From 07b20889e3052c7e77d6a6a54e7e83446eb1ba84 Mon Sep 17 00:00:00 2001 From: Ram Pai Date: Mon, 7 Nov 2005 17:19:07 -0500 Subject: [PATCH] beginning of the shared-subtree proper A private mount does not forward or receive propagation. This patch provides user the ability to convert any mount to private. Signed-off-by: Ram Pai Signed-off-by: Al Viro Signed-off-by: Linus Torvalds --- fs/pnode.c | 17 +++++++++++++++++ 1 file changed, 17 insertions(+) create mode 100644 fs/pnode.c (limited to 'fs/pnode.c') diff --git a/fs/pnode.c b/fs/pnode.c new file mode 100644 index 000000000000..aaa0dffda12a --- /dev/null +++ b/fs/pnode.c @@ -0,0 +1,17 @@ +/* + * linux/fs/pnode.c + * + * (C) Copyright IBM Corporation 2005. + * Released under GPL v2. + * Author : Ram Pai (linuxram@us.ibm.com) + * + */ +#include +#include +#include +#include "pnode.h" + +void change_mnt_propagation(struct vfsmount *mnt, int type) +{ + mnt->mnt_flags &= ~MNT_PNODE_MASK; +} -- cgit From 03e06e68ff76294e53ffa898cb844d2a997b043e Mon Sep 17 00:00:00 2001 From: Ram Pai Date: Mon, 7 Nov 2005 17:19:33 -0500 Subject: [PATCH] introduce shared mounts This creates shared mounts. A shared mount when bind-mounted to some mountpoint, propagates mount/umount events to each other. All the shared mounts that propagate events to each other belong to the same peer-group. Signed-off-by: Ram Pai Signed-off-by: Al Viro Signed-off-by: Linus Torvalds --- fs/pnode.c | 13 ++++++++++++- 1 file changed, 12 insertions(+), 1 deletion(-) (limited to 'fs/pnode.c') diff --git a/fs/pnode.c b/fs/pnode.c index aaa0dffda12a..1e22165ea41f 100644 --- a/fs/pnode.c +++ b/fs/pnode.c @@ -11,7 +11,18 @@ #include #include "pnode.h" +/* return the next shared peer mount of @p */ +static inline struct vfsmount *next_peer(struct vfsmount *p) +{ + return list_entry(p->mnt_share.next, struct vfsmount, mnt_share); +} + void change_mnt_propagation(struct vfsmount *mnt, int type) { - mnt->mnt_flags &= ~MNT_PNODE_MASK; + if (type == MS_SHARED) { + mnt->mnt_flags |= MNT_SHARED; + } else { + list_del_init(&mnt->mnt_share); + mnt->mnt_flags &= ~MNT_PNODE_MASK; + } } -- cgit From b90fa9ae8f51f098ee480bbaabd6867992e9fc58 Mon Sep 17 00:00:00 2001 From: Ram Pai Date: Mon, 7 Nov 2005 17:19:50 -0500 Subject: [PATCH] shared mount handling: bind and rbind Implement handling of MS_BIND in presense of shared mounts (see Documentation/sharedsubtree.txt in the end of patch series for detailed description). Signed-off-by: Ram Pai Signed-off-by: Al Viro Signed-off-by: Linus Torvalds --- fs/pnode.c | 81 +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++- 1 file changed, 80 insertions(+), 1 deletion(-) (limited to 'fs/pnode.c') diff --git a/fs/pnode.c b/fs/pnode.c index 1e22165ea41f..2d572b88e6f6 100644 --- a/fs/pnode.c +++ b/fs/pnode.c @@ -20,9 +20,88 @@ static inline struct vfsmount *next_peer(struct vfsmount *p) void change_mnt_propagation(struct vfsmount *mnt, int type) { if (type == MS_SHARED) { - mnt->mnt_flags |= MNT_SHARED; + set_mnt_shared(mnt); } else { list_del_init(&mnt->mnt_share); mnt->mnt_flags &= ~MNT_PNODE_MASK; } } + +/* + * get the next mount in the propagation tree. + * @m: the mount seen last + * @origin: the original mount from where the tree walk initiated + */ +static struct vfsmount *propagation_next(struct vfsmount *m, + struct vfsmount *origin) +{ + m = next_peer(m); + if (m == origin) + return NULL; + return m; +} + +/* + * mount 'source_mnt' under the destination 'dest_mnt' at + * dentry 'dest_dentry'. And propagate that mount to + * all the peer and slave mounts of 'dest_mnt'. + * Link all the new mounts into a propagation tree headed at + * source_mnt. Also link all the new mounts using ->mnt_list + * headed at source_mnt's ->mnt_list + * + * @dest_mnt: destination mount. + * @dest_dentry: destination dentry. + * @source_mnt: source mount. + * @tree_list : list of heads of trees to be attached. + */ +int propagate_mnt(struct vfsmount *dest_mnt, struct dentry *dest_dentry, + struct vfsmount *source_mnt, struct list_head *tree_list) +{ + struct vfsmount *m, *child; + int ret = 0; + struct vfsmount *prev_dest_mnt = dest_mnt; + struct vfsmount *prev_src_mnt = source_mnt; + LIST_HEAD(tmp_list); + LIST_HEAD(umount_list); + + for (m = propagation_next(dest_mnt, dest_mnt); m; + m = propagation_next(m, dest_mnt)) { + int type = CL_PROPAGATION; + + if (IS_MNT_NEW(m)) + continue; + + if (IS_MNT_SHARED(m)) + type |= CL_MAKE_SHARED; + + if (!