From dd5cae6e9772ecc62fd374f7a8ec10eb51c96c4d Mon Sep 17 00:00:00 2001
From: Al Viro <viro@zeniv.linux.org.uk>
Date: Tue, 7 Apr 2009 12:21:18 -0400
Subject: Don't bother with check_mnt() in do_add_mount() on shrinkable ones

These guys are what we add as submounts; checks for "is that attached in
our namespace" are simply irrelevant for those and counterproductive for
use of private vfsmount trees a-la what NFS folks want.

Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
---
 fs/namespace.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

(limited to 'fs/namespace.c')

diff --git a/fs/namespace.c b/fs/namespace.c
index 134d494158d..88a904d5aa2 100644
--- a/fs/namespace.c
+++ b/fs/namespace.c
@@ -1698,7 +1698,7 @@ int do_add_mount(struct vfsmount *newmnt, struct path *path,
 	       follow_down(&path->mnt, &path->dentry))
 		;
 	err = -EINVAL;
-	if (!check_mnt(path->mnt))
+	if (!(mnt_flags & MNT_SHRINKABLE) && !check_mnt(path->mnt))
 		goto unlock;
 
 	/* Refuse the same filesystem on the same mount point */
-- 
cgit v1.2.3-70-g09d2


From 589ff870ed60a9ebdd5ec99ec3f5afe1282fe151 Mon Sep 17 00:00:00 2001
From: Al Viro <viro@zeniv.linux.org.uk>
Date: Sat, 18 Apr 2009 03:28:19 -0400
Subject: Switch collect_mounts() to struct path

Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
---
 fs/namespace.c      | 4 ++--
 include/linux/fs.h  | 2 +-
 kernel/audit_tree.c | 6 +++---
 3 files changed, 6 insertions(+), 6 deletions(-)

(limited to 'fs/namespace.c')

diff --git a/fs/namespace.c b/fs/namespace.c
index 88a904d5aa2..c85962206aa 100644
--- a/fs/namespace.c
+++ b/fs/namespace.c
@@ -1253,11 +1253,11 @@ Enomem:
 	return NULL;
 }
 
-struct vfsmount *collect_mounts(struct vfsmount *mnt, struct dentry *dentry)
+struct vfsmount *collect_mounts(struct path *path)
 {
 	struct vfsmount *tree;
 	down_write(&namespace_sem);
-	tree = copy_tree(mnt, dentry, CL_COPY_ALL | CL_PRIVATE);
+	tree = copy_tree(path->mnt, path->dentry, CL_COPY_ALL | CL_PRIVATE);
 	up_write(&namespace_sem);
 	return tree;
 }
diff --git a/include/linux/fs.h b/include/linux/fs.h
index 323b5ce474c..03fb2102b8f 100644
--- a/include/linux/fs.h
+++ b/include/linux/fs.h
@@ -1800,7 +1800,7 @@ extern struct vfsmount *kern_mount_data(struct file_system_type *, void *data);
 extern int may_umount_tree(struct vfsmount *);
 extern int may_umount(struct vfsmount *);
 extern long do_mount(char *, char *, char *, unsigned long, void *);
-extern struct vfsmount *collect_mounts(struct vfsmount *, struct dentry *);
+extern struct vfsmount *collect_mounts(struct path *);
 extern void drop_collected_mounts(struct vfsmount *);
 
 extern int vfs_statfs(struct dentry *, struct kstatfs *);
diff --git a/kernel/audit_tree.c b/kernel/audit_tree.c
index 6e7351739a8..1f6396d7668 100644
--- a/kernel/audit_tree.c
+++ b/kernel/audit_tree.c
@@ -568,7 +568,7 @@ void audit_trim_trees(void)
 		if (err)
 			goto skip_it;
 
