On Wednesday 23 February 2005 22:57, David S. Miller wrote:
>
> For the millionth time, you can't just delete the existing sysctls.
> That is a user visisble interface.
>
> Instead, you have to make them at least appear to select a
> single congestion control algorithm
[..]
Maybe this could be a starting point.
This version of hybla patch makes our sysctls mutually exclusive by
switching all off just before performing a "write" operation on one of
them.
However, there should be more than one cleaner way to do it ...
--
Daniele Lacamera
root at danielinux.net
diff -ruN linux-2.6.11-rc4/Documentation/networking/ip-sysctl.txt hybla-2.6.11-rc4/Documentation/networking/ip-sysctl.txt
--- linux-2.6.11-rc4/Documentation/networking/ip-sysctl.txt 2005-02-13 04:06:20.000000000 +0100
+++ hybla-2.6.11-rc4/Documentation/networking/ip-sysctl.txt 2005-02-22 13:50:38.000000000 +0100
@@ -349,6 +349,18 @@
window. Allows two flows sharing the same connection to converge
more rapidly.
Default: 1
+
+tcp_hybla - BOOLEAN
+ Enable TCP-Hybla congestion control algorithm.
+ TCP-Hybla is a sender-side only change that eliminates penalization of
+ long-RTT, large-bandwidth connections, like when satellite legs are
+ involved, expecially when sharing a common bottleneck with normal
+ terrestrial connections.
+ Default: 0
+
+tcp_hybla_rtt0 - INTEGER
+ Divisor to set up rtt0 value for hybla congestion control.
+ Default: 40 (= 1/40 sec == 25ms)
tcp_default_win_scale - INTEGER
Sets the minimum window scale TCP will negotiate for on all
diff -ruN linux-2.6.11-rc4/include/linux/sysctl.h hybla-2.6.11-rc4/include/linux/sysctl.h
--- linux-2.6.11-rc4/include/linux/sysctl.h 2005-02-13 04:06:53.000000000 +0100
+++ hybla-2.6.11-rc4/include/linux/sysctl.h 2005-02-24 01:41:45.000000000 +0100
@@ -344,6 +344,8 @@
NET_TCP_DEFAULT_WIN_SCALE=105,
NET_TCP_MODERATE_RCVBUF=106,
NET_TCP_TSO_WIN_DIVISOR=107,
+ NET_TCP_HYBLA=108,
+ NET_TCP_HYBLA_RTT0=109,
};
enum {
@@ -788,6 +790,8 @@
void __user *, size_t *, loff_t *);
extern int proc_dointvec(ctl_table *, int, struct file *,
void __user *, size_t *, loff_t *);
+extern int proc_switch_congctl(ctl_table *, int, struct file *,
+ void __user *, size_t *, loff_t *);
extern int proc_dointvec_bset(ctl_table *, int, struct file *,
void __user *, size_t *, loff_t *);
extern int proc_dointvec_minmax(ctl_table *, int, struct file *,
diff -ruN linux-2.6.11-rc4/include/linux/tcp.h hybla-2.6.11-rc4/include/linux/tcp.h
--- linux-2.6.11-rc4/include/linux/tcp.h 2005-02-13 04:06:23.000000000 +0100
+++ hybla-2.6.11-rc4/include/linux/tcp.h 2005-02-22 13:04:53.000000000 +0100
@@ -434,6 +434,16 @@
__u32 last_cwnd; /* the last snd_cwnd */
__u32 last_stamp; /* time when updated last_cwnd */
} bictcp;
