Scheduler dev

include/linux/tcp.h

@@ -199,6 +199,15 @@ static inline struct tcp_request_sock *tcp_rsk(const struct request_sock *req)
 
 struct tcp_md5sig_key;
 
+/*phuc*/
+typedef enum {
+        RIGHT_RATIO_SET,
+        LEFT_RATIO_SET,
+        SEARCH_RATE,
+	RIGHT_RATIO_FINE,
+	LEFT_RATIO_FINE
+}ratio_search_state;
+
 struct tcp_sock {
 	/* inet_connection_sock has to be the first member of tcp_sock */
 	struct inet_connection_sock	inet_conn;
@@ -273,6 +282,9 @@ struct tcp_sock {
 		u32 end_seq; /* Ending TCP sequence of the skb */
 		u32 last_delivered; /* tp->delivered at last reo_wnd adj */
 		u8 reo_wnd_steps;   /* Allowed reordering window */
+		/*Phuc*/
+		u8 reord;    /* reordering detected */
+		/****/
 #define TCP_RACK_RECOVERY_THRESH 16
 		u8 reo_wnd_persist:5, /* No. of recovery since last adj */
 		   dsack_seen:1, /* Whether DSACK seen after last adj */
@@ -494,6 +506,31 @@ struct tcp_sock {
 	char		mptcp_sched_name[MPTCP_SCHED_NAME_MAX];
 	char		mptcp_pm_name[MPTCP_PM_NAME_MAX];
 #endif /* CONFIG_MPTCP */
+/* shivanga: Ratio Scheduler Variables */
+    u32 last_probe_tstamp;
+    u32 in_probe;
+    u32 last_ac_rate;
+    u32 last_ratio;
+    u64 prev_tx_bytes;
+    u64 prev_tstamp;
+    u32 rate_est_val;
+    u32 rate_est_cnt;
+    u32 last_rate_search_start[5]; /* 5 because that's the search trigger threshold */
+    u32 init_buffer_size[2];
+    u32 last_buffer_size[2];
+    u8 buffer_threshold_cnt;
+    s32 buffer_trigger_threshold;
+    ratio_search_state search_state;
+    //u8 buf_size_acc;
+    struct sock *prev_sk;
+    unsigned int num_segments_flow_one;
+    int ratio_search_step;
+    unsigned int ratio_rate_sample;
+    bool run_started;
+    bool init_search; 
+/*Phuc: Ratio scheduler optimization*/
+    u16 head_length;
+    u32 num_acks_head[2];
 };
 
 enum tsq_enum {
@@ -536,7 +573,7 @@ struct tcp_timewait_sock {
 	/* The time we sent the last out-of-window ACK: */
 	u32			  tw_last_oow_ack_time;
 
-	int			  tw_ts_recent_stamp;
+	long			  tw_ts_recent_stamp;
 #ifdef CONFIG_TCP_MD5SIG
 	struct tcp_md5sig_key	  *tw_md5_key;
 #endif

include/net/mptcp.h

@@ -157,6 +157,7 @@ struct mptcp_options_received {
 
 struct mptcp_tcp_sock {
 	struct hlist_node node;
+	struct tcp_sock	*next;		/* Next subflow socket */
 	struct hlist_node cb_list;
 	struct mptcp_options_received rx_opt;
 
@@ -166,6 +167,7 @@ struct mptcp_tcp_sock {
 	u16	map_data_len;
 	u16	slave_sk:1,
 		fully_established:1,
+		establish_increased:1,
 		second_packet:1,
 		attached:1,
 		send_mp_fail:1,
@@ -187,7 +189,7 @@ struct mptcp_tcp_sock {
 	u8	rem_id;
 	u8	sk_err;
 
-#define MPTCP_SCHED_SIZE 16
+#define MPTCP_SCHED_SIZE 64
 	u8	mptcp_sched[MPTCP_SCHED_SIZE] __aligned(8);
 
 	int	init_rcv_wnd;
@@ -261,6 +263,8 @@ struct mptcp_sched_ops {
 struct mptcp_cb {
 	/* list of sockets in this multipath connection */
 	struct hlist_head conn_list;
+	/*phuc*/
+	struct tcp_sock *connection_list;
 	/* list of sockets that need a call to release_cb */
 	struct hlist_head callback_list;
 
