sctp: Update SWS avaoidance receiver side algorithm

We currently send window update SACKs every time we free up 1 PMTU
worth of data.  That a lot more SACKs then necessary.  Instead, we'll
now send back the actuall window every time we send a sack, and do
window-update SACKs when a fraction of the receive buffer has been
opened.  The fraction is controlled with a sysctl.

Signed-off-by: Vlad Yasevich <vladislav.yasevich@hp.com>

diff --git a/include/net/sctp/constants.h b/include/net/sctp/constants.h
index 58f714a..6390884 100644 (file)
--- a/include/net/sctp/constants.h
+++ b/include/net/sctp/constants.h
@@ -308,6 +308,10 @@ enum { SCTP_MAX_GABS = 16 };
 
 #define SCTP_DEFAULT_MINWINDOW 1500    /* default minimum rwnd size */
 #define SCTP_DEFAULT_MAXWINDOW 65535   /* default rwnd size */
+#define SCTP_DEFAULT_RWND_SHIFT  4     /* by default, update on 1/16 of
+                                        * rcvbuf, which is 1/8 of initial
+                                        * window
+                                        */
 #define SCTP_DEFAULT_MAXSEGMENT 1500   /* MTU size, this is the limit
                                          * to which we will raise the P-MTU.
                                         */

diff --git a/include/net/sctp/structs.h b/include/net/sctp/structs.h
index cd2e187..90876b6 100644 (file)
--- a/include/net/sctp/structs.h
+++ b/include/net/sctp/structs.h
@@ -231,6 +231,11 @@ extern struct sctp_globals {
        /* Flag to indicate whether computing and verifying checksum
         * is disabled. */
         int checksum_disable;
+
+       /* Threshold for rwnd update SACKS.  Receive buffer shifted this many
+        * bits is an indicator of when to send and window update SACK.
+        */
+       int rwnd_update_shift;
 } sctp_globals;
 
 #define sctp_rto_initial               (sctp_globals.rto_initial)
@@ -267,6 +272,7 @@ extern struct sctp_globals {
 #define sctp_prsctp_enable             (sctp_globals.prsctp_enable)
 #define sctp_auth_enable               (sctp_globals.auth_enable)
 #define sctp_checksum_disable          (sctp_globals.checksum_disable)
+#define sctp_rwnd_upd_shift            (sctp_globals.rwnd_update_shift)
 
 /* SCTP Socket type: UDP or TCP style. */
 typedef enum {

diff --git a/net/sctp/associola.c b/net/sctp/associola.c
index 8e755eb..37e9825 100644 (file)
--- a/net/sctp/associola.c
+++ b/net/sctp/associola.c
@@ -1383,8 +1383,9 @@ static inline int sctp_peer_needs_update(struct sctp_association *asoc)
        case SCTP_STATE_SHUTDOWN_RECEIVED:
        case SCTP_STATE_SHUTDOWN_SENT:
                if ((asoc->rwnd > asoc->a_rwnd) &&
-                   ((asoc->rwnd - asoc->a_rwnd) >=
-                    min_t(__u32, (asoc->base.sk->sk_rcvbuf >> 1), asoc->pathmtu)))
+                   ((asoc->rwnd - asoc->a_rwnd) >= max_t(__u32,
+                          (asoc->base.sk->sk_rcvbuf >> sctp_rwnd_upd_shift),
+                          asoc->pathmtu)))
                        return 1;
                break;
        default:

diff --git a/net/sctp/protocol.c b/net/sctp/protocol.c
index 08ef203..a3c8988 100644 (file)
--- a/net/sctp/protocol.c
+++ b/net/sctp/protocol.c
@@ -1258,6 +1258,9 @@ SCTP_STATIC __init int sctp_init(void)
        /* Set SCOPE policy to enabled */
        sctp_scope_policy = SCTP_SCOPE_POLICY_ENABLE;
 
+       /* Set the default rwnd update threshold */
+       sctp_rwnd_upd_shift             = SCTP_DEFAULT_RWND_SHIFT;
+
        sctp_sysctl_register();
 
        INIT_LIST_HEAD(&sctp_address_families);

diff --git a/net/sctp/sm_sideeffect.c b/net/sctp/sm_sideeffect.c
index eda4fe7..8ae6709 100644 (file)
--- a/net/sctp/sm_sideeffect.c
+++ b/net/sctp/sm_sideeffect.c
@@ -217,8 +217,7 @@ static int sctp_gen_sack(struct sctp_association *asoc, int force,
                sctp_add_cmd_sf(commands, SCTP_CMD_TIMER_RESTART,
                                SCTP_TO(SCTP_EVENT_TIMEOUT_SACK));
        } else {
-               if (asoc->a_rwnd > asoc->rwnd)
-                       asoc->a_rwnd = asoc->rwnd;
+               asoc->a_rwnd = asoc->rwnd;
                sack = sctp_make_sack(asoc);
                if (!sack)
                        goto nomem;

diff --git a/net/sctp/sysctl.c b/net/sctp/sysctl.c
index ab7151d..ae03ded 100644 (file)
--- a/net/sctp/sysctl.c
+++ b/net/sctp/sysctl.c
@@ -52,6 +52,7 @@ static int int_max = INT_MAX;
 static int sack_timer_min = 1;
 static int sack_timer_max = 500;
 static int addr_scope_max = 3; /* check sctp_scope_policy_t in include/net/sctp/constants.h for max entries */
+static int rwnd_scale_max = 16;
 
 extern int sysctl_sctp_mem[3];
 extern int sysctl_sctp_rmem[3];
@@ -284,6 +285,18 @@ static ctl_table sctp_table[] = {
                .extra1         = &zero,
                .extra2         = &addr_scope_max,
        },
+       {
+               .ctl_name       = CTL_UNNUMBERED,
+               .procname       = "rwnd_update_shift",
+               .data           = &sctp_rwnd_upd_shift,
+               .maxlen         = sizeof(int),
+               .mode           = 0644,
+               .proc_handler   = &proc_dointvec_minmax,
+               .strategy       = &sysctl_intvec,
+               .extra1         = &one,
+               .extra2         = &rwnd_scale_max,
+       },
+
        { .ctl_name = 0 }
 };