+
+ /* Tcp Hybla structure. */
+ struct{
+ __u32 snd_cwnd_cents; /* Keeps increment values when it is <1, <<7 */
+ __u32 rho; /* Rho parameter, integer part */
+ __u32 rho2; /* Rho * Rho, integer part */
+ __u32 rho_3ls; /* Rho parameter, <<3 */
+ __u32 rho2_7ls; /* Rho^2, <<7 */
+ __u32 minrtt; /* Minimum smoothed round trip time value seen */
+ } hybla;
};
static inline struct tcp_sock *tcp_sk(const struct sock *sk)
diff -ruN linux-2.6.11-rc4/include/net/tcp.h hybla-2.6.11-rc4/include/net/tcp.h
--- linux-2.6.11-rc4/include/net/tcp.h 2005-02-13 04:05:28.000000000 +0100
+++ hybla-2.6.11-rc4/include/net/tcp.h 2005-02-22 16:30:05.000000000 +0100
@@ -608,6 +608,8 @@
extern int sysctl_tcp_bic_low_window;
extern int sysctl_tcp_moderate_rcvbuf;
extern int sysctl_tcp_tso_win_divisor;
+extern int sysctl_tcp_hybla;
+extern int sysctl_tcp_hybla_rtt0;
extern atomic_t tcp_memory_allocated;
extern atomic_t tcp_sockets_allocated;
@@ -2017,4 +2019,37 @@
return (cwnd != 0);
}
+
+/*
+ * TCP HYBLA Functions and constants
+ */
+/* Hybla reference round trip time (default= 1/40 sec = 25 ms), expressed in jiffies */
+#define RTT0 (__u32) ((HZ/sysctl_tcp_hybla_rtt0))
+/*
+ * This is called in tcp_ipv4.c and
+ * tcp_minisocks.c when connection starts
+ */
+static inline void init_hybla(struct tcp_sock *tp)
+{
+ tp->hybla.rho = 0;
+ tp->hybla.rho2 = 0;
+ tp->hybla.rho_3ls = 0;
+ tp->hybla.rho2_7ls = 0;
+ tp->hybla.snd_cwnd_cents = 0;
+}
+/* This is called to refresh values for hybla parameters */
+static inline void hybla_recalc_param (struct tcp_sock *tp)
+{
+
+ tp->hybla.rho_3ls = (tp->srtt / RTT0);
+ if(tp->hybla.rho_3ls < 8)
+ tp->hybla.rho_3ls =8;
+
+ tp->hybla.rho=(tp->hybla.rho_3ls >> 3);
+ tp->hybla.rho2_7ls = ((tp->hybla.rho_3ls * tp->hybla.rho_3ls)<<1);
+ tp->hybla.rho2=(tp->hybla.rho2_7ls >>7);
+
+ if (sysctl_tcp_hybla)
+ tp->snd_cwnd_clamp = min_t (__u32, tp->snd_cwnd_clamp, tp->hybla.rho<<16);
+}
#endif /* _TCP_H */
diff -ruN linux-2.6.11-rc4/kernel/sysctl.c hybla-2.6.11-rc4/kernel/sysctl.c
--- linux-2.6.11-rc4/kernel/sysctl.c 2005-02-13 04:05:27.000000000 +0100
+++ hybla-2.6.11-rc4/kernel/sysctl.c 2005-02-24 01:41:16.000000000 +0100
@@ -66,6 +66,11 @@
extern int printk_ratelimit_burst;
extern int pid_max_min, pid_max_max;
+extern int sysctl_tcp_vegas_cong_avoid;
+extern int sysctl_tcp_bic;
+extern int sysctl_tcp_westwood;
+extern int sysctl_tcp_hybla;
+
#if defined(CONFIG_X86_LOCAL_APIC) && defined(CONFIG_X86)
int unknown_nmi_panic;
extern int proc_unknown_nmi_panic(ctl_table *, int, struct file *,
@@ -1595,6 +1600,23 @@
NULL,NULL);
}
+/**
+ * Used by tcp engine to realize a multi-switch for congestion
+ * control algorithm selection.
