diff_cb() does not handle large dnodes

Trying to 'zfs diff' a snapshot with large dnodes will incorrectly try
to access its interior slots when dnodesize > sizeof(dnode_phys_t).
This is normally not an issue because the interior slots are
zero-filled, which report_dnode() handles calling
report_free_dnode_range(). However this is not the case for encrypted
large dnodes or filesystem using many SA based xattrs where the extra data past
the legacy dnode size boundary is interpreted as a dnode_phys_t.

Signed-off-by: loli10K <ezomori.nozomu@gmail.com>

module/zfs/dmu_diff.c

@@ -21,6 +21,7 @@
 /*
  * Copyright (c) 2010, Oracle and/or its affiliates. All rights reserved.
  * Copyright (c) 2012, 2018 by Delphix. All rights reserved.
+ * Copyright (c) 2010, loli10K <ezomori.nozomu@gmail.com>. All rights reserved.
  */
 
 #include <sys/dmu.h>
@@ -131,7 +132,7 @@ diff_cb(spa_t *spa, zilog_t *zilog, const blkptr_t *bp,
 		dnode_phys_t *blk;
 		arc_buf_t *abuf;
 		arc_flags_t aflags = ARC_FLAG_WAIT;
-		int blksz = BP_GET_LSIZE(bp);
+		int epb = BP_GET_LSIZE(bp) >> DNODE_SHIFT;
 		int zio_flags = ZIO_FLAG_CANFAIL;
 		int i;
 
@@ -143,7 +144,7 @@ diff_cb(spa_t *spa, zilog_t *zilog, const blkptr_t *bp,
 			return (SET_ERROR(EIO));
 
 		blk = abuf->b_data;
-		for (i = 0; i < blksz >> DNODE_SHIFT; i++) {
+		for (i = 0; i < epb; i += blk[i].dn_extra_slots + 1) {
 			uint64_t dnobj = (zb->zb_blkid <<
 			    (DNODE_BLOCK_SHIFT - DNODE_SHIFT)) + i;
 			err = report_dnode(da, dnobj, blk+i);

tests/zfs-tests/tests/functional/cli_root/zfs_diff/zfs_diff_encrypted.ksh

@@ -24,14 +24,15 @@
 # 1. Create an encrypted dataset
 # 2. Create two snapshots of the dataset
 # 3. Perform 'zfs diff -Ft' and verify no errors occur
+# 4. Perform the same test on a dataset with large dnodes
 #
 
 verify_runnable "both"
 
 function cleanup
 {
-	datasetexists $TESTPOOL/$TESTFS1 && \
-		log_must zfs destroy -r $TESTPOOL/$TESTFS1
+	destroy_dataset "$TESTPOOL/$TESTFS1" "-r"
+	destroy_dataset "$TESTPOOL/$TESTFS2" "-r"
 }
 
 log_assert "'zfs diff' should work with encrypted datasets"
@@ -50,4 +51,13 @@ log_must zfs snapshot $TESTPOOL/$TESTFS1@snap2
 # 3. Perform 'zfs diff' and verify no errors occur
 log_must zfs diff -Ft $TESTPOOL/$TESTFS1@snap1 $TESTPOOL/$TESTFS1@snap2
 
+# 4. Perform the same test on a dataset with large dnodes
+log_must eval "echo 'password' | zfs create -o dnodesize=4k \
+	-o encryption=on -o keyformat=passphrase $TESTPOOL/$TESTFS2"
+MNTPOINT="$(get_prop mountpoint $TESTPOOL/$TESTFS2)"
+log_must zfs snapshot $TESTPOOL/$TESTFS2@snap1
+log_must touch "$MNTPOINT/file"
+log_must zfs snapshot $TESTPOOL/$TESTFS2@snap2
+log_must zfs diff -Ft $TESTPOOL/$TESTFS2@snap1 $TESTPOOL/$TESTFS2@snap2
+
 log_pass "'zfs diff' works with encrypted datasets"