-		root_mnt = collect_mounts(path.mnt, path.dentry);
+		root_mnt = collect_mounts(&path);
 		path_put(&path);
 		if (!root_mnt)
 			goto skip_it;
@@ -660,7 +660,7 @@ int audit_add_tree_rule(struct audit_krule *rule)
 	err = kern_path(tree->pathname, 0, &path);
 	if (err)
 		goto Err;
-	mnt = collect_mounts(path.mnt, path.dentry);
+	mnt = collect_mounts(&path);
 	path_put(&path);
 	if (!mnt) {
 		err = -ENOMEM;
@@ -720,7 +720,7 @@ int audit_tag_tree(char *old, char *new)
 	err = kern_path(new, 0, &path);
 	if (err)
 		return err;
-	tagged = collect_mounts(path.mnt, path.dentry);
+	tagged = collect_mounts(&path);
 	path_put(&path);
 	if (!tagged)
 		return -ENOMEM;
-- 
cgit v1.2.3-70-g09d2


From 9393bd07cf218ca51d0e627653f906a9d76a9131 Mon Sep 17 00:00:00 2001
From: Al Viro <viro@zeniv.linux.org.uk>
Date: Sat, 18 Apr 2009 13:58:15 -0400
Subject: switch follow_down()

Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
---
 fs/afs/mntpt.c         |  2 +-
 fs/autofs/dirhash.c    |  5 ++---
 fs/autofs4/autofs_i.h  |  6 +++---
 fs/autofs4/dev-ioctl.c |  2 +-
 fs/autofs4/expire.c    | 15 +++++++--------
 fs/autofs4/root.c      |  7 +++----
 fs/cifs/cifs_dfs_ref.c |  2 +-
 fs/namei.c             | 12 ++++++------
 fs/namespace.c         |  4 ++--
 fs/nfs/namespace.c     |  2 +-
 fs/nfsd/vfs.c          |  3 +--
 include/linux/namei.h  |  2 +-
 12 files changed, 29 insertions(+), 33 deletions(-)

(limited to 'fs/namespace.c')

diff --git a/fs/afs/mntpt.c b/fs/afs/mntpt.c
index 2b9e2d03a39..c52be53f694 100644
--- a/fs/afs/mntpt.c
+++ b/fs/afs/mntpt.c
@@ -244,7 +244,7 @@ static void *afs_mntpt_follow_link(struct dentry *dentry, struct nameidata *nd)
 	case -EBUSY:
 		/* someone else made a mount here whilst we were busy */
 		while (d_mountpoint(nd->path.dentry) &&
-		       follow_down(&nd->path.mnt, &nd->path.dentry))
+		       follow_down(&nd->path))
 			;
 		err = 0;
 	default:
diff --git a/fs/autofs/dirhash.c b/fs/autofs/dirhash.c
index 4eb4d8dfb2f..2316e944a10 100644
--- a/fs/autofs/dirhash.c
+++ b/fs/autofs/dirhash.c
@@ -85,13 +85,12 @@ struct autofs_dir_ent *autofs_expire(struct super_block *sb,
 		}
 		path.mnt = mnt;
 		path_get(&path);
-		if (!follow_down(&path.mnt, &path.dentry)) {
+		if (!follow_down(&path)) {
 			path_put(&path);
 			DPRINTK(("autofs: not expirable (not a mounted directory): %s\n", ent->name));
 			continue;
 		}
-		while (d_mountpoint(path.dentry) &&
-		       follow_down(&path.mnt, &path.dentry))
+		while (d_mountpoint(path.dentry) && follow_down(&path));
 			;
 		umount_ok = may_umount(path.mnt);
 		path_put(&path);
diff --git a/fs/autofs4/autofs_i.h b/fs/autofs4/autofs_i.h
index b7ff33c6310..8f7cdde4173 100644
--- a/fs/autofs4/autofs_i.h
+++ b/fs/autofs4/autofs_i.h
@@ -223,12 +223,12 @@ int autofs4_wait(struct autofs_sb_info *,struct dentry *, enum autofs_notify);
 int autofs4_wait_release(struct autofs_sb_info *,autofs_wqt_t,int);
 void autofs4_catatonic_mode(struct autofs_sb_info *);
 
