Skip to contentSkip to conversation

Code-removal out of mptcp-net-next

PATCH Superseded Christoph Paasch 2 participants 14 messages Dec 18, 2018 – Dec 19, 2018

Subsystems (4)

v3 of 4 Superseded Latest is v4 →

lore
Share & tools
lore

Christoph Paasch

author
#0 lore raw

v3 cover Code-removal out of mptcp-net-next

Series removing a lot of code out of mptcp-net-next.

Before:

$ git diff --shortstat  1464193107da..mptcp-net-next/master
 54 files changed, 14663 insertions(+), 375 deletions(-)
$ git diff --shortstat  1464193107da..HEAD
 50 files changed, 12217 insertions(+), 349 deletions(-)

v3: Minor left-over removals that were missing - identified by Matthieu
v2: Some minor fixes - see notes in the sub-patches 4/12 and 12/12

Christoph Paasch (12):
mptcp: Remove support for DSS-checksums
mptcp: Remove some left-over cruft
mptcp: Remove RPS-recording for flow-steering in MPTCP
mptcp: Remove support for subflow-priorities (aka backup)
mptcp: Remove support for sendpage()
mptcp: Remove reinjection of data across subflows
mptcp: Remove support for TCP SYN-cookies
mptcp: Remove support for TCP Fast Open
mptcp: Remove MPTCP-code out of do_ip_setsockopt
mptcp: Remove scheduler-selection support
mptcp: Remove ndiffports path manager
mptcp: Make path-management no more a module nor selectable through
socketops

 include/linux/tcp.h          |  15 +-
 include/net/mptcp.h          | 250 +++++++---------------
 include/net/netns/mptcp.h    |   8 +-
 include/net/tcp.h            |  13 +-
 include/uapi/linux/if.h      |   1 -
 include/uapi/linux/tcp.h     |   1 -
 net/core/dev.c               |   2 +-
 net/ipv4/Kconfig             |  19 --
 net/ipv4/af_inet.c           |   6 -
 net/ipv4/ip_sockglue.c       |  14 --
 net/ipv4/syncookies.c        |  47 +----
 net/ipv4/tcp.c               | 155 +-------------
 net/ipv4/tcp_fastopen.c      |  16 +-
 net/ipv4/tcp_input.c         |  29 +--
 net/ipv4/tcp_minisocks.c     |   2 +-
 net/ipv4/tcp_output.c        |   3 +-
 net/ipv6/syncookies.c        |  32 +--
 net/mptcp/Kconfig            |  81 -------
 net/mptcp/Makefile           |   5 +-
 net/mptcp/mptcp_ctrl.c       | 261 +----------------------
 net/mptcp/mptcp_fullmesh.c   | 488 ++++++++-----------------------------------
 net/mptcp/mptcp_input.c      | 343 ++----------------------------
 net/mptcp/mptcp_ipv4.c       |  34 +--
 net/mptcp/mptcp_ipv6.c       |  34 +--
 net/mptcp/mptcp_ndiffports.c | 169 ---------------
 net/mptcp/mptcp_output.c     | 472 ++---------------------------------------
 net/mptcp/mptcp_pm.c         | 226 --------------------
 net/mptcp/mptcp_sched.c      | 280 +++----------------------
 28 files changed, 293 insertions(+), 2713 deletions(-)
 delete mode 100644 net/mptcp/mptcp_ndiffports.c
 delete mode 100644 net/mptcp/mptcp_pm.c
Signature · 1 line
2.16.2
author
#1 lore raw

PATCH v3 1/12 mptcp: Remove support for DSS-checksums

To reduce the code-size and complexity, we remove support for DSS-checksums. If we receive a MP_CAPABLE with a DSS-checksum we simply fallback to regular TCP.

Signed-off-by Christoph Paasch <cpaasch(a)apple.com>
---
 include/linux/tcp.h      |   3 +-
 include/net/mptcp.h      |  31 ++------
 net/mptcp/mptcp_ctrl.c   |  16 -----
 net/mptcp/mptcp_input.c  | 184 ++---------------------------------------------
 net/mptcp/mptcp_output.c |  69 ++----------------
 5 files changed, 17 insertions(+), 286 deletions(-)
include/linux/tcp.h +1 -2
index 29f8ab30c7ed..f38bf1e18e58 100644
@@ -94,8 +94,7 @@ struct tcp_out_options {
 	struct tcp_fastopen_cookie *fastopen_cookie;	/* Fast open cookie */
 #ifdef CONFIG_MPTCP
 	u16	mptcp_options;	/* bit field of MPTCP related OPTION_* */
-	u8	dss_csum:1,	/* dss-checksum required? */
-		add_addr_v4:1,
+	u8	add_addr_v4:1,
 		add_addr_v6:1,
 		mptcp_ver:4;
 
include/net/mptcp.h +6 -25
index b8b750034a91..a512e9386296 100644
@@ -102,8 +102,7 @@ struct mptcp_request_sock {
 
 	u8				loc_id;
 	u8				rem_id; /* Address-id in the MP_JOIN */
-	u8				dss_csum:1,
-					is_sub:1, /* Is this a new subflow? */
+	u8				is_sub:1, /* Is this a new subflow? */
 					low_prio:1, /* Interface set to low-prio? */
 					rcv_low_prio:1,
 					mptcp_ver:4;
@@ -111,7 +110,6 @@ struct mptcp_request_sock {
 
 struct mptcp_options_received {
 	u16	saw_mpc:1,
-		dss_csum:1,
 		drop_me:1,
 
 		is_mp_join:1,
@@ -168,7 +166,6 @@ struct mptcp_tcp_sock {
 		fully_established:1,
 		second_packet:1,
 		attached:1,
-		send_mp_fail:1,
 		include_mpc:1,
 		mapping_present:1,
 		map_data_fin:1,
@@ -269,7 +266,6 @@ struct mptcp_cb {
 		in_time_wait:1,
 		list_rcvd:1, /* XXX TO REMOVE */
 		addr_signal:1, /* Path-manager wants us to call addr_signal */
-		dss_csum:1,
 		server_side:1,
 		infinite_mapping_rcv:1,
 		infinite_mapping_snd:1,
@@ -288,7 +284,6 @@ struct mptcp_cb {
 	 * accessed. Thus, the whole data-path is on a single cache-line.
 	 */
 
-	u64	csum_cutoff_seq;
 	u64	infinite_rcv_seq;
 
 	/***** Start of fields, used for connection closure */
@@ -364,11 +359,9 @@ struct mptcp_cb {
  * To get the total length, just add the different options together.
  */
 #define MPTCP_SUB_LEN_SEQ	10
-#define MPTCP_SUB_LEN_SEQ_CSUM	12
 #define MPTCP_SUB_LEN_SEQ_ALIGN	12
 
 #define MPTCP_SUB_LEN_SEQ_64		14
-#define MPTCP_SUB_LEN_SEQ_CSUM_64	16
 #define MPTCP_SUB_LEN_SEQ_64_ALIGN	16
 
 #define MPTCP_SUB_LEN_ACK	4
@@ -434,10 +427,9 @@ extern bool mptcp_init_failed;
 #define OPTION_DATA_ACK		(1 << 4)
 #define OPTION_ADD_ADDR		(1 << 5)
 #define OPTION_MP_JOIN		(1 << 6)
-#define OPTION_MP_FAIL		(1 << 7)
-#define OPTION_MP_FCLOSE	(1 << 8)
-#define OPTION_REMOVE_ADDR	(1 << 9)
-#define OPTION_MP_PRIO		(1 << 10)
+#define OPTION_MP_FCLOSE	(1 << 7)
+#define OPTION_REMOVE_ADDR	(1 << 8)
+#define OPTION_MP_PRIO		(1 << 9)
 
 /* MPTCP flags: both TX and RX */
 #define MPTCPHDR_SEQ		0x01 /* DSS.M option is present */
@@ -447,7 +439,6 @@ extern bool mptcp_init_failed;
 #define MPTCPHDR_ACK		0x08
 #define MPTCPHDR_SEQ64_SET	0x10 /* Did we received a 64-bit seq number?  */
 #define MPTCPHDR_SEQ64_OFO	0x20 /* Is it not in our circular array? */
-#define MPTCPHDR_DSS_CSUM	0x40
 /* MPTCP flags: TX only */
 #define MPTCPHDR_INF		0x08
 #define MPTCP_REINJECT		0x10 /* Did we reinject this segment? */
@@ -640,16 +631,15 @@ struct mp_prio {
 	__u8	addr_id;
 } __attribute__((__packed__));
 
-static inline int mptcp_sub_len_dss(const struct mp_dss *m, const int csum)
+static inline int mptcp_sub_len_dss(const struct mp_dss *m)
 {
-	return 4 + m->A * (4 + m->a * 4) + m->M * (10 + m->m * 4 + csum * 2);
+	return 4 + m->A * (4 + m->a * 4) + m->M * (10 + m->m * 4);
 }
 
 #define MPTCP_SYSCTL	1
 
 extern int sysctl_mptcp_enabled;
 extern int sysctl_mptcp_version;
-extern int sysctl_mptcp_checksum;
 extern int sysctl_mptcp_debug;
 extern int sysctl_mptcp_syn_retries;
 
@@ -693,10 +683,8 @@ enum
 	MPTCP_MIB_MPCAPABLEPASSIVEFALLBACK,/* Server-side fallback during 3-way handshake */
 	MPTCP_MIB_MPCAPABLEACTIVEFALLBACK, /* Client-side fallback during 3-way handshake */
 	MPTCP_MIB_MPCAPABLERETRANSFALLBACK,/* Client-side stopped sending MP_CAPABLE after too many SYN-retransmissions */
-	MPTCP_MIB_CSUMENABLED,		/* Created MPTCP-connection with DSS-checksum enabled */
 	MPTCP_MIB_RETRANSSEGS,		/* Segments retransmitted at the MPTCP-level */
 	MPTCP_MIB_MPFAILRX,		/* Received an MP_FAIL */
-	MPTCP_MIB_CSUMFAIL,		/* Received segment with invalid checksum */
 	MPTCP_MIB_FASTCLOSERX,		/* Recevied a FAST_CLOSE */
 	MPTCP_MIB_FASTCLOSETX,		/* Sent a FAST_CLOSE */
 	MPTCP_MIB_FBACKSUB,		/* Fallback upon ack without data-ack on new subflow */
@@ -921,9 +909,6 @@ static inline bool mptcp_can_sendpage(struct sock *sk)
 {
 	struct mptcp_tcp_sock *mptcp;
 
-	if (tcp_sk(sk)->mpcb->dss_csum)
-		return false;
-
 	mptcp_for_each_sub(tcp_sk(sk)->mpcb, mptcp) {
 		struct sock *sk_it = mptcp_to_sock(mptcp);
 
@@ -1058,7 +1043,6 @@ static inline int is_master_tp(const struct tcp_sock *tp)
 static inline void mptcp_init_mp_opt(struct mptcp_options_received *mopt)
 {
 	mopt->saw_mpc = 0;
-	mopt->dss_csum = 0;
 	mopt->drop_me = 0;
 
 	mopt->is_mp_join = 0;
@@ -1153,9 +1137,6 @@ static inline bool mptcp_can_sg(const struct sock *meta_sk)
 {
 	struct mptcp_tcp_sock *mptcp;
 
-	if (tcp_sk(meta_sk)->mpcb->dss_csum)
-		return false;
-
 	mptcp_for_each_sub(tcp_sk(meta_sk)->mpcb, mptcp) {
 		struct sock *sk = mptcp_to_sock(mptcp);
 
net/mptcp/mptcp_ctrl.c +0 -16
index 916438e31b59..1a04decc9578 100644
@@ -30,7 +30,6 @@
 #include <net/inet_common.h>
 #include <net/inet6_hashtables.h>
 #include <net/ipv6.h>
-#include <net/ip6_checksum.h>
 #include <net/mptcp.h>
 #include <net/mptcp_v4.h>
 #if IS_ENABLED(CONFIG_IPV6)
@@ -67,7 +66,6 @@ int sysctl_mptcp_enabled __read_mostly = 1;
 int sysctl_mptcp_version __read_mostly = 0;
 static int min_mptcp_version;
 static int max_mptcp_version = 1;
-int sysctl_mptcp_checksum __read_mostly = 1;
 int sysctl_mptcp_debug __read_mostly;
 EXPORT_SYMBOL(sysctl_mptcp_debug);
 int sysctl_mptcp_syn_retries __read_mostly = 3;
@@ -132,13 +130,6 @@ static struct ctl_table mptcp_table[] = {
 		.extra1 = &min_mptcp_version,
 		.extra2 = &max_mptcp_version,
 	},
-	{
-		.procname = "mptcp_checksum",
-		.data = &sysctl_mptcp_checksum,
-		.maxlen = sizeof(int),
-		.mode = 0644,
-		.proc_handler = &proc_dointvec
-	},
 	{
 		.procname = "mptcp_debug",
 		.data = &sysctl_mptcp_debug,
@@ -2078,7 +2069,6 @@ static int __mptcp_check_req_master(struct sock *child,
 	child_tp->mptcp->snt_isn = tcp_rsk(req)->snt_isn;
 	child_tp->mptcp->rcv_isn = tcp_rsk(req)->rcv_isn;
 
-	mpcb->dss_csum = mtreq->dss_csum;
 	mpcb->server_side = 1;
 
 	/* Needs to be done here additionally, because when accepting a
@@ -2457,13 +2447,10 @@ void mptcp_reqsk_init(struct request_sock *req, const struct sock *sk,
 		      const struct sk_buff *skb, bool want_cookie)
 {
 	struct mptcp_options_received mopt;
-	struct mptcp_request_sock *mtreq = mptcp_rsk(req);
 
 	mptcp_init_mp_opt(&mopt);
 	tcp_parse_mptcp_options(skb, &mopt);
 
-	mtreq->dss_csum = mopt.dss_csum;
-
 	if (want_cookie) {
 		if (!mptcp_reqsk_new_cookie(req, sk, &mopt, skb))
 			/* No key available - back to regular TCP */
@@ -2500,7 +2487,6 @@ void mptcp_cookies_reqsk_init(struct request_sock *req,
 	inet_rsk(req)->saw_mpc = 1;
 	mtreq->is_sub = 0;
 	inet_rsk(req)->mptcp_rqsk = 1;
-	mtreq->dss_csum = mopt->dss_csum;
 
 out:
 	spin_unlock(&mptcp_tk_hashlock);
@@ -2771,10 +2757,8 @@ static const struct snmp_mib mptcp_snmp_list[] = {
 	SNMP_MIB_ITEM("MPCapableFallbackACK", MPTCP_MIB_MPCAPABLEPASSIVEFALLBACK),
 	SNMP_MIB_ITEM("MPCapableFallbackSYNACK", MPTCP_MIB_MPCAPABLEACTIVEFALLBACK),
 	SNMP_MIB_ITEM("MPCapableRetransFallback", MPTCP_MIB_MPCAPABLERETRANSFALLBACK),
-	SNMP_MIB_ITEM("MPTCPCsumEnabled", MPTCP_MIB_CSUMENABLED),
 	SNMP_MIB_ITEM("MPTCPRetrans", MPTCP_MIB_RETRANSSEGS),
 	SNMP_MIB_ITEM("MPFailRX", MPTCP_MIB_MPFAILRX),
-	SNMP_MIB_ITEM("MPCsumFail", MPTCP_MIB_CSUMFAIL),
 	SNMP_MIB_ITEM("MPFastcloseRX", MPTCP_MIB_FASTCLOSERX),
 	SNMP_MIB_ITEM("MPFastcloseTX", MPTCP_MIB_FASTCLOSETX),
 	SNMP_MIB_ITEM("MPFallbackAckSub", MPTCP_MIB_FBACKSUB),
net/mptcp/mptcp_input.c +5 -179

244 lines not on this page · raw message

net/mptcp/mptcp_output.c +5 -64

147 lines not on this page · raw message

Signature · 1 line
2.16.2
author
#2 lore raw

PATCH v3 2/12 mptcp: Remove some left-over cruft

In reply to Christoph Paasch

The MPTCP-congestion controls have been removed. This here is a left-over.

