LU-15102 lnet: Reset ni_ping_count only on receive

[fs/lustre-release.git] / lnet / lnet / net_fault.c

diff --git a/lnet/lnet/net_fault.c b/lnet/lnet/net_fault.c
index ce68063..efe69d9 100644 (file)
--- a/lnet/lnet/net_fault.c
+++ b/lnet/lnet/net_fault.c
@@ -25,7 +25,6 @@
  */
 /*
  * This file is part of Lustre, http://www.lustre.org/
- * Lustre is a trademark of Sun Microsystems, Inc.
  *
  * lnet/lnet/net_fault.c
  *
@@ -88,7 +87,7 @@ lnet_fault_attr_match(struct lnet_fault_attr *attr, lnet_nid_t src,
            !lnet_fault_nid_match(attr->fa_local_nid, local_nid))
                return false;
 
-       if (!(attr->fa_msg_mask & (1 << type)))
+       if (!(attr->fa_msg_mask & BIT(type)))
                return false;
 
        /* NB: ACK and REPLY have no portal, but they should have been
@@ -202,12 +201,10 @@ lnet_drop_rule_del(lnet_nid_t src, lnet_nid_t dst)
 {
        struct lnet_drop_rule *rule;
        struct lnet_drop_rule *tmp;
-       struct list_head       zombies;
-       int                    n = 0;
+       LIST_HEAD(zombies);
+       int n = 0;
        ENTRY;
 
-       INIT_LIST_HEAD(&zombies);
-
        lnet_net_lock(LNET_LOCK_EX);
        list_for_each_entry_safe(rule, tmp, &the_lnet.ln_drop_rules, dr_link) {
                if (rule->dr_attr.fa_src != src && src != 0)
@@ -315,7 +312,7 @@ lnet_fault_match_health(enum lnet_msg_hstatus *hstatus, __u32 mask)
                return;
        }
 
-       if (mask & (1 << choice)) {
+       if (mask & BIT(choice)) {
                *hstatus = choice;
                return;
        }
@@ -324,7 +321,7 @@ lnet_fault_match_health(enum lnet_msg_hstatus *hstatus, __u32 mask)
        i = HSTATUS_END;
        best_delta = HSTATUS_END;
        while (i > 0) {
-               if (mask & (1 << i)) {
+               if (mask & BIT(i)) {
                        delta = choice - i;
                        if (delta < 0)
                                delta *= -1;
@@ -537,8 +534,9 @@ delay_rule_match(struct lnet_delay_rule *rule, lnet_nid_t src,
                lnet_nid_t dst, unsigned int type, unsigned int portal,
                struct lnet_msg *msg)
 {
-       struct lnet_fault_attr  *attr = &rule->dl_attr;
-       bool                     delay;
+       struct lnet_fault_attr *attr = &rule->dl_attr;
+       bool delay;
+       time64_t now = ktime_get_seconds();
 
        if (!lnet_fault_attr_match(attr, src, LNET_NID_ANY,
                                   dst, type, portal))
@@ -547,8 +545,6 @@ delay_rule_match(struct lnet_delay_rule *rule, lnet_nid_t src,
        /* match this rule, check delay rate now */
        spin_lock(&rule->dl_lock);
        if (rule->dl_delay_time != 0) { /* time based delay */
-               time64_t now = ktime_get_seconds();
-
                rule->dl_stat.fs_count++;
                delay = now >= rule->dl_delay_time;
                if (delay) {
@@ -590,11 +586,11 @@ delay_rule_match(struct lnet_delay_rule *rule, lnet_nid_t src,
        rule->dl_stat.u.delay.ls_delayed++;
 
        list_add_tail(&msg->msg_list, &rule->dl_msg_list);
-       msg->msg_delay_send = ktime_get_seconds() + attr->u.delay.la_latency;
+       msg->msg_delay_send = now + attr->u.delay.la_latency;
        if (rule->dl_msg_send == -1) {
                rule->dl_msg_send = msg->msg_delay_send;
                mod_timer(&rule->dl_timer,
-                         jiffies + cfs_time_seconds(rule->dl_msg_send));
+                         jiffies + cfs_time_seconds(attr->u.delay.la_latency));
        }
 
        spin_unlock(&rule->dl_lock);
@@ -640,7 +636,7 @@ delayed_msg_check(struct lnet_delay_rule *rule, bool all,
        struct lnet_msg *tmp;
        time64_t now = ktime_get_seconds();
 
