Skip to contentSkip to conversation

net/tls: fixes for NVMe-over-TLS

PATCH Superseded Hannes Reinecke 3 participants 16 messages Jul 21, 2023 – Jul 25, 2023

Subsystems (5)

v8 of 9 Superseded Latest is v9 →

lore
Share & tools
lore
#2 lore raw

PATCH v8 2/6 net/tls: handle MSG_EOR for tls_device TX flow

In reply to Hannes Reinecke

tls_push_data() MSG_MORE, but bails out on MSG_EOR. Seeing that MSG_EOR is basically the opposite of MSG_MORE this patch adds handling MSG_EOR by treating it as the absence of MSG_MORE.
Consequently we should return an error when both are set.

---
 net/tls/tls_device.c | 6 +++++-
 1 file changed, 5 insertions(+), 1 deletion(-)
net/tls/tls_device.c +5 -1
index 2021fe557e50..5df18f696d7f 100644
@@ -441,9 +441,13 @@ static int tls_push_data(struct sock *sk,
 	long timeo;
 
 	if (flags &
-	    ~(MSG_MORE | MSG_DONTWAIT | MSG_NOSIGNAL | MSG_SPLICE_PAGES))
+	    ~(MSG_MORE | MSG_DONTWAIT | MSG_NOSIGNAL |
+	      MSG_SPLICE_PAGES | MSG_EOR))
 		return -EOPNOTSUPP;
 
+	if ((flags & (MSG_MORE | MSG_EOR)) == (MSG_MORE | MSG_EOR))
+		return -EINVAL;
+
 	if (unlikely(sk->sk_err))
 		return -sk->sk_err;
 
Signature · 1 line
2.35.3
#3 lore raw

PATCH v8 3/6 selftests/net/tls: add test for MSG_EOR

In reply to Hannes Reinecke

As the recent patch is modifying the behaviour for TLS re MSG_EOR handling we should be having a test for it.

---
 tools/testing/selftests/net/tls.c | 11 +++++++++++
 1 file changed, 11 insertions(+)
tools/testing/selftests/net/tls.c +11 -0
index a3c57004344c..4b63708c6a81 100644
@@ -486,6 +486,17 @@ TEST_F(tls, msg_more_unsent)
 	EXPECT_EQ(recv(self->cfd, buf, send_len, MSG_DONTWAIT), -1);
 }
 
+TEST_F(tls, msg_eor)
+{
+	char const *test_str = "test_read";
+	int send_len = 10;
+	char buf[10];
+
+	EXPECT_EQ(send(self->fd, test_str, send_len, MSG_EOR), send_len);
+	EXPECT_EQ(recv(self->cfd, buf, send_len, MSG_WAITALL), send_len);
+	EXPECT_EQ(memcmp(buf, test_str, send_len), 0);
+}
+
 TEST_F(tls, sendmsg_single)
 {
 	struct msghdr msg;
Signature · 1 line
2.35.3
#5 lore raw

PATCH v8 5/6 net/tls: split tls_rx_reader_lock

In reply to Hannes Reinecke

Split tls_rx_reader_{lock,unlock} into an 'acquire/release' and the actual locking part.
With that we can use the tls_rx_reader_lock in situations where the socket is already locked.

---
 net/tls/tls_sw.c | 38 ++++++++++++++++++++++----------------
 1 file changed, 22 insertions(+), 16 deletions(-)
net/tls/tls_sw.c +22 -16
index 9aef45e870a5..d0636ea13009 100644
@@ -1848,13 +1848,10 @@ tls_read_flush_backlog(struct sock *sk, struct tls_prot_info *prot,
 	return sk_flush_backlog(sk);
 }
 
-static int tls_rx_reader_lock(struct sock *sk, struct tls_sw_context_rx *ctx,
-			      bool nonblock)
+static int tls_rx_reader_acquire(struct sock *sk, struct tls_sw_context_rx *ctx,
+				 bool nonblock)
 {
 	long timeo;
-	int err;
-
-	lock_sock(sk);
 
 	timeo = sock_rcvtimeo(sk, nonblock);
 
@@ -1868,26 +1865,30 @@ static int tls_rx_reader_lock(struct sock *sk, struct tls_sw_context_rx *ctx,
 			      !READ_ONCE(ctx->reader_present), &wait);
 		remove_wait_queue(&ctx->wq, &wait);
 