-static inline int autofs4_follow_mount(struct vfsmount **mnt, struct dentry **dentry)
+static inline int autofs4_follow_mount(struct path *path)
 {
 	int res = 0;
 
-	while (d_mountpoint(*dentry)) {
-		int followed = follow_down(mnt, dentry);
+	while (d_mountpoint(path->dentry)) {
+		int followed = follow_down(path);
 		if (!followed)
 			break;
 		res = 1;
diff --git a/fs/autofs4/dev-ioctl.c b/fs/autofs4/dev-ioctl.c
index 670407576b2..f3da2eb51f5 100644
--- a/fs/autofs4/dev-ioctl.c
+++ b/fs/autofs4/dev-ioctl.c
@@ -562,7 +562,7 @@ static int autofs_dev_ioctl_ismountpoint(struct file *fp,
 		err = have_submounts(path.dentry);
 
 		if (path.mnt->mnt_mountpoint != path.mnt->mnt_root) {
-			if (follow_down(&path.mnt, &path.dentry))
+			if (follow_down(&path))
 				magic = path.mnt->mnt_sb->s_magic;
 		}
 	}
diff --git a/fs/autofs4/expire.c b/fs/autofs4/expire.c
index 3077d8f1652..aa39ae83f01 100644
--- a/fs/autofs4/expire.c
+++ b/fs/autofs4/expire.c
@@ -48,19 +48,19 @@ static inline int autofs4_can_expire(struct dentry *dentry,
 static int autofs4_mount_busy(struct vfsmount *mnt, struct dentry *dentry)
 {
 	struct dentry *top = dentry;
+	struct path path = {.mnt = mnt, .dentry = dentry};
 	int status = 1;
 
 	DPRINTK("dentry %p %.*s",
 		dentry, (int)dentry->d_name.len, dentry->d_name.name);
 
-	mntget(mnt);
-	dget(dentry);
+	path_get(&path);
 
-	if (!follow_down(&mnt, &dentry))
+	if (!follow_down(&path))
 		goto done;
 
-	if (is_autofs4_dentry(dentry)) {
-		struct autofs_sb_info *sbi = autofs4_sbi(dentry->d_sb);
+	if (is_autofs4_dentry(path.dentry)) {
+		struct autofs_sb_info *sbi = autofs4_sbi(path.dentry->d_sb);
 
 		/* This is an autofs submount, we can't expire it */
 		if (autofs_type_indirect(sbi->type))
@@ -70,7 +70,7 @@ static int autofs4_mount_busy(struct vfsmount *mnt, struct dentry *dentry)
 		 * Otherwise it's an offset mount and we need to check
 		 * if we can umount its mount, if there is one.
 		 */
-		if (!d_mountpoint(dentry)) {
+		if (!d_mountpoint(path.dentry)) {
 			status = 0;
 			goto done;
 		}
@@ -86,8 +86,7 @@ static int autofs4_mount_busy(struct vfsmount *mnt, struct dentry *dentry)
 	status = 0;
 done:
 	DPRINTK("returning = %d", status);
-	dput(dentry);
-	mntput(mnt);
+	path_put(&path);
 	return status;
 }
 
