optimize recv_fix_encryption_hierarchy()

recv_fix_encryption_hierarchy() in its present state goes through all
stream filesystems, and for each one traverses the snapshots in order to
find one that exists locally. This happens by calling guid_to_name() for
each snapshot, which iterates through all children of the filesystem.
This results in CPU utilization of 100% for several minutes (for ~1000
filesystems on a Ryzen 4350G) for 1 thread at the end of a raw receive
(-w, regardless whether encrypted or not, dryrun or not).

Fix this by following a different logic: using the top_fs name, call
gather_nvlist() to gather the nvlists for all local filesystems. For
each one filesystem, go through the snapshots to find the corresponding
stream's filesystem (since we know the snapshots guid and can search
with it in stream_avl for the stream's fs). Then go on to fix the
encryption roots and locations as in its present state.

Avoiding guid_to_name() iteratively makes
recv_fix_encryption_hierarchy() significantly faster (from several
minutes to seconds for ~1000 filesystems on a Ryzen 4350G).

Another problem is the following: in case we have promoted a clone of
the filesystem outside the top filesystem specified in zfs send, zfs
receive does not fail but returns an error:
recv_incremental_replication() fails to find its origin and errors out
with needagain=1. This results in recv_fix_hierarchy() not being called
which may render some children of the top fs not mountable since their
encryption root was not updated. To circumvent this make
recv_incremental_replication() silently ignore this error.

Reviewed-by: Alexander Motin <[email protected]>
Reviewed-by: Brian Behlendorf <[email protected]>
Signed-off-by: George Amanakis <[email protected]>
Closes openzfs#16929

Author: gamanakis

lib/libzfs/libzfs_sendrecv.c

@@ -3393,66 +3393,78 @@ created_before(libzfs_handle_t *hdl, avl_tree_t *avl,
  */
 static int
 recv_fix_encryption_hierarchy(libzfs_handle_t *hdl, const char *top_zfs,
-    nvlist_t *stream_nv)
+    nvlist_t *stream_nv, avl_tree_t *stream_avl)
 {
 	int err;
 	nvpair_t *fselem = NULL;
-	nvlist_t *stream_fss;
+	nvlist_t *local_nv;
+	avl_tree_t *local_avl;
+	boolean_t recursive;
+
+	recursive = (nvlist_lookup_boolean(stream_nv, "not_recursive") ==
+	    ENOENT);
 
-	stream_fss = fnvlist_lookup_nvlist(stream_nv, "fss");
+	/* Using top_zfs, gather the nvlists for all local filesystems. */
+	if ((err = gather_nvlist(hdl, top_zfs, NULL, NULL,
+	    recursive, B_TRUE, B_FALSE, recursive, B_FALSE, B_FALSE, B_FALSE,
+	    B_FALSE, B_TRUE, &local_nv, &local_avl)) != 0)
+		return (err);
 
-	while ((fselem = nvlist_next_nvpair(stream_fss, fselem)) != NULL) {
+	/*
+	 * Go through the nvlists of the local filesystems and check for
+	 * encryption roots.
+	 */
+	while ((fselem = nvlist_next_nvpair(local_nv, fselem)) != NULL) {
 		zfs_handle_t *zhp = NULL;
 		uint64_t crypt;
-		nvlist_t *snaps, *props, *stream_nvfs = NULL;
-		nvpair_t *snapel = NULL;
+		nvlist_t *stream_props, *snaps, *stream_nvfs = NULL,
+		    *nvfs = NULL;
 		boolean_t is_encroot, is_clone, stream_encroot;
-		char *cp;
-		const char *stream_keylocation = NULL;
+		const char *stream_keylocation = NULL, *fsname;
 		char keylocation[MAXNAMELEN];
-		char fsname[ZFS_MAX_DATASET_NAME_LEN];
-
-		keylocation[0] = '\0';
-		stream_nvfs = fnvpair_value_nvlist(fselem);
-		snaps = fnvlist_lookup_nvlist(stream_nvfs, "snaps");
-		props = fnvlist_lookup_nvlist(stream_nvfs, "props");
-		stream_encroot = nvlist_exists(stream_nvfs, "is_encroot");
-
-		/* find a snapshot from the stream that exists locally */
-		err = ENOENT;
-		while ((snapel = nvlist_next_nvpair(snaps, snapel)) != NULL) {
-			uint64_t guid;
-
-			guid = fnvpair_value_uint64(snapel);
-			err = guid_to_name(hdl, top_zfs, guid, B_FALSE,
-			    fsname);
-			if (err == 0)
-				break;
-		}
-
-		if (err != 0)
-			continue;
-
-		cp = strchr(fsname, '@');
-		if (cp != NULL)
-			*cp = '\0';
+		nvpair_t *snapelem;
 
+		nvfs = fnvpair_value_nvlist(fselem);
+		snaps = fnvlist_lookup_nvlist(nvfs, "snaps");
+		fsname = fnvlist_lookup_string(nvfs, "name");
 		zhp = zfs_open(hdl, fsname, ZFS_TYPE_DATASET);
 		if (zhp == NULL) {
 			err = ENOENT;
 			goto error;
 		}
 
-		crypt = zfs_prop_get_int(zhp, ZFS_PROP_ENCRYPTION);
-		is_clone = zhp->zfs_dmustats.dds_origin[0] != '\0';
-		(void) zfs_crypto_get_encryption_root(zhp, &is_encroot, NULL);
-
 		/* we don't need to do anything for unencrypted datasets */
+		crypt = zfs_prop_get_int(zhp, ZFS_PROP_ENCRYPTION);
 		if (crypt == ZIO_CRYPT_OFF) {
 			zfs_close(zhp);
 			continue;
 		}
 
