Allow mounting datasets more than once

Currently mounting an already mounted zfs dataset results in an
error, whereas it is typically allowed with other filesystems.
This causes some bad interactions with mount namespaces. Take
this sequence for example:

- Create a dataset
- Create a snapshot of the dataset
- Create a clone of the snapshot
- Create a new mount namespace
- Rename the original dataset

The rename results in unmounting and remounting the clone in the
original mount namespace, however the remount fails because the
dataset is still mounted in the new mount namespace. (Note that
this means the mount in the new mount namespace is never being
unmounted, so perhaps the unmount/remount of the clone isn't
actually necessary.)

The problem here is a result of the way mounting is implemented
in the kernel module. Since it is not mounting block devices it
uses mount_nodev() instead of the usual mount_bdev(). However,
mount_nodev() is written for filesystems for which each mount is
a new instance (i.e. a new super block), and zfs should be able
to detect when a mount request can be satisfied using an existing
super block.

Change zpl_mount() to call sget() directly with it's own test
callback. Passing the objset_t object as the fs data allows
checking if a superblock already exists for the dataset, and in
that case we just need to return a new reference for the sb's
root dentry.

Reviewed-by: Brian Behlendorf <[email protected]>
Reviewed-by: Tom Caputi <[email protected]>
Signed-off-by: Alek Pinchuk <[email protected]>
Signed-off-by: Seth Forshee <[email protected]>
Closes openzfs#5796
Closes openzfs#7207

config/kernel-fst-mount.m4

@@ -0,0 +1,28 @@
+dnl #
+dnl # 2.6.38 API change
+dnl # The .get_sb callback has been replaced by a .mount callback
+dnl # in the file_system_type structure.
+dnl #
+AC_DEFUN([ZFS_AC_KERNEL_FST_MOUNT], [
+        AC_MSG_CHECKING([whether fst->mount() exists])
+        ZFS_LINUX_TRY_COMPILE([
+                #include <linux/fs.h>
+
+                static struct dentry *
+                mount(struct file_system_type *fs_type, int flags,
+                    const char *osname, void *data) {
+                        struct dentry *d = NULL;
+                        return (d);
+                }
+
+                static struct file_system_type fst __attribute__ ((unused)) = {
+                        .mount = mount,
+                };
+        ],[
+        ],[
+                AC_MSG_RESULT(yes)
+                AC_DEFINE(HAVE_FST_MOUNT, 1, [fst->mount() exists])
+        ],[
+                AC_MSG_RESULT(no)
+        ])
+])

config/kernel.m4

@@ -103,7 +103,7 @@ AC_DEFUN([ZFS_AC_CONFIG_KERNEL], [
 	ZFS_AC_KERNEL_TRUNCATE_SETSIZE
 	ZFS_AC_KERNEL_6ARGS_SECURITY_INODE_INIT_SECURITY
 	ZFS_AC_KERNEL_CALLBACK_SECURITY_INODE_INIT_SECURITY
-	ZFS_AC_KERNEL_MOUNT_NODEV
+	ZFS_AC_KERNEL_FST_MOUNT
 	ZFS_AC_KERNEL_SHRINK
 	ZFS_AC_KERNEL_SHRINK_CONTROL_HAS_NID
 	ZFS_AC_KERNEL_S_INSTANCES_LIST_HEAD

include/linux/vfs_compat.h

@@ -182,6 +182,30 @@ zpl_bdi_destroy(struct super_block *sb)
 }
 #endif
 
+/*
+ * 4.14 adds SB_* flag definitions, define them to MS_* equivalents
+ * if not set.
+ */
+#ifndef	SB_RDONLY
+#define	SB_RDONLY	MS_RDONLY
+#endif
+
+#ifndef	SB_SILENT
+#define	SB_SILENT	MS_SILENT
+#endif
+
+#ifndef	SB_ACTIVE
+#define	SB_ACTIVE	MS_ACTIVE
+#endif
+
+#ifndef	SB_POSIXACL
+#define	SB_POSIXACL	MS_POSIXACL
+#endif
+
+#ifndef	SB_MANDLOCK
+#define	SB_MANDLOCK	MS_MANDLOCK
+#endif
+
 /*
  * 2.6.38 API change,
  * LOOKUP_RCU flag introduced to distinguish rcu-walk from ref-walk cases.

module/zfs/zpl_super.c

@@ -248,14 +248,61 @@ zpl_fill_super(struct super_block *sb, void *data, int silent)
 	return (error);
 }
 
