Inital merge from Linux source tree 2.6.31.9 to 2.6.32 for aNCR

include/linux/sysctl.h

@@ -434,6 +434,9 @@ enum
 	NET_TCP_ALLOWED_CONG_CONTROL=123,
 	NET_TCP_MAX_SSTHRESH=124,
 	NET_TCP_FRTO_RESPONSE=125,
+	NET_TCP_REORDER=126,
+	NET_TCP_AVAIL_REORDER=127,
+	NET_TCP_ALLOWED_REORDER=128,
 };
 
 enum {

include/linux/tcp.h

@@ -50,7 +50,7 @@ struct tcphdr {
 		fin:1;
 #else
 #error	"Adjust your <asm/byteorder.h> defines"
-#endif	
+#endif
 	__be16	window;
 	__sum16	check;
 	__be16	urg_ptr;
@@ -61,14 +61,14 @@ struct tcphdr {
  *  (union is compatible to any of its members)
  *  This means this part of the code is -fstrict-aliasing safe now.
  */
-union tcp_word_hdr { 
+union tcp_word_hdr {
 	struct tcphdr hdr;
 	__be32 		  words[5];
-}; 
+};
 
-#define tcp_flag_word(tp) ( ((union tcp_word_hdr *)(tp))->words [3]) 
+#define tcp_flag_word(tp) ( ((union tcp_word_hdr *)(tp))->words [3])
 
-enum { 
+enum {
 	TCP_FLAG_CWR = __cpu_to_be32(0x00800000),
 	TCP_FLAG_ECE = __cpu_to_be32(0x00400000),
 	TCP_FLAG_URG = __cpu_to_be32(0x00200000),
@@ -79,7 +79,7 @@ enum {
 	TCP_FLAG_FIN = __cpu_to_be32(0x00010000),
 	TCP_RESERVED_BITS = __cpu_to_be32(0x0F000000),
 	TCP_DATA_OFFSET = __cpu_to_be32(0xF0000000)
-}; 
+};
 
 /* TCP socket options */
 #define TCP_NODELAY		1	/* Turn off Nagle's algorithm. */
@@ -96,6 +96,8 @@ enum {
 #define TCP_QUICKACK		12	/* Block/reenable quick acks */
 #define TCP_CONGESTION		13	/* Congestion control algorithm */
 #define TCP_MD5SIG		14	/* TCP MD5 Signature (RFC2385) */
+#define TCP_REORDER		15 /* Reordering Algorithm */
+#define TCP_REORDER_MODE		16	/* meaning of "mode" depends on reorder module */
 
 #define TCPI_OPT_TIMESTAMPS	1
 #define TCPI_OPT_SACK		2
@@ -157,6 +159,11 @@ struct tcp_info
 	__u32	tcpi_rcv_space;
 
 	__u32	tcpi_total_retrans;
+	__u32	tcpi_total_fast_retrans;
+	__u32	tcpi_total_rto_retrans;
+	__u32   tcpi_total_dsacks;
+	__u32	tcpi_dupthresh;
+	__u32	tcpi_last_reor_sample;
 };
 
 /* for TCP_MD5SIG socket option */
@@ -244,6 +251,13 @@ static inline struct tcp_request_sock *tcp_rsk(const struct request_sock *req)
 	return (struct tcp_request_sock *)req;
 }
 
+struct reorder_sample {
+	struct list_head list;
+	u32 seq;
+	int factor;
+	int sample;
+};
+
 struct tcp_sock {
 	/* inet_connection_sock has to be the first member of tcp_sock */
 	struct inet_connection_sock	inet_conn;
@@ -330,6 +344,7 @@ struct tcp_sock {
 	u32	snd_cwnd_clamp; /* Do not allow snd_cwnd to grow above this */
 	u32	snd_cwnd_used;
 	u32	snd_cwnd_stamp;
+    u32 current_cwnd;
 
  	u32	rcv_wnd;	/* Current receiver window		*/
 	u32	write_seq;	/* Tail(+1) of data held in tcp send buffer */