-       if (!all && cfs_time_seconds(rule->dl_msg_send) > now)
+       if (!all && rule->dl_msg_send > now)
                return;
 
        spin_lock(&rule->dl_lock);
@@ -663,7 +659,8 @@ delayed_msg_check(struct lnet_delay_rule *rule, bool all,
                                 struct lnet_msg, msg_list);
                rule->dl_msg_send = msg->msg_delay_send;
                mod_timer(&rule->dl_timer,
-                         jiffies + cfs_time_seconds(rule->dl_msg_send));
+                         jiffies +
+                         cfs_time_seconds(msg->msg_delay_send - now));
        }
        spin_unlock(&rule->dl_lock);
 }
@@ -679,6 +676,20 @@ delayed_msg_process(struct list_head *msg_list, bool drop)
                int             rc;
 
                msg = list_entry(msg_list->next, struct lnet_msg, msg_list);
+
+               if (msg->msg_sending) {
+                       /* Delayed send */
+                       list_del_init(&msg->msg_list);
+                       ni = msg->msg_txni;
+                       CDEBUG(D_NET, "TRACE: msg %p %s -> %s : %s\n", msg,
+                              libcfs_nidstr(&ni->ni_nid),
+                              libcfs_nidstr(&msg->msg_txpeer->lpni_nid),
+                              lnet_msgtyp2str(msg->msg_type));
+                       lnet_ni_send(ni, msg);
+                       continue;
+               }
+
+               /* Delayed receive */
                LASSERT(msg->msg_rxpeer != NULL);
                LASSERT(msg->msg_rxni != NULL);
 
@@ -703,6 +714,7 @@ delayed_msg_process(struct list_head *msg_list, bool drop)
                        case LNET_CREDIT_OK:
                                lnet_ni_recv(ni, msg->msg_private, msg, 0,
                                             0, msg->msg_len, msg->msg_len);
+                               /* fallthrough */
                        case LNET_CREDIT_WAIT:
                                continue;
                        default: /* failures */
@@ -723,10 +735,9 @@ delayed_msg_process(struct list_head *msg_list, bool drop)
 void
 lnet_delay_rule_check(void)
 {
-       struct lnet_delay_rule  *rule;
-       struct list_head         msgs;
+       struct lnet_delay_rule *rule;
+       LIST_HEAD(msgs);
 
-       INIT_LIST_HEAD(&msgs);
        while (1) {
                if (list_empty(&delay_dd.dd_sched_rules))
                        break;
@@ -888,16 +899,13 @@ int
 lnet_delay_rule_del(lnet_nid_t src, lnet_nid_t dst, bool shutdown)
 {
        struct lnet_delay_rule *rule;
-       struct lnet_delay_rule  *tmp;
-       struct list_head        rule_list;
-       struct list_head        msg_list;
-       int                     n = 0;
-       bool                    cleanup;
+       struct lnet_delay_rule *tmp;
+       LIST_HEAD(rule_list);
+       LIST_HEAD(msg_list);
+       int n = 0;
+       bool cleanup;
        ENTRY;
 
-       INIT_LIST_HEAD(&rule_list);
-       INIT_LIST_HEAD(&msg_list);
-
        if (shutdown)
                src = dst = 0;
 
@@ -1080,10 +1088,10 @@ lnet_fault_ctl(int opc, struct libcfs_ioctl_data *data)
 int
 lnet_fault_init(void)
 {
-       CLASSERT(LNET_PUT_BIT == 1 << LNET_MSG_PUT);
-       CLASSERT(LNET_ACK_BIT == 1 << LNET_MSG_ACK);
-       CLASSERT(LNET_GET_BIT == 1 << LNET_MSG_GET);
-       CLASSERT(LNET_REPLY_BIT == 1 << LNET_MSG_REPLY);
+       BUILD_BUG_ON(LNET_PUT_BIT != BIT(LNET_MSG_PUT));
+       BUILD_BUG_ON(LNET_ACK_BIT != BIT(LNET_MSG_ACK));
+       BUILD_BUG_ON(LNET_GET_BIT != BIT(LNET_MSG_GET));
+       BUILD_BUG_ON(LNET_REPLY_BIT != BIT(LNET_MSG_REPLY));
 
        mutex_init(&delay_dd.dd_mutex);
        spin_lock_init(&delay_dd.dd_lock);