From 616511d039af402670de8500d0e24495113a9cab Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Mon, 22 Jun 2009 15:09:13 -0400 Subject: VFS: Uninline the function put_mnt_ns() In order to allow modules to use it without having to export vfsmount_lock. Signed-off-by: Trond Myklebust Signed-off-by: Linus Torvalds --- fs/namespace.c | 8 ++++++-- 1 file changed, 6 insertions(+), 2 deletions(-) (limited to 'fs/namespace.c') diff --git a/fs/namespace.c b/fs/namespace.c index 2dd333b0fe7f..6645846f2056 100644 --- a/fs/namespace.c +++ b/fs/namespace.c @@ -2246,10 +2246,14 @@ void __init mnt_init(void) init_mount_tree(); } -void __put_mnt_ns(struct mnt_namespace *ns) +void put_mnt_ns(struct mnt_namespace *ns) { - struct vfsmount *root = ns->root; + struct vfsmount *root; LIST_HEAD(umount_list); + + if (!atomic_dec_and_lock(&ns->count, &vfsmount_lock)) + return; + root = ns->root; ns->root = NULL; spin_unlock(&vfsmount_lock); down_write(&namespace_sem); -- cgit v1.2.3 From cf8d2c11cb77f129675478792122f50827e5b0ae Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Mon, 22 Jun 2009 15:09:13 -0400 Subject: VFS: Add VFS helper functions for setting up private namespaces The purpose of this patch is to improve the remote mount path lookup support for distributed filesystems such as the NFSv4 client. When given a mount command of the form "mount server:/foo/bar /mnt", the NFSv4 client is required to look up the filehandle for "server:/", and then look up each component of the remote mount path "foo/bar" in order to find the directory that is actually going to be mounted on /mnt. Following that remote mount path may involve following symlinks, crossing server-side mount points and even following referrals to filesystem volumes on other servers. Since the standard VFS path lookup code already supports walking paths that contain all these features (using in-kernel automounts for following referrals) we would like to be able to reuse that rather than duplicate the full path traversal functionality in the NFSv4 client code. This patch therefore defines a VFS helper function create_mnt_ns(), that sets up a temporary filesystem namespace and attaches a root filesystem to it. It exports the create_mnt_ns() and put_mnt_ns() function for use by filesystem modules. Signed-off-by: Trond Myklebust Signed-off-by: Linus Torvalds --- fs/namespace.c | 45 +++++++++++++++++++++++++++++++++++++-------- 1 file changed, 37 insertions(+), 8 deletions(-) (limited to 'fs/namespace.c') diff --git a/fs/namespace.c b/fs/namespace.c index 6645846f2056..a7bea8c8bd46 100644 --- a/fs/namespace.c +++ b/fs/namespace.c @@ -1937,6 +1937,21 @@ dput_out: return retval; } +static struct mnt_namespace *alloc_mnt_ns(void) +{ + struct mnt_namespace *new_ns; + + new_ns = kmalloc(sizeof(struct mnt_namespace), GFP_KERNEL); + if (!new_ns) + return ERR_PTR(-ENOMEM); + atomic_set(&new_ns->count, 1); + new_ns->root = NULL; + INIT_LIST_HEAD(&new_ns->list); + init_waitqueue_head(&new_ns->poll); + new_ns->event = 0; + return new_ns; +} + /* * Allocate a new namespace structure and populate it with contents * copied from the namespace of the passed in task structure. @@ -1948,14 +1963,9 @@ static struct mnt_namespace *dup_mnt_ns(struct mnt_namespace *mnt_ns, struct