(child = copy_tree(source_mnt, source_mnt->mnt_root, + type))) { + ret = -ENOMEM; + list_splice(tree_list, tmp_list.prev); + goto out; + } + + if (is_subdir(dest_dentry, m->mnt_root)) { + mnt_set_mountpoint(m, dest_dentry, child); + list_add_tail(&child->mnt_hash, tree_list); + } else { + /* + * This can happen if the parent mount was bind mounted + * on some subdirectory of a shared/slave mount. + */ + list_add_tail(&child->mnt_hash, &tmp_list); + } + prev_dest_mnt = m; + prev_src_mnt = child; + } +out: + spin_lock(&vfsmount_lock); + while (!list_empty(&tmp_list)) { + child = list_entry(tmp_list.next, struct vfsmount, mnt_hash); + list_del_init(&child->mnt_hash); + umount_tree(child, &umount_list); + } + spin_unlock(&vfsmount_lock); + release_mounts(&umount_list); + return ret; +} -- cgit From a05964f3917c7c55368c229d7985f8e7c9977e97 Mon Sep 17 00:00:00 2001 From: Ram Pai Date: Mon, 7 Nov 2005 17:20:17 -0500 Subject: [PATCH] shared mounts handling: umount An unmount of a mount creates a umount event on the parent. If the parent is a shared mount, it gets propagated to all mounts in the peer group. Signed-off-by: Ram Pai Signed-off-by: Al Viro Signed-off-by: Linus Torvalds --- fs/pnode.c | 87 +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++- 1 file changed, 86 insertions(+), 1 deletion(-) (limited to 'fs/pnode.c') diff --git a/fs/pnode.c b/fs/pnode.c index 2d572b88e6f6..7bc942d047cd 100644 --- a/fs/pnode.c +++ b/fs/pnode.c @@ -99,9 +99,94 @@ out: while (!list_empty(&tmp_list)) { child = list_entry(tmp_list.next, struct vfsmount, mnt_hash); list_del_init(&child->mnt_hash); - umount_tree(child, &umount_list); + umount_tree(child, 0, &umount_list); } spin_unlock(&vfsmount_lock); release_mounts(&umount_list); return ret; } + +/* + * return true if the refcount is greater than count + */ +static inline int do_refcount_check(struct vfsmount *mnt, int count) +{ + int mycount = atomic_read(&mnt->mnt_count); + return (mycount > count); +} + +/* + * check if the mount 'mnt' can be unmounted successfully. + * @mnt: the mount to be checked for unmount + * NOTE: unmounting 'mnt' would naturally propagate to all + * other mounts its parent propagates to. + * Check if any of these mounts that **do not have submounts** + * have more references than 'refcnt'. If so return busy. + */ +int propagate_mount_busy(struct vfsmount *mnt, int refcnt) +{ + struct vfsmount *m, *child; + struct vfsmount *parent = mnt->mnt_parent; + int ret = 0; + + if (mnt == parent) + return do_refcount_check(mnt, refcnt); + + /* + * quickly check if the current mount can be unmounted. + * If not, we don't have to go checking for all other + * mounts + */ + if (!list_empty(&mnt->mnt_mounts) || do_refcount_check(mnt, refcnt)) + return 1; + + for (m = propagation_next(parent, parent); m; + m = propagation_next(m, parent)) { + child = __lookup_mnt(m, mnt->mnt_mountpoint, 0); + if (child && list_empty(&child->mnt_mounts) && + (ret = do_refcount_check(child, 1))) + break; + } + return ret; +} + +/* + * NOTE: unmounting 'mnt' naturally propagates to all other mounts its + * parent propagates to. + */ +static void __propagate_umount(struct vfsmount *mnt) +{ + struct vfsmount *parent = mnt->mnt_parent; + struct vfsmount *m; + + BUG_ON(parent == mnt); + + for (m = propagation_next(parent, parent); m; + m = propagation_next(m, parent)) { + + struct vfsmount *child = __lookup_mnt(m, + mnt->mnt_mountpoint, 0); + /* + * umount the child only if the child has no + * other children + */ + if (child && list_empty(&child->mnt_mounts)) { + list_del(&child->mnt_hash); + list_add_tail(&child->mnt_hash, &mnt->mnt_hash); + } + } +} + +/* + * collect all mounts that receive propagation from the mount in @list, + * and return these additional mounts in the same list. + * @list: the list of mounts to be unmounted. + */ +int propagate_umount(struct list_head *list) +{ + struct vfsmount *mnt; + + list_for_each_entry(mnt, list, mnt_hash) + __propagate_umount(mnt); + return 0; +} -- cgit From a58b0eb8e64b78d9315a5491955e78b1391d42e5 Mon Sep 17 00:00:00 2001 From: Ram Pai Date: Mon, 7 Nov 2005 17:20:48 -0500 Subject: [PATCH] introduce slave mounts A slave mount always has a master mount from which it receives mount/umount events. Unlike shared mount the event propagation does not flow from the slave mount to the master. Signed-off-by: Ram Pai Signed-off-by: Al Viro Signed-off-by: Linus Torvalds --- fs/pnode.c | 54 +++++++++++++++++++++++++++++++++++++++++++++++++++--- 1 file changed, 51 insertions(+), 3 deletions(-) (limited to 'fs/pnode.c') diff --git a/fs/pnode.c b/fs/pnode.c index 7bc942d047cd..f73eba24f1d1 100644 --- a/fs/pnode.c +++ b/fs/pnode.c @@ -17,13 +17,61 @@ static inline struct vfsmount *next_peer(struct vfsmount *p) return list_entry(p->mnt_share.next, struct vfsmount, mnt_share); } +static int do_make_slave(struct vfsmount *mnt) +{ + struct vfsmount *peer_mnt = mnt, *master = mnt->mnt_master; + struct vfsmount *slave_mnt; + + /* + * slave 'mnt' to a peer mount that has the + * same root dentry. If none is available than + * slave it to anything that is available. + */ + while ((peer_mnt = next_peer(peer_mnt)) != mnt && + peer_mnt->mnt_root != mnt->mnt_root) ; + + if (peer_mnt == mnt) { + peer_mnt = next_peer(mnt); + if (peer_mnt == mnt) + peer_mnt = NULL; + } + list_del_init(&mnt->mnt_share); + + if (peer_mnt) + master = peer_mnt; + + if (master) { + list_for_each_entry(slave_mnt, &mnt->mnt_slave_list, mnt_slave) + slave_mnt->mnt_master = master; + list_del(&mnt->mnt_slave); + list_add(&mnt->mnt_slave, &master->mnt_slave_list); + list_splice(&mnt->mnt_slave_list, master->mnt_slave_list.prev); + INIT_LIST_HEAD(&mnt->mnt_slave_list); + } else { + struct list_head *p = &mnt->mnt_slave_list; + while (!list_empty(p)) { + slave_mnt = list_entry(p->next, + struct vfsmount, mnt_slave); + list_del_init(&slave_mnt->mnt_slave); + slave_mnt->mnt_master = NULL; + } + } + mnt->mnt_master = master; + CLEAR_MNT_SHARED(mnt); + INIT_LIST_HEAD(&mnt->mnt_slave_list); + return 0; +} + void change_mnt_propagation(struct vfsmount *mnt, int type) { if (type == MS_SHARED) { set_mnt_shared(mnt); - } else { - list_del_init(&mnt->mnt_share); - mnt->mnt_flags &= ~MNT_PNODE_MASK; + return; + } + do_make_slave(mnt); + if (type != MS_SLAVE) { + list_del_init(&mnt->mnt_slave); + mnt->mnt_master = NULL; } } -- cgit From 5afe00221389998a25d611dc7941c06580c29eb6 Mon Sep 17 00:00:00 2001 From: Ram Pai Date: Mon, 7 Nov 2005 17:21:01 -0500 Subject: [PATCH] handling of slave mounts This makes bind, rbind, move, clone namespace and umount operations aware of the semantics of slave mount (see Documentation/sharedsubtree.txt in the last patch of the series for detailed description). Signed-off-by: Ram Pai Signed-off-by: Al Viro Signed-off-by: Linus Torvalds --- fs/pnode.c | 81 +++++++++++++++++++++++++++++++++++++++++++++++++++++++------- 1 file changed, 72 insertions(+), 9 deletions(-) (limited to 'fs/pnode.c') diff --git a/fs/pnode.c b/fs/pnode.c index f73eba24f1d1..3e266c5a3071 100644 --- a/fs/pnode.c +++ b/fs/pnode.c @@ -17,6 +17,16 @@ static inline struct vfsmount *next_peer(struct vfsmount *p) return list_entry(p->mnt_share.next, struct