diff --git a/fs/autofs4/root.c b/fs/autofs4/root.c
index e383bf0334f..b96a3c57359 100644
--- a/fs/autofs4/root.c
+++ b/fs/autofs4/root.c
@@ -181,7 +181,7 @@ static void *autofs4_follow_link(struct dentry *dentry, struct nameidata *nd)
 		nd->flags);
 	/*
 	 * For an expire of a covered direct or offset mount we need
-	 * to beeak out of follow_down() at the autofs mount trigger
+	 * to break out of follow_down() at the autofs mount trigger
 	 * (d_mounted--), so we can see the expiring flag, and manage
 	 * the blocking and following here until the expire is completed.
 	 */
@@ -190,7 +190,7 @@ static void *autofs4_follow_link(struct dentry *dentry, struct nameidata *nd)
 		if (ino->flags & AUTOFS_INF_EXPIRING) {
 			spin_unlock(&sbi->fs_lock);
 			/* Follow down to our covering mount. */
-			if (!follow_down(&nd->path.mnt, &nd->path.dentry))
+			if (!follow_down(&nd->path))
 				goto done;
 			goto follow;
 		}
@@ -230,8 +230,7 @@ follow:
 	 * to follow it.
 	 */
 	if (d_mountpoint(dentry)) {
-		if (!autofs4_follow_mount(&nd->path.mnt,
-					  &nd->path.dentry)) {
+		if (!autofs4_follow_mount(&nd->path)) {
 			status = -ENOENT;
 			goto out_error;
 		}
diff --git a/fs/cifs/cifs_dfs_ref.c b/fs/cifs/cifs_dfs_ref.c
index 83d62759c7c..3bb11be8b6a 100644
--- a/fs/cifs/cifs_dfs_ref.c
+++ b/fs/cifs/cifs_dfs_ref.c
@@ -275,7 +275,7 @@ static int add_mount_helper(struct vfsmount *newmnt, struct nameidata *nd,
 	case -EBUSY:
 		/* someone else made a mount here whilst we were busy */
 		while (d_mountpoint(nd->path.dentry) &&
-		       follow_down(&nd->path.mnt, &nd->path.dentry))
+		       follow_down(&nd->path))
 			;
 		err = 0;
 	default:
diff --git a/fs/namei.c b/fs/namei.c
index 8c1f48ae68e..4d49a3eee6d 100644
--- a/fs/namei.c
+++ b/fs/namei.c
@@ -731,16 +731,16 @@ static void follow_mount(struct vfsmount **mnt, struct dentry **dentry)
 /* no need for dcache_lock, as serialization is taken care in
  * namespace.c
  */
-int follow_down(struct vfsmount **mnt, struct dentry **dentry)
+int follow_down(struct path *path)
 {
 	struct vfsmount *mounted;
 
-	mounted = lookup_mnt(*mnt, *dentry);
+	mounted = lookup_mnt(path->mnt, path->dentry);
 	if (mounted) {
-		dput(*dentry);
-		mntput(*mnt);
-		*mnt = mounted;
-		*dentry = dget(mounted->mnt_root);
+		dput(path->dentry);
+		mntput(path->mnt);
+		path->mnt = mounted;
+		path->dentry = dget(mounted->mnt_root);
 		return 1;
 	}
 	return 0;
diff --git a/fs/namespace.c b/fs/namespace.c
index c85962206aa..ba5237be1cf 100644
--- a/fs/namespace.c
+++ b/fs/namespace.c
@@ -1601,7 +1601,7 @@ static int do_move_mount(struct path *path, char *old_name)
 
 	down_write(&namespace_sem);
 	while (d_mountpoint(path->dentry) &&
-	       follow_down(&path->mnt, &path->dentry))
+	       follow_down(path))
 		;
 	err = -EINVAL;
 	if (!check_mnt(path->mnt) || !check_mnt(old_path.mnt))
@@ -1695,7 +1695,7 @@ int do_add_mount(struct vfsmount *newmnt, struct path *path,
 	down_write(&namespace_sem);
 	/* Something was mounted here while we slept */
 	while (d_mountpoint(path->dentry) &&
-	       follow_down(&path->mnt, &path->dentry))
+	       follow_down(path))
 		;
 	err = -EINVAL;
 	if (!(mnt_flags & MNT_SHRINKABLE) && !check_mnt(path->mnt))
diff --git a/fs/nfs/namespace.c b/fs/nfs/namespace.c
index 64a288ee046..f01caec8446 100644
--- a/fs/nfs/namespace.c
+++ b/fs/nfs/namespace.c
@@ -154,7 +154,7 @@ out_err:
 	goto out;
 out_follow:
 	while (d_mountpoint(nd->path.dentry) &&
-	       follow_down(&nd->path.mnt, &nd->path.dentry))
+	       follow_down(&nd->path))
 		;
 	err = 0;
 	goto out;
diff --git a/fs/nfsd/vfs.c b/fs/nfsd/vfs.c
index 7b2b3f77532..99f83575359 100644
--- a/fs/nfsd/vfs.c
+++ b/fs/nfsd/vfs.c
@@ -105,8 +105,7 @@ nfsd_cross_mnt(struct svc_rqst *rqstp, struct dentry **dpp,
 			    .dentry = dget(dentry)};
 	int err = 0;
 
