L2ARC: Stop rebuild before setting spa_final_txg

Without doing that there is a race window on export when history log write by completed rebuild dirties transaction beyond final, triggering assertion. Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov> Reviewed-by: George Amanakis <gamanakis@gmail.com> Signed-off-by: Alexander Motin <mav@FreeBSD.org> Sponsored by: iXsystems, Inc. Closes openzfs#16714 Closes openzfs#16782
truenas · Dec 2, 2024 · 0463f3d · 0463f3d
1 parent b36b97a
commit 0463f3d
Show file tree

Hide file tree

Showing 3 changed files with 35 additions and 3 deletions.
diff --git a/include/sys/arc.h b/include/sys/arc.h
@@ -347,6 +347,7 @@ void l2arc_fini(void);
 void l2arc_start(void);
 void l2arc_stop(void);
 void l2arc_spa_rebuild_start(spa_t *spa);
+void l2arc_spa_rebuild_stop(spa_t *spa);
 
 #ifndef _KERNEL
 extern boolean_t arc_watch;

diff --git a/module/zfs/arc.c b/module/zfs/arc.c
@@ -9846,6 +9846,37 @@ l2arc_spa_rebuild_start(spa_t *spa)
 	}
 }
 
+void
+l2arc_spa_rebuild_stop(spa_t *spa)
+{
+	ASSERT(MUTEX_HELD(&spa_namespace_lock) ||
+	    spa->spa_export_thread == curthread);
+
+	for (int i = 0; i < spa->spa_l2cache.sav_count; i++) {
+		l2arc_dev_t *dev =
+		    l2arc_vdev_get(spa->spa_l2cache.sav_vdevs[i]);
+		if (dev == NULL)
+			continue;
+		mutex_enter(&l2arc_rebuild_thr_lock);
+		dev->l2ad_rebuild_cancel = B_TRUE;
+		mutex_exit(&l2arc_rebuild_thr_lock);
+	}
+	for (int i = 0; i < spa->spa_l2cache.sav_count; i++) {
+		l2arc_dev_t *dev =
+		    l2arc_vdev_get(spa->spa_l2cache.sav_vdevs[i]);
+		if (dev == NULL)
+			continue;
+		mutex_enter(&l2arc_rebuild_thr_lock);
+		if (dev->l2ad_rebuild_began == B_TRUE) {
+			while (dev->l2ad_rebuild == B_TRUE) {
+				cv_wait(&l2arc_rebuild_thr_cv,
+				    &l2arc_rebuild_thr_lock);
+			}
+		}
+		mutex_exit(&l2arc_rebuild_thr_lock);
+	}
+}
+
 /*
  * Main entry point for L2ARC rebuilding.
  */
@@ -9854,12 +9885,12 @@ l2arc_dev_rebuild_thread(void *arg)
 {
 	l2arc_dev_t *dev = arg;
 
-	VERIFY(!dev->l2ad_rebuild_cancel);
 	VERIFY(dev->l2ad_rebuild);
 	(void) l2arc_rebuild(dev);
 	mutex_enter(&l2arc_rebuild_thr_lock);
 	dev->l2ad_rebuild_began = B_FALSE;
 	dev->l2ad_rebuild = B_FALSE;
+	cv_signal(&l2arc_rebuild_thr_cv);
 	mutex_exit(&l2arc_rebuild_thr_lock);
 
 	thread_exit();
@@ -10010,8 +10041,6 @@ l2arc_rebuild(l2arc_dev_t *dev)
 		for (;;) {
 			mutex_enter(&l2arc_rebuild_thr_lock);
 			if (dev->l2ad_rebuild_cancel) {
-				dev->l2ad_rebuild = B_FALSE;
-				cv_signal(&l2arc_rebuild_thr_cv);
 				mutex_exit(&l2arc_rebuild_thr_lock);
 				err = SET_ERROR(ECANCELED);
 				goto out;

diff --git a/module/zfs/spa.c b/module/zfs/spa.c
@@ -2081,6 +2081,7 @@ spa_unload(spa_t *spa)
 			vdev_trim_stop_all(root_vdev, VDEV_TRIM_ACTIVE);
 			vdev_autotrim_stop_all(spa);
 			vdev_rebuild_stop_all(spa);
+			l2arc_spa_rebuild_stop(spa);
 		}
 	}
 
@@ -7115,6 +7116,7 @@ spa_export_common(const char *pool, int new_state, nvlist_t **oldconfig,
 		vdev_trim_stop_all(rvd, VDEV_TRIM_ACTIVE);
 		vdev_autotrim_stop_all(spa);
 		vdev_rebuild_stop_all(spa);
+		l2arc_spa_rebuild_stop(spa);
 
 		/*
 		 * We want this to be reflected on every label,