vfsmount, mnt_share); } +static inline struct vfsmount *first_slave(struct vfsmount *p) +{ + return list_entry(p->mnt_slave_list.next, struct vfsmount, mnt_slave); +} + +static inline struct vfsmount *next_slave(struct vfsmount *p) +{ + return list_entry(p->mnt_slave.next, struct vfsmount, mnt_slave); +} + static int do_make_slave(struct vfsmount *mnt) { struct vfsmount *peer_mnt = mnt, *master = mnt->mnt_master; @@ -83,10 +93,64 @@ void change_mnt_propagation(struct vfsmount *mnt, int type) static struct vfsmount *propagation_next(struct vfsmount *m, struct vfsmount *origin) { - m = next_peer(m); - if (m == origin) - return NULL; - return m; + /* are there any slaves of this mount? */ + if (!IS_MNT_NEW(m) && !list_empty(&m->mnt_slave_list)) + return first_slave(m); + + while (1) { + struct vfsmount *next; + struct vfsmount *master = m->mnt_master; + + if ( master == origin->mnt_master ) { + next = next_peer(m); + return ((next == origin) ? NULL : next); + } else if (m->mnt_slave.next != &master->mnt_slave_list) + return next_slave(m); + + /* back at master */ + m = master; + } +} + +/* + * return the source mount to be used for cloning + * + * @dest the current destination mount + * @last_dest the last seen destination mount + * @last_src the last seen source mount + * @type return CL_SLAVE if the new mount has to be + * cloned as a slave. + */ +static struct vfsmount *get_source(struct vfsmount *dest, + struct vfsmount *last_dest, + struct vfsmount *last_src, + int *type) +{ + struct vfsmount *p_last_src = NULL; + struct vfsmount *p_last_dest = NULL; + *type = CL_PROPAGATION;; + + if (IS_MNT_SHARED(dest)) + *type |= CL_MAKE_SHARED; + + while (last_dest != dest->mnt_master) { + p_last_dest = last_dest; + p_last_src = last_src; + last_dest = last_dest->mnt_master; + last_src = last_src->mnt_master; + } + + if (p_last_dest) { + do { + p_last_dest = next_peer(p_last_dest); + } while (IS_MNT_NEW(p_last_dest)); + } + + if (dest != p_last_dest) { + *type |= CL_SLAVE; + return last_src; + } else + return p_last_src; } /* @@ -114,16 +178,15 @@ int propagate_mnt(struct vfsmount *dest_mnt, struct dentry *dest_dentry, for (m = propagation_next(dest_mnt, dest_mnt); m; m = propagation_next(m, dest_mnt)) { - int type = CL_PROPAGATION; + int type; + struct vfsmount *source; if (IS_MNT_NEW(m)) continue; - if (IS_MNT_SHARED(m)) - type |= CL_MAKE_SHARED; + source = get_source(m, prev_dest_mnt, prev_src_mnt, &type); - if (!(child = copy_tree(source_mnt, source_mnt->mnt_root, - type))) { + if (!(child = copy_tree(source, source->mnt_root, type))) { ret = -ENOMEM; list_splice(tree_list, tmp_list.prev); goto out; -- cgit From 9676f0c6389b62bd6b24d77d4b3abdbcfa32d0f2 Mon Sep 17 00:00:00 2001 From: Ram Pai Date: Mon, 7 Nov 2005 17:21:20 -0500 Subject: [PATCH] unbindable mounts An unbindable mount does not forward or receive propagation. Also unbindable mount disallows bind mounts. The semantics is as follows. Bind semantics: It is invalid to bind mount an unbindable mount. Move semantics: It is invalid to move an unbindable mount under shared mount. Clone-namespace semantics: If a mount is unbindable in the parent namespace, the corresponding cloned mount in the child namespace becomes unbindable too. Note: there is subtle difference, unbindable mounts cannot be bind mounted but can be cloned during clone-namespace. Signed-off-by: Ram Pai Signed-off-by: Al Viro Signed-off-by: Linus Torvalds --- fs/pnode.c | 2 ++ 1 file changed, 2 insertions(+) (limited to 'fs/pnode.c') diff --git a/fs/pnode.c b/fs/pnode.c index 3e266c5a3071..aeeec8ba8dd2 100644 --- a/fs/pnode.c +++ b/fs/pnode.c @@ -82,6 +82,8 @@ void change_mnt_propagation(struct vfsmount *mnt, int type) if (type != MS_SLAVE) { list_del_init(&mnt->mnt_slave); mnt->mnt_master = NULL; + if (type == MS_UNBINDABLE) + mnt->mnt_flags |= MNT_UNBINDABLE; } } -- cgit