-	while (follow_down(&path.mnt, &path.dentry) &&
-	       d_mountpoint(path.dentry))
+	while (d_mountpoint(path.dentry) && follow_down(&path))
 		;
 
 	exp2 = rqst_exp_get_by_name(rqstp, &path);
diff --git a/include/linux/namei.h b/include/linux/namei.h
index 9cd5a717be3..d870ae2faed 100644
--- a/include/linux/namei.h
+++ b/include/linux/namei.h
@@ -78,7 +78,7 @@ extern void release_open_intent(struct nameidata *);
 extern struct dentry *lookup_one_len(const char *, struct dentry *, int);
 extern struct dentry *lookup_one_noperm(const char *, struct dentry *);
 
-extern int follow_down(struct vfsmount **, struct dentry **);
+extern int follow_down(struct path *);
 extern int follow_up(struct path *);
 
 extern struct dentry *lock_rename(struct dentry *, struct dentry *);
-- 
cgit v1.2.3-70-g09d2


From 1c755af4df75996b0dd4b7e6cacaf9d57a6ef2ef Mon Sep 17 00:00:00 2001
From: Al Viro <viro@zeniv.linux.org.uk>
Date: Sat, 18 Apr 2009 14:06:57 -0400
Subject: switch lookup_mnt()

Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
---
 fs/namei.c             | 6 +++---
 fs/namespace.c         | 4 ++--
 include/linux/dcache.h | 2 +-
 3 files changed, 6 insertions(+), 6 deletions(-)

(limited to 'fs/namespace.c')

diff --git a/fs/namei.c b/fs/namei.c
index c006bc61d1e..527119afb6a 100644
--- a/fs/namei.c
+++ b/fs/namei.c
@@ -702,7 +702,7 @@ static int __follow_mount(struct path *path)
 {
 	int res = 0;
 	while (d_mountpoint(path->dentry)) {
-		struct vfsmount *mounted = lookup_mnt(path->mnt, path->dentry);
+		struct vfsmount *mounted = lookup_mnt(path);
 		if (!mounted)
 			break;
 		dput(path->dentry);
@@ -718,7 +718,7 @@ static int __follow_mount(struct path *path)
 static void follow_mount(struct path *path)
 {
 	while (d_mountpoint(path->dentry)) {
-		struct vfsmount *mounted = lookup_mnt(path->mnt, path->dentry);
+		struct vfsmount *mounted = lookup_mnt(path);
 		if (!mounted)
 			break;
 		dput(path->dentry);
@@ -735,7 +735,7 @@ int follow_down(struct path *path)
 {
 	struct vfsmount *mounted;
 