-#ifdef HAVE_MOUNT_NODEV
+static int
+zpl_test_super(struct super_block *s, void *data)
+{
+	zfsvfs_t *zfsvfs = s->s_fs_info;
+	objset_t *os = data;
+
+	if (zfsvfs == NULL)
+		return (0);
+
+	return (os == zfsvfs->z_os);
+}
+
+static struct super_block *
+zpl_mount_impl(struct file_system_type *fs_type, int flags, zfs_mnt_t *zm)
+{
+	struct super_block *s;
+	objset_t *os;
+	int err;
+
+	err = dmu_objset_hold(zm->mnt_osname, FTAG, &os);
+	if (err)
+		return (ERR_PTR(-err));
+
+	s = zpl_sget(fs_type, zpl_test_super, set_anon_super, flags, os);
+	dmu_objset_rele(os, FTAG);
+	if (IS_ERR(s))
+		return (ERR_CAST(s));
+
+	if (s->s_root == NULL) {
+		err = zpl_fill_super(s, zm, flags & SB_SILENT ? 1 : 0);
+		if (err) {
+			deactivate_locked_super(s);
+			return (ERR_PTR(err));
+		}
+		s->s_flags |= SB_ACTIVE;
+	} else if ((flags ^ s->s_flags) & SB_RDONLY) {
+		deactivate_locked_super(s);
+		return (ERR_PTR(-EBUSY));
+	}
+
+	return (s);
+}
+
+#ifdef HAVE_FST_MOUNT
 static struct dentry *
 zpl_mount(struct file_system_type *fs_type, int flags,
     const char *osname, void *data)
 {
 	zfs_mnt_t zm = { .mnt_osname = osname, .mnt_data = data };
 
-	return (mount_nodev(fs_type, flags, &zm, zpl_fill_super));
+	struct super_block *sb = zpl_mount_impl(fs_type, flags, &zm);
+	if (IS_ERR(sb))
+		return (ERR_CAST(sb));
+
+	return (dget(sb->s_root));
 }
 #else
 static int
@@ -264,9 +311,15 @@ zpl_get_sb(struct file_system_type *fs_type, int flags,
 {
 	zfs_mnt_t zm = { .mnt_osname = osname, .mnt_data = data };
 
-	return (get_sb_nodev(fs_type, flags, &zm, zpl_fill_super, mnt));
+	struct super_block *sb = zpl_mount_impl(fs_type, flags, &zm);
+	if (IS_ERR(sb))
+		return (PTR_ERR(sb));
+
+	(void) simple_set_mnt(mnt, sb);
+
+	return (0);
 }
-#endif /* HAVE_MOUNT_NODEV */
+#endif /* HAVE_FST_MOUNT */
 
 static void
 zpl_kill_sb(struct super_block *sb)
@@ -333,10 +386,10 @@ const struct super_operations zpl_super_operations = {
 struct file_system_type zpl_fs_type = {
 	.owner			= THIS_MODULE,
 	.name			= ZFS_DRIVER,
-#ifdef HAVE_MOUNT_NODEV
+#ifdef HAVE_FST_MOUNT
 	.mount			= zpl_mount,
 #else
 	.get_sb			= zpl_get_sb,
-#endif /* HAVE_MOUNT_NODEV */
+#endif /* HAVE_FST_MOUNT */
 	.kill_sb		= zpl_kill_sb,
 };

tests/runfiles/linux.run

@@ -174,7 +174,7 @@ tests = ['zfs_mount_001_pos', 'zfs_mount_002_pos', 'zfs_mount_003_pos',
     'zfs_mount_004_pos', 'zfs_mount_005_pos', 'zfs_mount_007_pos',
     'zfs_mount_008_pos', 'zfs_mount_009_neg', 'zfs_mount_010_neg',
     'zfs_mount_011_neg', 'zfs_mount_012_neg', 'zfs_mount_all_001_pos',
-    'zfs_mount_encrypted', 'zfs_mount_remount']
+    'zfs_mount_encrypted', 'zfs_mount_remount', 'zfs_multi_mount']
 tags = ['functional', 'cli_root', 'zfs_mount']
 