-		if (timeo <= 0) {
-			err = -EAGAIN;
-			goto err_unlock;
-		}
-		if (signal_pending(current)) {
-			err = sock_intr_errno(timeo);
-			goto err_unlock;
-		}
+		if (timeo <= 0)
+			return -EAGAIN;
+		if (signal_pending(current))
+			return sock_intr_errno(timeo);
 	}
 
 	WRITE_ONCE(ctx->reader_present, 1);
 
 	return 0;
+}
 
-err_unlock:
-	release_sock(sk);
+static int tls_rx_reader_lock(struct sock *sk, struct tls_sw_context_rx *ctx,
+			      bool nonblock)
+{
+	int err;
+
+	lock_sock(sk);
+	err = tls_rx_reader_acquire(sk, ctx, nonblock);
+	if (err)
+		release_sock(sk);
 	return err;
 }
 
-static void tls_rx_reader_unlock(struct sock *sk, struct tls_sw_context_rx *ctx)
+static void tls_rx_reader_release(struct sock *sk, struct tls_sw_context_rx *ctx)
 {
 	if (unlikely(ctx->reader_contended)) {
 		if (wq_has_sleeper(&ctx->wq))
@@ -1899,6 +1900,11 @@ static void tls_rx_reader_unlock(struct sock *sk, struct tls_sw_context_rx *ctx)
 	}
 
 	WRITE_ONCE(ctx->reader_present, 0);
+}
+
+static void tls_rx_reader_unlock(struct sock *sk, struct tls_sw_context_rx *ctx)
+{
+	tls_rx_reader_release(sk, ctx);
 	release_sock(sk);
 }
 
Signature · 1 line
2.35.3
#0 lore raw

v8 cover net/tls: fixes for NVMe-over-TLS

Hi all,

here are some small fixes to get NVMe-over-TLS up and running. The first set are just minor modifications to have MSG_EOR handled for TLS, but the second set implements the ->read_sock() callback for tls_sw which I guess could do with some reviews.

As usual, comments and reviews are welcome.

Changes to the original submission:
- Add a testcase for MSG_EOR handling

Changes to v2:
- Bail out on conflicting message flags
- Rework flag handling

Changes to v3:
- Return -EINVAL on conflicting flags
- Rebase on top of net-next

Changes to v4:
- Add tlx_rx_reader_lock() to read_sock
- Add MSG_EOR handling to tls_sw_readpages()

Changes to v5:
- Rebase to latest upstream
- Split tls_rx_reader_lock() as suggested by Sagi

Changes to v6:
- Fixup tls_strp_read_copyin() to avoid infinite recursion in tls_read_sock()
- Rework tls_read_sock() to read all available data

Changes to v7:
- Include reviews from Jakub

Hannes Reinecke (6):
net/tls: handle MSG_EOR for tls_sw TX flow net/tls: handle MSG_EOR for tls_device TX flow selftests/net/tls: add test for MSG_EOR net/tls: Use tcp_read_sock() instead of ops->read_sock() net/tls: split tls_rx_reader_lock
net/tls: implement ->read_sock()

 net/tls/tls.h                     |   2 +
 net/tls/tls_device.c              |   6 +-
 net/tls/tls_main.c                |   2 +
 net/tls/tls_strp.c                |   3 +-
 net/tls/tls_sw.c                  | 132 ++++++++++++++++++++++++++----
 tools/testing/selftests/net/tls.c |  11 +++
 6 files changed, 136 insertions(+), 20 deletions(-)
Signature · 1 line
2.35.3
#1 lore raw

PATCH v8 1/6 net/tls: handle MSG_EOR for tls_sw TX flow

tls_sw_sendmsg() already handles MSG_MORE, but bails out on MSG_EOR.
Seeing that MSG_EOR is basically the opposite of MSG_MORE this patch adds handling MSG_EOR by treating it as the negation of MSG_MORE.
And erroring out if MSG_EOR is specified with MSG_MORE.