-	mounted = lookup_mnt(path->mnt, path->dentry);
+	mounted = lookup_mnt(path);
 	if (mounted) {
 		dput(path->dentry);
 		mntput(path->mnt);
diff --git a/fs/namespace.c b/fs/namespace.c
index ba5237be1cf..b94ad3d685f 100644
--- a/fs/namespace.c
+++ b/fs/namespace.c
@@ -442,11 +442,11 @@ struct vfsmount *__lookup_mnt(struct vfsmount *mnt, struct dentry *dentry,
  * lookup_mnt increments the ref count before returning
  * the vfsmount struct.
  */
-struct vfsmount *lookup_mnt(struct vfsmount *mnt, struct dentry *dentry)
+struct vfsmount *lookup_mnt(struct path *path)
 {
 	struct vfsmount *child_mnt;
 	spin_lock(&vfsmount_lock);
-	if ((child_mnt = __lookup_mnt(mnt, dentry, 1)))
+	if ((child_mnt = __lookup_mnt(path->mnt, path->dentry, 1)))
 		mntget(child_mnt);
 	spin_unlock(&vfsmount_lock);
 	return child_mnt;
diff --git a/include/linux/dcache.h b/include/linux/dcache.h
index 97978004338..72ce2ae8859 100644
--- a/include/linux/dcache.h
+++ b/include/linux/dcache.h
@@ -370,7 +370,7 @@ static inline int d_mountpoint(struct dentry *dentry)
 	return dentry->d_mounted;
 }
 
-extern struct vfsmount *lookup_mnt(struct vfsmount *, struct dentry *);
+extern struct vfsmount *lookup_mnt(struct path *);
 extern struct dentry *lookup_create(struct nameidata *nd, int is_dir);
 
 extern int sysctl_vfs_cache_pressure;
-- 
cgit v1.2.3-70-g09d2


From d3ef3d7351ccfbef3e5d926efc5ee332136f40d4 Mon Sep 17 00:00:00 2001
From: "npiggin@suse.de" <npiggin@suse.de>
Date: Sun, 26 Apr 2009 20:25:54 +1000
Subject: fs: mnt_want_write speedup

This patch speeds up lmbench lat_mmap test by about 8%. lat_mmap is set up
basically to mmap a 64MB file on tmpfs, fault in its pages, then unmap it.
A microbenchmark yes, but it exercises some important paths in the mm.

Before:
 avg = 501.9
 std = 14.7773

After:
 avg = 462.286
 std = 5.46106

(50 runs of each, stddev gives a reasonable confidence, but there is quite
a bit of variation there still)

It does this by removing the complex per-cpu locking and counter-cache and
replaces it with a percpu counter in struct vfsmount. This makes the code
much simpler, and avoids spinlocks (although the msync is still pretty
costly, unfortunately). It results in about 900 bytes smaller code too. It
does increase the size of a vfsmount, however.

It should also give a speedup on large systems if CPUs are frequently operating
on different mounts (because the existing scheme has to operate on an atomic in
the struct vfsmount when switching between mounts). But I'm most interested in
the single threaded path performance for the moment.

[AV: minor cleanup]

Cc: Dave Hansen <haveblue@us.ibm.com>
Signed-off-by: Nick Piggin <npiggin@suse.de>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
---
 fs/namespace.c        | 268 +++++++++++++++++---------------------------------
 include/linux/mount.h |  21 ++--
 2 files changed, 106 insertions(+), 183 deletions(-)

(limited to 'fs/namespace.c')

diff --git a/fs/namespace.c b/fs/namespace.c
index b94ad3d685f..22ae06ad751 100644
--- a/fs/namespace.c
+++ b/fs/namespace.c
@@ -131,10 +131,20 @@ struct vfsmount *alloc_vfsmnt(const char *name)
 		INIT_LIST_HEAD(&mnt->mnt_share);
 		INIT_LIST_HEAD(&mnt->mnt_slave_list);
 		INIT_LIST_HEAD(&mnt->mnt_slave);
-		atomic_set(&mnt->__mnt_writers, 0);
+#ifdef CONFIG_SMP
+		mnt->mnt_writers = alloc_percpu(int);
+		if (!mnt->mnt_writers)
+			goto out_free_devname;
+#else
+		mnt->mnt_writers = 0;
+#endif
 	}
 	return mnt;
 
+#ifdef CONFIG_SMP
+out_free_devname:
+	kfree(mnt->mnt_devname);
+#endif
 out_free_id:
 	mnt_free_id(mnt);
 out_free_cache:
@@ -171,65 +181,38 @@ int __mnt_is_readonly(struct vfsmount *mnt)
 }
 EXPORT_SYMBOL_GPL(__mnt_is_readonly);
 
