WLAN subsystem: Sysctl support for key TCP/IP parameters

It has been observed that default values for some of key tcp/ip
parameters are affecting the tput/performance of the system. Hence
extending configuration capabilities to TCP/Ip stack through
sysctl interface

Change-Id: I4287e9103769535f43e0934bac08435a524ee6a4
CRs-Fixed: 507581
Signed-off-by: Ravi Joshi <ravij@codeaurora.org>

include/net/tcp.h

@@ -141,6 +141,9 @@ extern void tcp_time_wait(struct sock *sk, int state, int timeo);
 						 * most likely due to retrans in 3WHS.
 						 */
 
+/* Number of full MSS to receive before Acking RFC2581 */
+#define TCP_DELACK_SEG          1
+
 #define TCP_RESOURCE_PROBE_INTERVAL ((unsigned)(HZ/2U)) /* Maximal interval between probes
 					                 * for local resources.
 					                 */
@@ -289,6 +292,11 @@ extern int sysctl_tcp_limit_output_bytes;
 extern int sysctl_tcp_challenge_ack_limit;
 
 extern atomic_long_t tcp_memory_allocated;
+
+/* sysctl variables for controlling various tcp parameters */
+extern int sysctl_tcp_delack_seg;
+extern int sysctl_tcp_use_userconfig;
+
 extern struct percpu_counter tcp_sockets_allocated;
 extern int tcp_memory_pressure;
 
@@ -386,6 +394,12 @@ extern ssize_t tcp_splice_read(struct socket *sk, loff_t *ppos,
 			       struct pipe_inode_info *pipe, size_t len,
 			       unsigned int flags);
 
+/* sysctl master controller */
+extern int tcp_use_userconfig_sysctl_handler(struct ctl_table *, int,
+				void __user *, size_t *, loff_t *);
+extern int tcp_proc_delayed_ack_control(struct ctl_table *, int,
+				void __user *, size_t *, loff_t *);
+
 static inline void tcp_dec_quickack_mode(struct sock *sk,
 					 const unsigned int pkts)
 {

net/ipv4/sysctl_net_ipv4.c

@@ -38,6 +38,10 @@ static int ip_ttl_min = 1;
 static int ip_ttl_max = 255;
 static int ip_ping_group_range_min[] = { 0, 0 };
 static int ip_ping_group_range_max[] = { GID_T_MAX, GID_T_MAX };
+static int tcp_delack_seg_min = TCP_DELACK_MIN;
+static int tcp_delack_seg_max = 60;
+static int tcp_use_userconfig_min;
+static int tcp_use_userconfig_max = 1;
 
 /* Update system visible IP port range */
 static void set_local_port_range(int range[2])
@@ -771,6 +775,25 @@ static struct ctl_table ipv4_table[] = {
 		.proc_handler	= proc_dointvec_minmax,
 		.extra1		= &one
 	},
+	{
+		.procname	= "tcp_delack_seg",
+		.data		= &sysctl_tcp_delack_seg,
+		.maxlen		= sizeof(sysctl_tcp_delack_seg),
+		.mode		= 0644,
+		.proc_handler	= tcp_proc_delayed_ack_control,
+		.extra1		= &tcp_delack_seg_min,
+		.extra2		= &tcp_delack_seg_max,
+	},
+	{
+		.procname       = "tcp_use_userconfig",
+		.data           = &sysctl_tcp_use_userconfig,
+		.maxlen         = sizeof(sysctl_tcp_use_userconfig),
+		.mode           = 0644,
+		.proc_handler   = tcp_use_userconfig_sysctl_handler,
+		.extra1		= &tcp_use_userconfig_min,
+		.extra2		= &tcp_use_userconfig_max,
+	},
+
 	{ }
 };
 

net/ipv4/tcp.c

@@ -298,6 +298,13 @@ EXPORT_SYMBOL(sysctl_tcp_wmem);
 atomic_long_t tcp_memory_allocated;	/* Current allocated memory. */
 EXPORT_SYMBOL(tcp_memory_allocated);
 
