diff mbox series

[v6,mptcp-next,1/5] mptcp: MP_FAIL suboption sending

Message ID 277ac1e6d1fde4c180eba3f1bb1846ea58679915.1627464017.git.geliangtang@xiaomi.com (mailing list archive)
State Superseded, archived
Headers show
Series MP_FAIL support | expand

Commit Message

Geliang Tang July 28, 2021, 9:35 a.m. UTC 
From: Geliang Tang <geliangtang@xiaomi.com>

This patch added the MP_FAIL suboption sending support.

Add a new flag named send_mp_fail in struct mptcp_subflow_context. If
this flag is set, send out MP_FAIL suboption.

Add a new member fail_seq in struct mptcp_out_options to save the data
sequence number to put into the MP_FAIL suboption.

An MP_FAIL option could be included in a RST or on the subflow-level
ACK.

Signed-off-by: Geliang Tang <geliangtang@xiaomi.com>
---
 include/net/mptcp.h  |  5 +++-
 net/mptcp/options.c  | 61 +++++++++++++++++++++++++++++++++++++++++---
 net/mptcp/protocol.h |  3 +++
 3 files changed, 64 insertions(+), 5 deletions(-)

Comments

Paolo Abeni July 28, 2021, 10:31 a.m. UTC | #1 
Hello,

On Wed, 2021-07-28 at 17:35 +0800, Geliang Tang wrote:
> From: Geliang Tang <geliangtang@xiaomi.com>
> 
> This patch added the MP_FAIL suboption sending support.
> 
> Add a new flag named send_mp_fail in struct mptcp_subflow_context. If
> this flag is set, send out MP_FAIL suboption.
> 
> Add a new member fail_seq in struct mptcp_out_options to save the data
> sequence number to put into the MP_FAIL suboption.
> 
> An MP_FAIL option could be included in a RST or on the subflow-level
> ACK.
> 
> Signed-off-by: Geliang Tang <geliangtang@xiaomi.com>
> ---
>  include/net/mptcp.h  |  5 +++-
>  net/mptcp/options.c  | 61 +++++++++++++++++++++++++++++++++++++++++---
>  net/mptcp/protocol.h |  3 +++
>  3 files changed, 64 insertions(+), 5 deletions(-)
> 
> diff --git a/include/net/mptcp.h b/include/net/mptcp.h
> index 3236010afa29..6026bbefbffd 100644
> --- a/include/net/mptcp.h
> +++ b/include/net/mptcp.h
> @@ -74,7 +74,10 @@ struct mptcp_out_options {
>  			struct mptcp_addr_info addr;
>  			u64 ahmac;
>  		};
> -		struct mptcp_ext ext_copy;
> +		struct {
> +			struct mptcp_ext ext_copy;
> +			u64 fail_seq;
> +		};
>  		struct {
>  			u32 nonce;
>  			u32 token;
> diff --git a/net/mptcp/options.c b/net/mptcp/options.c
> index 753d6ac43bff..2b15063c8009 100644
> --- a/net/mptcp/options.c
> +++ b/net/mptcp/options.c
> @@ -763,7 +763,7 @@ static bool mptcp_established_options_mp_prio(struct sock *sk,
>  	return true;
>  }
>  
> -static noinline void mptcp_established_options_rst(struct sock *sk, struct sk_buff *skb,
> +static noinline bool mptcp_established_options_rst(struct sock *sk, struct sk_buff *skb,
>  						   unsigned int *size,
>  						   unsigned int remaining,
>  						   struct mptcp_out_options *opts)
> @@ -771,12 +771,36 @@ static noinline void mptcp_established_options_rst(struct sock *sk, struct sk_bu
>  	const struct mptcp_subflow_context *subflow = mptcp_subflow_ctx(sk);
>  
>  	if (remaining < TCPOLEN_MPTCP_RST)
> -		return;
> +		return false;
>  
>  	*size = TCPOLEN_MPTCP_RST;
>  	opts->suboptions |= OPTION_MPTCP_RST;
>  	opts->reset_transient = subflow->reset_transient;
>  	opts->reset_reason = subflow->reset_reason;
> +
> +	return true;
> +}
> +
> +static bool mptcp_established_options_mp_fail(struct sock *sk,
> +					      unsigned int *size,
> +					      unsigned int remaining,
> +					      struct mptcp_out_options *opts)
> +{
> +	struct mptcp_subflow_context *subflow = mptcp_subflow_ctx(sk);
> +
> +	if (!subflow->send_mp_fail)
> +		return false;
> +
> +	if (remaining < TCPOLEN_MPTCP_FAIL)
> +		return false;
> +
> +	*size = TCPOLEN_MPTCP_FAIL;
> +	opts->suboptions |= OPTION_MPTCP_FAIL;
> +	opts->fail_seq = subflow->map_seq;
> +
> +	pr_debug("MP_FAIL fail_seq=%llu", opts->fail_seq);
> +
> +	return true;
>  }
>  
>  bool mptcp_established_options(struct sock *sk, struct sk_buff *skb,
> @@ -795,15 +819,30 @@ bool mptcp_established_options(struct sock *sk, struct sk_buff *skb,
>  		return false;
>  
>  	if (unlikely(skb && TCP_SKB_CB(skb)->tcp_flags & TCPHDR_RST)) {
> -		mptcp_established_options_rst(sk, skb, size, remaining, opts);
> +		if (mptcp_established_options_mp_fail(sk, &opt_size, remaining, opts)) {
> +			*size += opt_size;
> +			remaining -= opt_size;
> +		}
> +		if (mptcp_established_options_rst(sk, skb, &opt_size, remaining, opts)) {
> +			*size += opt_size;
> +			remaining -= opt_size;
> +		}
>  		return true;
>  	}
>  
>  	snd_data_fin = mptcp_data_fin_enabled(msk);
>  	if (mptcp_established_options_mp(sk, skb, snd_data_fin, &opt_size, remaining, opts))
>  		ret = true;
> -	else if (mptcp_established_options_dss(sk, skb, snd_data_fin, &opt_size, remaining, opts))
> +	else if (mptcp_established_options_dss(sk, skb, snd_data_fin, &opt_size, remaining, opts)) {
>  		ret = true;
> +		if (opts->ext_copy.use_ack) {