@@ -285,9 +289,13 @@ struct mptcp_cb {
 		rcv_hiseq_index:1, /* Index in rcv_high_order of rcv_nxt */
 		tcp_ca_explicit_set:1; /* was meta CC set by app? */
 
+	/* socket count in this connection */
+	u8 cnt_subflows;
+	u8 cnt_established;
 #define MPTCP_SCHED_DATA_SIZE 8
 	u8 mptcp_sched[MPTCP_SCHED_DATA_SIZE] __aligned(8);
-	const struct mptcp_sched_ops *sched_ops;
+	//const struct mptcp_sched_ops *sched_ops;
+	struct mptcp_sched_ops *sched_ops;
 
 	struct sk_buff_head reinject_queue;
 	/* First cache-line boundary is here minus 8 bytes. But from the
@@ -343,6 +351,11 @@ struct mptcp_cb {
 	u32 orig_window_clamp;
 
 	struct tcp_info	*master_info;
+
+	/* swetank */
+	u32 cnt_in_order;
+	u32 cnt_out_of_order;
+	/* end:swetankk */
 };
 
 #define MPTCP_VERSION_0 0
@@ -459,6 +472,18 @@ extern bool mptcp_init_failed;
 #define MPTCPHDR_INF		0x08
 #define MPTCP_REINJECT		0x10 /* Did we reinject this segment? */
 
+/* swetankk */
+#define MPTCP_SCHED_PROBE
+#ifdef CONFIG_MPTCP_QUEUE_PROBE
+    #define MPTCP_QUEUE_PROBE
+#endif
+
+#ifdef MPTCP_QUEUE_PROBE
+	#define MPTCP_RCV_QUEUE 0x00
+	#define MPTCP_OFO_QUEUE 0x01
+#endif
+/* end: swetankk */
+
 struct mptcp_option {
 	__u8	kind;
 	__u8	len;
@@ -647,6 +672,35 @@ struct mp_prio {
 	__u8	addr_id;
 } __attribute__((__packed__));
 
+/* swetankk */
+#ifdef MPTCP_SCHED_PROBE
+struct mptcp_sched_probe {
+    unsigned long id;
+    struct sock *sk;
+    bool selector_reject;
+    bool found_unused_reject;
+    bool def_unavailable;
+    bool temp_unavailable;
+	bool srtt_reject;
+    bool selected;
+    int split;
+    int skblen;
+    u32 tx_bytes;
+    u32 trans_start;
+};
+#endif
+#ifdef MPTCP_QUEUE_PROBE
+struct mptcp_queue_probe {
+	u8 q_id;
+	struct tcp_sock *meta_tp;
+	u32 skb_seq;
+	u32 skb_end_seq;
+	u8 op_id;
+	u32 q_size;
+};
+#endif
+/* end:swetankk */
+
 static inline int mptcp_sub_len_dss(const struct mp_dss *m, const int csum)
 {
 	return 4 + m->A * (4 + m->a * 4) + m->M * (10 + m->m * 4 + csum * 2);
@@ -659,6 +713,22 @@ extern int sysctl_mptcp_version;
 extern int sysctl_mptcp_checksum;
 extern int sysctl_mptcp_debug;
 extern int sysctl_mptcp_syn_retries;
+/* swetankk */
+extern int sysctl_mptcp_scheduler_optimizations_disabled;
+extern int sysctl_mptcp_set_backup;
+/* end: swetankk */
+/* shivanga */
+extern int sysctl_num_segments_flow_one;
+extern int sysctl_mptcp_rate_sample;
+extern int sysctl_mptcp_ratio_trigger_search;
+extern int sysctl_mptcp_ratio_search_step;
+extern int sysctl_mptcp_trigger_threshold;
+extern int sysctl_mptcp_probe_interval_secs;
+extern int sysctl_mptcp_ratio_static;
+
+#define MPTCP_BYTES_NOT_SENT_MAX 165
+//#define MPTCP_RATE_MAX 50
+/* end: shivanga */
 
 extern struct workqueue_struct *mptcp_wq;
 
@@ -668,6 +738,7 @@ extern struct workqueue_struct *mptcp_wq;
 			pr_err(fmt, ##args);					\
 	} while (0)
 