---
 net/tls/tls_sw.c | 5 ++++-
 1 file changed, 4 insertions(+), 1 deletion(-)
net/tls/tls_sw.c +4 -1
index 53f944e6d8ef..9aef45e870a5 100644
@@ -984,6 +984,9 @@ static int tls_sw_sendmsg_locked(struct sock *sk, struct msghdr *msg,
 	int ret = 0;
 	int pending;
 
+	if (!eor && (msg->msg_flags & MSG_EOR))
+		return -EINVAL;
+
 	if (unlikely(msg->msg_controllen)) {
 		ret = tls_process_cmsg(sk, msg, &record_type);
 		if (ret) {
@@ -1193,7 +1196,7 @@ int tls_sw_sendmsg(struct sock *sk, struct msghdr *msg, size_t size)
 	int ret;
 
 	if (msg->msg_flags & ~(MSG_MORE | MSG_DONTWAIT | MSG_NOSIGNAL |
-			       MSG_CMSG_COMPAT | MSG_SPLICE_PAGES |
+			       MSG_CMSG_COMPAT | MSG_SPLICE_PAGES | MSG_EOR |
 			       MSG_SENDPAGE_NOPOLICY))
 		return -EOPNOTSUPP;
 
Signature · 1 line
2.35.3
#4 lore raw

PATCH v8 4/6 net/tls: Use tcp_read_sock() instead of ops->read_sock()

In reply to Hannes Reinecke

TLS resets the protocol operations, so the read_sock() callback might be changed, too.
In this case using sock->ops->readsock() in tls_strp_read_copyin() will enter an infinite recursion if the read_sock() callback is calling tls_rx_rec_wait() which will call into sock->ops->readsock() via tls_strp_read_copyin().
But as tls_strp_read_copyin() is supposed to produce data from the consumed socket and that socket is always a TCP socket we can call tcp_read_sock() directly without having to deal with callbacks.

---
 net/tls/tls_strp.c | 3 +--
 1 file changed, 1 insertion(+), 2 deletions(-)
net/tls/tls_strp.c +1 -2
index f37f4a0fcd3c..ca1e0e198ceb 100644
@@ -369,7 +369,6 @@ static int tls_strp_copyin(read_descriptor_t *desc, struct sk_buff *in_skb,
 
 static int tls_strp_read_copyin(struct tls_strparser *strp)
 {
-	struct socket *sock = strp->sk->sk_socket;
 	read_descriptor_t desc;
 
 	desc.arg.data = strp;
@@ -377,7 +376,7 @@ static int tls_strp_read_copyin(struct tls_strparser *strp)
 	desc.count = 1; /* give more than one skb per call */
 
 	/* sk should be locked here, so okay to do read_sock */
-	sock->ops->read_sock(strp->sk, &desc, tls_strp_copyin);
+	tcp_read_sock(strp->sk, &desc, tls_strp_copyin);
 
 	return desc.error;
 }
Signature · 1 line
2.35.3
#6 lore raw

PATCH v8 6/6 net/tls: implement ->read_sock()

In reply to Hannes Reinecke

Implement ->read_sock() function for use with nvme-tcp.

---
 net/tls/tls.h      |  2 ++
 net/tls/tls_main.c |  2 ++
 net/tls/tls_sw.c   | 89 ++++++++++++++++++++++++++++++++++++++++++++++
 3 files changed, 93 insertions(+)
net/tls/tls.h +2 -0
index 86cef1c68e03..7e4d45537deb 100644
@@ -110,6 +110,8 @@ bool tls_sw_sock_is_readable(struct sock *sk);
 ssize_t tls_sw_splice_read(struct socket *sock, loff_t *ppos,
 			   struct pipe_inode_info *pipe,
 			   size_t len, unsigned int flags);
+int tls_sw_read_sock(struct sock *sk, read_descriptor_t *desc,
+		     sk_read_actor_t read_actor);
 
 int tls_device_sendmsg(struct sock *sk, struct msghdr *msg, size_t size);
 void tls_device_splice_eof(struct socket *sock);
net/tls/tls_main.c +2 -0
index b6896126bb92..7dbb8cd8f809 100644
@@ -962,10 +962,12 @@ static void build_proto_ops(struct proto_ops ops[TLS_NUM_CONFIG][TLS_NUM_CONFIG]
 	ops[TLS_BASE][TLS_SW  ] = ops[TLS_BASE][TLS_BASE];
 	ops[TLS_BASE][TLS_SW  ].splice_read	= tls_sw_splice_read;
 	ops[TLS_BASE][TLS_SW  ].poll		= tls_sk_poll;
+	ops[TLS_BASE][TLS_SW  ].read_sock	= tls_sw_read_sock;
 
 	ops[TLS_SW  ][TLS_SW  ] = ops[TLS_SW  ][TLS_BASE];
 	ops[TLS_SW  ][TLS_SW  ].splice_read	= tls_sw_splice_read;
 	ops[TLS_SW  ][TLS_SW  ].poll		= tls_sk_poll;
+	ops[TLS_SW  ][TLS_SW  ].read_sock	= tls_sw_read_sock;
 
