From: Olivier Houchard <ohouchard@haproxy.com>
Date: Fri, 2 May 2025 11:29:05 +0000 (+0000)
Subject: MEDIUM: peers: Give up if we fail to take locks in hot path
X-Git-Tag: v3.2-dev14~10
X-Git-Url: http://git.ipfire.org/cgi-bin/gitweb.cgi?a=commitdiff_plain;h=b3ad7b6371e10cee4e0c23f80191e62d1da7415f;p=thirdparty%2Fhaproxy.git

MEDIUM: peers: Give up if we fail to take locks in hot path

In peer_send_msgs(), give up in order to retry later if we failed at
getting the update read lock.
Similarly, in __process_running_peer_sync(), give up and just reschedule
the task if we failed to get the peer lock. There is an heavy contention
on both those locks, so we could spend a lot of time trying to get them.
This helps getting peers perform better under heavy load.
---

diff --git a/src/peers.c b/src/peers.c
index abc00026e..f13d9d393 100644
--- a/src/peers.c
+++ b/src/peers.c
@@ -2639,7 +2639,10 @@ static inline int peer_send_msgs(struct appctx *appctx,
 			if (!(peer->flags & PEER_F_TEACH_PROCESS)) {
 				int must_send;
 
-				HA_RWLOCK_RDLOCK(STK_TABLE_UPDT_LOCK, &st->table->updt_lock);
+				if (HA_RWLOCK_TRYRDLOCK(STK_TABLE_UPDT_LOCK, &st->table->updt_lock)) {
+					applet_have_more_data(appctx);
+					return -1;
+				}
 				must_send = (peer->learnstate == PEER_LR_ST_NOTASSIGNED) && (st->last_pushed != st->table->localupdate);
 				HA_RWLOCK_RDUNLOCK(STK_TABLE_UPDT_LOCK, &st->table->updt_lock);
 
@@ -3413,6 +3416,7 @@ static void __process_running_peer_sync(struct task *task, struct peers *peers,
 {
 	struct peer *peer;
 	struct shared_table *st;
+	int must_resched = 0;
 
 	/* resync timeout set to TICK_ETERNITY means we just start
 	 * a new process and timer was not initialized.
@@ -3440,7 +3444,10 @@ static void __process_running_peer_sync(struct task *task, struct peers *peers,
 
 	/* For each session */
 	for (peer = peers->remote; peer; peer = peer->next) {
-		HA_SPIN_LOCK(PEER_LOCK, &peer->lock);
+		if (HA_SPIN_TRYLOCK(PEER_LOCK, &peer->lock) != 0) {
+			must_resched = 1;
+			continue;
+		}
 
 		sync_peer_learn_state(peers, peer);
 		sync_peer_app_state(peers, peer);
@@ -3567,12 +3574,13 @@ static void __process_running_peer_sync(struct task *task, struct peers *peers,
 		HA_ATOMIC_OR(&peers->flags, PEERS_F_RESYNC_REMOTE_FINISHED|PEERS_F_DBG_RESYNC_REMOTETIMEOUT);
 	}
 
-	if ((peers->flags & PEERS_RESYNC_STATEMASK) != PEERS_RESYNC_FINISHED) {
+	if (!must_resched && (peers->flags & PEERS_RESYNC_STATEMASK) != PEERS_RESYNC_FINISHED) {
 		/* Resync not finished*/
 		/* reschedule task to resync timeout if not expired, to ended resync if needed */
 		if (!tick_is_expired(peers->resync_timeout, now_ms))
 			task->expire = tick_first(task->expire, peers->resync_timeout);
-	}
+	} else if (must_resched)
+		task_wakeup(task, TASK_WOKEN_OTHER);
 }
 
 /* Process the sync task for a stopping process. It is called from process_peer_sync() only */