+/* Iterates over all subflows */
 static inline struct sock *mptcp_to_sock(const struct mptcp_tcp_sock *mptcp)
 {
 	return (struct sock *)mptcp->tp;
@@ -676,7 +747,6 @@ static inline struct sock *mptcp_to_sock(const struct mptcp_tcp_sock *mptcp)
 #define mptcp_for_each_sub(__mpcb, __mptcp)					\
 	hlist_for_each_entry_rcu(__mptcp, &((__mpcb)->conn_list), node)
 
-/* Must be called with the appropriate lock held */
 #define mptcp_for_each_sub_safe(__mpcb, __mptcp, __tmp)				\
 	hlist_for_each_entry_safe(__mptcp, __tmp, &((__mpcb)->conn_list), node)
 
@@ -903,6 +973,15 @@ bool subflow_is_backup(const struct tcp_sock *tp);
 struct sock *get_available_subflow(struct sock *meta_sk, struct sk_buff *skb,
 				   bool zero_wnd_test);
 extern struct mptcp_sched_ops mptcp_sched_default;
+/* swetankk */
+#ifdef MPTCP_SCHED_PROBE
+extern void mptcp_sched_probe_init(struct mptcp_sched_probe *sprobe);
+extern struct mptcp_sched_probe* mptcp_sched_probe_log_hook(struct mptcp_sched_probe* sprobe, bool selected, unsigned long sched_probe_id, struct sock *sk);
+#endif
+#ifdef MPTCP_QUEUE_PROBE
+extern struct mptcp_queue_probe* mptcp_queue_probe_log_hook(u8 q_id, struct tcp_sock *meta_tp, struct sk_buff *skb, u8 op_id);
+#endif
+/* end: swetankk */
 
 /* Initializes function-pointers and MPTCP-flags */
 static inline void mptcp_init_tcp_sock(struct sock *sk)

make_mptcp.sh

@@ -0,0 +1,4 @@
+make -j8
+make modules -j8
+make modules_install 
+make install

make_mptcp_ratio_only.sh

@@ -0,0 +1,9 @@
+echo "Unloading mptcp_ratio..."
+rmmod mptcp_ratio
+echo "Buiding mptcp submodules..."
+make -C . M=net/mptcp
+echo "Copying mptcp_ratio.ko..."
+cp net/mptcp/mptcp_ratio.ko /lib/modules/4.19.224/kernel/net/mptcp/mptcp_ratio.ko
+echo "Reloading mptcp_ratio..."
+modprobe mptcp_ratio
+echo "Done!"

net/Kconfig

@@ -340,6 +340,50 @@ config NET_PKTGEN
 	  To compile this code as a module, choose M here: the
 	  module will be called pktgen.
 
+config NET_TCPPROBE
+	tristate "TCP connection probing"
+	depends on INET && PROC_FS && KPROBES
+	---help---
+	This module allows for capturing the changes to TCP connection
+	state in response to incoming packets. It is used for debugging
+	TCP congestion avoidance modules. If you don't understand
+	what was just said, you don't need it: say N.
+
+	Documentation on how to use TCP connection probing can be found
+	at:
+
+	  http://www.linuxfoundation.org/collaborate/workgroups/networking/tcpprobe
+
+	To compile this code as a module, choose M here: the
+	module will be called tcp_probe.
+
+config NET_MPTCP_SCHED_PROBE
+	tristate "MPTCP scheduler probing"
+	depends on INET && PROC_FS && KPROBES && (MPTCP=y)
+	---help---
+	This module allows for probing the MPTCP scheduler decision and state. For 
+	now, only the RTT scheduler is supported. If you don't understand what was 
+	just said, you don't need it: say N. Design heavily borrowed from the famous
+	tcp_probe module: 
+
+		http://www.linuxfoundation.org/collaborate/workgroups/networking/tcpprobe
+
+	To compile this code as a module, choose M here: the module will be called
+	mptcp_sched_probe.
+
+config NET_MPTCP_QUEUE_PROBE
+    tristate "MPTCP meta recv and ofo queue probing"
+    depends on INET && PROC_FS && KPROBES && (MPTCP=y) && (MPTCP_QUEUE_PROBE=y)
+    ---help---
+    This module allows for probing the MPTCP meta-level receive and out-of-order
+	queues. If you don't understand what was just said, you don't need it: say N. 
+	Design heavily borrowed from the famous tcp_probe module:
+
+        http://www.linuxfoundation.org/collaborate/workgroups/networking/tcpprobe
+
+    To compile this code as a module, choose M here: the module will be called
+    mptcp_queue_probe.
+
 config NET_DROP_MONITOR
 	tristate "Network packet drop alerting service"
 	depends on INET && TRACEPOINTS