 #ifdef CONFIG_TLS_DEVICE
 	ops[TLS_HW  ][TLS_BASE] = ops[TLS_BASE][TLS_BASE];
net/tls/tls_sw.c +89 -0

95 lines not on this page · raw message

Signature · 1 line
2.35.3
maintainer Reviewed-by on the v8 series
#9 lore raw
In reply to Hannes Reinecke
Quoted 4 lines · On Fri, 21 Jul 2023 16:35:17 +0200 Hannes Reinecke wrote:

here are some small fixes to get NVMe-over-TLS up and running. The first set are just minor modifications to have MSG_EOR handled for TLS, but the second set implements the ->read_sock() callback for tls_sw which I guess could do with some reviews.

Sagi, I _think_ a stable branch with this should be doable, would you like one, or no rush?

#10 lore raw
Quoted 10 lines · Jakub Kicinski
Quoted 4 lines · On Fri, 21 Jul 2023 16:35:17 +0200 Hannes Reinecke wrote:

here are some small fixes to get NVMe-over-TLS up and running. The first set are just minor modifications to have MSG_EOR handled for TLS, but the second set implements the ->read_sock() callback for tls_sw which I guess could do with some reviews.

Sagi, I _think_ a stable branch with this should be doable, would you like one, or no rush?

I guess a stable branch would not be too bad; I've got another set of patches for the NVMe side, too.
Sagi?

Cheers,

Hannes

Signature · 5 lines
Dr. Hannes Reinecke                Kernel Storage Architect
[email protected]                              +49 911 74053 688
SUSE Software Solutions GmbH, Maxfeldstr. 5, 90409 Nürnberg
HRB 36809 (AG Nürnberg), Geschäftsführer: Ivo Totev, Andrew
Myers, Andrew McDonald, Martje Boudien Moerman
maintainer
#11 lore raw
Quoted 6 lines
Quoted 6 lines
Quoted 5 lines

here are some small fixes to get NVMe-over-TLS up and running. The first set are just minor modifications to have MSG_EOR handled for TLS, but the second set implements the ->read_sock() callback for tls_sw which I guess could do with some reviews.

Quoted 8 lines
Quoted 4 lines

Sagi, I _think_ a stable branch with this should be doable, would you like one, or no rush?

I guess a stable branch would not be too bad; I've got another set of patches for the NVMe side, too.
Sagi?

I don't think there is a real need for this to go to stable, nothing is using it. Perhaps the MSG_EOR patches can go to stable in case there is some userspace code that wants to rely on it.

maintainer
#7 lore raw

Re: [PATCH 6/6] net/tls: implement ->read_sock()

In reply to Hannes Reinecke
Quoted 91 lines · Hannes Reinecke

Implement ->read_sock() function for use with nvme-tcp.

---
  net/tls/tls.h      |  2 ++
  net/tls/tls_main.c |  2 ++
  net/tls/tls_sw.c   | 89 ++++++++++++++++++++++++++++++++++++++++++++++
  3 files changed, 93 insertions(+)
net/tls/tls.h +2 -0
index 86cef1c68e03..7e4d45537deb 100644
@@ -110,6 +110,8 @@ bool tls_sw_sock_is_readable(struct sock *sk);
  ssize_t tls_sw_splice_read(struct socket *sock, loff_t *ppos,
  			   struct pipe_inode_info *pipe,
  			   size_t len, unsigned int flags);
+int tls_sw_read_sock(struct sock *sk, read_descriptor_t *desc,
+		     sk_read_actor_t read_actor);
 
  int tls_device_sendmsg(struct sock *sk, struct msghdr *msg, size_t size);
  void tls_device_splice_eof(struct socket *sock);
net/tls/tls_main.c +2 -0
index b6896126bb92..7dbb8cd8f809 100644
@@ -962,10 +962,12 @@ static void build_proto_ops(struct proto_ops ops[TLS_NUM_CONFIG][TLS_NUM_CONFIG]
  	ops[TLS_BASE][TLS_SW  ] = ops[TLS_BASE][TLS_BASE];
  	ops[TLS_BASE][TLS_SW  ].splice_read	= tls_sw_splice_read;
  	ops[TLS_BASE][TLS_SW  ].poll		= tls_sk_poll;
+	ops[TLS_BASE][TLS_SW  ].read_sock	= tls_sw_read_sock;
 
