Fix zed/udev thrashing due to autoexpand=on

Flushing the device and invaliding the page cache to ensure a
consistent view of the device is only needed when the device
was successfully expanded.  Unconditionally performing these
operations is safe but it has the side effect of triggering
another udev change event which will be handled by the ZED.

Signed-off-by: Brian Behlendorf <[email protected]>
Issue openzfs#7366

lib/libzfs/libzfs_pool.c

@@ -2461,12 +2461,15 @@ zpool_relabel_disk(libzfs_handle_t *hdl, const char *path, const char *msg)
 	 *
 	 * Also, we don't call efi_rescan() - that would just return EBUSY.
 	 * The module will do it for us in vdev_disk_open().
+	 *
+	 * On success flush the buffers to disk and invalidate the page cache
+	 * to ensure a consistent view of the relabled device.
 	 */
 	error = efi_use_whole_disk(fd);
-
-	/* Flush the buffers to disk and invalidate the page cache. */
-	(void) fsync(fd);
-	(void) ioctl(fd, BLKFLSBUF);
+	if (error == 0) {
+		(void) fsync(fd);
+		(void) ioctl(fd, BLKFLSBUF);
+	}
 
 	(void) close(fd);
 	if (error && error != VT_ENOSPC) {