vfsmount *rootmnt = NULL, *pwdmnt = NULL; struct vfsmount *p, *q; - new_ns = kmalloc(sizeof(struct mnt_namespace), GFP_KERNEL); - if (!new_ns) - return ERR_PTR(-ENOMEM); - - atomic_set(&new_ns->count, 1); - INIT_LIST_HEAD(&new_ns->list); - init_waitqueue_head(&new_ns->poll); - new_ns->event = 0; + new_ns = alloc_mnt_ns(); + if (IS_ERR(new_ns)) + return new_ns; down_write(&namespace_sem); /* First pass: copy the tree topology */ @@ -2019,6 +2029,24 @@ struct mnt_namespace *copy_mnt_ns(unsigned long flags, struct mnt_namespace *ns, return new_ns; } +/** + * create_mnt_ns - creates a private namespace and adds a root filesystem + * @mnt: pointer to the new root filesystem mountpoint + */ +struct mnt_namespace *create_mnt_ns(struct vfsmount *mnt) +{ + struct mnt_namespace *new_ns; + + new_ns = alloc_mnt_ns(); + if (!IS_ERR(new_ns)) { + mnt->mnt_ns = new_ns; + new_ns->root = mnt; + list_add(&new_ns->list, &new_ns->root->mnt_list); + } + return new_ns; +} +EXPORT_SYMBOL(create_mnt_ns); + SYSCALL_DEFINE5(mount, char __user *, dev_name, char __user *, dir_name, char __user *, type, unsigned long, flags, void __user *, data) { @@ -2264,3 +2292,4 @@ void put_mnt_ns(struct mnt_namespace *ns) release_mounts(&umount_list); kfree(ns); } +EXPORT_SYMBOL(put_mnt_ns); -- cgit v1.2.3 From 3b22edc5730b87d360ee7dd7143397ba09b73a47 Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Tue, 23 Jun 2009 17:29:49 -0400 Subject: VFS: Switch init_mount_tree() to use the new create_mnt_ns() helper Eliminates some duplicated code... Signed-off-by: Trond Myklebust Signed-off-by: Al Viro --- fs/namespace.c | 11 ++--------- 1 file changed, 2 insertions(+), 9 deletions(-) (limited to 'fs/namespace.c') diff --git a/fs/namespace.c b/fs/namespace.c index a7bea8c8bd46..4a86b8595164 100644 --- a/fs/namespace.c +++ b/fs/namespace.c @@ -2222,16 +2222,9 @@ static void __init init_mount_tree(void) mnt = do_kern_mount("rootfs", 0, "rootfs", NULL); if (IS_ERR(mnt)) panic("Can't create rootfs"); - ns = kmalloc(sizeof(*ns), GFP_KERNEL); - if (!ns) + ns = create_mnt_ns(mnt); + if (IS_ERR(ns)) panic("Can't allocate initial namespace"); - atomic_set(&ns->count, 1); - INIT_LIST_HEAD(&ns->list); - init_waitqueue_head(&ns->poll); - ns->event = 0; - list_add(&mnt->mnt_list, &ns->list); - ns->root = mnt; - mnt->mnt_ns = ns; init_task.nsproxy->mnt_ns = ns; get_mnt_ns(ns); -- cgit v1.2.3 From f21f62208a6f60e2e05440b2e438d9541822dc4d Mon Sep 17 00:00:00 2001 From: Al Viro Date: Wed, 24 Jun 2009 03:12:00 -0400 Subject: ... and the same for vfsmount id/mount group id Signed-off-by: Al Viro --- fs/namespace.c | 26 ++++++++++++++++++++++---- 1 file changed, 22 insertions(+), 4 deletions(-) (limited to 'fs/namespace.c') diff --git a/fs/namespace.c b/fs/namespace.c index 4a86b8595164..3dc283fd4716 100644 --- a/fs/namespace.c +++ b/fs/namespace.c @@ -42,6 +42,8 @@ __cacheline_aligned_in_smp DEFINE_SPINLOCK(vfsmount_lock); static int event; static DEFINE_IDA(mnt_id_ida); static DEFINE_IDA(mnt_group_ida); +static int mnt_id_start = 0; +static int mnt_group_start = 1; static struct list_head *mount_hashtable __read_mostly; static struct kmem_cache *mnt_cache __read_mostly; @@ -69,7 +71,9 @@ static int mnt_alloc_id(struct vfsmount *mnt) retry: ida_pre_get(&mnt_id_ida, GFP_KERNEL); spin_lock(&vfsmount_lock); - res = ida_get_new(&mnt_id_ida, &mnt->mnt_id); + res = ida_get_new_above(&mnt_id_ida, mnt_id_start, &mnt->mnt_id); + if (!res) + mnt_id_start = mnt->mnt_id + 1; spin_unlock(&vfsmount_lock); if (res == -EAGAIN) goto retry; @@ -79,8 +83,11 @@ retry: static void mnt_free_id(struct vfsmount *mnt) { + int id = mnt->mnt_id; spin_lock(&vfsmount_lock); - ida_remove(&mnt_id_ida, mnt->mnt_id); + ida_remove(&mnt_id_ida, id); + if (mnt_id_start > id) + mnt_id_start = id; spin_unlock(&vfsmount_lock); } @@ -91,10 +98,18 @@ static void mnt_free_id(struct vfsmount *mnt) */ static int mnt_alloc_group_id(struct vfsmount *mnt) { + int res; + if (!ida_pre_get(&mnt_group_ida, GFP_KERNEL)) return -ENOMEM; - return ida_get_new_above(&mnt_group_ida, 1, &mnt->mnt_group_id); + res = ida_get_new_above(&mnt_group_ida, + mnt_group_start, + &mnt->mnt_group_id); + if (!res) + mnt_group_start = mnt->mnt_group_id + 1; + + return res; } /* @@ -102,7 +117,10 @@ static int mnt_alloc_group_id(struct vfsmount *mnt) */ void mnt_release_group_id(struct vfsmount *mnt) { - ida_remove(&mnt_group_ida, mnt->mnt_group_id); + int id = mnt->mnt_group_id; + ida_remove(&mnt_group_ida, id); + if (mnt_group_start > id) + mnt_group_start = id; mnt->mnt_group_id = 0; } -- cgit v1.2.3 From b43f3cbd21ffbd719fd4fa6642bfe6af255ded34 Mon Sep 17 00:00:00 2001 From: Alexey Dobriyan Date: Wed, 8 Jul 2009 01:54:37 +0400 Subject: headers: mnt_namespace.h redux Fix various silly problems wrt mnt_namespace.h: - exit_mnt_ns() isn't used, remove it - done that, sched.h and nsproxy.h inclusions aren't needed - mount.h inclusion was need for vfsmount_lock, but no longer - remove mnt_namespace.h inclusion from files which don't use anything from mnt_namespace.h Signed-off-by: Alexey Dobriyan Signed-off-by: Linus Torvalds --- fs/namespace.c | 1 + 1 file changed, 1 insertion(+) (limited to 'fs/namespace.c') diff --git a/fs/namespace.c b/fs/namespace.c index 3dc283fd4716..277c28a63ead 100644 --- a/fs/namespace.c +++ b/fs/namespace.c @@ -22,6 +22,7 @@ #include #include #include +#include #include #include #include -- cgit v1.2.3 From 2d8dd38a5aa0cc2490bbad9b75e77fa154e1d145 Mon Sep 17 00:00:00 2001 From: OGAWA Hirofumi Date: Thu, 6 Aug 2009 15:07:39 -0700 Subject: vfs: mnt_want_write_file(): fix special file handling I suspect that mnt_want_write_file() may have wrong assumption. I think mnt_want_write_file() is assuming it increments ->mnt_writers if (file->f_mode & FMODE_WRITE). But, if it's special_file(), it is false? Signed-off-by: OGAWA Hirofumi Acked-by: Dave Hansen Cc: Al Viro Cc: Nick Piggin Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/namespace.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) (limited to 'fs/namespace.c') diff --git a/fs/namespace.c b/fs/namespace.c index 277c28a63ead..7230787d18b0 100644 --- a/fs/namespace.c +++ b/fs/namespace.c @@ -316,7 +316,8 @@ EXPORT_SYMBOL_GPL(mnt_clone_write); */ int mnt_want_write_file(struct file *file) { - if (!(file->f_mode & FMODE_WRITE)) + struct inode *inode = file->f_dentry->d_inode; + if (!(file->f_mode & FMODE_WRITE) || special_file(inode->i_mode)) return mnt_want_write(file->f_path.mnt); else return mnt_clone_write(file->f_path.mnt); -- cgit v1.2.3