-struct mnt_writer {
-	/*
-	 * If holding multiple instances of this lock, they
-	 * must be ordered by cpu number.
-	 */
-	spinlock_t lock;
-	struct lock_class_key lock_class; /* compiles out with !lockdep */
-	unsigned long count;
-	struct vfsmount *mnt;
-} ____cacheline_aligned_in_smp;
-static DEFINE_PER_CPU(struct mnt_writer, mnt_writers);
+static inline void inc_mnt_writers(struct vfsmount *mnt)
+{
+#ifdef CONFIG_SMP
+	(*per_cpu_ptr(mnt->mnt_writers, smp_processor_id()))++;
+#else
+	mnt->mnt_writers++;
+#endif
+}
 
-static int __init init_mnt_writers(void)
+static inline void dec_mnt_writers(struct vfsmount *mnt)
 {
-	int cpu;
-	for_each_possible_cpu(cpu) {
-		struct mnt_writer *writer = &per_cpu(mnt_writers, cpu);
-		spin_lock_init(&writer->lock);
-		lockdep_set_class(&writer->lock, &writer->lock_class);
-		writer->count = 0;
-	}
-	return 0;
+#ifdef CONFIG_SMP
+	(*per_cpu_ptr(mnt->mnt_writers, smp_processor_id()))--;
+#else
+	mnt->mnt_writers--;
+#endif
 }
-fs_initcall(init_mnt_writers);
 
-static void unlock_mnt_writers(void)
+static unsigned int count_mnt_writers(struct vfsmount *mnt)
 {
+#ifdef CONFIG_SMP
+	unsigned int count = 0;
 	int cpu;
-	struct mnt_writer *cpu_writer;
 
 	for_each_possible_cpu(cpu) {
-		cpu_writer = &per_cpu(mnt_writers, cpu);
-		spin_unlock(&cpu_writer->lock);
+		count += *per_cpu_ptr(mnt->mnt_writers, cpu);
 	}
-}
 
-static inline void __clear_mnt_count(struct mnt_writer *cpu_writer)
-{
-	if (!cpu_writer->mnt)
-		return;
-	/*
-	 * This is in case anyone ever leaves an invalid,
-	 * old ->mnt and a count of 0.
-	 */
-	if (!cpu_writer->count)
-		return;
-	atomic_add(cpu_writer->count, &cpu_writer->mnt->__mnt_writers);
-	cpu_writer->count = 0;
-}
- /*
- * must hold cpu_writer->lock
- */
-static inline void use_cpu_writer_for_mount(struct mnt_writer *cpu_writer,
-					  struct vfsmount *mnt)
-{
-	if (cpu_writer->mnt == mnt)
-		return;
-	__clear_mnt_count(cpu_writer);
-	cpu_writer->mnt = mnt;
+	return count;
+#else
+	return mnt->mnt_writers;
+#endif
 }
 
 /*
@@ -253,75 +236,34 @@ static inline void use_cpu_writer_for_mount(struct mnt_writer *cpu_writer,
 int mnt_want_write(struct vfsmount *mnt)
 {
 	int ret = 0;
-	struct mnt_writer *cpu_writer;
 
-	cpu_writer = &get_cpu_var(mnt_writers);
-	spin_lock(&cpu_writer->lock);
+	preempt_disable();
+	inc_mnt_writers(mnt);
+	/*
+	 * The store to inc_mnt_writers must be visible before we pass
+	 * MNT_WRITE_HOLD loop below, so that the slowpath can see our
+	 * incremented count after it has set MNT_WRITE_HOLD.
+	 */
+	smp_mb();
+	while (mnt->mnt_flags & MNT_WRITE_HOLD)
+		cpu_relax();
+	/*
+	 * After the slowpath clears MNT_WRITE_HOLD, mnt_is_readonly will
+	 * be set to match its requirements. So we must not load that until
+	 * MNT_WRITE_HOLD is cleared.
+	 */
+	smp_rmb();
 	if (__mnt_is_readonly(mnt)) {
+		dec_mnt_writers(mnt);
 		ret = -EROFS;
 		goto out;
 	}
-	use_cpu_writer_for_mount(cpu_writer, mnt);
-	cpu_writer->count++;
 out:
-	spin_unlock(&cpu_writer->lock);
-	put_cpu_var(mnt_writers);
+	preempt_enable();
 	return ret;
 }
 EXPORT_SYMBOL_GPL(mnt_want_write);
 