Signed-off-by Christoph Paasch <cpaasch(a)apple.com>
---
 net/ipv4/Kconfig | 19 -------------------
 1 file changed, 19 deletions(-)
net/ipv4/Kconfig +0 -19
index 3620a29db1ef..32cae39cdff6 100644
@@ -717,21 +717,6 @@ choice
 	config DEFAULT_BBR
 		bool "BBR" if TCP_CONG_BBR=y
 
-	config DEFAULT_LIA
-		bool "Lia" if TCP_CONG_LIA=y
-
-	config DEFAULT_OLIA
-		bool "Olia" if TCP_CONG_OLIA=y
-
-	config DEFAULT_WVEGAS
-		bool "Wvegas" if TCP_CONG_WVEGAS=y
-
-	config DEFAULT_BALIA
-		bool "Balia" if TCP_CONG_BALIA=y
-
-	config DEFAULT_MCTCPDESYNC
-		bool "Mctcpdesync (EXPERIMENTAL)" if TCP_CONG_MCTCPDESYNC=y
-
 	config DEFAULT_RENO
 		bool "Reno"
 endchoice
@@ -752,10 +737,6 @@ config DEFAULT_TCP_CONG
 	default "vegas" if DEFAULT_VEGAS
 	default "westwood" if DEFAULT_WESTWOOD
 	default "veno" if DEFAULT_VENO
-	default "lia" if DEFAULT_LIA
-	default "olia" if DEFAULT_OLIA
-	default "wvegas" if DEFAULT_WVEGAS
-	default "balia" if DEFAULT_BALIA
 	default "reno" if DEFAULT_RENO
 	default "dctcp" if DEFAULT_DCTCP
 	default "cdg" if DEFAULT_CDG
Signature · 1 line
2.16.2
author
#3 lore raw

PATCH v3 3/12 mptcp: Remove RPS-recording for flow-steering in MPTCP

In reply to Christoph Paasch

It's an optimization. We don't need it!

Signed-off-by Christoph Paasch <cpaasch(a)apple.com>
---
 net/ipv4/af_inet.c |  6 ------
 net/ipv4/tcp.c     | 34 ----------------------------------
 2 files changed, 40 deletions(-)
net/ipv4/af_inet.c +0 -6
index 6b510da0086d..e95f6ac342ff 100644
@@ -744,12 +744,6 @@ int inet_accept(struct socket *sock, struct socket *newsock, int flags,
 	sock_rps_record_flow(sk2);
 
 	if (sk2->sk_protocol == IPPROTO_TCP && mptcp(tcp_sk(sk2))) {
-		struct mptcp_tcp_sock *mptcp;
-
-		mptcp_for_each_sub(tcp_sk(sk2)->mpcb, mptcp) {
-			sock_rps_record_flow(mptcp_to_sock(mptcp));
-		}
-
 		if (tcp_sk(sk2)->mpcb->master_sk) {
 			struct sock *sk_it = tcp_sk(sk2)->mpcb->master_sk;
 
net/ipv4/tcp.c +0 -34
index b8a60db7fb7a..213be064adde 100644
@@ -821,16 +821,6 @@ ssize_t tcp_splice_read(struct socket *sock, loff_t *ppos,
 
 	lock_sock(sk);
 
-#ifdef CONFIG_MPTCP
-	if (mptcp(tcp_sk(sk))) {
-		struct mptcp_tcp_sock *mptcp;
-
-		mptcp_for_each_sub(tcp_sk(sk)->mpcb, mptcp) {
-			sock_rps_record_flow(mptcp_to_sock(mptcp));
-		}
-	}
-#endif
-
 	timeo = sock_rcvtimeo(sk, sock->file->f_flags & O_NONBLOCK);
 	while (tss.len) {
 		ret = __tcp_splice_read(sk, &tss);
@@ -995,8 +985,6 @@ ssize_t do_tcp_sendpages(struct sock *sk, struct page *page, int offset,
 	}
 
 	if (mptcp(tp)) {
-		struct mptcp_tcp_sock *mptcp;
-
 		/* We must check this with socket-lock hold because we iterate
 		 * over the subflows.
 		 */
@@ -1009,10 +997,6 @@ ssize_t do_tcp_sendpages(struct sock *sk, struct page *page, int offset,
 			lock_sock(sk);
 			return ret;
 		}
-
-		mptcp_for_each_sub(tp->mpcb, mptcp) {
-			sock_rps_record_flow(mptcp_to_sock(mptcp));
-		}
 	}
 
 	sk_clear_bit(SOCKWQ_ASYNC_NOSPACE, sk);
@@ -1290,14 +1274,6 @@ int tcp_sendmsg_locked(struct sock *sk, struct msghdr *msg, size_t size)
 			goto do_error;
 	}
 
-	if (mptcp(tp)) {
-		struct mptcp_tcp_sock *mptcp;
-
-		mptcp_for_each_sub(tp->mpcb, mptcp) {
-			sock_rps_record_flow(mptcp_to_sock(mptcp));
-		}
-	}
-
 	if (unlikely(tp->repair)) {
 		if (tp->repair_queue == TCP_RECV_QUEUE) {
 			copied = tcp_send_rcvq(sk, msg, size);
@@ -2022,16 +1998,6 @@ int tcp_recvmsg(struct sock *sk, struct msghdr *msg, size_t len, int nonblock,
 
 	lock_sock(sk);
 
-#ifdef CONFIG_MPTCP
-	if (mptcp(tp)) {
-		struct mptcp_tcp_sock *mptcp;
-
-		mptcp_for_each_sub(tp->mpcb, mptcp) {
-			sock_rps_record_flow(mptcp_to_sock(mptcp));
-		}
-	}
-#endif
-
 	err = -ENOTCONN;
 	if (sk->sk_state == TCP_LISTEN)
 		goto out;
Signature · 1 line
2.16.2
author
#4 lore raw

PATCH v3 4/12 mptcp: Remove support for subflow-priorities (aka backup)

In reply to Christoph Paasch
Signed-off-by Christoph Paasch <cpaasch(a)apple.com>

---

Notes:
v2: * Make sure that we detect NETDEV_CHANGE events unsetting the IFF_NOMULTIPATH flag

 include/linux/tcp.h          |   1 -
 include/net/mptcp.h          |  43 +-----------
 include/uapi/linux/if.h      |   1 -
 net/core/dev.c               |   2 +-
 net/mptcp/mptcp_ctrl.c       |   6 --
 net/mptcp/mptcp_fullmesh.c   | 154 +++++++++++--------------------------------
 net/mptcp/mptcp_input.c      |  36 ----------
 net/mptcp/mptcp_ipv4.c       |   5 +-
 net/mptcp/mptcp_ipv6.c       |   5 +-
 net/mptcp/mptcp_ndiffports.c |   4 +-
 net/mptcp/mptcp_output.c     |  27 +-------
 net/mptcp/mptcp_pm.c         |   2 +-
 net/mptcp/mptcp_sched.c      |  29 +-------
 13 files changed, 49 insertions(+), 266 deletions(-)
include/linux/tcp.h +0 -1
index f38bf1e18e58..cc896f842c31 100644
@@ -109,7 +109,6 @@ struct tcp_out_options {
 			__u32	sender_nonce;
 					/* random number of the sender */
 			__u32	token;	/* token for mptcp */
-			u8	low_prio:1;
 		} mp_join_syns;
 	};
 
include/net/mptcp.h +1 -42

117 lines not on this page · raw message

include/uapi/linux/if.h +0 -1
index 7255e08393db..51540df75bf8 100644
@@ -133,7 +133,6 @@ enum net_device_flags {
 #endif /* __UAPI_DEF_IF_NET_DEVICE_FLAGS_LOWER_UP_DORMANT_ECHO */
 
 #define IFF_NOMULTIPATH	0x80000		/* Disable for MPTCP 		*/
-#define IFF_MPBACKUP	0x100000	/* Use as backup path for MPTCP */
 
 #define IFF_VOLATILE	(IFF_LOOPBACK|IFF_POINTOPOINT|IFF_BROADCAST|IFF_ECHO|\
 		IFF_MASTER|IFF_SLAVE|IFF_RUNNING|IFF_LOWER_UP|IFF_DORMANT)
net/core/dev.c +1 -1
index 95cff2426ea1..c5cafaac7f58 100644
@@ -7487,7 +7487,7 @@ int __dev_change_flags(struct net_device *dev, unsigned int flags)
 
 	dev->flags = (flags & (IFF_DEBUG | IFF_NOTRAILERS | IFF_NOARP |
 			       IFF_DYNAMIC | IFF_MULTICAST | IFF_PORTSEL |
-			       IFF_AUTOMEDIA | IFF_NOMULTIPATH | IFF_MPBACKUP)) |
+			       IFF_AUTOMEDIA | IFF_NOMULTIPATH)) |
 		     (dev->flags & (IFF_UP | IFF_VOLATILE | IFF_PROMISC |
 				    IFF_ALLMULTI));
 
net/mptcp/mptcp_ctrl.c +0 -6
index 1a04decc9578..d367bbaa03a5 100644
@@ -2217,11 +2217,6 @@ struct sock *mptcp_check_req_child(struct sock *meta_sk,
 		goto teardown;
 	}
 
-	/* The child is a clone of the meta socket, we must now reset
-	 * some of the fields
-	 */
-	child_tp->mptcp->rcv_low_prio = mtreq->rcv_low_prio;
-
 	/* We should allow proper increase of the snd/rcv-buffers. Thus, we
 	 * use the original values instead of the bloated up ones from the
 	 * clone.
@@ -2437,7 +2432,6 @@ void mptcp_join_reqsk_init(const struct mptcp_cb *mpcb,
 	mtreq->mptcp_hash_tmac = *(u64 *)mptcp_hash_mac;
 
 	mtreq->rem_id = mopt.rem_id;
-	mtreq->rcv_low_prio = mopt.low_prio;
 	inet_rsk(req)->saw_mpc = 1;
 
 	MPTCP_INC_STATS(sock_net(mpcb->meta_sk), MPTCP_MIB_JOINSYNRX);
net/mptcp/mptcp_fullmesh.c +38 -116

325 lines not on this page · raw message

net/mptcp/mptcp_input.c +0 -36
index 9a4d1609b3b9..652da0439644 100644
@@ -1503,14 +1503,12 @@ void mptcp_parse_options(const uint8_t *ptr, int opsize,
 		case MPTCP_SUB_LEN_JOIN_SYN:
 			mopt->is_mp_join = 1;
 			mopt->saw_mpc = 1;
-			mopt->low_prio = mpjoin->b;
 			mopt->rem_id = mpjoin->addr_id;
 			mopt->mptcp_rem_token = mpjoin->u.syn.token;
 			mopt->mptcp_recv_nonce = mpjoin->u.syn.nonce;
 			break;
 		case MPTCP_SUB_LEN_JOIN_SYNACK:
 			mopt->saw_mpc = 1;
-			mopt->low_prio = mpjoin->b;
 			mopt->rem_id = mpjoin->addr_id;
 			mopt->mptcp_recv_tmac = mpjoin->u.synack.mac;
 			mopt->mptcp_recv_nonce = mpjoin->u.synack.nonce;
@@ -1613,26 +1611,6 @@ void mptcp_parse_options(const uint8_t *ptr, int opsize,
 		mopt->saw_rem_addr = 1;
 		mopt->rem_addr_ptr = ptr;
 		break;
-	case MPTCP_SUB_PRIO:
-	{
-		const struct mp_prio *mpprio = (struct mp_prio *)ptr;
-
-		if (opsize != MPTCP_SUB_LEN_PRIO &&
-		    opsize != MPTCP_SUB_LEN_PRIO_ADDR) {
-			mptcp_debug("%s: mp_prio: bad option size %d\n",
-				    __func__, opsize);
-			break;
-		}
-
-		mopt->saw_low_prio = 1;
-		mopt->low_prio = mpprio->b;
-
-		if (opsize == MPTCP_SUB_LEN_PRIO_ADDR) {
-			mopt->saw_low_prio = 2;
-			mopt->prio_addr_id = mpprio->addr_id;
-		}
-		break;
-	}
 	case MPTCP_SUB_FAIL:
 		if (opsize != MPTCP_SUB_LEN_FAIL) {
 			mptcp_debug("%s: mp_fail: bad option size %d\n",
@@ -1963,19 +1941,6 @@ bool mptcp_handle_options(struct sock *sk, const struct tcphdr *th,
 		mopt->more_rem_addr = 0;
 		mopt->saw_rem_addr = 0;
 	}
-	if (mopt->saw_low_prio) {
-		if (mopt->saw_low_prio == 1) {
-			tp->mptcp->rcv_low_prio = mopt->low_prio;
-		} else {
-			struct mptcp_tcp_sock *mptcp;
-
-			mptcp_for_each_sub(tp->mpcb, mptcp) {
-				if (mptcp->rem_id == mopt->prio_addr_id)
-					mptcp->rcv_low_prio = mopt->low_prio;
-			}
-		}
-		mopt->saw_low_prio = 0;
-	}
 
 	mptcp_data_ack(sk, skb);
 
