Optimize txg_kick() process (#12274)

Use dp_dirty_pertxg[] for txg_kick(), instead of dp_dirty_total in
original code. Extra parameter "txg" is added for txg_kick(), thus it
knows which txg to kick. Also txg_kick() call is moved from
dsl_pool_need_dirty_delay() to dsl_pool_dirty_space() so that we can
know the txg number assigned for txg_kick().

Some unnecessary code regarding dp_dirty_total in txg_sync_thread() is
also cleaned up.

Reviewed-by: Brian Behlendorf <[email protected]>
Reviewed-by: Matthew Ahrens <[email protected]>
Reviewed-by: Alexander Motin <[email protected]>
Signed-off-by: jxdking <[email protected]>
Closes #12274

3 files changed, 34 insertions, 29 deletions

diff --git a/include/sys/txg.h b/include/sys/txg.h
index 22158bd1a..f38f0006c 100644
--- a/include/sys/txg.h
+++ b/include/sys/txg.h
@@ -78,7 +78,7 @@ extern void txg_register_callbacks(txg_handle_t *txghp, list_t *tx_callbacks);
 
 extern void txg_delay(struct dsl_pool *dp, uint64_t txg, hrtime_t delta,
     hrtime_t resolution);
-extern void txg_kick(struct dsl_pool *dp);
+extern void txg_kick(struct dsl_pool *dp, uint64_t txg);
 
 /*
  * Wait until the given transaction group has finished syncing.
diff --git a/module/zfs/dsl_pool.c b/module/zfs/dsl_pool.c
index 456ef5372..ba1cb96a7 100644
--- a/module/zfs/dsl_pool.c
+++ b/module/zfs/dsl_pool.c
@@ -902,18 +902,26 @@ dsl_pool_need_dirty_delay(dsl_pool_t *dp)
 {
 	uint64_t delay_min_bytes =
 	    zfs_dirty_data_max * zfs_delay_min_dirty_percent / 100;
-	uint64_t dirty_min_bytes =
-	    zfs_dirty_data_max * zfs_dirty_data_sync_percent / 100;
-	uint64_t dirty;
 
 	mutex_enter(&dp->dp_lock);
-	dirty = dp->dp_dirty_total;
+	uint64_t dirty = dp->dp_dirty_total;
 	mutex_exit(&dp->dp_lock);
-	if (dirty > dirty_min_bytes)
-		txg_kick(dp);
+
 	return (dirty > delay_min_bytes);
 }
 
+static boolean_t
+dsl_pool_need_dirty_sync(dsl_pool_t *dp, uint64_t txg)
+{
+	ASSERT(MUTEX_HELD(&dp->dp_lock));
+
+	uint64_t dirty_min_bytes =
+	    zfs_dirty_data_max * zfs_dirty_data_sync_percent / 100;
+	uint64_t dirty = dp->dp_dirty_pertxg[txg & TXG_MASK];
+
+	return (dirty > dirty_min_bytes);
+}
+
 void
 dsl_pool_dirty_space(dsl_pool_t *dp, int64_t space, dmu_tx_t *tx)
 {
@@ -921,7 +929,12 @@ dsl_pool_dirty_space(dsl_pool_t *dp, int64_t space, dmu_tx_t *tx)
 		mutex_enter(&dp->dp_lock);
 		dp->dp_dirty_pertxg[tx->tx_txg & TXG_MASK] += space;
 		dsl_pool_dirty_delta(dp, space);
+		boolean_t needsync = !dmu_tx_is_syncing(tx) &&
+		    dsl_pool_need_dirty_sync(dp, tx->tx_txg);
 		mutex_exit(&dp->dp_lock);
+
+		if (needsync)
+			txg_kick(dp, tx->tx_txg);
 	}
 }
 
