Skip to contentSkip to conversation

tcp: Disable header prediction for MD5.

PATCH Applied Kuniyuki Iwashima 3 participants 6 messages Aug 3, 2023 – Aug 5, 2023

Subsystems (3)
lore
Share & tools
lore
#0 lore raw

v3 cover tcp: Disable header prediction for MD5.

The 1st patch disable header prediction for MD5 flow and the 2nd patch updates the stale comment in tcp_parse_options().

Changes:
v3:
* Disable header prediction instead of enabling
* Add the 2nd patch

v2: https://lore.kernel.org/netdev/[email protected]/
* Update function graph

v1: https://lore.kernel.org/all/[email protected]/

Kuniyuki Iwashima (2):
tcp: Disable header prediction for MD5 flow.
tcp: Update stale comment for MD5 in tcp_parse_options().

 net/ipv4/tcp_input.c     | 5 ++---
 net/ipv4/tcp_minisocks.c | 2 --
 net/ipv4/tcp_output.c    | 5 -----
 3 files changed, 2 insertions(+), 10 deletions(-)
Signature · 1 line
2.30.2
#1 lore raw

PATCH v3 1/2 tcp: Disable header prediction for MD5 flow.

TCP socket saves the minimum required header length in tcp_header_len of struct tcp_sock, and later the value is used in __tcp_fast_path_on() to generate a part of TCP header in tcp_sock(sk)->pred_flags.

In tcp_rcv_established(), if the incoming packet has the same pattern with pred_flags, we enter the fast path and skip full option parsing.

The MD5 option is parsed in tcp_v[46]_rcv(), so we need not parse it again later in tcp_rcv_established() unless other options exist. We add TCPOLEN_MD5SIG_ALIGNED to tcp_header_len in two paths to avoid the slow path.

For passive open connections with MD5, we add TCPOLEN_MD5SIG_ALIGNED to tcp_header_len in tcp_create_openreq_child() after 3WHS.

On the other hand, we do it in tcp_connect_init() for active open connections. However, the value is overwritten while processing SYN+ACK or crossed SYN in tcp_rcv_synsent_state_process().

These two cases will have the wrong value in pred_flags and never go into the fast path.

We could update tcp_header_len in tcp_rcv_synsent_state_process(), but a test with slightly modified netperf which uses MD5 for each flow shows that the slow path is actually a bit faster than the fast path.

On c5.4xlarge EC2 instance (16 vCPU, 32 GiB mem)

  $ for i in {1..10}; do
  ./super_netperf $(nproc) -H localhost -l 10 -- -m 256 -M 256;
  done

Avg of 10
* 36e68eadd303 : 10.376 Gbps
* all fast path : 10.374 Gbps (patch v2, See Link)
* all slow path : 10.394 Gbps

The header prediction is not worth adding complexity for MD5, so let's disable it for MD5.

---
v3:
* Disable header prediction

v2: https://lore.kernel.org/netdev/[email protected]/
* Update function graph

v1: https://lore.kernel.org/netdev/[email protected]/

---
 net/ipv4/tcp_minisocks.c | 2 --
 net/ipv4/tcp_output.c    | 5 -----
 2 files changed, 7 deletions(-)
net/ipv4/tcp_minisocks.c +0 -2
index c8f2aa003387..ddba3b67f7c8 100644
@@ -570,8 +570,6 @@ struct sock *tcp_create_openreq_child(const struct sock *sk,
 	newtp->tsoffset = treq->ts_off;
 #ifdef CONFIG_TCP_MD5SIG
 	newtp->md5sig_info = NULL;	/*XXX*/
-	if (treq->af_specific->req_md5_lookup(sk, req_to_sk(req)))
-		newtp->tcp_header_len += TCPOLEN_MD5SIG_ALIGNED;
 #endif
 	if (skb->len >= TCP_MSS_DEFAULT + newtp->tcp_header_len)
 		newicsk->icsk_ack.last_seg_size = skb->len - newtp->tcp_header_len;
net/ipv4/tcp_output.c +0 -5
index 51d8638d4b4c..c5412ee77fc8 100644
@@ -3741,11 +3741,6 @@ static void tcp_connect_init(struct sock *sk)
 	if (READ_ONCE(sock_net(sk)->ipv4.sysctl_tcp_timestamps))
 		tp->tcp_header_len += TCPOLEN_TSTAMP_ALIGNED;
 