 [tests/functional/cli_root/zfs_program]

tests/zfs-tests/tests/functional/cli_root/zfs_mount/Makefile.am

@@ -16,7 +16,8 @@ dist_pkgdata_SCRIPTS = \
 	zfs_mount_012_neg.ksh \
 	zfs_mount_encrypted.ksh \
 	zfs_mount_all_001_pos.ksh \
-	zfs_mount_remount.ksh
+	zfs_mount_remount.ksh \
+	zfs_multi_mount.ksh
 
 dist_pkgdata_DATA = \
 	zfs_mount.cfg \

tests/zfs-tests/tests/functional/cli_root/zfs_mount/zfs_multi_mount.ksh

@@ -0,0 +1,104 @@
+#!/bin/ksh -p
+#
+# CDDL HEADER START
+#
+# This file and its contents are supplied under the terms of the
+# Common Development and Distribution License ("CDDL"), version 1.0.
+# You may only use this file in accordance with the terms of version
+# 1.0 of the CDDL.
+#
+# A full copy of the text of the CDDL should have accompanied this
+# source.  A copy is of the CDDL is also available via the Internet
+# at http://www.illumos.org/license/CDDL.
+#
+# CDDL HEADER END
+#
+
+#
+# Copyright(c) 2018 Datto Inc.
+#
+
+. $STF_SUITE/include/libtest.shlib
+
+#
+# DESCRIPTION:
+# Verify multi mount functionality
+#
+# STRATEGY:
+# 1. Create fs
+# 2. Create and hold open file in filesystem
+# 3. Lazy unmount
+# 4. Verify remounting fs that was lazily unmounted is possible
+# 5. Verify multiple mounts of the same dataset are possible
+# 6. Verify bind mount doesn't prevent rename
+#
+
+verify_runnable "both"
+
+function cleanup
+{
+	ismounted $MNTPFS && log_must umount $MNTPFS
+	ismounted $MNTPFS2 && log_must umount $MNTPFS2
+	ismounted $MNTPFS3 && log_must umount $MNTPFS3
+	ismounted $MNTPFS4 && log_must umount $MNTPFS4
+	ismounted $RENAMEMNT && log_must umount $RENAMEMNT
+	datasetexists $TESTDS && log_must destroy_dataset "$TESTDS" "-f"
+}
+log_onexit cleanup
+
+log_assert "Verify multiple mounts into one namespace are possible"
+
+# 1. Create fs
+TESTDS="$TESTPOOL/multi-mount-test"
+log_must zfs create $TESTDS
+
+# 2. Create and hold open file in filesystem
+MNTPFS="$(get_prop mountpoint $TESTDS)"
+FILENAME="$MNTPFS/file"
+log_must mkfile 128k $FILENAME
+log_must exec 9<> $FILENAME # open file
+
+# 3. Lazy umount
+log_must umount -l $MNTPFS
+if [ -f $FILENAME ]; then
+	log_fail "Lazy unmount failed"
+fi
+
+# 4. Verify remounting fs that was lazily unmounted is possible
+log_must zfs mount $TESTDS
+if [ ! -f $FILENAME ]; then
+	log_fail "Lazy remount failed"
+fi
+log_must exec 9>&- # close fd
+
+# 5. Verify multiple mounts of the same dataset are possible
+MNTPFS2="$MNTPFS-second"
+FILENAME="$MNTPFS2/file"
+log_must mkdir $MNTPFS2
+log_must mount -t zfs -o zfsutil $TESTDS $MNTPFS2
+if [ ! -f $FILENAME ]; then
+	log_fail "First multi mount failed"
+fi
+
+MNTPFS3="$MNTPFS-third"
+FILENAME="$MNTPFS3/file"
+log_must mkdir $MNTPFS3
+log_must mount -t zfs -o zfsutil $TESTDS $MNTPFS3
+if [ ! -f $FILENAME ]; then
+	log_fail "Second multi mount failed"
+fi
+
+# 6. Verify bind mount doesn't prevent rename
+RENAMEFS="$TESTDS-newname"
+MNTPFS4="$MNTPFS-fourth"
+log_must mkdir $MNTPFS4
+log_must mount --bind $MNTPFS $MNTPFS4
+log_must zfs rename $TESTDS $RENAMEFS
+RENAMEMNT="$(get_prop mountpoint $RENAMEFS)"
+FILENAME="$RENAMEMNT/file"
+if [ ! -f $FILENAME ]; then
+	log_fail "Rename failed"
+fi
+log_must zfs rename $RENAMEFS $TESTDS
+
+log_pass "Multiple mounts are possible"