@@ -2077,7 +2042,6 @@ int mptcp_rcv_synsent_state_process(struct sock *sk, struct sock **skptr,
 		 * until the 4th ack arrives.
 		 */
 		tp->mptcp->pre_established = 1;
-		tp->mptcp->rcv_low_prio = tp->mptcp->rx_opt.low_prio;
 
 		mptcp_hmac_sha1((u8 *)&mpcb->mptcp_loc_key,
 				(u8 *)&mpcb->mptcp_rem_key,
net/mptcp/mptcp_ipv4.c +1 -4
index c808d10205f5..00c54f421ba7 100644
@@ -104,7 +104,6 @@ static int mptcp_v4_join_init_req(struct request_sock *req, const struct sock *m
 	const struct mptcp_cb *mpcb = tcp_sk(meta_sk)->mpcb;
 	union inet_addr addr;
 	int loc_id;
-	bool low_prio = false;
 
 	/* We need to do this as early as possible. Because, if we fail later
 	 * (e.g., get_local_id), then reqsk_free tries to remove the
@@ -120,11 +119,10 @@ static int mptcp_v4_join_init_req(struct request_sock *req, const struct sock *m
 						    tcp_hdr(skb)->source,
 						    tcp_hdr(skb)->dest);
 	addr.ip = inet_rsk(req)->ir_loc_addr;
-	loc_id = mpcb->pm_ops->get_local_id(AF_INET, &addr, sock_net(meta_sk), &low_prio);
+	loc_id = mpcb->pm_ops->get_local_id(AF_INET, &addr, sock_net(meta_sk));
 	if (loc_id == -1)
 		return -1;
 	mtreq->loc_id = loc_id;
-	mtreq->low_prio = low_prio;
 
 	mptcp_join_reqsk_init(mpcb, req, skb);
 
@@ -292,7 +290,6 @@ int mptcp_init4_subsockets(struct sock *meta_sk, const struct mptcp_loc4 *loc,
 	}
 
 	tp->mptcp->slave_sk = 1;
-	tp->mptcp->low_prio = loc->low_prio;
 
 	/* Initializing the timer for an MPTCP subflow */
 	timer_setup(&tp->mptcp->mptcp_ack_timer, mptcp_ack_handler, 0);
net/mptcp/mptcp_ipv6.c +1 -4
index 2d730379d5ed..7249d171c9ef 100644
@@ -133,7 +133,6 @@ static int mptcp_v6_join_init_req(struct request_sock *req, const struct sock *m
 	const struct mptcp_cb *mpcb = tcp_sk(meta_sk)->mpcb;
 	union inet_addr addr;
 	int loc_id;
-	bool low_prio = false;
 
 	/* We need to do this as early as possible. Because, if we fail later
 	 * (e.g., get_local_id), then reqsk_free tries to remove the
@@ -149,11 +148,10 @@ static int mptcp_v6_join_init_req(struct request_sock *req, const struct sock *m
 						    tcp_hdr(skb)->source,
 						    tcp_hdr(skb)->dest);
 	addr.in6 = inet_rsk(req)->ir_v6_loc_addr;
-	loc_id = mpcb->pm_ops->get_local_id(AF_INET6, &addr, sock_net(meta_sk), &low_prio);
+	loc_id = mpcb->pm_ops->get_local_id(AF_INET6, &addr, sock_net(meta_sk));
 	if (loc_id == -1)
 		return -1;
 	mtreq->loc_id = loc_id;
-	mtreq->low_prio = low_prio;
 
 	mptcp_join_reqsk_init(mpcb, req, skb);
 
@@ -319,7 +317,6 @@ int mptcp_init6_subsockets(struct sock *meta_sk, const struct mptcp_loc6 *loc,
 	}
 
 	tp->mptcp->slave_sk = 1;
-	tp->mptcp->low_prio = loc->low_prio;
 
 	/* Initializing the timer for an MPTCP subflow */
 	timer_setup(&tp->mptcp->mptcp_ack_timer, mptcp_ack_handler, 0);
net/mptcp/mptcp_ndiffports.c +1 -3
index 58acf50df752..9faef10d8b64 100644
@@ -61,7 +61,6 @@ static void create_subflow_worker(struct work_struct *work)
 
 			loc.addr.s_addr = inet_sk(meta_sk)->inet_saddr;
 			loc.loc4_id = 0;
-			loc.low_prio = 0;
 			if (mpcb->master_sk)
 				loc.if_idx = mpcb->master_sk->sk_bound_dev_if;
 			else
@@ -79,7 +78,6 @@ static void create_subflow_worker(struct work_struct *work)
 
 			loc.addr = inet6_sk(meta_sk)->saddr;
 			loc.loc6_id = 0;
-			loc.low_prio = 0;
 			if (mpcb->master_sk)
 				loc.if_idx = mpcb->master_sk->sk_bound_dev_if;
 			else
@@ -128,7 +126,7 @@ static void ndiffports_create_subflows(struct sock *meta_sk)
 }
 
 static int ndiffports_get_local_id(sa_family_t family, union inet_addr *addr,
-				   struct net *net, bool *low_prio)
+				   struct net *net)
 {
 	return 0;
 }
net/mptcp/mptcp_output.c +2 -25
index f9bce20259e9..af695fafaab2 100644
@@ -960,7 +960,6 @@ void mptcp_syn_options(const struct sock *sk, struct tcp_out_options *opts,
 		opts->mptcp_options |= OPTION_MP_JOIN | OPTION_TYPE_SYN;
 		*remaining -= MPTCP_SUB_LEN_JOIN_SYN_ALIGN;
 		opts->mp_join_syns.token = mpcb->mptcp_rem_token;
-		opts->mp_join_syns.low_prio  = tp->mptcp->low_prio;
 		opts->addr_id = tp->mptcp->loc_id;
 		opts->mp_join_syns.sender_nonce = tp->mptcp->mptcp_loc_nonce;
 	}
@@ -984,7 +983,6 @@ void mptcp_synack_options(struct request_sock *req,
 		opts->mp_join_syns.sender_truncated_mac =
 				mtreq->mptcp_hash_tmac;
 		opts->mp_join_syns.sender_nonce = mtreq->mptcp_loc_nonce;
-		opts->mp_join_syns.low_prio = mtreq->low_prio;
 		opts->addr_id = mtreq->loc_id;
 		*remaining -= MPTCP_SUB_LEN_JOIN_SYNACK_ALIGN;
 	}
@@ -1085,15 +1083,6 @@ void mptcp_established_options(struct sock *sk, struct sk_buff *skb,
 	    mpcb->mptcp_ver < MPTCP_VERSION_1)
 		mpcb->pm_ops->addr_signal(sk, size, opts, skb);
 
-	if (unlikely(tp->mptcp->send_mp_prio) &&
-	    MAX_TCP_OPTION_SPACE - *size >= MPTCP_SUB_LEN_PRIO_ALIGN) {
-		opts->options |= OPTION_MPTCP;
-		opts->mptcp_options |= OPTION_MP_PRIO;
-		if (skb)
-			tp->mptcp->send_mp_prio = 0;
-		*size += MPTCP_SUB_LEN_PRIO_ALIGN;
-	}
-
 	return;
 }
 
@@ -1149,7 +1138,7 @@ void mptcp_options_write(__be32 *ptr, struct tcp_sock *tp,
 			mpj->len = MPTCP_SUB_LEN_JOIN_SYN;
 			mpj->u.syn.token = opts->mp_join_syns.token;
 			mpj->u.syn.nonce = opts->mp_join_syns.sender_nonce;
-			mpj->b = opts->mp_join_syns.low_prio;
+			mpj->b = 0;
 			mpj->addr_id = opts->addr_id;
 			ptr += MPTCP_SUB_LEN_JOIN_SYN_ALIGN >> 2;
 		} else if (OPTION_TYPE_SYNACK & opts->mptcp_options) {
@@ -1157,7 +1146,7 @@ void mptcp_options_write(__be32 *ptr, struct tcp_sock *tp,
 			mpj->u.synack.mac =
 				opts->mp_join_syns.sender_truncated_mac;
 			mpj->u.synack.nonce = opts->mp_join_syns.sender_nonce;
-			mpj->b = opts->mp_join_syns.low_prio;
+			mpj->b = 0;
 			mpj->addr_id = opts->addr_id;
 			ptr += MPTCP_SUB_LEN_JOIN_SYNACK_ALIGN >> 2;
 		} else if (OPTION_TYPE_ACK & opts->mptcp_options) {
@@ -1252,18 +1241,6 @@ void mptcp_options_write(__be32 *ptr, struct tcp_sock *tp,
 		else
 			ptr += mptcp_write_dss_data_seq(tp, skb, ptr);
 	}
-	if (unlikely(OPTION_MP_PRIO & opts->mptcp_options)) {
-		struct mp_prio *mpprio = (struct mp_prio *)ptr;
-
-		mpprio->kind = TCPOPT_MPTCP;
-		mpprio->len = MPTCP_SUB_LEN_PRIO;
-		mpprio->sub = MPTCP_SUB_PRIO;
-		mpprio->rsv = 0;
-		mpprio->b = tp->mptcp->low_prio;
-		mpprio->addr_id = TCPOPT_NOP;
-
-		ptr += MPTCP_SUB_LEN_PRIO_ALIGN >> 2;
-	}
 }
 
 /* Sends the datafin */
net/mptcp/mptcp_pm.c +1 -1
index 08e3f05a9e65..fbdf8eb58d17 100644
@@ -35,7 +35,7 @@ static DEFINE_SPINLOCK(mptcp_pm_list_lock);
 static LIST_HEAD(mptcp_pm_list);
 
 static int mptcp_default_id(sa_family_t family, union inet_addr *addr,
-			    struct net *net, bool *low_prio)
+			    struct net *net)
 {
 	return 0;
 }
net/mptcp/mptcp_sched.c +3 -26
index 3662bdf97c0d..7d4804f8ba91 100644
@@ -113,22 +113,11 @@ static int mptcp_dont_reinject_skb(const struct tcp_sock *tp, const struct sk_bu
 		mptcp_pi_to_flag(tp->mptcp->path_index) & TCP_SKB_CB(skb)->path_mask;
 }
 
-static bool subflow_is_backup(const struct tcp_sock *tp)
-{
-	return tp->mptcp->rcv_low_prio || tp->mptcp->low_prio;
-}
-
-static bool subflow_is_active(const struct tcp_sock *tp)
-{
-	return !tp->mptcp->rcv_low_prio && !tp->mptcp->low_prio;
-}
-
 /* Generic function to iterate over used and unused subflows and to select the
  * best one
  */
 static struct sock
 *get_subflow_from_selectors(struct mptcp_cb *mpcb, struct sk_buff *skb,
-			    bool (*selector)(const struct tcp_sock *),
 			    bool zero_wnd_test, bool *force)
 {
 	struct sock *bestsk = NULL;
@@ -142,10 +131,6 @@ static struct sock
 		struct tcp_sock *tp = tcp_sk(sk);
 		bool unused = false;
 
-		/* First, we choose only the wanted sks */
-		if (!(*selector)(tp))
-			continue;
-
 		if (!mptcp_dont_reinject_skb(tp, skb))
 			unused = true;
 		else if (found_unused)
@@ -206,8 +191,6 @@ static struct sock
  * a given MSS. If all subflows are found to be busy, NULL is returned
  * The flow is selected based on the shortest RTT.
  * If all paths have full cong windows, we simply return NULL.
- *
- * Additionally, this function is aware of the backup-subflows.
  */
 static struct sock *get_available_subflow(struct sock *meta_sk,
 					  struct sk_buff *skb,
@@ -233,22 +216,16 @@ static struct sock *get_available_subflow(struct sock *meta_sk,
 
 	/* Find the best subflow */
 restart:
-	sk = get_subflow_from_selectors(mpcb, skb, &subflow_is_active,
-					zero_wnd_test, &force);
+	sk = get_subflow_from_selectors(mpcb, skb, zero_wnd_test, &force);
 	if (force)
 		/* one unused active sk or one NULL sk when there is at least
 		 * one temporally unavailable unused active sk
 		 */
 		return sk;
 
-	sk = get_subflow_from_selectors(mpcb, skb, &subflow_is_backup,
-					zero_wnd_test, &force);
 	if (!force && skb) {
-		/* one used backup sk or one NULL sk where there is no one
-		 * temporally unavailable unused backup sk
-		 *
-		 * the skb passed through all the available active and backups
-		 * sks, so clean the path mask
+		/* the skb passed through all the available subflows, so clean
+		 * the path mask.
 		 */
 		TCP_SKB_CB(skb)->path_mask = 0;
 
Signature · 1 line
2.16.2
author
#5 lore raw

PATCH v3 5/12 mptcp: Remove support for sendpage()

In reply to Christoph Paasch
Signed-off-by Christoph Paasch <cpaasch(a)apple.com>
---
 include/net/mptcp.h | 18 ------------------
 net/ipv4/tcp.c      | 21 ++-------------------
 2 files changed, 2 insertions(+), 37 deletions(-)
include/net/mptcp.h +0 -18
index e01a642466c0..9e7dd681fb71 100644
@@ -868,20 +868,6 @@ struct request_sock *rev_mptcp_rsk(const struct mptcp_request_sock *req)
 	return (struct request_sock *)req;
 }
 
