sk, ctx);
 	return copied ? : err;
 
-- 
2.53.0


  parent reply	other threads:[~2026-03-05 21:14 UTC|newest]

Thread overview: 10+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2026-03-05 21:13 [PATCH v1 0/6] TLS read_sock performance scalability Chuck Lever
2026-03-05 21:13 ` [PATCH v1 1/6] tls: Fix dangling skb pointer in tls_sw_read_sock() Chuck Lever
2026-03-05 22:19   ` Jakub Kicinski
2026-03-06 14:33     ` Chuck Lever
2026-03-05 21:13 ` [PATCH v1 2/6] tls: Factor tls_strp_msg_release() from tls_strp_msg_done() Chuck Lever
2026-03-05 21:13 ` Chuck Lever [this message]
2026-03-05 21:14 ` [PATCH v1 4/6] tls: Flush backlog before tls_rx_rec_wait in read_sock Chuck Lever
2026-03-05 21:14 ` [PATCH v1 5/6] tls: Restructure tls_sw_read_sock() into submit/deliver phases Chuck Lever
2026-03-05 21:14 ` [PATCH v1 6/6] tls: Enable batch async decryption in read_sock Chuck Lever
2026-03-10  3:14 ` [PATCH v1 0/6] TLS read_sock performance scalability Jakub Kicinski

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=20260305211402.39408-4-cel@kernel.org \
    --to=cel@kernel.org \
    --cc=alistair.francis@wdc.com \
    --cc=chuck.lever@oracle.com \
    --cc=hare@suse.de \
    --cc=john.fastabend@gmail.com \
    --cc=kernel-tls-handshake@lists.linux.dev \
    --cc=kuba@kernel.org \
    --cc=netdev@vger.kernel.org \
    --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