I *think* we could drop this check as the RFC says:

"""
   it will send back an MP_FAIL option on
   the subflow-level ACK,
"""

And to me subflow-level ACK really means TCP ack, that is any packet on
a TCP-established subflow will do.

Anyhow I think we can adjust the above with a squash-to patch after
this series is merged, as it already went through several iterations
and overall it LGTM.

While at that a few 'unlikely()' annotation will help ;)

/P
diff mbox series

Patch

diff --git a/include/net/mptcp.h b/include/net/mptcp.h
index 3236010afa29..6026bbefbffd 100644
--- a/include/net/mptcp.h
+++ b/include/net/mptcp.h
@@ -74,7 +74,10 @@  struct mptcp_out_options {
 			struct mptcp_addr_info addr;
 			u64 ahmac;
 		};
-		struct mptcp_ext ext_copy;
+		struct {
+			struct mptcp_ext ext_copy;
+			u64 fail_seq;
+		};
 		struct {
 			u32 nonce;
 			u32 token;
diff --git a/net/mptcp/options.c b/net/mptcp/options.c
index 753d6ac43bff..2b15063c8009 100644
--- a/net/mptcp/options.c
+++ b/net/mptcp/options.c
@@ -763,7 +763,7 @@  static bool mptcp_established_options_mp_prio(struct sock *sk,
 	return true;
 }
 
-static noinline void mptcp_established_options_rst(struct sock *sk, struct sk_buff *skb,
+static noinline bool mptcp_established_options_rst(struct sock *sk, struct sk_buff *skb,
 						   unsigned int *size,
 						   unsigned int remaining,
 						   struct mptcp_out_options *opts)
@@ -771,12 +771,36 @@  static noinline void mptcp_established_options_rst(struct sock *sk, struct sk_bu
 	const struct mptcp_subflow_context *subflow = mptcp_subflow_ctx(sk);
 
 	if (remaining < TCPOLEN_MPTCP_RST)
-		return;
+		return false;
 
 	*size = TCPOLEN_MPTCP_RST;
 	opts->suboptions |= OPTION_MPTCP_RST;
 	opts->reset_transient = subflow->reset_transient;
 	opts->reset_reason = subflow->reset_reason;
+
+	return true;
+}
+
+static bool mptcp_established_options_mp_fail(struct sock *sk,
+					      unsigned int *size,
+					      unsigned int remaining,
+					      struct mptcp_out_options *opts)
+{
+	struct mptcp_subflow_context *subflow = mptcp_subflow_ctx(sk);
+
+	if (!subflow->send_mp_fail)
+		return false;
+
+	if (remaining < TCPOLEN_MPTCP_FAIL)
+		return false;
+
+	*size = TCPOLEN_MPTCP_FAIL;
+	opts->suboptions |= OPTION_MPTCP_FAIL;
+	opts->fail_seq = subflow->map_seq;
+
+	pr_debug("MP_FAIL fail_seq=%llu", opts->fail_seq);
+
+	return true;
 }
 