-static inline bool mptcp_can_sendpage(struct sock *sk)
-{
-	struct mptcp_tcp_sock *mptcp;
-
-	mptcp_for_each_sub(tcp_sk(sk)->mpcb, mptcp) {
-		struct sock *sk_it = mptcp_to_sock(mptcp);
-
-		if (!(sk_it->sk_route_caps & NETIF_F_SG))
-			return false;
-	}
-
-	return true;
-}
-
 static inline void mptcp_push_pending_frames(struct sock *meta_sk)
 {
 	/* We check packets out and send-head here. TCP only checks the
@@ -1366,10 +1352,6 @@ static inline int mptcp_rcv_synsent_state_process(struct sock *sk,
 {
 	return 0;
 }
-static inline bool mptcp_can_sendpage(struct sock *sk)
-{
-	return false;
-}
 static inline int mptcp_init_tw_sock(struct sock *sk,
 				     struct tcp_timewait_sock *tw)
 {
net/ipv4/tcp.c +2 -19
index 213be064adde..87f7616aaf93 100644
@@ -977,28 +977,12 @@ ssize_t do_tcp_sendpages(struct sock *sk, struct page *page, int offset,
 	 * is fully established.
 	 */
 	if (((1 << sk->sk_state) & ~(TCPF_ESTABLISHED | TCPF_CLOSE_WAIT)) &&
-	    !tcp_passive_fastopen(mptcp(tp) && tp->mpcb->master_sk ?
-				  tp->mpcb->master_sk : sk)) {
+	    !tcp_passive_fastopen(sk)) {
 		err = sk_stream_wait_connect(sk, &timeo);
 		if (err != 0)
 			goto out_err;
 	}
 
-	if (mptcp(tp)) {
-		/* We must check this with socket-lock hold because we iterate
-		 * over the subflows.
-		 */
-		if (!mptcp_can_sendpage(sk)) {
-			ssize_t ret;
-
-			release_sock(sk);
-			ret = sock_no_sendpage(sk->sk_socket, page, offset,
-					       size, flags);
-			lock_sock(sk);
-			return ret;
-		}
-	}
-
 	sk_clear_bit(SOCKWQ_ASYNC_NOSPACE, sk);
 
 	mss_now = tcp_send_mss(sk, &size_goal, flags);
@@ -1117,8 +1101,7 @@ EXPORT_SYMBOL_GPL(do_tcp_sendpages);
 int tcp_sendpage_locked(struct sock *sk, struct page *page, int offset,
 			size_t size, int flags)
 {
-	/* If MPTCP is enabled, we check it later after establishment */
-	if (!mptcp(tcp_sk(sk)) && !(sk->sk_route_caps & NETIF_F_SG))
+	if (!(sk->sk_route_caps & NETIF_F_SG) || mptcp(tcp_sk(sk)))
 		return sock_no_sendpage_locked(sk, page, offset, size, flags);
 
 	tcp_rate_check_app_limited(sk);  /* is sending application-limited? */
Signature · 1 line
2.16.2
author
#6 lore raw

PATCH v3 6/12 mptcp: Remove reinjection of data across subflows

In reply to Christoph Paasch

It's an "optimisation" that we don't need in net-next.

Signed-off-by Christoph Paasch <cpaasch(a)apple.com>
---
 include/linux/tcp.h        |   3 -
 include/net/mptcp.h        |   9 --
 net/mptcp/mptcp_ctrl.c     |  10 +-
 net/mptcp/mptcp_fullmesh.c |  12 --
 net/mptcp/mptcp_input.c    |  20 +--
 net/mptcp/mptcp_output.c   | 358 +--------------------------------------------
 net/mptcp/mptcp_sched.c    |  32 ++--
 7 files changed, 21 insertions(+), 423 deletions(-)
include/linux/tcp.h +0 -3
index cc896f842c31..6dd9c45c52d6 100644
@@ -470,9 +470,6 @@ struct tcp_sock {
 		request_mptcp:1, /* Did we send out an MP_CAPABLE?
 				  * (this speeds up mptcp_doit() in tcp_recvmsg)
 				  */
-		pf:1, /* Potentially Failed state: when this flag is set, we
-		       * stop using the subflow
-		       */
 		mp_killed:1, /* Killed with a tcp_done in mptcp? */
 		was_meta_sk:1,	/* This was a meta sk (in case of reuse) */
 		is_master_sk:1,
include/net/mptcp.h +0 -9
index 9e7dd681fb71..5e3fc3e7adba 100644
@@ -265,12 +265,6 @@ struct mptcp_cb {
 	u8 mptcp_sched[MPTCP_SCHED_DATA_SIZE] __aligned(8);
 	struct mptcp_sched_ops *sched_ops;
 
-	struct sk_buff_head reinject_queue;
-	/* First cache-line boundary is here minus 8 bytes. But from the
-	 * reinject-queue only the next and prev pointers are regularly
-	 * accessed. Thus, the whole data-path is on a single cache-line.
-	 */
-
 	u64	infinite_rcv_seq;
 
 	/***** Start of fields, used for connection closure */
@@ -421,7 +415,6 @@ extern bool mptcp_init_failed;
 #define MPTCPHDR_SEQ64_OFO	0x20 /* Is it not in our circular array? */
 /* MPTCP flags: TX only */
 #define MPTCPHDR_INF		0x08
-#define MPTCP_REINJECT		0x10 /* Did we reinject this segment? */
 
 struct mptcp_option {
 	__u8	kind;
@@ -719,7 +712,6 @@ int mptcp_add_sock(struct sock *meta_sk, struct sock *sk, u8 loc_id, u8 rem_id,
 		   gfp_t flags);
 void mptcp_del_sock(struct sock *sk);
 void mptcp_update_metasocket(const struct sock *meta_sk);
-void mptcp_reinject_data(struct sock *orig_sk, int clone_it);
 void mptcp_update_sndbuf(const struct tcp_sock *tp);
 void mptcp_send_fin(struct sock *meta_sk);
 void mptcp_send_active_reset(struct sock *meta_sk, gfp_t priority);
@@ -1256,7 +1248,6 @@ static inline int is_master_tp(const struct tcp_sock *tp)
 }
 static inline void mptcp_del_sock(const struct sock *sk) {}
 static inline void mptcp_update_metasocket(const struct sock *meta_sk) {}
-static inline void mptcp_reinject_data(struct sock *orig_sk, int clone_it) {}
 static inline void mptcp_update_sndbuf(const struct tcp_sock *tp) {}
 static inline void mptcp_clean_rtx_infinite(const struct sk_buff *skb,
 					    const struct sock *sk) {}
net/mptcp/mptcp_ctrl.c +1 -9
index d367bbaa03a5..4d11c9f6ceb6 100644
@@ -561,7 +561,7 @@ struct sock *mptcp_select_ack_sock(const struct sock *meta_sk)
 		struct tcp_sock *tp = tcp_sk(sk);
 		u32 elapsed;
 
-		if (!mptcp_sk_can_send_ack(sk) || tp->pf)
+		if (!mptcp_sk_can_send_ack(sk) || inet_csk(sk)->icsk_ca_state == TCP_CA_Loss)
 			continue;
 
 		elapsed = keepalive_time_elapsed(tp);
@@ -693,8 +693,6 @@ void mptcp_destroy_sock(struct sock *sk)
 		struct mptcp_tcp_sock *mptcp;
 		struct hlist_node *tmp;
 
-		__skb_queue_purge(&tcp_sk(sk)->mpcb->reinject_queue);
-
 		/* We have to close all remaining subflows. Normally, they
 		 * should all be about to get closed. But, if the kernel is
 		 * forcing a closure (e.g., tcp_write_err), the subflows might
@@ -1230,7 +1228,6 @@ static int mptcp_alloc_mpcb(struct sock *meta_sk, __u64 remote_key,
 	meta_tp->was_meta_sk = 0;
 
 	/* Initialize the queues */
-	skb_queue_head_init(&mpcb->reinject_queue);
 	master_tp->out_of_order_queue = RB_ROOT;
 	master_sk->tcp_rtx_queue = RB_ROOT;
 	INIT_LIST_HEAD(&master_tp->tsq_node);
@@ -1461,9 +1458,6 @@ void mptcp_del_sock(struct sock *sk)
 	tp->mptcp->attached = 0;
 	mpcb->path_index_bits &= ~(1 << tp->mptcp->path_index);
 
-	if (!tcp_write_queue_empty(sk) || !tcp_rtx_queue_empty(sk))
-		mptcp_reinject_data(sk, 0);
-
 	if (is_master_tp(tp)) {
 		struct sock *meta_sk = mptcp_meta_sk(sk);
 		struct tcp_sock *meta_tp = tcp_sk(meta_sk);
@@ -1921,8 +1915,6 @@ void mptcp_disconnect(struct sock *sk)
 	struct tcp_sock *tp = tcp_sk(sk);
 	struct hlist_node *tmp;
 
-	__skb_queue_purge(&tp->mpcb->reinject_queue);
-
 	if (tp->inside_tk_table)
 		mptcp_hash_remove_bh(tp);
 
net/mptcp/mptcp_fullmesh.c +0 -12
index 4dad457246b1..2446f75b0b2c 100644
@@ -923,12 +923,6 @@ static void mptcp_address_worker(struct work_struct *work)
 					    !ipv6_addr_equal(&inet6_sk(sk)->saddr, &event->addr.in6))
 						continue;
 
-					/* Reinject, so that pf = 1 and so we
-					 * won't select this one as the
-					 * ack-sock.
-					 */
-					mptcp_reinject_data(sk, 0);
-
 					/* We announce the removal of this id */
 					announce_remove_addr(tcp_sk(sk)->mptcp->loc_id, meta_sk);
 
@@ -1440,12 +1434,6 @@ static void full_mesh_release_sock(struct sock *meta_sk)
 		}
 
 		if (shall_remove) {
-			/* Reinject, so that pf = 1 and so we
-			 * won't select this one as the
-			 * ack-sock.
-			 */
-			mptcp_reinject_data(sk, 0);
-
 			announce_remove_addr(tcp_sk(sk)->mptcp->loc_id,
 					     meta_sk);
 
net/mptcp/mptcp_input.c +1 -19
index 652da0439644..f247e9ef2853 100644
@@ -92,7 +92,7 @@ static inline int mptcp_tso_acked_reinject(const struct sock *meta_sk,
 /* Cleans the meta-socket retransmission queue and the reinject-queue. */
 static void mptcp_clean_rtx_queue(struct sock *meta_sk, u32 prior_snd_una)
 {
-	struct sk_buff *skb, *tmp, *next;
+	struct sk_buff *skb, *next;
 	struct tcp_sock *meta_tp = tcp_sk(meta_sk);
 	struct mptcp_cb *mpcb = meta_tp->mpcb;
 	bool acked = false;
@@ -149,20 +149,6 @@ static void mptcp_clean_rtx_queue(struct sock *meta_sk, u32 prior_snd_una)
 		}
 		sk_wmem_free_skb(meta_sk, skb);
 	}
-	/* Remove acknowledged data from the reinject queue */
-	skb_queue_walk_safe(&mpcb->reinject_queue, skb, tmp) {
-		if (before(meta_tp->snd_una, TCP_SKB_CB(skb)->end_seq)) {
-			if (tcp_skb_pcount(skb) == 1 ||
-			    !after(meta_tp->snd_una, TCP_SKB_CB(skb)->seq))
-				break;
-
-			mptcp_tso_acked_reinject(meta_sk, skb);
-			break;
-		}
-
-		__skb_unlink(skb, &mpcb->reinject_queue);
-		__kfree_skb(skb);
-	}
 
 	if (likely(between(meta_tp->snd_up, prior_snd_una, meta_tp->snd_una)))
 		meta_tp->snd_up = meta_tp->snd_una;
@@ -1242,9 +1228,6 @@ static void mptcp_data_ack(struct sock *sk, const struct sk_buff *skb)
 	u32 nwin, data_ack, data_seq;
 	u16 data_len = 0;
 
-	/* A valid packet came in - subflow is operational again */
-	tp->pf = 0;
-
 	/* Even if there is no data-ack, we stop retransmitting.
 	 * Except if this is a SYN/ACK. Then it is just a retransmission
 	 */
@@ -1400,7 +1383,6 @@ static void mptcp_send_reset_rem_id(const struct mptcp_cb *mpcb, u8 rem_id)
 		struct sock *sk_it = mptcp_to_sock(mptcp);
 
 		if (tcp_sk(sk_it)->mptcp->rem_id == rem_id) {
-			mptcp_reinject_data(sk_it, 0);
 			mptcp_send_reset(sk_it);
 		}
 	}
net/mptcp/mptcp_output.c +6 -352

413 lines not on this page · raw message

net/mptcp/mptcp_sched.c +13 -19
index 7d4804f8ba91..7f19b4b4bea0 100644
@@ -29,7 +29,7 @@ static bool mptcp_is_def_unavailable(struct sock *sk)
 	if (tp->mptcp->pre_established)
 		return true;
 
-	if (tp->pf)
+	if (inet_csk(sk)->icsk_ca_state == TCP_CA_Loss)
 		return true;
 
 	return false;
@@ -333,25 +333,19 @@ static struct sk_buff *__mptcp_next_segment(struct sock *meta_sk, int *reinject)
 	if (mpcb->infinite_mapping_snd || mpcb->send_infinite_mapping)
 		return tcp_send_head(meta_sk);
 
-	skb = skb_peek(&mpcb->reinject_queue);
+	skb = tcp_send_head(meta_sk);
 
-	if (skb) {
-		*reinject = 1;
-	} else {
-		skb = tcp_send_head(meta_sk);
-
-		if (!skb && meta_sk->sk_socket &&
-		    test_bit(SOCK_NOSPACE, &meta_sk->sk_socket->flags) &&
-		    sk_stream_wspace(meta_sk) < sk_stream_min_wspace(meta_sk)) {
-			struct sock *subsk = get_available_subflow(meta_sk, NULL,
-								   false);
-			if (!subsk)
-				return NULL;
-
-			skb = mptcp_rcv_buf_optimization(subsk, 0);
-			if (skb)
-				*reinject = -1;
-		}
+	if (!skb && meta_sk->sk_socket &&
+	    test_bit(SOCK_NOSPACE, &meta_sk->sk_socket->flags) &&
+	    sk_stream_wspace(meta_sk) < sk_stream_min_wspace(meta_sk)) {
+		struct sock *subsk = get_available_subflow(meta_sk, NULL,
+							   false);
+		if (!subsk)
+			return NULL;
+
+		skb = mptcp_rcv_buf_optimization(subsk, 0);
+		if (skb)
+			*reinject = -1;
 	}
 	return skb;
 }
Signature · 1 line
2.16.2
author
#7 lore raw