include/net/inet_connection_sock.h

@@ -95,6 +95,7 @@ struct inet_connection_sock {
 	__u32			  icsk_rto;
 	__u32			  icsk_pmtu_cookie;
 	const struct tcp_congestion_ops *icsk_ca_ops;
+	const struct tcp_reorder_ops *icsk_ro_ops;
 	const struct inet_connection_sock_af_ops *icsk_af_ops;
 	unsigned int		  (*icsk_sync_mss)(struct sock *sk, u32 pmtu);
 	__u8			  icsk_ca_state;
@@ -126,7 +127,9 @@ struct inet_connection_sock {
 		int		  probe_size;
 	} icsk_mtup;
 	u32			  icsk_ca_priv[16];
+	u32			  icsk_ro_priv[5];
 #define ICSK_CA_PRIV_SIZE	(16 * sizeof(u32))
+#define ICSK_RO_PRIV_SIZE   (5  * sizeof(u32))
 };
 
 #define ICSK_TIME_RETRANS	1	/* Retransmit timer */
@@ -144,6 +147,11 @@ static inline void *inet_csk_ca(const struct sock *sk)
 	return (void *)inet_csk(sk)->icsk_ca_priv;
 }
 
+static inline void *inet_csk_ro(const struct sock *sk)
+{
+	return (void *)inet_csk(sk)->icsk_ro_priv;
+}
+
 extern struct sock *inet_csk_clone(struct sock *sk,
 				   const struct request_sock *req,
 				   const gfp_t priority);

include/net/tcp.h

@@ -468,9 +468,11 @@ extern __u32 cookie_v6_init_sequence(struct sock *sk, struct sk_buff *skb,
 extern void __tcp_push_pending_frames(struct sock *sk, unsigned int cur_mss,
 				      int nonagle);
 extern int tcp_may_send_now(struct sock *sk);
-extern int tcp_retransmit_skb(struct sock *, struct sk_buff *);
+extern int tcp_retransmit_skb(struct sock *, struct sk_buff *, int fast_rexmit);
+//extern int tcp_retransmit_skb(struct sock *, struct sk_buff *);
 extern void tcp_retransmit_timer(struct sock *sk);
-extern void tcp_xmit_retransmit_queue(struct sock *);
+extern void tcp_xmit_retransmit_queue(struct sock *, int fast_rexmit);
+//extern void tcp_xmit_retransmit_queue(struct sock *);
 extern void tcp_simple_retransmit(struct sock *);
 extern int tcp_trim_head(struct sock *, struct sk_buff *, u32);
 extern int tcp_fragment(struct sock *, struct sk_buff *, u32, unsigned int);
@@ -653,6 +655,68 @@ static inline int tcp_skb_mss(const struct sk_buff *skb)
 	return skb_shinfo(skb)->gso_size;
 }
 