diff --git a/module/zfs/txg.c b/module/zfs/txg.c
index c55b1d8f9..c9eb84bbd 100644
--- a/module/zfs/txg.c
+++ b/module/zfs/txg.c
@@ -499,14 +499,6 @@ txg_wait_callbacks(dsl_pool_t *dp)
 }
 
 static boolean_t
-txg_is_syncing(dsl_pool_t *dp)
-{
-	tx_state_t *tx = &dp->dp_tx;
-	ASSERT(MUTEX_HELD(&tx->tx_sync_lock));
-	return (tx->tx_syncing_txg != 0);
-}
-
-static boolean_t
 txg_is_quiescing(dsl_pool_t *dp)
 {
 	tx_state_t *tx = &dp->dp_tx;
@@ -539,8 +531,6 @@ txg_sync_thread(void *arg)
 		clock_t timeout = zfs_txg_timeout * hz;
 		clock_t timer;
 		uint64_t txg;
-		uint64_t dirty_min_bytes =
-		    zfs_dirty_data_max * zfs_dirty_data_sync_percent / 100;
 
 		/*
 		 * We sync when we're scanning, there's someone waiting
@@ -551,8 +541,7 @@ txg_sync_thread(void *arg)
 		while (!dsl_scan_active(dp->dp_scan) &&
 		    !tx->tx_exiting && timer > 0 &&
 		    tx->tx_synced_txg >= tx->tx_sync_txg_waiting &&
-		    !txg_has_quiesced_to_sync(dp) &&
-		    dp->dp_dirty_total < dirty_min_bytes) {
+		    !txg_has_quiesced_to_sync(dp)) {
 			dprintf("waiting; tx_synced=%llu waiting=%llu dp=%p\n",
 			    (u_longlong_t)tx->tx_synced_txg,
 			    (u_longlong_t)tx->tx_sync_txg_waiting, dp);
@@ -566,6 +555,11 @@ txg_sync_thread(void *arg)
 		 * prompting it to do so if necessary.
 		 */
 		while (!tx->tx_exiting && !txg_has_quiesced_to_sync(dp)) {
+			if (txg_is_quiescing(dp)) {
+				txg_thread_wait(tx, &cpr,
+				    &tx->tx_quiesce_done_cv, 0);
+				continue;
+			}
 			if (tx->tx_quiesce_txg_waiting < tx->tx_open_txg+1)
 				tx->tx_quiesce_txg_waiting = tx->tx_open_txg+1;
 			cv_broadcast(&tx->tx_quiesce_more_cv);
@@ -791,24 +785,22 @@ txg_wait_open(dsl_pool_t *dp, uint64_t txg, boolean_t should_quiesce)
 }
 
 /*
- * If there isn't a txg syncing or in the pipeline, push another txg through
- * the pipeline by quiescing the open txg.
+ * Pass in the txg number that should be synced.
  */
 void
-txg_kick(dsl_pool_t *dp)
+txg_kick(dsl_pool_t *dp, uint64_t txg)
 {
 	tx_state_t *tx = &dp->dp_tx;
 
 	ASSERT(!dsl_pool_config_held(dp));
 
+	if (tx->tx_sync_txg_waiting >= txg)
+		return;
+
 	mutex_enter(&tx->tx_sync_lock);
-	if (!txg_is_syncing(dp) &&
-	    !txg_is_quiescing(dp) &&
-	    tx->tx_quiesce_txg_waiting <= tx->tx_open_txg &&
-	    tx->tx_sync_txg_waiting <= tx->tx_synced_txg &&
-	    tx->tx_quiesced_txg <= tx->tx_synced_txg) {
-		tx->tx_quiesce_txg_waiting = tx->tx_open_txg + 1;
-		cv_broadcast(&tx->tx_quiesce_more_cv);
+	if (tx->tx_sync_txg_waiting < txg) {
+		tx->tx_sync_txg_waiting = txg;
+		cv_broadcast(&tx->tx_sync_more_cv);
 	}
 	mutex_exit(&tx->tx_sync_lock);
 }