  	ops[TLS_SW  ][TLS_SW  ] = ops[TLS_SW  ][TLS_BASE];
  	ops[TLS_SW  ][TLS_SW  ].splice_read	= tls_sw_splice_read;
  	ops[TLS_SW  ][TLS_SW  ].poll		= tls_sk_poll;
+	ops[TLS_SW  ][TLS_SW  ].read_sock	= tls_sw_read_sock;
 
  #ifdef CONFIG_TLS_DEVICE
  	ops[TLS_HW  ][TLS_BASE] = ops[TLS_BASE][TLS_BASE];
net/tls/tls_sw.c +40 -0
index d0636ea13009..f7ffbe7620cb 100644
@@ -2202,6 +2202,95 @@ ssize_t tls_sw_splice_read(struct socket *sock,  loff_t *ppos,
  	goto splice_read_end;
  }
 
+int tls_sw_read_sock(struct sock *sk, read_descriptor_t *desc,
+		     sk_read_actor_t read_actor)
+{
+	struct tls_context *tls_ctx = tls_get_ctx(sk);
+	struct tls_sw_context_rx *ctx = tls_sw_ctx_rx(tls_ctx);
+	struct strp_msg *rxm = NULL;
+	struct sk_buff *skb = NULL;
+	struct sk_psock *psock;
+	struct tls_msg *tlm;
+	ssize_t copied = 0;
+	int err, used;
+
+	psock = sk_psock_get(sk);
+	if (psock) {
+		sk_psock_put(sk, psock);
+		return -EINVAL;
+	}
+	err = tls_rx_reader_acquire(sk, ctx, true);
+	if (err < 0)
+		return err;
+
+	/* If crypto failed the connection is broken */
+	err = ctx->async_wait.err;
+	if (err)
+		goto read_sock_end;
+
+	do {
+		if (!skb_queue_empty(&ctx->rx_list)) {
+			skb = __skb_dequeue(&ctx->rx_list);
+			rxm = strp_msg(skb);
+			tlm = tls_msg(skb);
+		} else {
+			struct tls_decrypt_arg darg;
+
+			err = tls_rx_rec_wait(sk, NULL, true, true);
+			if (err <= 0)
+				goto read_sock_end;
+
+			memset(&darg.inargs, 0, sizeof(darg.inargs));
+

net/tls/tls_sw.c ssize_t tls_sw_splice_read(struct socket *sock, loff_t *ppos, · line 2254

Quoted 10 lines
net/tls/tls_sw.c +10 -0
+			rxm = strp_msg(tls_strp_msg(ctx));
+			tlm = tls_msg(tls_strp_msg(ctx));
+
+			err = tls_rx_one_record(sk, NULL, &darg);
+			if (err < 0) {
+				tls_err_abort(sk, -EBADMSG);
+				goto read_sock_end;
+			}
+
+			sk_flush_backlog(sk);

Question,
Based on Jakub's comment, the flush is better spaced out. Why not just do it once at the end? Or alternatively, call tls_read_flush_backlog() ? Or just count by hand every 4 records or 128K (and once in the end)?

I don't really know what would be the impact though, but you are effectively releasing and re-acquiring the socket flushing the backlog every record...

#8 lore raw

Re: [PATCH 6/6] net/tls: implement ->read_sock()

Quoted 109 lines · Sagi Grimberg
Quoted 89 lines · Hannes Reinecke

Implement ->read_sock() function for use with nvme-tcp.

