dbuf_sync_leaf: check DB_READ in state assertions

Block cloning introduced a new state transition from DB_NOFILL to
DB_READ. This occurs when a block is cloned and then read on the
current txg.

In this case, the clone will move the dbuf to DB_NOFILL, and then the
read will be issued for the overidden block pointer. If that read is
still outstanding when it comes time to write, the dbuf will be in
DB_READ, which is not handled by the checks in dbuf_sync_leaf, thus
tripping the assertions.

This updates those checks to allow DB_READ as a valid state iff the
dirty record is for a BRT write and there is a override block pointer.
This is a safe situation because the block already exists, so there's
nothing that could change from underneath the read.

Reviewed-by: Brian Behlendorf <[email protected]>
Reviewed-by: Kay Pedersen <[email protected]>
Signed-off-by: Rob Norris <[email protected]>
Original-patch-by: Kay Pedersen <[email protected]>
Sponsored-By: OpenDrives Inc.
Sponsored-By: Klara Inc.
Closes #15050

1 files changed, 9 insertions, 0 deletions

diff --git a/module/zfs/dbuf.c b/module/zfs/dbuf.c
index fbeac866a..b7453578a 100644
--- a/module/zfs/dbuf.c
+++ b/module/zfs/dbuf.c
@@ -4457,6 +4457,15 @@ dbuf_sync_leaf(dbuf_dirty_record_t *dr, dmu_tx_t *tx)
 	} else if (db->db_state == DB_FILL) {
 		/* This buffer was freed and is now being re-filled */
 		ASSERT(db->db.db_data != dr->dt.dl.dr_data);
+	} else if (db->db_state == DB_READ) {
+		/*
+		 * This buffer has a clone we need to write, and an in-flight
+		 * read on the BP we're about to clone. Its safe to issue the
+		 * write here because the read has already been issued and the
+		 * contents won't change.
+		 */
+		ASSERT(dr->dt.dl.dr_brtwrite &&
+		    dr->dt.dl.dr_override_state == DR_OVERRIDDEN);
 	} else {
 		ASSERT(db->db_state == DB_CACHED || db->db_state == DB_NOFILL);
 	}