PATCH v3 7/12 mptcp: Remove support for TCP SYN-cookies

In reply to Christoph Paasch
Signed-off-by Christoph Paasch <cpaasch(a)apple.com>

---

Notes:
v3: * More cleanup removing further cookie-things

 include/net/mptcp.h      | 14 ++------
 include/net/tcp.h        | 13 +++----
 net/ipv4/syncookies.c    | 47 ++++---------------------
 net/ipv4/tcp_input.c     |  2 +-
 net/ipv4/tcp_minisocks.c |  2 +-
 net/ipv6/syncookies.c    | 32 ++++-------------
 net/mptcp/mptcp_ctrl.c   | 92 +++---------------------------------------------
 net/mptcp/mptcp_ipv4.c   | 29 ++++-----------
 net/mptcp/mptcp_ipv6.c   | 29 ++++-----------
 9 files changed, 40 insertions(+), 220 deletions(-)
include/net/mptcp.h +3 -11
index 5e3fc3e7adba..32b1fecb70bc 100644
@@ -739,8 +739,7 @@ int mptcp_create_master_sk(struct sock *meta_sk, __u64 remote_key,
 			   __u8 mptcp_ver, u32 window);
 int mptcp_check_req_fastopen(struct sock *child, struct request_sock *req);
 int mptcp_check_req_master(struct sock *sk, struct sock *child,
-			   struct request_sock *req, const struct sk_buff *skb,
-			   int drop);
+			   struct request_sock *req, const struct sk_buff *skb);
 struct sock *mptcp_check_req_child(struct sock *meta_sk,
 				   struct sock *child,
 				   struct request_sock *req,
@@ -802,14 +801,11 @@ void mptcp_join_reqsk_init(const struct mptcp_cb *mpcb,
 			   const struct request_sock *req,
 			   struct sk_buff *skb);
 void mptcp_reqsk_init(struct request_sock *req, const struct sock *sk,
-		      const struct sk_buff *skb, bool want_cookie);
+		      const struct sk_buff *skb);
 int mptcp_conn_request(struct sock *sk, struct sk_buff *skb);
 void mptcp_enable_sock(struct sock *sk);
 void mptcp_disable_sock(struct sock *sk);
 void mptcp_disable_static_key(void);
-void mptcp_cookies_reqsk_init(struct request_sock *req,
-			      struct mptcp_options_received *mopt,
-			      struct sk_buff *skb);
 void mptcp_sock_destruct(struct sock *sk);
 int mptcp_finish_handshake(struct sock *child, struct sk_buff *skb);
 int mptcp_get_info(const struct sock *meta_sk, char __user *optval, int optlen);
@@ -1285,8 +1281,7 @@ static inline int mptcp_check_req_fastopen(struct sock *child,
 static inline int mptcp_check_req_master(const struct sock *sk,
 					 const struct sock *child,
 					 const struct request_sock *req,
-					 const struct sk_buff *skb,
-					 int drop)
+					 const struct sk_buff *skb)
 {
 	return 1;
 }
@@ -1357,9 +1352,6 @@ static inline void mptcp_remove_shortcuts(const struct mptcp_cb *mpcb,
 					  const struct sk_buff *skb) {}
 static inline void mptcp_init_tcp_sock(struct sock *sk) {}
 static inline void mptcp_disable_static_key(void) {}
-static inline void mptcp_cookies_reqsk_init(struct request_sock *req,
-					    struct mptcp_options_received *mopt,
-					    struct sk_buff *skb) {}
 static inline void mptcp_fin(struct sock *meta_sk) {}
 static inline bool mptcp_can_new_subflow(const struct sock *meta_sk)
 {
include/net/tcp.h +5 -8
index c4a9297da7a0..f206ea426018 100644
@@ -644,8 +644,7 @@ static inline u32 tcp_cookie_time(void)
 
 u32 __cookie_v4_init_sequence(const struct iphdr *iph, const struct tcphdr *th,
 			      u16 *mssp);
-__u32 cookie_v4_init_sequence(struct request_sock *req, const struct sock *sk,
-			      const struct sk_buff *skb, __u16 *mss);
+__u32 cookie_v4_init_sequence(const struct sk_buff *skb, __u16 *mss);
 u64 cookie_init_timestamp(struct request_sock *req);
 bool cookie_timestamp_decode(const struct net *net,
 			     struct tcp_options_received *opt);
@@ -659,8 +658,7 @@ struct sock *cookie_v6_check(struct sock *sk, struct sk_buff *skb);
 
 u32 __cookie_v6_init_sequence(const struct ipv6hdr *iph,
 			      const struct tcphdr *th, u16 *mssp);
-__u32 cookie_v6_init_sequence(struct request_sock *req, const struct sock *sk,
-			      const struct sk_buff *skb, __u16 *mss);
+__u32 cookie_v6_init_sequence(const struct sk_buff *skb, __u16 *mss);
 #endif
 /* tcp_output.c */
 
@@ -2086,8 +2084,8 @@ struct tcp_request_sock_ops {
 			struct sk_buff *skb,
 			bool want_cookie);
 #ifdef CONFIG_SYN_COOKIES
-	__u32 (*cookie_init_seq)(struct request_sock *req, const struct sock *sk,
-				 const struct sk_buff *skb, __u16 *mss);
+	__u32 (*cookie_init_seq)(const struct sk_buff *skb,
+				 __u16 *mss);
 #endif
 	struct dst_entry *(*route_req)(const struct sock *sk, struct flowi *fl,
 				       const struct request_sock *req);
@@ -2101,13 +2099,12 @@ struct tcp_request_sock_ops {
 
 #ifdef CONFIG_SYN_COOKIES
 static inline __u32 cookie_init_sequence(const struct tcp_request_sock_ops *ops,
-					 struct request_sock *req,
 					 const struct sock *sk, struct sk_buff *skb,
 					 __u16 *mss)
 {
 	tcp_synq_overflow(sk);
 	__NET_INC_STATS(sock_net(sk), LINUX_MIB_SYNCOOKIESSENT);
-	return ops->cookie_init_seq(req, sk, skb, mss);
+	return ops->cookie_init_seq(skb, mss);
 }
 #else
 static inline __u32 cookie_init_sequence(const struct tcp_request_sock_ops *ops,
net/ipv4/syncookies.c +7 -40

104 lines not on this page · raw message

net/ipv4/tcp_input.c +1 -1
index c2d5fa27ee8d..ea36a23267cf 100644
@@ -6674,7 +6674,7 @@ int tcp_conn_request(struct request_sock_ops *rsk_ops,
 	tcp_ecn_create_request(req, skb, sk, dst);
 
 	if (want_cookie) {
-		isn = cookie_init_sequence(af_ops, req, sk, skb, &req->mss);
+		isn = cookie_init_sequence(af_ops, sk, skb, &req->mss);
 		req->cookie_ts = tmp_opt.tstamp_ok;
 		if (!tmp_opt.tstamp_ok)
 			inet_rsk(req)->ecn_ok = 0;
net/ipv4/tcp_minisocks.c +1 -1
index 90912094135b..987316e23f45 100644
@@ -838,7 +838,7 @@ struct sock *tcp_check_req(struct sock *sk, struct sk_buff *skb,
 		goto listen_overflow;
 
 	if (own_req && !is_meta_sk(sk)) {
-		int ret = mptcp_check_req_master(sk, child, req, skb, 1);
+		int ret = mptcp_check_req_master(sk, child, req, skb);
 		if (ret < 0)
 			goto listen_overflow;
 
net/ipv6/syncookies.c +7 -25
index bb86437a7003..fa3739f6de48 100644
@@ -20,8 +20,6 @@
 #include <linux/kernel.h>
 #include <net/secure_seq.h>
 #include <net/ipv6.h>
-#include <net/mptcp.h>
-#include <net/mptcp_v6.h>
 #include <net/tcp.h>
 
 #define COOKIEBITS 24	/* Upper bits store count */
@@ -113,8 +111,7 @@ u32 __cookie_v6_init_sequence(const struct ipv6hdr *iph,
 }
 EXPORT_SYMBOL_GPL(__cookie_v6_init_sequence);
 
-__u32 cookie_v6_init_sequence(struct request_sock *req, const struct sock *sk,
-			      const struct sk_buff *skb, __u16 *mssp)
+__u32 cookie_v6_init_sequence(const struct sk_buff *skb, __u16 *mssp)
 {
 	const struct ipv6hdr *iph = ipv6_hdr(skb);
 	const struct tcphdr *th = tcp_hdr(skb);
@@ -136,7 +133,6 @@ EXPORT_SYMBOL_GPL(__cookie_v6_check);
 struct sock *cookie_v6_check(struct sock *sk, struct sk_buff *skb)
 {
 	struct tcp_options_received tcp_opt;
-	struct mptcp_options_received mopt;
 	struct inet_request_sock *ireq;
 	struct tcp_request_sock *treq;
 	struct ipv6_pinfo *np = inet6_sk(sk);
@@ -166,8 +162,7 @@ struct sock *cookie_v6_check(struct sock *sk, struct sk_buff *skb)
 
 	/* check for timestamp cookie support */
 	memset(&tcp_opt, 0, sizeof(tcp_opt));
-	mptcp_init_mp_opt(&mopt);
-	tcp_parse_options(sock_net(sk), skb, &tcp_opt, &mopt, 0, NULL, NULL);
+	tcp_parse_options(sock_net(sk), skb, &tcp_opt, NULL, 0, NULL, NULL);
 
 	if (tcp_opt.saw_tstamp && tcp_opt.rcv_tsecr) {
 		tsoff = secure_tcpv6_ts_off(sock_net(sk),
@@ -180,27 +175,14 @@ struct sock *cookie_v6_check(struct sock *sk, struct sk_buff *skb)
 		goto out;
 
 	ret = NULL;
-#ifdef CONFIG_MPTCP
-	if (mopt.saw_mpc)
-		req = inet_reqsk_alloc(&mptcp6_request_sock_ops, sk, false);
-	else
-#endif
-		req = inet_reqsk_alloc(&tcp6_request_sock_ops, sk, false);
+	req = inet_reqsk_alloc(&tcp6_request_sock_ops, sk, false);
 	if (!req)
 		goto out;
 
 	ireq = inet_rsk(req);
-	ireq->mptcp_rqsk = 0;
-	ireq->saw_mpc = 0;
 	treq = tcp_rsk(req);
 	treq->tfo_listener = false;
 
-	/* Must be done before anything else, as it initializes
-	 * hash_entry of the MPTCP request-sock.
-	 */
-	if (mopt.saw_mpc)
-		mptcp_cookies_reqsk_init(req, &mopt, skb);
-
 	if (security_inet_conn_request(sk, skb, req))
 		goto out_free;
 
@@ -264,10 +246,10 @@ struct sock *cookie_v6_check(struct sock *sk, struct sk_buff *skb)
 	}
 
 	req->rsk_window_clamp = tp->window_clamp ? :dst_metric(dst, RTAX_WINDOW);
-	tp->ops->select_initial_window(sk, tcp_full_space(sk), req->mss,
-				       &req->rsk_rcv_wnd, &req->rsk_window_clamp,
-				       ireq->wscale_ok, &rcv_wscale,
-				       dst_metric(dst, RTAX_INITRWND));
+	tcp_select_initial_window(sk, tcp_full_space(sk), req->mss,
+				  &req->rsk_rcv_wnd, &req->rsk_window_clamp,
+				  ireq->wscale_ok, &rcv_wscale,
+				  dst_metric(dst, RTAX_INITRWND));
 
 	ireq->rcv_wscale = rcv_wscale;
 	ireq->ecn_ok = cookie_ecn_ok(&tcp_opt, sock_net(sk), dst);
net/mptcp/mptcp_ctrl.c +4 -88

125 lines not on this page · raw message

net/mptcp/mptcp_ipv4.c +6 -23
index 00c54f421ba7..1f09cba6dc51 100644
@@ -71,31 +71,17 @@ static int mptcp_v4_init_req(struct request_sock *req, const struct sock *sk,
 
 	mptcp_rsk(req)->hash_entry.pprev = NULL;
 	mptcp_rsk(req)->is_sub = 0;
-	inet_rsk(req)->mptcp_rqsk = 1;
+	inet_rsk(req)->mptcp_rqsk = 0;
 
-	/* In case of SYN-cookies, we wait for the isn to be generated - it is
-	 * input to the key-generation.
-	 */
-	if (!want_cookie)
-		mptcp_reqsk_init(req, sk, skb, false);
+	/* SYN-cookies are not supported */
+	if (!want_cookie) {
+		inet_rsk(req)->mptcp_rqsk = 1;
+		mptcp_reqsk_init(req, sk, skb);
+	}
 
 	return 0;
 }
 
-#ifdef CONFIG_SYN_COOKIES
-static u32 mptcp_v4_cookie_init_seq(struct request_sock *req, const struct sock *sk,
-				    const struct sk_buff *skb, __u16 *mssp)
-{
-	__u32 isn = cookie_v4_init_sequence(req, sk, skb, mssp);
-
-	tcp_rsk(req)->snt_isn = isn;
-
-	mptcp_reqsk_init(req, sk, skb, true);
-
-	return isn;
-}
-#endif
-
 /* May be called without holding the meta-level lock */
 static int mptcp_v4_join_init_req(struct request_sock *req, const struct sock *meta_sk,
 				  struct sk_buff *skb, bool want_cookie)
@@ -381,9 +367,6 @@ int mptcp_pm_v4_init(void)
 
 	mptcp_request_sock_ipv4_ops = tcp_request_sock_ipv4_ops;
 	mptcp_request_sock_ipv4_ops.init_req = mptcp_v4_init_req;
-#ifdef CONFIG_SYN_COOKIES
-	mptcp_request_sock_ipv4_ops.cookie_init_seq = mptcp_v4_cookie_init_seq;
-#endif
 	mptcp_join_request_sock_ipv4_ops = tcp_request_sock_ipv4_ops;
 	mptcp_join_request_sock_ipv4_ops.init_req = mptcp_v4_join_init_req;
 
net/mptcp/mptcp_ipv6.c +6 -23
index 7249d171c9ef..d138560572ad 100644
@@ -100,31 +100,17 @@ static int mptcp_v6_init_req(struct request_sock *req, const struct sock *sk,
 
 	mptcp_rsk(req)->hash_entry.pprev = NULL;
 	mptcp_rsk(req)->is_sub = 0;
-	inet_rsk(req)->mptcp_rqsk = 1;
+	inet_rsk(req)->mptcp_rqsk = 0;
 