 bool mptcp_established_options(struct sock *sk, struct sk_buff *skb,
@@ -795,15 +819,30 @@  bool mptcp_established_options(struct sock *sk, struct sk_buff *skb,
 		return false;
 
 	if (unlikely(skb && TCP_SKB_CB(skb)->tcp_flags & TCPHDR_RST)) {
-		mptcp_established_options_rst(sk, skb, size, remaining, opts);
+		if (mptcp_established_options_mp_fail(sk, &opt_size, remaining, opts)) {
+			*size += opt_size;
+			remaining -= opt_size;
+		}
+		if (mptcp_established_options_rst(sk, skb, &opt_size, remaining, opts)) {
+			*size += opt_size;
+			remaining -= opt_size;
+		}
 		return true;
 	}
 
 	snd_data_fin = mptcp_data_fin_enabled(msk);
 	if (mptcp_established_options_mp(sk, skb, snd_data_fin, &opt_size, remaining, opts))
 		ret = true;
-	else if (mptcp_established_options_dss(sk, skb, snd_data_fin, &opt_size, remaining, opts))
+	else if (mptcp_established_options_dss(sk, skb, snd_data_fin, &opt_size, remaining, opts)) {
 		ret = true;
+		if (opts->ext_copy.use_ack) {
+			if (mptcp_established_options_mp_fail(sk, &opt_size, remaining, opts)) {
+				*size += opt_size;
+				remaining -= opt_size;
+				return true;
+			}
+		}
+	}
 
 	/* we reserved enough space for the above options, and exceeding the
 	 * TCP option space would be fatal
@@ -1210,6 +1249,20 @@  static u16 mptcp_make_csum(const struct mptcp_ext *mpext)
 void mptcp_write_options(__be32 *ptr, const struct tcp_sock *tp,
 			 struct mptcp_out_options *opts)
 {
+	if (unlikely(OPTION_MPTCP_FAIL & opts->suboptions)) {
+		const struct sock *ssk = (const struct sock *)tp;
+		struct mptcp_subflow_context *subflow;
+
+		subflow = mptcp_subflow_ctx(ssk);
+		subflow->send_mp_fail = 0;
+
+		*ptr++ = mptcp_option(MPTCPOPT_MP_FAIL,
+				      TCPOLEN_MPTCP_FAIL,
+				      0, 0);
+		put_unaligned_be64(opts->fail_seq, ptr);
+		ptr += 2;
+	}
+
 	/* RST is mutually exclusive with everything else */
 	if (unlikely(OPTION_MPTCP_RST & opts->suboptions)) {
 		*ptr++ = mptcp_option(MPTCPOPT_RST,
diff --git a/net/mptcp/protocol.h b/net/mptcp/protocol.h
index e8a36ff52af6..b389fec18c89 100644
--- a/net/mptcp/protocol.h
+++ b/net/mptcp/protocol.h
@@ -27,6 +27,7 @@ 
 #define OPTION_MPTCP_PRIO	BIT(9)
 #define OPTION_MPTCP_RST	BIT(10)
 #define OPTION_MPTCP_DSS	BIT(11)
+#define OPTION_MPTCP_FAIL	BIT(12)
 
 /* MPTCP option subtypes */
 #define MPTCPOPT_MP_CAPABLE	0
@@ -68,6 +69,7 @@ 
 #define TCPOLEN_MPTCP_PRIO_ALIGN	4
 #define TCPOLEN_MPTCP_FASTCLOSE		12
 #define TCPOLEN_MPTCP_RST		4
+#define TCPOLEN_MPTCP_FAIL		12
 
 #define TCPOLEN_MPTCP_MPC_ACK_DATA_CSUM	(TCPOLEN_MPTCP_DSS_CHECKSUM + TCPOLEN_MPTCP_MPC_ACK_DATA)
 
@@ -429,6 +431,7 @@  struct mptcp_subflow_context {
 		mpc_map : 1,
 		backup : 1,
 		send_mp_prio : 1,
+		send_mp_fail : 1,
 		rx_eof : 1,
 		can_ack : 1,        /* only after processing the remote a key */
 		disposable : 1,	    /* ctx can be free at ulp release time */