+/*
+ * Interface for adding new TCP segment reordering handlers
+ */
+#define TCP_REORDER_NAME_MAX 16
+#define TCP_REORDER_MAX 128
+#define TCP_REORDER_BUF_MAX  (TCP_REORDER_NAME_MAX*TCP_REORDER_MAX)
+
+#define TCP_REORDER_NON_RESTRICTED 0x1
+
+struct tcp_reorder_ops {
+	struct list_head	list;
+	unsigned long flags;
+
+	/* initialize private data (optional) */
+	void (*init)(struct sock *sk);
+	/* cleanup private data  (optional) */
+	void (*release)(struct sock *sk);
+
+	/* return dupack threshold (required) */
+	u32 (*dupthresh)(struct sock *sk);
+	u32 (*moddupthresh)(struct sock *sk);
+	/* update the mode of operation (required) */
+	void (*update_mode)(struct sock *sk, int val);
+	/* allow cwnd moderation in disorder state [bool] (required) */
+	int allow_moderation;
+	/* allow head timeout to trigger fast recovery [bool] (required) */
+	int allow_head_to;
+
+	/* a new sack'ed segment (optional) */
+	void (*new_sack)(struct sock *sk);
+	/* a non-retransmitted SACK hole was filled (optional) */
+	void (*sack_hole_filled)(struct sock *sk, int flag);
+	/* state machine will start now (optional) */
+	void (*sm_starts)(struct sock *sk, int flag, int acked);
+	/* recovery phase starts (optional) */
+	void (*recovery_starts)(struct sock *sk, int flag);
+	void (*recovery_ends)(struct sock *sk, int flag);
+	/* reordering event with a certain degree was detected (optional) */
+	void (*reorder_detected)(struct sock *sk, int length);
+	/* reordering event with a certain factor was detected (optional) */
+	void (*reorder_detected_factor)(struct sock *sk, int factor);
+	/* a RTO timeout happened (optional) */
+	void (*rto_happened)(struct sock *sk);
+
+	char 		name[TCP_REORDER_NAME_MAX];
+	struct module	*owner;
+};
+
+extern int tcp_register_reorder(struct tcp_reorder_ops *ro);
+extern void tcp_unregister_reorder(struct tcp_reorder_ops *ro);
+extern void tcp_init_reorder(struct sock *sk);
+extern void tcp_cleanup_reorder(struct sock *sk);
+extern int tcp_set_default_reorder(const char *name);
+extern void tcp_get_default_reorder(char *name);
+extern void tcp_get_available_reorder(char *buf, size_t maxlen);
+extern void tcp_get_allowed_reorder(char *buf, size_t maxlen);
+extern int tcp_set_allowed_reorder(char *val);
+extern int tcp_set_reorder(struct sock *sk, const char *name);
+extern u32 tcp_native_dupthresh(struct sock *sk);
+extern struct tcp_reorder_ops tcp_init_reorder_ops;
+extern struct tcp_reorder_ops tcp_native;
+
 /* Events passed to congestion control interface */
 enum tcp_ca_event {
 	CA_EVENT_TX_START,	/* first transmit when no packets in flight */

kernel/sysctl_check.c

@@ -390,6 +390,9 @@ static const struct trans_ctl_table trans_net_ipv4_table[] = {
 	{ NET_TCP_ALLOWED_CONG_CONTROL,		"tcp_allowed_congestion_control" },
 	{ NET_TCP_MAX_SSTHRESH,			"tcp_max_ssthresh" },
 	{ NET_TCP_FRTO_RESPONSE,		"tcp_frto_response" },
+	{ NET_TCP_REORDER,		"tcp_reorder"}
+	{ NET_TCP_AVAIL_REORDER,		"tcp_available_reorder"}
+	{ NET_TCP_ALLOWED_REORDER,		"tcp_allowed_reorder"}
 	{ 2088 /* NET_IPQ_QMAX */,		"ip_queue_maxlen" },
 	{}
 };

nameonly.log

@@ -0,0 +1,26 @@
+This file contains the list of files that were originally added/modified to the
+linux source tree for aNCR implementation
+
+config-meshnode
+config-vmeshnode-pae
+include/linux/sysctl.h
+include/linux/tcp.h
+include/net/inet_connection_sock.h
+include/net/tcp.h
+kernel/sysctl_check.c
+net/ipv4/Kconfig
+net/ipv4/Makefile
+net/ipv4/sysctl_net_ipv4.c
+net/ipv4/tcp.c
+net/ipv4/tcp_ancr.c
+net/ipv4/tcp_cong.c
+net/ipv4/tcp_input.c
+net/ipv4/tcp_ipv4.c
+net/ipv4/tcp_leungma.c
+net/ipv4/tcp_minisocks.c
+net/ipv4/tcp_ncr.c
+net/ipv4/tcp_noreor.c
+net/ipv4/tcp_output.c
+net/ipv4/tcp_reorder.c
+net/ipv4/tcp_timer.c
+net/ipv6/tcp_ipv6.c

net/ipv4/Kconfig

@@ -627,3 +627,74 @@ config TCP_MD5SIG
 
 	  If unsure, say N.
 