-	/* In case of SYN-cookies, we wait for the isn to be generated - it is
-	 * input to the key-generation.
-	 */
-	if (!want_cookie)
-		mptcp_reqsk_init(req, sk, skb, false);
+	/* SYN-cookies are not supported */
+	if (!want_cookie) {
+		inet_rsk(req)->mptcp_rqsk = 1;
+		mptcp_reqsk_init(req, sk, skb);
+	}
 
 	return 0;
 }
 
-#ifdef CONFIG_SYN_COOKIES
-static u32 mptcp_v6_cookie_init_seq(struct request_sock *req, const struct sock *sk,
-				    const struct sk_buff *skb, __u16 *mssp)
-{
-	__u32 isn = cookie_v6_init_sequence(req, sk, skb, mssp);
-
-	tcp_rsk(req)->snt_isn = isn;
-
-	mptcp_reqsk_init(req, sk, skb, true);
-
-	return isn;
-}
-#endif
-
 /* May be called without holding the meta-level lock */
 static int mptcp_v6_join_init_req(struct request_sock *req, const struct sock *meta_sk,
 				  struct sk_buff *skb, bool want_cookie)
@@ -427,9 +413,6 @@ int mptcp_pm_v6_init(void)
 
 	mptcp_request_sock_ipv6_ops = tcp_request_sock_ipv6_ops;
 	mptcp_request_sock_ipv6_ops.init_req = mptcp_v6_init_req;
-#ifdef CONFIG_SYN_COOKIES
-	mptcp_request_sock_ipv6_ops.cookie_init_seq = mptcp_v6_cookie_init_seq;
-#endif
 
 	mptcp_join_request_sock_ipv6_ops = tcp_request_sock_ipv6_ops;
 	mptcp_join_request_sock_ipv6_ops.init_req = mptcp_v6_join_init_req;
Signature · 1 line
2.16.2
author
#8 lore raw

PATCH v3 8/12 mptcp: Remove support for TCP Fast Open

In reply to Christoph Paasch
Signed-off-by Christoph Paasch <cpaasch(a)apple.com>
---
 include/linux/tcp.h      |  2 +-
 include/net/mptcp.h      |  9 +-----
 net/ipv4/tcp.c           |  7 ++---
 net/ipv4/tcp_fastopen.c  | 16 +++--------
 net/ipv4/tcp_input.c     | 23 ++--------------
 net/mptcp/mptcp_ctrl.c   | 53 ------------------------------------
 net/mptcp/mptcp_input.c  | 71 ------------------------------------------------
 net/mptcp/mptcp_output.c |  5 +---
 8 files changed, 13 insertions(+), 173 deletions(-)
include/linux/tcp.h +1 -1
index 6dd9c45c52d6..01321044047c 100644
@@ -58,7 +58,7 @@ static inline unsigned int tcp_optlen(const struct sk_buff *skb)
 /* TCP Fast Open */
 #define TCP_FASTOPEN_COOKIE_MIN	4	/* Min Fast Open Cookie size in bytes */
 #define TCP_FASTOPEN_COOKIE_MAX	16	/* Max Fast Open Cookie size in bytes */
-#define TCP_FASTOPEN_COOKIE_SIZE 4	/* the size employed by this impl. */
+#define TCP_FASTOPEN_COOKIE_SIZE 8	/* the size employed by this impl. */
 
 /* TCP Fast Open Cookie as stored in memory */
 struct tcp_fastopen_cookie {
include/net/mptcp.h +1 -8
index 32b1fecb70bc..136f18ce1d8e 100644
@@ -737,7 +737,6 @@ void mptcp_close(struct sock *meta_sk, long timeout);
 bool mptcp_doit(struct sock *sk);
 int mptcp_create_master_sk(struct sock *meta_sk, __u64 remote_key,
 			   __u8 mptcp_ver, u32 window);
-int mptcp_check_req_fastopen(struct sock *child, struct request_sock *req);
 int mptcp_check_req_master(struct sock *sk, struct sock *child,
 			   struct request_sock *req, const struct sk_buff *skb);
 struct sock *mptcp_check_req_child(struct sock *meta_sk,
@@ -1049,8 +1048,7 @@ static inline void mptcp_check_rcvseq_wrap(struct tcp_sock *meta_tp,
 
 static inline int mptcp_sk_can_send(const struct sock *sk)
 {
-	return tcp_passive_fastopen(sk) ||
-	       ((1 << sk->sk_state) & (TCPF_ESTABLISHED | TCPF_CLOSE_WAIT) &&
+	return ((1 << sk->sk_state) & (TCPF_ESTABLISHED | TCPF_CLOSE_WAIT) &&
 		!tcp_sk(sk)->mptcp->pre_established);
 }
 
@@ -1273,11 +1271,6 @@ static inline bool mptcp_doit(struct sock *sk)
 {
 	return false;
 }
-static inline int mptcp_check_req_fastopen(struct sock *child,
-					   struct request_sock *req)
-{
-	return 1;
-}
 static inline int mptcp_check_req_master(const struct sock *sk,
 					 const struct sock *child,
 					 const struct request_sock *req,
net/ipv4/tcp.c +3 -4
index 87f7616aaf93..b2dac73903be 100644
@@ -1233,7 +1233,7 @@ int tcp_sendmsg_locked(struct sock *sk, struct msghdr *msg, size_t size)
 	}
 
 	if (unlikely(flags & MSG_FASTOPEN || inet_sk(sk)->defer_connect) &&
-	    !tp->repair) {
+	    !tp->repair && !sock_flag(sk, SOCK_MPTCP)) {
 		err = tcp_sendmsg_fastopen(sk, msg, &copied_syn, size);
 		if (err == -EINPROGRESS && copied_syn > 0)
 			goto out;
@@ -1250,8 +1250,7 @@ int tcp_sendmsg_locked(struct sock *sk, struct msghdr *msg, size_t size)
 	 * is fully established.
 	 */
 	if (((1 << sk->sk_state) & ~(TCPF_ESTABLISHED | TCPF_CLOSE_WAIT)) &&
-	    !tcp_passive_fastopen(mptcp(tp) && tp->mpcb->master_sk ?
-				  tp->mpcb->master_sk : sk)) {
+	    !tcp_passive_fastopen(sk)) {
 		err = sk_stream_wait_connect(sk, &timeo);
 		if (err != 0)
 			goto do_error;
@@ -3098,7 +3097,7 @@ static int do_tcp_setsockopt(struct sock *sk, int level,
 
 	case TCP_FASTOPEN:
 		if (val >= 0 && ((1 << sk->sk_state) & (TCPF_CLOSE |
-		    TCPF_LISTEN))) {
+		    TCPF_LISTEN)) && !sock_flag(sk, SOCK_MPTCP)) {
 			tcp_fastopen_init_key_once(net);
 
 			fastopen_queue_tune(sk, val);
net/ipv4/tcp_fastopen.c +4 -12
index 370fe7d4b579..018a48477355 100644
@@ -9,7 +9,6 @@
 #include <linux/rculist.h>
 #include <net/inetpeer.h>
 #include <net/tcp.h>
-#include <net/mptcp.h>
 
 void tcp_fastopen_init_key_once(struct net *net)
 {
@@ -219,7 +218,7 @@ static struct sock *tcp_fastopen_create_child(struct sock *sk,
 {
 	struct tcp_sock *tp;
 	struct request_sock_queue *queue = &inet_csk(sk)->icsk_accept_queue;
-	struct sock *child, *meta_sk;
+	struct sock *child;
 	bool own_req;
 
 	req->num_retrans = 0;
@@ -259,22 +258,15 @@ static struct sock *tcp_fastopen_create_child(struct sock *sk,
 
 	refcount_set(&req->rsk_refcnt, 2);
 
+	/* Now finish processing the fastopen child socket. */
+	tcp_init_transfer(child, BPF_SOCK_OPS_PASSIVE_ESTABLISHED_CB);
+
 	tp->rcv_nxt = TCP_SKB_CB(skb)->seq + 1;
 
 	tcp_fastopen_add_skb(child, skb);
 
 	tcp_rsk(req)->rcv_nxt = tp->rcv_nxt;
 	tp->rcv_wup = tp->rcv_nxt;
-
-	meta_sk = child;
-	if (!mptcp_check_req_fastopen(meta_sk, req)) {
-		child = tcp_sk(meta_sk)->mpcb->master_sk;
-		tp = tcp_sk(child);
-	}
-
-	/* Now finish processing the fastopen child socket. */
-	tcp_init_transfer(child, BPF_SOCK_OPS_PASSIVE_ESTABLISHED_CB);
-
 	/* tcp_conn_request() is sending the SYNACK,
 	 * and queues the child into listener accept queue.
 	 */
net/ipv4/tcp_input.c +3 -20
index ea36a23267cf..49ac1b821ac1 100644
@@ -5784,17 +5784,10 @@ static bool tcp_rcv_fastopen_synack(struct sock *sk, struct sk_buff *synack,
 				    struct tcp_fastopen_cookie *cookie)
 {
 	struct tcp_sock *tp = tcp_sk(sk);
-	struct sk_buff *data = NULL;
+	struct sk_buff *data = tp->syn_data ? tcp_rtx_queue_head(sk) : NULL;
 	u16 mss = tp->rx_opt.mss_clamp, try_exp = 0;
 	bool syn_drop = false;
 
-	if (tp->syn_data) {
-		if (mptcp(tp))
-			data = tcp_write_queue_head(mptcp_meta_sk(sk));
-		else
-			data = tcp_rtx_queue_head(sk);
-	}
-
 	if (mss == tp->rx_opt.user_mss) {
 		struct tcp_options_received opt;
 
@@ -5825,11 +5818,7 @@ static bool tcp_rcv_fastopen_synack(struct sock *sk, struct sk_buff *synack,
 
 	tcp_fastopen_cache_set(sk, mss, cookie, syn_drop, try_exp);
 
-	/* In mptcp case, we do not rely on "retransmit", but instead on
-	 * "transmit", because if fastopen data is not acked, the retransmission
-	 * becomes the first MPTCP data (see mptcp_rcv_synsent_fastopen).
-	 */
-	if (data && !mptcp(tp)) { /* Retransmit unacked data in SYN */
+	if (data) { /* Retransmit unacked data in SYN */
 		skb_rbtree_walk_from(data) {
 			if (__tcp_retransmit_skb(sk, data, 1))
 				break;
@@ -6689,18 +6678,12 @@ int tcp_conn_request(struct request_sock_ops *rsk_ops,
 		fastopen_sk = tcp_try_fastopen(sk, skb, req, &foc, dst);
 	}
 	if (fastopen_sk) {
-		struct sock *meta_sk = fastopen_sk;
-
-		if (mptcp(tcp_sk(fastopen_sk)))
-			meta_sk = mptcp_meta_sk(fastopen_sk);
 		af_ops->send_synack(fastopen_sk, dst, &fl, req,
 				    &foc, TCP_SYNACK_FASTOPEN);
 		/* Add the child socket directly into the accept queue */
-		inet_csk_reqsk_queue_add(sk, req, meta_sk);
+		inet_csk_reqsk_queue_add(sk, req, fastopen_sk);
 		sk->sk_data_ready(sk);
 		bh_unlock_sock(fastopen_sk);
-		if (meta_sk != fastopen_sk)
-			bh_unlock_sock(meta_sk);
 		sock_put(fastopen_sk);
 	} else {
 		tcp_rsk(req)->tfo_listener = false;
net/mptcp/mptcp_ctrl.c +0 -53

59 lines not on this page · raw message

net/mptcp/mptcp_input.c +0 -71

89 lines not on this page · raw message

net/mptcp/mptcp_output.c +1 -4
index 26cf41f75922..2d84eb2edd97 100644
@@ -1292,12 +1292,9 @@ void mptcp_meta_retransmit_timer(struct sock *meta_sk)
 
 void mptcp_sub_retransmit_timer(struct sock *sk)
 {
-	struct tcp_sock *tp = tcp_sk(sk);
-
 	tcp_retransmit_timer(sk);
 
-	if (!tp->fastopen_rsk)
-		mptcp_set_rto(sk);
+	mptcp_set_rto(sk);
 }
 
 /* Modify values to an mptcp-level for the initial window of new subflows */
Signature · 1 line
2.16.2
author
#9 lore raw

PATCH v3 9/12 mptcp: Remove MPTCP-code out of do_ip_setsockopt

In reply to Christoph Paasch
Signed-off-by Christoph Paasch <cpaasch(a)apple.com>
---
 net/ipv4/ip_sockglue.c | 14 --------------
 1 file changed, 14 deletions(-)
net/ipv4/ip_sockglue.c +0 -14
index 4760ebe590e1..7c864ad41e6d 100644
@@ -753,20 +753,6 @@ static int do_ip_setsockopt(struct sock *sk, int level,
 			inet->tos = val;
 			sk->sk_priority = rt_tos2priority(val);
 			sk_dst_reset(sk);
-			/* Update TOS on mptcp subflow */
-			if (is_meta_sk(sk)) {
-				struct mptcp_tcp_sock *mptcp;
-
-				mptcp_for_each_sub(tcp_sk(sk)->mpcb, mptcp) {
-					struct sock *sk_it = mptcp_to_sock(mptcp);
-
-					if (inet_sk(sk_it)->tos != inet_sk(sk)->tos) {
-						inet_sk(sk_it)->tos = inet_sk(sk)->tos;
-						sk_it->sk_priority = sk->sk_priority;
-						sk_dst_reset(sk_it);
-					}
-				}
-			}
 		}
 		break;
 	case IP_TTL:
Signature · 1 line
2.16.2
author
#10 lore raw