---
  net/tls/tls.h      |  2 ++
  net/tls/tls_main.c |  2 ++
  net/tls/tls_sw.c   | 89 ++++++++++++++++++++++++++++++++++++++++++++++

  3 files changed, 93 insertions(+)
net/tls/tls.h +2 -0
index 86cef1c68e03..7e4d45537deb 100644
@@ -110,6 +110,8 @@ bool tls_sw_sock_is_readable(struct sock *sk);
   ssize_t tls_sw_splice_read(struct socket *sock, loff_t *ppos,
                  struct pipe_inode_info *pipe,
                  size_t len, unsigned int flags);
+int tls_sw_read_sock(struct sock *sk, read_descriptor_t *desc,
+             sk_read_actor_t read_actor);
   int tls_device_sendmsg(struct sock *sk, struct msghdr *msg, size_t size);
   void tls_device_splice_eof(struct socket *sock);
net/tls/tls_main.c +0 -0
index b6896126bb92..7dbb8cd8f809 100644
@@ -962,10 +962,12 @@ static void build_proto_ops(struct proto_ops 

ops[TLS_NUM_CONFIG][TLS_NUM_CONFIG] ops[TLS_BASE][TLS_SW  ] = ops[TLS_BASE][TLS_BASE]; ops[TLS_BASE][TLS_SW  ].splice_read    = tls_sw_splice_read; ops[TLS_BASE][TLS_SW  ].poll        = tls_sk_poll; +    ops[TLS_BASE][TLS_SW  ].read_sock    = tls_sw_read_sock; ops[TLS_SW  ][TLS_SW  ] = ops[TLS_SW  ][TLS_BASE]; ops[TLS_SW  ][TLS_SW  ].splice_read    = tls_sw_splice_read; ops[TLS_SW  ][TLS_SW  ].poll        = tls_sk_poll; +    ops[TLS_SW  ][TLS_SW  ].read_sock    = tls_sw_read_sock; #ifdef CONFIG_TLS_DEVICE
      ops[TLS_HW  ][TLS_BASE] = ops[TLS_BASE][TLS_BASE];

net/tls/tls_sw.c +40 -0
index d0636ea13009..f7ffbe7620cb 100644
@@ -2202,6 +2202,95 @@ ssize_t tls_sw_splice_read(struct socket 
 *sock,  loff_t *ppos,
       goto splice_read_end;
   }
+int tls_sw_read_sock(struct sock *sk, read_descriptor_t *desc,
+             sk_read_actor_t read_actor)
+{
+    struct tls_context *tls_ctx = tls_get_ctx(sk);
+    struct tls_sw_context_rx *ctx = tls_sw_ctx_rx(tls_ctx);
+    struct strp_msg *rxm = NULL;
+    struct sk_buff *skb = NULL;
+    struct sk_psock *psock;
+    struct tls_msg *tlm;
+    ssize_t copied = 0;
+    int err, used;
+
+    psock = sk_psock_get(sk);
+    if (psock) {
+        sk_psock_put(sk, psock);
+        return -EINVAL;
+    }
+    err = tls_rx_reader_acquire(sk, ctx, true);
+    if (err < 0)
+        return err;
+
+    /* If crypto failed the connection is broken */
+    err = ctx->async_wait.err;
+    if (err)
+        goto read_sock_end;
+
+    do {
+        if (!skb_queue_empty(&ctx->rx_list)) {
+            skb = __skb_dequeue(&ctx->rx_list);
+            rxm = strp_msg(skb);
+            tlm = tls_msg(skb);
+        } else {
+            struct tls_decrypt_arg darg;
+
+            err = tls_rx_rec_wait(sk, NULL, true, true);
+            if (err <= 0)
+                goto read_sock_end;
+
+            memset(&darg.inargs, 0, sizeof(darg.inargs));
+

net/tls/tls_sw.c ssize_t tls_sw_splice_read(struct socket · line 2254

Quoted 10 lines
net/tls/tls_sw.c +10 -0
+            rxm = strp_msg(tls_strp_msg(ctx));
+            tlm = tls_msg(tls_strp_msg(ctx));
+
+            err = tls_rx_one_record(sk, NULL, &darg);
+            if (err < 0) {
+                tls_err_abort(sk, -EBADMSG);
+                goto read_sock_end;
+            }
+
+            sk_flush_backlog(sk);

Question,
Based on Jakub's comment, the flush is better spaced out. Why not just do it once at the end? Or alternatively, call tls_read_flush_backlog() ? Or just count by hand every 4 records or 128K (and once in the end)?

Quoted 3 lines

I don't really know what would be the impact though, but you are effectively releasing and re-acquiring the socket flushing the backlog every record...

I really have no idea.
I'll see to modify it to use tls_read_flush_backlog().