+int sysctl_tcp_delack_seg __read_mostly = TCP_DELACK_SEG;
+EXPORT_SYMBOL(sysctl_tcp_delack_seg);
+
+int sysctl_tcp_use_userconfig __read_mostly;
+EXPORT_SYMBOL(sysctl_tcp_use_userconfig);
+
+
 /*
  * Current number of TCP sockets.
  */
@@ -1336,8 +1343,11 @@ void tcp_cleanup_rbuf(struct sock *sk, int copied)
 		   /* Delayed ACKs frequently hit locked sockets during bulk
 		    * receive. */
 		if (icsk->icsk_ack.blocked ||
-		    /* Once-per-two-segments ACK was not sent by tcp_input.c */
-		    tp->rcv_nxt - tp->rcv_wup > icsk->icsk_ack.rcv_mss ||
+		    /* Once-per-sysctl_tcp_delack_seg segments
+			  * ACK was not sent by tcp_input.c
+			  */
+		    tp->rcv_nxt - tp->rcv_wup > (icsk->icsk_ack.rcv_mss) *
+						sysctl_tcp_delack_seg ||
 		    /*
 		     * If this read emptied read buffer, we send ACK, if
 		     * connection is not bidirectional, user drained

net/ipv4/tcp_input.c

@@ -4757,7 +4757,8 @@ static void __tcp_ack_snd_check(struct sock *sk, int ofo_possible)
 	struct tcp_sock *tp = tcp_sk(sk);
 
 	    /* More than one full frame received... */
-	if (((tp->rcv_nxt - tp->rcv_wup) > inet_csk(sk)->icsk_ack.rcv_mss &&
+	if (((tp->rcv_nxt - tp->rcv_wup) > (inet_csk(sk)->icsk_ack.rcv_mss) *
+					sysctl_tcp_delack_seg &&
 	     /* ... and right edge of window advances far enough.
 	      * (tcp_recvmsg() will send ACK otherwise). Or...
 	      */

net/ipv4/tcp_timer.c

@@ -32,6 +32,42 @@ int sysctl_tcp_retries2 __read_mostly = TCP_RETR2;
 int sysctl_tcp_orphan_retries __read_mostly;
 int sysctl_tcp_thin_linear_timeouts __read_mostly;
 
+static void tcp_write_timer(unsigned long);
+static void tcp_delack_timer(unsigned long);
+static void tcp_keepalive_timer(unsigned long data);
+
+/*Function to reset tcp_ack related sysctl on resetting master control */
+void set_tcp_default(void)
+{
+	sysctl_tcp_delack_seg	= TCP_DELACK_SEG;
+}
+
+/*sysctl handler for tcp_ack realted master control */
+int tcp_proc_delayed_ack_control(ctl_table *table, int write,
+			void __user *buffer, size_t *length, loff_t *ppos)
+{
+	int ret = proc_dointvec_minmax(table, write, buffer, length, ppos);
+
+	/* The ret value will be 0 if the input validation is successful
+	 * and the values are written to sysctl table. If not, the stack
+	 * will continue to work with currently configured values
+	 */
+	return ret;
+}
+
+/*sysctl handler for tcp_ack realted master control */
+int tcp_use_userconfig_sysctl_handler(ctl_table *table, int write,
+			void __user *buffer, size_t *length, loff_t *ppos)
+{
+	int ret = proc_dointvec_minmax(table, write, buffer, length, ppos);
+
+	if (write && ret == 0) {
+		if (!sysctl_tcp_use_userconfig)
+			set_tcp_default();
+	}
+	return ret;
+}
+
 static void tcp_write_err(struct sock *sk)
 {
 	sk->sk_err = sk->sk_err_soft ? : ETIMEDOUT;