PATCH v3 10/12 mptcp: Remove scheduler-selection support

In reply to Christoph Paasch
Signed-off-by Christoph Paasch <cpaasch(a)apple.com>
---
 include/linux/tcp.h      |   3 -
 include/net/mptcp.h      |  41 ++-------
 include/uapi/linux/tcp.h |   1 -
 net/ipv4/tcp.c           |  46 ----------
 net/mptcp/Kconfig        |  32 -------
 net/mptcp/mptcp_ctrl.c   |  35 +-------
 net/mptcp/mptcp_output.c |   7 +-
 net/mptcp/mptcp_sched.c  | 223 ++++-------------------------------------------
 8 files changed, 28 insertions(+), 360 deletions(-)
include/linux/tcp.h +0 -3
index 01321044047c..02af5c56d2e6 100644
@@ -476,17 +476,14 @@ struct tcp_sock {
 		close_it:1,	/* Must close socket in mptcp_data_ready? */
 		closing:1,
 		mptcp_ver:4,
-		mptcp_sched_setsockopt:1,
 		mptcp_pm_setsockopt:1,
 		record_master_info:1;
 	struct mptcp_tcp_sock *mptcp;
 #ifdef CONFIG_MPTCP
-#define MPTCP_SCHED_NAME_MAX 16
 #define MPTCP_PM_NAME_MAX 16
 	struct hlist_nulls_node tk_table;
 	u32		mptcp_loc_token;
 	u64		mptcp_loc_key;
-	char		mptcp_sched_name[MPTCP_SCHED_NAME_MAX];
 	char		mptcp_pm_name[MPTCP_PM_NAME_MAX];
 #endif /* CONFIG_MPTCP */
 };
include/net/mptcp.h +9 -32

69 lines not on this page · raw message

include/uapi/linux/tcp.h +0 -1
index 072ce786a0fd..2fcfa2257cc9 100644
@@ -134,7 +134,6 @@ enum {
 #define TCP_CM_INQ		TCP_INQ
 
 #define MPTCP_ENABLED		42
-#define MPTCP_SCHEDULER		43
 #define MPTCP_PATH_MANAGER	44
 #define MPTCP_INFO		45
 
net/ipv4/tcp.c +0 -46

58 lines not on this page · raw message

net/mptcp/Kconfig +0 -32
index 9c490ec8fba8..860e9e17f76f 100644
@@ -57,35 +57,3 @@ config DEFAULT_MPTCP_PM
 	default "ndiffports" if DEFAULT_NDIFFPORTS
 	default "default"
 
-menuconfig MPTCP_SCHED_ADVANCED
-	bool "MPTCP: advanced scheduler control"
-	depends on MPTCP=y
-	---help---
-	  Support for selection of different schedulers. You should choose 'Y' here,
-	  if you want to choose a different scheduler than the default one.
-
-if MPTCP_SCHED_ADVANCED
-
-choice
-	prompt "Default MPTCP Scheduler"
-	default DEFAULT
-	help
-	  Select the Scheduler of your choice
-
-	config DEFAULT_SCHEDULER
-		bool "Default"
-		---help---
-		  This is the default scheduler, sending first on the subflow
-		  with the lowest RTT.
-
-endchoice
-endif
-
-config DEFAULT_MPTCP_SCHED
-	string
-	depends on (MPTCP=y)
-	default "default" if DEFAULT_SCHEDULER
-	default "roundrobin" if DEFAULT_ROUNDROBIN
-	default "redundant" if DEFAULT_REDUNDANT
-	default "default"
-
net/mptcp/mptcp_ctrl.c +1 -34
index b7befa77d76a..766d73d505d5 100644
@@ -94,25 +94,6 @@ static int proc_mptcp_path_manager(struct ctl_table *ctl, int write,
 	return ret;
 }
 
-static int proc_mptcp_scheduler(struct ctl_table *ctl, int write,
-				void __user *buffer, size_t *lenp,
-				loff_t *ppos)
-{
-	char val[MPTCP_SCHED_NAME_MAX];
-	struct ctl_table tbl = {
-		.data = val,
-		.maxlen = MPTCP_SCHED_NAME_MAX,
-	};
-	int ret;
-
-	mptcp_get_default_scheduler(val);
-
-	ret = proc_dostring(&tbl, write, buffer, lenp, ppos);
-	if (write && ret == 0)
-		ret = mptcp_set_default_scheduler(val);
-	return ret;
-}
-
 static struct ctl_table mptcp_table[] = {
 	{
 		.procname = "mptcp_enabled",
@@ -150,12 +131,6 @@ static struct ctl_table mptcp_table[] = {
 		.maxlen		= MPTCP_PM_NAME_MAX,
 		.proc_handler	= proc_mptcp_path_manager,
 	},
-	{
-		.procname	= "mptcp_scheduler",
-		.mode		= 0644,
-		.maxlen		= MPTCP_SCHED_NAME_MAX,
-		.proc_handler	= proc_mptcp_scheduler,
-	},
 	{ }
 };
 
@@ -604,7 +579,6 @@ static void mptcp_mpcb_put(struct mptcp_cb *mpcb)
 {
 	if (refcount_dec_and_test(&mpcb->mpcb_refcnt)) {
 		mptcp_cleanup_path_manager(mpcb);
-		mptcp_cleanup_scheduler(mpcb);
 		kfree(mpcb->master_info);
 		kmem_cache_free(mptcp_cb_cache, mpcb);
 	}
@@ -1276,7 +1250,6 @@ static int mptcp_alloc_mpcb(struct sock *meta_sk, __u64 remote_key,
 	refcount_set(&mpcb->mpcb_refcnt, 1);
 
 	mptcp_init_path_manager(mpcb);
-	mptcp_init_scheduler(mpcb);
 
 	if (!try_module_get(inet_csk(master_sk)->icsk_ca_ops->owner))
 		tcp_assign_congestion_control(master_sk);
@@ -1344,8 +1317,7 @@ int mptcp_add_sock(struct sock *meta_sk, struct sock *sk, u8 loc_id, u8 rem_id,
 
 	tp->mptcp->loc_id = loc_id;
 	tp->mptcp->rem_id = rem_id;
-	if (mpcb->sched_ops->init)
-		mpcb->sched_ops->init(sk);
+	mptcp_sched_init(sk);
 
 	/* The corresponding sock_put is in mptcp_sock_destruct(). It cannot be
 	 * included in mptcp_del_sock(), because the mpcb must remain alive
@@ -2825,17 +2797,12 @@ void __init mptcp_init(void)
 	if (mptcp_register_path_manager(&mptcp_pm_default))
 		goto register_pm_failed;
 
-	if (mptcp_register_scheduler(&mptcp_sched_default))
-		goto register_sched_failed;
-
 	pr_info("MPTCP: Unstable branch");
 
 	mptcp_init_failed = false;
 
 	return;
 
-register_sched_failed:
-	mptcp_unregister_path_manager(&mptcp_pm_default);
 register_pm_failed:
 	unregister_net_sysctl_table(mptcp_sysctl);
 register_sysctl_failed:
net/mptcp/mptcp_output.c +3 -4
index 2d84eb2edd97..00a3674ad693 100644
@@ -306,7 +306,7 @@ int mptcp_write_wakeup(struct sock *meta_sk, int mib)
 	    before(TCP_SKB_CB(skb)->seq, tcp_wnd_end(meta_tp))) {
 		unsigned int mss;
 		unsigned int seg_size = tcp_wnd_end(meta_tp) - TCP_SKB_CB(skb)->seq;
-		struct sock *subsk = meta_tp->mpcb->sched_ops->get_subflow(meta_sk, skb, true);
+		struct sock *subsk = mptcp_get_available_subflow(meta_sk, skb, true);
 		struct tcp_sock *subtp;
 
 		WARN_ON(TCP_SKB_CB(skb)->sacked);
@@ -395,8 +395,7 @@ bool mptcp_write_xmit(struct sock *meta_sk, unsigned int mss_now, int nonagle,
 
 	tcp_mstamp_refresh(meta_tp);
 
-	while ((skb = mpcb->sched_ops->next_segment(meta_sk, &reinject, &subsk,
-						    &sublimit))) {
+	while ((skb = mptcp_next_segment(meta_sk, &reinject, &subsk, &sublimit))) {
 		enum tcp_queue tcp_queue = TCP_FRAG_IN_WRITE_QUEUE;
 		unsigned int limit;
 
@@ -1111,7 +1110,7 @@ int mptcp_retransmit_skb(struct sock *meta_sk, struct sk_buff *skb)
 	/* We need to make sure that the retransmitted segment can be sent on a
 	 * subflow right now. If it is too big, it needs to be fragmented.
 	 */
-	subsk = meta_tp->mpcb->sched_ops->get_subflow(meta_sk, skb, false);
+	subsk = mptcp_get_available_subflow(meta_sk, skb, false);
 	if (!subsk) {
 		/* We want to increase icsk_retransmits, thus return 0, so that
 		 * mptcp_meta_retransmit_timer enters the desired branch.
net/mptcp/mptcp_sched.c +15 -208

279 lines not on this page · raw message

Signature · 1 line
2.16.2
author
#11 lore raw

PATCH v3 11/12 mptcp: Remove ndiffports path manager

In reply to Christoph Paasch
Signed-off-by Christoph Paasch <cpaasch(a)apple.com>
---
 net/mptcp/Kconfig            |  12 ----
 net/mptcp/Makefile           |   1 -
 net/mptcp/mptcp_ndiffports.c | 167 -------------------------------------------
 3 files changed, 180 deletions(-)
 delete mode 100644 net/mptcp/mptcp_ndiffports.c
net/mptcp/Kconfig +0 -12
index 860e9e17f76f..60c07fde5dec 100644
@@ -23,14 +23,6 @@ config MPTCP_FULLMESH
 	---help---
 	  This path-management module will create a full-mesh among all IP-addresses.
 
-config MPTCP_NDIFFPORTS
-	tristate "MPTCP ndiff-ports"
-	depends on MPTCP=y
-	---help---
-	  This path-management module will create multiple subflows between the same
-	  pair of IP-addresses, modifying the source-port. You can set the number
-	  of subflows via the mptcp_ndiffports-sysctl.
-
 choice
 	prompt "Default MPTCP Path-Manager"
 	default DEFAULT
@@ -40,9 +32,6 @@ choice
 	config DEFAULT_FULLMESH
 		bool "Full mesh" if MPTCP_FULLMESH=y
 
-	config DEFAULT_NDIFFPORTS
-		bool "ndiff-ports" if MPTCP_NDIFFPORTS=y
-
 	config DEFAULT_DUMMY
 		bool "Default"
 
@@ -54,6 +43,5 @@ config DEFAULT_MPTCP_PM
 	string
 	default "default" if DEFAULT_DUMMY
 	default "fullmesh" if DEFAULT_FULLMESH
-	default "ndiffports" if DEFAULT_NDIFFPORTS
 	default "default"
 
net/mptcp/Makefile +0 -1
index 8e28132c6c2a..433fd17b2987 100644
@@ -9,7 +9,6 @@ mptcp-y := mptcp_ctrl.o mptcp_ipv4.o mptcp_pm.o \
 	   mptcp_output.o mptcp_input.o mptcp_sched.o
 
 obj-$(CONFIG_MPTCP_FULLMESH) += mptcp_fullmesh.o
-obj-$(CONFIG_MPTCP_NDIFFPORTS) += mptcp_ndiffports.o
 
 mptcp-$(subst m,y,$(CONFIG_IPV6)) += mptcp_ipv6.o
 
net/mptcp/mptcp_ndiffports.c deleted +0 -167

167 lines not on this page · raw message

Signature · 1 line
2.16.2
author
#12 lore raw

PATCH v3 12/12 mptcp: Make path-management no more a module nor selectable through socketops

In reply to Christoph Paasch
Signed-off-by Christoph Paasch <cpaasch(a)apple.com>

---

Notes:
v2:
* Added fullmesh_initialized flag to indicate that we went through _new_session() and it's all setup. Necessary because the removal got rid of the function-pointers which were guarding that implicitly.
* Fixed some oversights in the removal of mptcp_become_fully_estab().

 include/linux/tcp.h        |   3 -
 include/net/mptcp.h        |  93 ++++++++-----
 include/net/netns/mptcp.h  |   8 +-
 net/ipv4/tcp.c             |  47 -------
 net/ipv4/tcp_input.c       |   4 +-
 net/ipv4/tcp_output.c      |   3 +-
 net/mptcp/Kconfig          |  37 -----
 net/mptcp/Makefile         |   4 +-
 net/mptcp/mptcp_ctrl.c     |  49 +------
 net/mptcp/mptcp_fullmesh.c | 326 +++++++--------------------------------------
 net/mptcp/mptcp_input.c    |  32 ++---
 net/mptcp/mptcp_ipv4.c     |   2 +-
 net/mptcp/mptcp_ipv6.c     |   2 +-
 net/mptcp/mptcp_output.c   |   8 +-
 net/mptcp/mptcp_pm.c       | 226 -------------------------------
 15 files changed, 136 insertions(+), 708 deletions(-)
 delete mode 100644 net/mptcp/mptcp_pm.c
include/linux/tcp.h +0 -3
index 02af5c56d2e6..84ec344fe184 100644
@@ -476,15 +476,12 @@ struct tcp_sock {
 		close_it:1,	/* Must close socket in mptcp_data_ready? */
 		closing:1,
 		mptcp_ver:4,
-		mptcp_pm_setsockopt:1,
 		record_master_info:1;
 	struct mptcp_tcp_sock *mptcp;
 #ifdef CONFIG_MPTCP
-#define MPTCP_PM_NAME_MAX 16
 	struct hlist_nulls_node tk_table;
 	u32		mptcp_loc_token;
 	u64		mptcp_loc_key;
-	char		mptcp_pm_name[MPTCP_PM_NAME_MAX];
 #endif /* CONFIG_MPTCP */
 };
 
include/net/mptcp.h +58 -35

130 lines not on this page · raw message

include/net/netns/mptcp.h +2 -6
index 6680f3bbcfc8..c9dd96de4db5 100644
@@ -32,12 +32,8 @@
 
 #include <linux/compiler.h>
 
-enum {
-	MPTCP_PM_FULLMESH = 0,
-	MPTCP_PM_MAX
-};
-
 struct mptcp_mib;
+struct mptcp_fm_ns;
 
 struct netns_mptcp {
 	DEFINE_SNMP_STAT(struct mptcp_mib, mptcp_statistics);
@@ -46,7 +42,7 @@ struct netns_mptcp {
 	struct proc_dir_entry *proc_net_mptcp;
 #endif
 
-	void *path_managers[MPTCP_PM_MAX];
+	struct mptcp_fm_ns *fm_ns;
 };
 
 #endif /* __NETNS_MPTCP_H__ */
net/ipv4/tcp.c +0 -47

59 lines not on this page · raw message

net/ipv4/tcp_input.c +2 -2
index 49ac1b821ac1..5cf9161db844 100644
@@ -6177,7 +6177,7 @@ int tcp_rcv_state_process(struct sock *sk, struct sk_buff *skb)
 			 * addresses, which can only be done after the third ack
 			 * of the 3-way handshake.
 			 */
-			mptcp_update_metasocket(tp->meta_sk);
+			full_mesh_new_session(tp->meta_sk);
 		}
 		if (queued >= 0)
 			return queued;
@@ -6280,7 +6280,7 @@ int tcp_rcv_state_process(struct sock *sk, struct sk_buff *skb)
 		 */
 		if (mptcp(tp)) {
 			if (is_master_tp(tp))
-				mptcp_update_metasocket(mptcp_meta_sk(sk));
+				full_mesh_new_session(mptcp_meta_sk(sk));
 			else
 				tcp_send_ack(sk);
 		}
net/ipv4/tcp_output.c +1 -2
index a1cb6a3c0dda..e41db7407a90 100644
@@ -921,8 +921,7 @@ void tcp_release_cb(struct sock *sk)
 		__sock_put(sk);
 	}
 	if (flags & TCPF_PATH_MANAGER_DEFERRED) {
-		if (tcp_sk(sk)->mpcb->pm_ops->release_sock)
-			tcp_sk(sk)->mpcb->pm_ops->release_sock(sk);
+		full_mesh_release_sock(sk);
 		__sock_put(sk);
 	}
 	if (flags & TCPF_SUB_DEFERRED)
net/mptcp/Kconfig +0 -37
index 60c07fde5dec..c3b835e40d8b 100644
@@ -8,40 +8,3 @@ config MPTCP
           This replaces the normal TCP stack with a Multipath TCP stack,
           able to use several paths at once.
 