+		is_clone = zhp->zfs_dmustats.dds_origin[0] != '\0';
+		(void) zfs_crypto_get_encryption_root(zhp, &is_encroot, NULL);
+		keylocation[0] = '\0';
+
+		/*
+		 * Go through the snapshots of the local filesystem and find
+		 * the stream's filesystem.
+		 */
+		for (snapelem = nvlist_next_nvpair(snaps, NULL);
+		    snapelem; snapelem = nvlist_next_nvpair(snaps, snapelem)) {
+			uint64_t thisguid;
+
+			thisguid = fnvpair_value_uint64(snapelem);
+			stream_nvfs = fsavl_find(stream_avl, thisguid, NULL);
+
+			if (stream_nvfs != NULL)
+				break;
+		}
+
+		if (stream_nvfs == NULL)
+			continue;
+
+		stream_props = fnvlist_lookup_nvlist(stream_nvfs, "props");
+		stream_encroot = nvlist_exists(stream_nvfs, "is_encroot");
+
 		/*
 		 * If the dataset is flagged as an encryption root, was not
 		 * received as a clone and is not currently an encryption root,
@@ -3468,7 +3480,7 @@ recv_fix_encryption_hierarchy(libzfs_handle_t *hdl, const char *top_zfs,
 				}
 			}
 
-			stream_keylocation = fnvlist_lookup_string(props,
+			stream_keylocation = fnvlist_lookup_string(stream_props,
 			    zfs_prop_to_name(ZFS_PROP_KEYLOCATION));
 
 			/*
@@ -3535,14 +3547,14 @@ recv_incremental_replication(libzfs_handle_t *hdl, const char *tofs,
 	boolean_t needagain, progress, recursive;
 	const char *s1, *s2;
 
+	if (flags->dryrun)
+		return (0);
+
 	fromsnap = fnvlist_lookup_string(stream_nv, "fromsnap");
 
 	recursive = (nvlist_lookup_boolean(stream_nv, "not_recursive") ==
 	    ENOENT);
 
-	if (flags->dryrun)
-		return (0);
-
 again:
 	needagain = progress = B_FALSE;
 
@@ -4016,9 +4028,9 @@ zfs_receive_package(libzfs_handle_t *hdl, int fd, const char *destname,
 		    stream_nv, stream_avl, NULL);
 	}
 
-	if (raw && softerr == 0 && *top_zfs != NULL) {
+	if (raw && *top_zfs != NULL && !flags->dryrun) {
 		softerr = recv_fix_encryption_hierarchy(hdl, *top_zfs,
-		    stream_nv);
+		    stream_nv, stream_avl);
 	}
 
 out:

tests/zfs-tests/tests/functional/rsend/send_encrypted_hierarchy.ksh

@@ -61,16 +61,17 @@ log_must eval "zfs receive -d -F $POOL2 < $BACKDIR/fs-before-R"
 dstds=$(get_dst_ds $POOL/$FS $POOL2)
 log_must cmp_ds_subs $POOL/$FS $dstds
 
-log_must verify_encryption_root $POOL/$FS $POOL/$FS
-log_must verify_keylocation $POOL/$FS "prompt"
-log_must verify_origin $POOL/$FS "-"
+log_must verify_encryption_root $POOL2/$FS $POOL2/$FS
+log_must verify_keylocation $POOL2/$FS "prompt"
+log_must verify_origin $POOL2/$FS "-"
 
-log_must verify_encryption_root $POOL/clone $POOL/$FS
-log_must verify_keylocation $POOL/clone "none"
-log_must verify_origin $POOL/clone "$POOL/$FS@snap"
+log_must verify_encryption_root $POOL2/clone $POOL2/$FS
+log_must verify_keylocation $POOL2/clone "none"
+log_must verify_origin $POOL2/clone "$POOL2/$FS@snap"
 
 log_must verify_encryption_root $POOL/$FS/child $POOL/$FS
-log_must verify_keylocation $POOL/$FS/child "none"
+log_must verify_encryption_root $POOL2/$FS/child $POOL2/$FS
+log_must verify_keylocation $POOL2/$FS/child "none"
 
 # Alter the hierarchy and re-send
 log_must eval "echo $PASSPHRASE1 | zfs change-key -o keyformat=passphrase" \
@@ -93,4 +94,20 @@ log_must verify_origin $POOL/clone "-"
 log_must verify_encryption_root $POOL/$FS/child $POOL/$FS/child
 log_must verify_keylocation $POOL/$FS/child "prompt"
 
+log_must verify_encryption_root $POOL2 "-"
+log_must verify_encryption_root $POOL2/clone $POOL2/clone
+log_must verify_encryption_root $POOL2/$FS $POOL2/clone
+log_must verify_encryption_root $POOL2/$FS/child $POOL2/$FS/child
+
+log_must verify_keylocation $POOL2 "none"
+log_must verify_keylocation $POOL2/clone "prompt"
+log_must verify_keylocation $POOL2/$FS "none"
+log_must verify_keylocation $POOL2/$FS/child "prompt"
+
+log_must verify_origin $POOL2 "-"
+log_must verify_origin $POOL2/clone "-"
+log_must verify_origin $POOL2/$FS "$POOL2/clone@snap"
+log_must verify_origin $POOL2/$FS/child "-"
+log_must zfs list
+
 log_pass "Raw recursive sends preserve filesystem structure."