+ */
+int proc_switch_congctl(ctl_table *table, int write, struct file *filp,
+ void __user *buffer, size_t *lenp, loff_t *ppos)
+{
+ if(write){
+ sysctl_tcp_vegas_cong_avoid=0;
+ sysctl_tcp_westwood=0;
+ sysctl_tcp_bic=0;
+ sysctl_tcp_hybla=0;
+ }
+ return do_proc_dointvec(table,write,filp,buffer,lenp,ppos,
+ NULL,NULL);
+}
+
#define OP_SET 0
#define OP_AND 1
#define OP_OR 2
@@ -2216,6 +2238,7 @@
* exception granted :-)
*/
EXPORT_SYMBOL(proc_dointvec);
+EXPORT_SYMBOL(proc_switch_congctl);
EXPORT_SYMBOL(proc_dointvec_jiffies);
EXPORT_SYMBOL(proc_dointvec_minmax);
EXPORT_SYMBOL(proc_dointvec_userhz_jiffies);
diff -ruN linux-2.6.11-rc4/net/ipv4/sysctl_net_ipv4.c hybla-2.6.11-rc4/net/ipv4/sysctl_net_ipv4.c
--- linux-2.6.11-rc4/net/ipv4/sysctl_net_ipv4.c 2005-02-13 04:07:01.000000000 +0100
+++ hybla-2.6.11-rc4/net/ipv4/sysctl_net_ipv4.c 2005-02-24 01:40:49.000000000 +0100
@@ -608,7 +608,7 @@
.data = &sysctl_tcp_westwood,
.maxlen = sizeof(int),
.mode = 0644,
- .proc_handler = &proc_dointvec,
+ .proc_handler = &proc_switch_congctl,
},
{
.ctl_name = NET_TCP_VEGAS,
@@ -616,7 +616,7 @@
.data = &sysctl_tcp_vegas_cong_avoid,
.maxlen = sizeof(int),
.mode = 0644,
- .proc_handler = &proc_dointvec,
+ .proc_handler = &proc_switch_congctl,
},
{
.ctl_name = NET_TCP_VEGAS_ALPHA,
@@ -648,7 +648,7 @@
.data = &sysctl_tcp_bic,
.maxlen = sizeof(int),
.mode = 0644,
- .proc_handler = &proc_dointvec,
+ .proc_handler = &proc_switch_congctl,
},
{
.ctl_name = NET_TCP_BIC_FAST_CONVERGENCE,
@@ -682,6 +682,22 @@
.mode = 0644,
.proc_handler = &proc_dointvec,
},
+ {
+ .ctl_name = NET_TCP_HYBLA,
+ .procname = "tcp_hybla",
+ .data = &sysctl_tcp_hybla,
+ .maxlen = sizeof(int),
+ .mode = 0644,
+ .proc_handler = &proc_switch_congctl,
+ },
+ {
+ .ctl_name = NET_TCP_HYBLA_RTT0,
+ .procname = "tcp_hybla_rtt0",
+ .data = &sysctl_tcp_hybla_rtt0,
+ .maxlen = sizeof(int),
+ .mode = 0644,
+ .proc_handler = &proc_dointvec,
+ },
{ .ctl_name = 0 }
};
diff -ruN linux-2.6.11-rc4/net/ipv4/tcp.c hybla-2.6.11-rc4/net/ipv4/tcp.c
--- linux-2.6.11-rc4/net/ipv4/tcp.c 2005-02-13 04:05:50.000000000 +0100
+++ hybla-2.6.11-rc4/net/ipv4/tcp.c 2005-02-22 16:30:41.000000000 +0100
@@ -1813,6 +1813,10 @@
if (!(sk->sk_userlocks & SOCK_BINDADDR_LOCK))
inet_reset_saddr(sk);
+
+ /* [TCP HYBLA] Reset values on disconnect */
+ if (sysctl_tcp_hybla)
+ init_hybla(tp);
sk->sk_shutdown = 0;
sock_reset_flag(sk, SOCK_DONE);
diff -ruN linux-2.6.11-rc4/net/ipv4/tcp_input.c hybla-2.6.11-rc4/net/ipv4/tcp_input.c
--- linux-2.6.11-rc4/net/ipv4/tcp_input.c 2005-02-13 04:07:01.000000000 +0100
+++ hybla-2.6.11-rc4/net/ipv4/tcp_input.c 2005-02-22 13:58:49.000000000 +0100
@@ -62,6 +62,7 @@
* engine. Lots of bugs are found.