-#ifdef CONFIG_TCP_MD5SIG
-	if (tp->af_specific->md5_lookup(sk, sk))
-		tp->tcp_header_len += TCPOLEN_MD5SIG_ALIGNED;
-#endif
-
 	/* If user gave his TCP_MAXSEG, record it to clamp */
 	if (tp->rx_opt.user_mss)
 		tp->rx_opt.mss_clamp = tp->rx_opt.user_mss;
Signature · 1 line
2.30.2
Reviewed-by on v3 1/2
#2 lore raw

Re: [PATCH v3 net-next 1/2] tcp: Disable header prediction for MD5 flow.

Quoted 38 lines · Kuniyuki Iwashima

TCP socket saves the minimum required header length in tcp_header_len of struct tcp_sock, and later the value is used in __tcp_fast_path_on() to generate a part of TCP header in tcp_sock(sk)->pred_flags.

In tcp_rcv_established(), if the incoming packet has the same pattern with pred_flags, we enter the fast path and skip full option parsing.

The MD5 option is parsed in tcp_v[46]_rcv(), so we need not parse it again later in tcp_rcv_established() unless other options exist. We add TCPOLEN_MD5SIG_ALIGNED to tcp_header_len in two paths to avoid the slow path.

For passive open connections with MD5, we add TCPOLEN_MD5SIG_ALIGNED to tcp_header_len in tcp_create_openreq_child() after 3WHS.

On the other hand, we do it in tcp_connect_init() for active open connections. However, the value is overwritten while processing SYN+ACK or crossed SYN in tcp_rcv_synsent_state_process().

These two cases will have the wrong value in pred_flags and never go into the fast path.

We could update tcp_header_len in tcp_rcv_synsent_state_process(), but a test with slightly modified netperf which uses MD5 for each flow shows that the slow path is actually a bit faster than the fast path.

On c5.4xlarge EC2 instance (16 vCPU, 32 GiB mem)

  $ for i in {1..10}; do
  ./super_netperf $(nproc) -H localhost -l 10 -- -m 256 -M 256;
  done

Avg of 10
* 36e68eadd303 : 10.376 Gbps
* all fast path : 10.374 Gbps (patch v2, See Link)
* all slow path : 10.394 Gbps

Quoted 6 lines

The header prediction is not worth adding complexity for MD5, so let's disable it for MD5.

---

#3 lore raw

PATCH v3 2/2 tcp: Update stale comment for MD5 in tcp_parse_options().

In reply to Kuniyuki Iwashima

Since commit 9ea88a153001 ("tcp: md5: check md5 signature without socket lock"), the MD5 option is checked in tcp_v[46]_rcv().

---
 net/ipv4/tcp_input.c | 5 ++---
 1 file changed, 2 insertions(+), 3 deletions(-)
net/ipv4/tcp_input.c +2 -3
index 670c3dab24f2..2995802126e4 100644
@@ -4126,9 +4126,8 @@ void tcp_parse_options(const struct net *net,
 				break;
 #ifdef CONFIG_TCP_MD5SIG
 			case TCPOPT_MD5SIG:
-				/*
-				 * The MD5 Hash has already been
-				 * checked (see tcp_v{4,6}_do_rcv()).
+				/* The MD5 Hash has already been
+				 * checked (see tcp_v{4,6}_rcv()).
 				 */
 				break;
 #endif
Signature · 1 line
2.30.2
#5 lore raw
In reply to Kuniyuki Iwashima
This series was applied to netdev/net-next.git (main) by Jakub Kicinski <[email protected]>:

Hello:

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

Quoted 10 lines · On Thu, 3 Aug 2023 15:45:50 -0700 you wrote:

The 1st patch disable header prediction for MD5 flow and the 2nd patch updates the stale comment in tcp_parse_options().

Changes:
v3:
* Disable header prediction instead of enabling
* Add the 2nd patch

[...]

Here is the summary with links:
- [v3,net-next,1/2] tcp: Disable header prediction for MD5 flow.

    https://git.kernel.org/netdev/net-next/c/d0f2b7a9ca0a
  - [v3,net-next,2/2] tcp: Update stale comment for MD5 in tcp_parse_options().
    https://git.kernel.org/netdev/net-next/c/b20515368932

You are awesome, thank you!

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