Cheers,

Hannes

Signature · 5 lines
Dr. Hannes Reinecke                Kernel Storage Architect
[email protected]                              +49 911 74053 688
SUSE Software Solutions GmbH, Maxfeldstr. 5, 90409 Nürnberg
HRB 36809 (AG Nürnberg), Geschäftsführer: Ivo Totev, Andrew
Myers, Andrew McDonald, Martje Boudien Moerman
maintainer
#12 lore raw
In reply to Sagi Grimberg
Quoted 5 lines · On Mon, 24 Jul 2023 10:21:52 +0300 Sagi Grimberg wrote:
Quoted 5 lines
Quoted 4 lines

Sagi, I _think_ a stable branch with this should be doable, would you like one, or no rush?

Quoted 7 lines
Quoted 3 lines

I guess a stable branch would not be too bad; I've got another set of patches for the NVMe side, too.
Sagi?

I don't think there is a real need for this to go to stable, nothing is using it. Perhaps the MSG_EOR patches can go to stable in case there is some userspace code that wants to rely on it.

I'm probably using the wrong word. I mean a branch based on -rc3 that's not going to get rebased so the commits IDs match and we can both pull it in. Not stable as in Greg KH.

maintainer
#13 lore raw
Quoted 10 lines
Quoted 10 lines
Quoted 9 lines
Quoted 5 lines

Sagi, I _think_ a stable branch with this should be doable, would you like one, or no rush?

I guess a stable branch would not be too bad; I've got another set of patches for the NVMe side, too.
Sagi?

Quoted 7 lines
Quoted 3 lines

I don't think there is a real need for this to go to stable, nothing is using it. Perhaps the MSG_EOR patches can go to stable in case there is some userspace code that wants to rely on it.

I'm probably using the wrong word. I mean a branch based on -rc3 that's not going to get rebased so the commits IDs match and we can both pull it in. Not stable as in Greg KH.

Are you aiming this for 6.5 ? We are unlikely to get the nvme bits in this round. I also don't think there is a conflict so the nvme bits can go in for 6.6 and later the nvme tree will pull the tls updates.

maintainer
#14 lore raw
Quoted 7 lines · On Mon, 24 Jul 2023 22:44:49 +0300 Sagi Grimberg wrote:
Quoted 3 lines

I'm probably using the wrong word. I mean a branch based on -rc3 that's not going to get rebased so the commits IDs match and we can both pull it in. Not stable as in Greg KH.

Are you aiming this for 6.5 ? We are unlikely to get the nvme bits in this round. I also don't think there is a conflict so the nvme bits can go in for 6.6 and later the nvme tree will pull the tls updates.

Great, less work :)

Let's see a v9 with the flushing improved and we'll apply to net-next directly.

#15 lore raw
In reply to Sagi Grimberg
Quoted 14 lines · Sagi Grimberg
Quoted 14 lines
Quoted 13 lines
Quoted 9 lines
Quoted 5 lines

Sagi, I _think_ a stable branch with this should be doable, would you like one, or no rush?

I guess a stable branch would not be too bad; I've got another set of patches for the NVMe side, too.
Sagi?

I don't think there is a real need for this to go to stable, nothing is using it. Perhaps the MSG_EOR patches can go to stable in case there is some userspace code that wants to rely on it.

Quoted 7 lines
Quoted 3 lines

I'm probably using the wrong word. I mean a branch based on -rc3 that's not going to get rebased so the commits IDs match and we can both pull it in. Not stable as in Greg KH.

Are you aiming this for 6.5 ? We are unlikely to get the nvme bits in this round. I also don't think there is a conflict so the nvme bits can go in for 6.6 and later the nvme tree will pull the tls updates.

I really would love to get this into 6.5, as then we can do a clean rebase of the NVMe development tree off 6.5, knowing that we won't have to touch the networking. Otherwise you have to be sure to fork the 6.6 development branch off the right point.

But then I'm not doing the forking, so really all I care is that the networking bits are present in the 6.6 NVMe development branch ...

Cheers,

Hannes

Signature · 5 lines
Dr. Hannes Reinecke                Kernel Storage Architect
[email protected]                              +49 911 74053 688
SUSE Software Solutions GmbH, Maxfeldstr. 5, 90409 Nürnberg
HRB 36809 (AG Nürnberg), Geschäftsführer: Ivo Totev, Andrew
Myers, Andrew McDonald, Martje Boudien Moerman