ZIL: Call brt_pending_add() replaying TX_CLONE_RANGE

zil_claim_clone_range() takes references on cloned blocks before ZIL
replay.  Later zil_free_clone_range() drops them after replay or on
dataset destroy.  The total balance is neutral.  It means on actual
replay we must take additional references, which would stay in BRT.

Without this blocks could be freed prematurely when either original
file or its clone are destroyed.  I've observed BRT being emptied
and the feature being desctivated after ZIL replay completion, which
should not have happened.  With the patch I see expected stats.

Signed-off-by:	Alexander Motin <mav@FreeBSD.org>
Sponsored by:	iXsystems, Inc.

include/sys/dmu.h

@@ -1086,8 +1086,7 @@ int dmu_offset_next(objset_t *os, uint64_t object, boolean_t hole,
 int dmu_read_l0_bps(objset_t *os, uint64_t object, uint64_t offset,
     uint64_t length, struct blkptr *bps, size_t *nbpsp);
 int dmu_brt_clone(objset_t *os, uint64_t object, uint64_t offset,
-    uint64_t length, dmu_tx_t *tx, const struct blkptr *bps, size_t nbps,
-    boolean_t replay);
+    uint64_t length, dmu_tx_t *tx, const struct blkptr *bps, size_t nbps);
 
 /*
  * Initial setup and final teardown.

module/zfs/dmu.c

@@ -2286,7 +2286,7 @@ dmu_read_l0_bps(objset_t *os, uint64_t object, uint64_t offset, uint64_t length,
 
 int
 dmu_brt_clone(objset_t *os, uint64_t object, uint64_t offset, uint64_t length,
-    dmu_tx_t *tx, const blkptr_t *bps, size_t nbps, boolean_t replay)
+    dmu_tx_t *tx, const blkptr_t *bps, size_t nbps)
 {
 	spa_t *spa;
 	dmu_buf_t **dbp, *dbuf;
@@ -2360,10 +2360,8 @@ dmu_brt_clone(objset_t *os, uint64_t object, uint64_t offset, uint64_t length,
 		 * When data in embedded into BP there is no need to create
 		 * BRT entry as there is no data block. Just copy the BP as
 		 * it contains the data.
-		 * Also, when replaying ZIL we don't want to bump references
-		 * in the BRT as it was already done during ZIL claim.
 		 */
-		if (!replay && !BP_IS_HOLE(bp) && !BP_IS_EMBEDDED(bp)) {
+		if (!BP_IS_HOLE(bp) && !BP_IS_EMBEDDED(bp)) {
 			brt_pending_add(spa, bp, tx);
 		}
 	}

module/zfs/zfs_vnops.c

@@ -1327,7 +1327,7 @@ zfs_clone_range(znode_t *inzp, uint64_t *inoffp, znode_t *outzp,
 		}
 
 		error = dmu_brt_clone(outos, outzp->z_id, outoff, size, tx,
-		    bps, nbps, B_FALSE);
+		    bps, nbps);
 		if (error != 0) {
 			dmu_tx_commit(tx);
 			break;
@@ -1461,7 +1461,7 @@ zfs_clone_range_replay(znode_t *zp, uint64_t off, uint64_t len, uint64_t blksz,
 	if (zp->z_blksz < blksz)
 		zfs_grow_blocksize(zp, blksz, tx);
 
-	dmu_brt_clone(zfsvfs->z_os, zp->z_id, off, len, tx, bps, nbps, B_TRUE);
+	dmu_brt_clone(zfsvfs->z_os, zp->z_id, off, len, tx, bps, nbps);
 
 	zfs_tstamp_update_setup(zp, CONTENT_MODIFIED, mtime, ctime);