-menuconfig MPTCP_PM_ADVANCED
-	bool "MPTCP: advanced path-manager control"
-	depends on MPTCP=y
-	---help---
-	  Support for selection of different path-managers. You should choose 'Y' here,
-	  because otherwise you will not actively create new MPTCP-subflows.
-
-if MPTCP_PM_ADVANCED
-
-config MPTCP_FULLMESH
-	tristate "MPTCP Full-Mesh Path-Manager"
-	depends on MPTCP=y
-	---help---
-	  This path-management module will create a full-mesh among all IP-addresses.
-
-choice
-	prompt "Default MPTCP Path-Manager"
-	default DEFAULT
-	help
-	  Select the Path-Manager of your choice
-
-	config DEFAULT_FULLMESH
-		bool "Full mesh" if MPTCP_FULLMESH=y
-
-	config DEFAULT_DUMMY
-		bool "Default"
-
-endchoice
-
-endif
-
-config DEFAULT_MPTCP_PM
-	string
-	default "default" if DEFAULT_DUMMY
-	default "fullmesh" if DEFAULT_FULLMESH
-	default "default"
-
net/mptcp/Makefile +1 -3
index 433fd17b2987..87901e6b3843 100644
@@ -5,10 +5,8 @@
 
 obj-$(CONFIG_MPTCP) += mptcp.o
 
-mptcp-y := mptcp_ctrl.o mptcp_ipv4.o mptcp_pm.o \
+mptcp-y := mptcp_ctrl.o mptcp_ipv4.o mptcp_fullmesh.o \
 	   mptcp_output.o mptcp_input.o mptcp_sched.o
 
-obj-$(CONFIG_MPTCP_FULLMESH) += mptcp_fullmesh.o
-
 mptcp-$(subst m,y,$(CONFIG_IPV6)) += mptcp_ipv6.o
 
net/mptcp/mptcp_ctrl.c +3 -46

103 lines not on this page · raw message

net/mptcp/mptcp_fullmesh.c +50 -276

531 lines not on this page · raw message

net/mptcp/mptcp_input.c +13 -19
index 88c7a64ef26c..68c10ac0072b 100644
@@ -44,15 +44,6 @@ static inline bool before64(const u64 seq1, const u64 seq2)
 /* is seq1 > seq2 ? */
 #define after64(seq1, seq2)	before64(seq2, seq1)
 
-static inline void mptcp_become_fully_estab(struct sock *sk)
-{
-	tcp_sk(sk)->mptcp->fully_established = 1;
-
-	if (is_master_tp(tcp_sk(sk)) &&
-	    tcp_sk(sk)->mpcb->pm_ops->fully_established)
-		tcp_sk(sk)->mpcb->pm_ops->fully_established(mptcp_meta_sk(sk));
-}
-
 /* Similar to tcp_tso_acked without any memory accounting */
 static inline int mptcp_tso_acked_reinject(const struct sock *meta_sk,
 					   struct sk_buff *skb)
@@ -436,8 +427,11 @@ static int mptcp_prevalidate_skb(struct sock *sk, struct sk_buff *skb)
 	 */
 	if (!tp->mptcp->fully_established) {
 		tp->mptcp->init_rcv_wnd -= skb->len;
-		if (tp->mptcp->init_rcv_wnd < 0)
-			mptcp_become_fully_estab(sk);
+		if (tp->mptcp->init_rcv_wnd < 0) {
+			tp->mptcp->fully_established = 1;
+			if (is_master_tp(tp))
+				full_mesh_create_subflows(tp->meta_sk);
+		}
 	}
 
 	return 0;
@@ -1243,11 +1237,14 @@ static void mptcp_data_ack(struct sock *sk, const struct sk_buff *skb)
 		return;
 
 	if (unlikely(!tp->mptcp->fully_established) &&
-	    tp->mptcp->snt_isn + 1 != TCP_SKB_CB(skb)->ack_seq)
+	    tp->mptcp->snt_isn + 1 != TCP_SKB_CB(skb)->ack_seq) {
 		/* As soon as a subflow-data-ack (not acking syn, thus snt_isn + 1)
 		 * includes a data-ack, we are fully established
 		 */
-		mptcp_become_fully_estab(sk);
+		tp->mptcp->fully_established = 1;
+		if (is_master_tp(tp))
+			full_mesh_create_subflows(meta_sk);
+	}
 
 	/* After we did the subflow-only processing (stopping timer and marking
 	 * subflow as established), check if we can proceed with MPTCP-level
@@ -1757,8 +1754,7 @@ static void mptcp_handle_add_addr(const unsigned char *ptr, struct sock *sk)
 		return;
 	}
 
-	if (mpcb->pm_ops->add_raddr)
-		mpcb->pm_ops->add_raddr(mpcb, &addr, family, port, mpadd->addr_id);
+	full_mesh_add_raddr(mpcb, &addr, family, port, mpadd->addr_id);
 
 	MPTCP_INC_STATS(sock_net(sk), MPTCP_MIB_ADDADDRRX);
 }
@@ -1773,8 +1769,7 @@ static void mptcp_handle_rem_addr(const unsigned char *ptr, struct sock *sk)
 	for (i = 0; i <= mprem->len - MPTCP_SUB_LEN_REMOVE_ADDR; i++) {
 		rem_id = (&mprem->addrs_id)[i];
 
-		if (mpcb->pm_ops->rem_raddr)
-			mpcb->pm_ops->rem_raddr(mpcb, rem_id);
+		full_mesh_rem_raddr(mpcb, rem_id);
 		mptcp_send_reset_rem_id(mpcb, rem_id);
 
 		MPTCP_INC_STATS(sock_net(sk), MPTCP_MIB_REMADDRSUB);
@@ -1877,8 +1872,7 @@ static inline void mptcp_path_array_check(struct sock *meta_sk)
 
 	if (unlikely(mpcb->list_rcvd)) {
 		mpcb->list_rcvd = 0;
-		if (mpcb->pm_ops->new_remote_address)
-			mpcb->pm_ops->new_remote_address(meta_sk);
+		full_mesh_create_subflows(meta_sk);
 	}
 }
 
net/mptcp/mptcp_ipv4.c +1 -1
index 1f09cba6dc51..6f52a3a75ff3 100644
@@ -105,7 +105,7 @@ static int mptcp_v4_join_init_req(struct request_sock *req, const struct sock *m
 						    tcp_hdr(skb)->source,
 						    tcp_hdr(skb)->dest);
 	addr.ip = inet_rsk(req)->ir_loc_addr;
-	loc_id = mpcb->pm_ops->get_local_id(AF_INET, &addr, sock_net(meta_sk));
+	loc_id = full_mesh_get_local_id(AF_INET, &addr, sock_net(meta_sk));
 	if (loc_id == -1)
 		return -1;
 	mtreq->loc_id = loc_id;
net/mptcp/mptcp_ipv6.c +1 -1
index d138560572ad..ceaa3cd9e40e 100644
@@ -134,7 +134,7 @@ static int mptcp_v6_join_init_req(struct request_sock *req, const struct sock *m
 						    tcp_hdr(skb)->source,
 						    tcp_hdr(skb)->dest);
 	addr.in6 = inet_rsk(req)->ir_v6_loc_addr;
-	loc_id = mpcb->pm_ops->get_local_id(AF_INET6, &addr, sock_net(meta_sk));
+	loc_id = full_mesh_get_local_id(AF_INET6, &addr, sock_net(meta_sk));
 	if (loc_id == -1)
 		return -1;
 	mtreq->loc_id = loc_id;
net/mptcp/mptcp_output.c +4 -4
index 00a3674ad693..b3978a7e9254 100644
@@ -709,9 +709,9 @@ void mptcp_established_options(struct sock *sk, struct sk_buff *skb,
 		*size += MPTCP_SUB_LEN_JOIN_ACK_ALIGN;
 	}
 
-	if (unlikely(mpcb->addr_signal) && mpcb->pm_ops->addr_signal &&
+	if (unlikely(mpcb->addr_signal) &&
 	    mpcb->mptcp_ver >= MPTCP_VERSION_1 && skb && !mptcp_is_data_seq(skb)) {
-		mpcb->pm_ops->addr_signal(sk, size, opts, skb);
+		full_mesh_addr_signal(sk, size, opts, skb);
 
 		if (opts->add_addr_v6)
 			/* Skip subsequent options */
@@ -734,9 +734,9 @@ void mptcp_established_options(struct sock *sk, struct sk_buff *skb,
 		*size += MPTCP_SUB_LEN_DSS_ALIGN;
 	}
 
-	if (unlikely(mpcb->addr_signal) && mpcb->pm_ops->addr_signal &&
+	if (unlikely(mpcb->addr_signal) &&
 	    mpcb->mptcp_ver < MPTCP_VERSION_1)
-		mpcb->pm_ops->addr_signal(sk, size, opts, skb);
+		full_mesh_addr_signal(sk, size, opts, skb);
 
 	return;
 }
net/mptcp/mptcp_pm.c deleted +0 -226

226 lines not on this page · raw message

Signature · 1 line
2.16.2
#13 lore raw
In reply to Christoph Paasch

Hello,

Thanks Christoph for the v3!

For those who are interested by doing this review in Gerrit, I just sent this v3:

https://review.gerrithub.io/q/topic:%22cpaasch-remove-extras-v3%22+(status:open%20OR%20status:merged)

I also forgot to mention that I did the review for the v2 in Gerrit too.

Cheers,
Matthieu

Quoted 64 lines · Christoph Paasch

Series removing a lot of code out of mptcp-net-next.

Before:

$ git diff --shortstat  1464193107da..mptcp-net-next/master
 54 files changed, 14663 insertions(+), 375 deletions(-)
$ git diff --shortstat  1464193107da..HEAD
 50 files changed, 12217 insertions(+), 349 deletions(-)

v3: Minor left-over removals that were missing - identified by Matthieu
v2: Some minor fixes - see notes in the sub-patches 4/12 and 12/12

Christoph Paasch (12):
mptcp: Remove support for DSS-checksums
mptcp: Remove some left-over cruft
mptcp: Remove RPS-recording for flow-steering in MPTCP
mptcp: Remove support for subflow-priorities (aka backup)
mptcp: Remove support for sendpage()
mptcp: Remove reinjection of data across subflows
mptcp: Remove support for TCP SYN-cookies
mptcp: Remove support for TCP Fast Open
mptcp: Remove MPTCP-code out of do_ip_setsockopt
mptcp: Remove scheduler-selection support
mptcp: Remove ndiffports path manager
mptcp: Make path-management no more a module nor selectable through
socketops

 include/linux/tcp.h          |  15 +-
 include/net/mptcp.h          | 250 +++++++---------------
 include/net/netns/mptcp.h    |   8 +-
 include/net/tcp.h            |  13 +-
 include/uapi/linux/if.h      |   1 -
 include/uapi/linux/tcp.h     |   1 -
 net/core/dev.c               |   2 +-
 net/ipv4/Kconfig             |  19 --
 net/ipv4/af_inet.c           |   6 -
 net/ipv4/ip_sockglue.c       |  14 --
 net/ipv4/syncookies.c        |  47 +----
 net/ipv4/tcp.c               | 155 +-------------
 net/ipv4/tcp_fastopen.c      |  16 +-
 net/ipv4/tcp_input.c         |  29 +--
 net/ipv4/tcp_minisocks.c     |   2 +-
 net/ipv4/tcp_output.c        |   3 +-
 net/ipv6/syncookies.c        |  32 +--
 net/mptcp/Kconfig            |  81 -------
 net/mptcp/Makefile           |   5 +-
 net/mptcp/mptcp_ctrl.c       | 261 +----------------------
 net/mptcp/mptcp_fullmesh.c   | 488 ++++++++-----------------------------------
 net/mptcp/mptcp_input.c      | 343 ++----------------------------
 net/mptcp/mptcp_ipv4.c       |  34 +--
 net/mptcp/mptcp_ipv6.c       |  34 +--
 net/mptcp/mptcp_ndiffports.c | 169 ---------------
 net/mptcp/mptcp_output.c     | 472 ++---------------------------------------
 net/mptcp/mptcp_pm.c         | 226 --------------------
 net/mptcp/mptcp_sched.c      | 280 +++----------------------
 28 files changed, 293 insertions(+), 2713 deletions(-)
 delete mode 100644 net/mptcp/mptcp_ndiffports.c
 delete mode 100644 net/mptcp/mptcp_pm.c

--
2.16.2

Signature · 5 lines
Matthieu Baerts | R&D Engineer
matthieu.baerts(a)tessares.net
Tessares SA | Hybrid Access Solutions
www.tessares.net
1 Avenue Jean Monnet, 1348 Louvain-la-Neuve, Belgium