* Pasi Sarolahti: F-RTO for dealing with spurious RTOs
* Angelo Dell'Aera: TCP Westwood+ support
+ * Daniele Lacamera: TCP Hybla Congestion control support
*/
#include <linux/config.h>
@@ -89,6 +90,8 @@
int sysctl_tcp_nometrics_save;
int sysctl_tcp_westwood;
int sysctl_tcp_vegas_cong_avoid;
+int sysctl_tcp_hybla=0;
+int sysctl_tcp_hybla_rtt0=40;
int sysctl_tcp_moderate_rcvbuf = 1;
@@ -595,6 +598,29 @@
tp->vegas.cntRTT++;
}
+/*
+ * [TCP HYBLA] Update Values, if necessary, when a new
+ * smoothed RTT Estimation becomes available
+ */
+static void hybla_update_rtt(struct tcp_sock *tp, long m)
+{
+ /* This sets rho to the smallest RTT received. */
+ if (tp->srtt!=0){
+ /* Recalculate rho only if this srtt is the lowest */
+ if (tp->srtt < tp->hybla.minrtt){
+ hybla_recalc_param(tp);
+ /* update minimum rtt */
+ tp->hybla.minrtt = tp->srtt;
+ }
+ } else {
+ /* 1st Rho measurement */
+ hybla_recalc_param(tp);
+ /* set minimum rtt as this is the 1st ever seen */
+ tp->hybla.minrtt = tp->srtt;
+ tp->snd_cwnd=tp->hybla.rho;
+ }
+}
+
/* Called to compute a smoothed rtt estimate. The data fed to this
* routine either comes from timestamps, or from segments that were
* known _not_ to have been retransmitted [see Karn/Partridge
@@ -669,6 +695,8 @@
}
tcp_westwood_update_rtt(tp, tp->srtt >> 3);
+ if(sysctl_tcp_hybla)
+ hybla_update_rtt(tp,mrtt);
}
/* Calculate rto without backoff. This is the second half of Van Jacobson's
@@ -808,6 +836,11 @@
else
cwnd = (tp->mss_cache_std > 1095) ? 3 : 4;
}
+ /* Hybla initial Window value set. */
+ if (sysctl_tcp_hybla){
+ hybla_recalc_param(tp);
+ cwnd=max_t(__u32, 2U, tp->hybla.rho);
+ }
return min_t(__u32, cwnd, tp->snd_cwnd_clamp);
}
@@ -2322,12 +2355,153 @@
tp->snd_cwnd_stamp = tcp_time_stamp;
}
+/***
+ ** TCP-HYBLA Congestion control algorithm, based on:
+ ** C.Caini, R.Firrincieli, "TCP-Hybla: A TCP Enhancement
+ ** for Heterogeneous Networks",
+ ** International Journal on satellite Communications,
+ ** September 2004
+ ***/
+static __inline__ __u32 hybla_slowstart_fraction_increment(__u32 odds){
+ switch (odds) {
+ case 0:
+ return 128;
+ case 1:
+ return 139;
+ case 2:
+ return 152;
+ case 3:
+ return 165;
+ case 4:
+ return 181;
+ case 5:
+ return 197;
+ case 6:
+ return 215;
+ case 7:
+ return 234;
+ default:
+ return 128;
+
+ }
+}
+static __inline__ void hybla_fractions_cong_avoid(struct tcp_sock *tp)
+{
+ __u32 increment;
+ __u32 odd;
+ __u32 rho_fractions;
+ //__u8 is_slowstart=0;
+ __u32 window, ssthresh;
+
+ if (tp->hybla.rho==0)
+ hybla_recalc_param(tp);
+
+ ssthresh = tp->snd_ssthresh;
+ window=tp->snd_cwnd;
+ rho_fractions=tp->hybla.rho_3ls - (tp->hybla.rho << 3);
+
+ if (window < ssthresh){
+ return;
+ } else {
+ /*** congestion avoidance
+ *** INC = RHO^2 / W
+ *** as long as increment is estimated as (rho<<7)/window
+ *** it already is <<7 and we can easily count its fractions.
+ ***/
+ increment =(tp->hybla.rho2_7ls/window);
+ odd = increment % 128;
+ tp->snd_cwnd_cnt++;
+ }
+ tp->hybla.snd_cwnd_cents += odd;
+
+}
+
+ /* TCP Hybla main routine.
+ * This is the algorithm behavior:
+ * o Recalc Hybla parameters if min_rtt has changed
+ * o Give cwnd a new value based on the model proposed
+ * o remember increments <1
+ */
+static __inline__ void tcp_hybla_cong_avoid(struct tcp_sock *tp)
+{
+ __u32 increment;
+ __u32 odd;
+ __u32 rho_fractions;
+ __u32 window,clamp, ssthresh;
+ __u8 is_slowstart=0;
+
+ if (tp->hybla.rho==0)
+ hybla_recalc_param(tp);
+
+ clamp = tp->snd_cwnd_clamp ;
+ window = tp->snd_cwnd;
+ ssthresh = tp->snd_ssthresh;
+ rho_fractions=tp->hybla.rho_3ls - (tp->hybla.rho << 3);
+
+ if (window < ssthresh){
+ /*** slow start
+ *** INC = 2^RHO - 1
+ *** This is done by splitting the rho parameter
+ *** into 2 parts: an integer part and a fraction part.