net/mptcp/Kconfig

@@ -109,6 +109,20 @@ config MPTCP_REDUNDANT
 	  This scheduler sends all packets redundantly over all subflows to decreases
 	  latency and jitter on the cost of lower throughput.
 
+config MPTCP_RTT
+    tristate "MPTCP RTT"
+    depends on (MPTCP=y)
+    ---help---
+      This is a round-trip time (RTT) based scheduler (copy of the default one). It sends 
+	  packets to the flow with the lower RTT.
+
+config MPTCP_RATIO
+    tristate "MPTCP Ratio"
+    depends on (MPTCP=y)
+    ---help---
+      This is a quota based scheduler that assigns packets to subflows based on given ratio.
+      It based on the design of the ROUNDROBIN scheduler which gives equal quota to all subflows.
+
 choice
 	prompt "Default MPTCP Scheduler"
 	default DEFAULT_SCHEDULER
@@ -133,6 +147,18 @@ choice
 		  This is the redundant scheduler, sending packets redundantly over
 		  all the subflows.
 
+	config DEFAULT_RTT
+         bool "RTT" if MPTCP_RTT=y
+         ---help---
+         This is a copy of default RTT-based scheduler. The only difference is that
+		 it is built as a seperate module, making it easier to make changes to it.
+
+    config DEFAULT_RATIO
+         bool "Ratio" if MPTCP_RATIO=y
+         ---help---
+         This is a quota based scheduler that assigns packets to subflows based on given ratio.
+         It based on the design of the ROUNDROBIN scheduler which gives equal quota to all subflows.
+
 endchoice
 endif
 
@@ -142,5 +168,12 @@ config DEFAULT_MPTCP_SCHED
 	default "default" if DEFAULT_SCHEDULER
 	default "roundrobin" if DEFAULT_ROUNDROBIN
 	default "redundant" if DEFAULT_REDUNDANT
+	default "rtt" if DEFAULT_RTT
+    	default "ratio" if DEFAULT_RATIO
 	default "default"
 
+config MPTCP_QUEUE_PROBE
+        bool "MPTCP Meta-Queues Instrumentation"
+        depends on (MPTCP=y)
+        ---help---
+          This adds instrumentation in the MPTCP stack to trace meta rcv and ofo queues.

net/mptcp/Makefile

@@ -6,7 +6,7 @@
 obj-$(CONFIG_MPTCP) += mptcp.o
 
 mptcp-y := mptcp_ctrl.o mptcp_ipv4.o mptcp_pm.o \
-	   mptcp_output.o mptcp_input.o mptcp_sched.o
+	   mptcp_output.o mptcp_input.o mptcp_sched.o mptcp_sched_probe_log_hook.o
 
 obj-$(CONFIG_TCP_CONG_LIA) += mptcp_coupled.o
 obj-$(CONFIG_TCP_CONG_OLIA) += mptcp_olia.o
@@ -20,5 +20,9 @@ obj-$(CONFIG_MPTCP_NETLINK) += mptcp_netlink.o
 obj-$(CONFIG_MPTCP_ROUNDROBIN) += mptcp_rr.o
 obj-$(CONFIG_MPTCP_REDUNDANT) += mptcp_redundant.o
 obj-$(CONFIG_MPTCP_BLEST) += mptcp_blest.o
+obj-$(CONFIG_MPTCP_RTT) += mptcp_rtt.o
+obj-$(CONFIG_MPTCP_RATIO) += mptcp_ratio.o
+obj-$(CONFIG_NET_MPTCP_SCHED_PROBE) += mptcp_sched_probe.o
+obj-$(CONFIG_NET_MPTCP_QUEUE_PROBE) += mptcp_queue_probe.o
 
 mptcp-$(subst m,y,$(CONFIG_IPV6)) += mptcp_ipv6.o