+menuconfig TCP_REORDER_ADVANCED
+	bool "TCP: advanced reordering algorithms"
+	---help---
+	  Support for selection of various TCP reordering algorithms.
+
+	  Nearly all users can safely say no here, and a safe default
+	  selection will be made (native Linux algorithm).
+
+	  If unsure, say N.
+
+if TCP_REORDER_ADVANCED
+
+config TCP_REORDER_NCR
+	tristate "TCP-NCR"
+	depends on EXPERIMENTAL
+	default n
+	---help---
+	TCP-NCR improves the robustness of TCP to non-congestion events.
+	The algorithm is described in RFC4653
+
+	For further details see:
+	  http://tools.ietf.org/html/rfc4653
+
+config TCP_REORDER_ANCR
+	tristate "TCP-ANCR"
+	depends on EXPERIMENTAL
+	default n
+	---help---
+	TCP-ANCR improves the robustness of TCP to reordering.
+	It is based on TCP-NCR (http://tools.ietf.org/html/rfc4653)
+	and the native Linux behavior.
+
+config TCP_REORDER_LEUNGMA
+	tristate "TCP-LEUNGMA"
+	depends on EXPERIMENTAL
+	default n
+	---help---
+	TCP-LEUNGMA is based on this paper by Leung
+	and Ma (http://www.eee.hku.hk/~kcleung/papers/journals/
+	TCP_reordering:JCN_2005/JCN04-3-107.pdf)
+
+choice
+	prompt "Default TCP reordering algorithm"
+	default DEFAULT_REORDER_NATIVE
+	help
+	  Select the TCP reordering algorithm that will be used by default
+	  for all connections.
+
+	config DEFAULT_REORDER_NCR
+		bool "ncr" if TCP_REORDER_NCR=y
+
+	config DEFAULT_REORDER_ANCR
+		bool "ancr" if TCP_REORDER_ANCR=y
+
+	config DEFAULT_REORDER_LEUNGMA
+		bool "leungma" if TCP_REORDER_LEUNGMA=y
+
+	config DEFAULT_REORDER_NATIVE
+		bool "native"
+
+endchoice
+
+endif
+
+config DEFAULT_TCP_REORDER
+	string
+	default "ncr" if DEFAULT_REORDER_NCR
+	default "ancr" if DEFAULT_REORDER_ANCR
+	default "leungma" if DEFAULT_REORDER_LEUNGMA
+	default "native"
+

net/ipv4/Makefile

@@ -7,7 +7,7 @@ obj-y     := route.o inetpeer.o protocol.o \
 	     ip_output.o ip_sockglue.o inet_hashtables.o \
 	     inet_timewait_sock.o inet_connection_sock.o \
 	     tcp.o tcp_input.o tcp_output.o tcp_timer.o tcp_ipv4.o \
-	     tcp_minisocks.o tcp_cong.o \
+	     tcp_minisocks.o tcp_cong.o tcp_reorder.o \
 	     datagram.o raw.o udp.o udplite.o \
 	     arp.o icmp.o devinet.o af_inet.o  igmp.o \
 	     fib_frontend.o fib_semantics.o \
@@ -48,6 +48,10 @@ obj-$(CONFIG_TCP_CONG_SCALABLE) += tcp_scalable.o
 obj-$(CONFIG_TCP_CONG_LP) += tcp_lp.o
 obj-$(CONFIG_TCP_CONG_YEAH) += tcp_yeah.o
 obj-$(CONFIG_TCP_CONG_ILLINOIS) += tcp_illinois.o
+obj-$(CONFIG_TCP_REORDER_NCR) += tcp_ncr.o
+obj-$(CONFIG_TCP_REORDER_ANCR) += tcp_ancr.o
+obj-$(CONFIG_TCP_REORDER_LEUNGMA) += tcp_leungma.o
+#obj-$(CONFIG_TCP_REORDER_NONE) += tcp_noreor.o
 obj-$(CONFIG_NETLABEL) += cipso_ipv4.o
 
 obj-$(CONFIG_XFRM) += xfrm4_policy.o xfrm4_state.o xfrm4_input.o \