+ *** Inrement<<7 is estimated by doing:
+ *** [2^(int+fract)]<<7
+ *** that is equal to:
+ *** (2^int) * [(2^fract) <<7]
+ *** 2^int is straightly computed as 1<<int,
+ *** while we will use hybla_slowstart_fraction_increment() to
+ *** calculate 2^fract in a <<7 value.
+ ***/
+ is_slowstart=1;
+ increment =( (1 << tp->hybla.rho) * hybla_slowstart_fraction_increment(rho_fractions) ) - 128;
+ odd = increment % 128;
+ window += (increment >> 7);
+ } else {
+ /*** congestion avoidance
+ *** INC = RHO^2 / W
+ *** as long as increment is estimated as (rho<<7)/window
+ *** it already is <<7 and we can easily count its fractions.
+ ***/
+ increment =(tp->hybla.rho2_7ls/window);
+ odd = increment % 128;
+ window += (increment >> 7);
+
+ if (increment < 128)
+ tp->snd_cwnd_cnt++;
+ }
+ tp->hybla.snd_cwnd_cents += odd;
+
+ /***
+ *** check when fractions goes >=128
+ *** and increase cwnd by 1.
+ ***/
+ while( tp->hybla.snd_cwnd_cents >= 128){
+ window++;
+ tp->hybla.snd_cwnd_cents -= 128;
+ tp->snd_cwnd_cnt = 0;
+ }
+ /***
+ *** clamp down slowstart cwnd to ssthresh value.
+ ***/
+ if (is_slowstart)
+ window = min_t(__u32, window, ssthresh);
+
+ tp->snd_cwnd = min_t (__u32, window, clamp);
+
+ tp->snd_cwnd_stamp=tcp_time_stamp;
+
+}
+
static inline void tcp_cong_avoid(struct tcp_sock *tp, u32 ack, u32 seq_rtt)
{
- if (tcp_vegas_enabled(tp))
+ if (tcp_vegas_enabled(tp)){
vegas_cong_avoid(tp, ack, seq_rtt);
- else
- reno_cong_avoid(tp);
+ return;
+ }
+ if (sysctl_tcp_hybla){
+ tcp_hybla_cong_avoid(tp);
+ return;
+ }
+ reno_cong_avoid(tp);
}
/* Restart timer after forward progress on connection.
diff -ruN linux-2.6.11-rc4/net/ipv4/tcp_ipv4.c hybla-2.6.11-rc4/net/ipv4/tcp_ipv4.c
--- linux-2.6.11-rc4/net/ipv4/tcp_ipv4.c 2005-02-13 04:05:51.000000000 +0100
+++ hybla-2.6.11-rc4/net/ipv4/tcp_ipv4.c 2005-02-22 13:19:02.000000000 +0100
@@ -2055,6 +2055,9 @@
* efficiently to them. -DaveM
*/
tp->snd_cwnd = 2;
+
+ /* Reset hybla parameters on socket initialization. */
+ init_hybla(tp);
/* See draft-stevens-tcpca-spec-01 for discussion of the
* initialization of these values.
diff -ruN linux-2.6.11-rc4/net/ipv4/tcp_minisocks.c hybla-2.6.11-rc4/net/ipv4/tcp_minisocks.c
--- linux-2.6.11-rc4/net/ipv4/tcp_minisocks.c 2005-02-13 04:07:01.000000000 +0100
+++ hybla-2.6.11-rc4/net/ipv4/tcp_minisocks.c 2005-02-22 13:17:07.000000000 +0100
@@ -784,6 +784,9 @@
newtp->dsack = 0;
newtp->eff_sacks = 0;
+
+ /* Reset hybla parameters on socket initialization. */
+ init_hybla(newtp);
newtp->probes_out = 0;
newtp->num_sacks = 0;