-static void lock_mnt_writers(void)
-{
-	int cpu;
-	struct mnt_writer *cpu_writer;
-
-	for_each_possible_cpu(cpu) {
-		cpu_writer = &per_cpu(mnt_writers, cpu);
-		spin_lock(&cpu_writer->lock);
-		__clear_mnt_count(cpu_writer);
-		cpu_writer->mnt = NULL;
-	}
-}
-
-/*
- * These per-cpu write counts are not guaranteed to have
- * matched increments and decrements on any given cpu.
- * A file open()ed for write on one cpu and close()d on
- * another cpu will imbalance this count.  Make sure it
- * does not get too far out of whack.
- */
-static void handle_write_count_underflow(struct vfsmount *mnt)
-{
-	if (atomic_read(&mnt->__mnt_writers) >=
-	    MNT_WRITER_UNDERFLOW_LIMIT)
-		return;
-	/*
-	 * It isn't necessary to hold all of the locks
-	 * at the same time, but doing it this way makes
-	 * us share a lot more code.
-	 */
-	lock_mnt_writers();
-	/*
-	 * vfsmount_lock is for mnt_flags.
-	 */
-	spin_lock(&vfsmount_lock);
-	/*
-	 * If coalescing the per-cpu writer counts did not
-	 * get us back to a positive writer count, we have
-	 * a bug.
-	 */
-	if ((atomic_read(&mnt->__mnt_writers) < 0) &&
-	    !(mnt->mnt_flags & MNT_IMBALANCED_WRITE_COUNT)) {
-		WARN(1, KERN_DEBUG "leak detected on mount(%p) writers "
-				"count: %d\n",
-			mnt, atomic_read(&mnt->__mnt_writers));
-		/* use the flag to keep the dmesg spam down */
-		mnt->mnt_flags |= MNT_IMBALANCED_WRITE_COUNT;
-	}
-	spin_unlock(&vfsmount_lock);
-	unlock_mnt_writers();
-}
-
 /**
  * mnt_drop_write - give up write access to a mount
  * @mnt: the mount on which to give up write access
@@ -332,37 +274,9 @@ static void handle_write_count_underflow(struct vfsmount *mnt)
  */
 void mnt_drop_write(struct vfsmount *mnt)
 {
-	int must_check_underflow = 0;
-	struct mnt_writer *cpu_writer;
-
-	cpu_writer = &get_cpu_var(mnt_writers);
-	spin_lock(&cpu_writer->lock);
-
-	use_cpu_writer_for_mount(cpu_writer, mnt);
-	if (cpu_writer->count > 0) {
-		cpu_writer->count--;
-	} else {
-		must_check_underflow = 1;
-		atomic_dec(&mnt->__mnt_writers);
-	}
-
-	spin_unlock(&cpu_writer->lock);
-	/*
-	 * Logically, we could call this each time,
-	 * but the __mnt_writers cacheline tends to
-	 * be cold, and makes this expensive.
-	 */
-	if (must_check_underflow)
-		handle_write_count_underflow(mnt);
-	/*
-	 * This could be done right after the spinlock
-	 * is taken because the spinlock keeps us on
-	 * the cpu, and disables preemption.  However,
-	 * putting it here bounds the amount that
-	 * __mnt_writers can underflow.  Without it,
-	 * we could theoretically wrap __mnt_writers.
-	 */
-	put_cpu_var(mnt_writers);
+	preempt_disable();
+	dec_mnt_writers(mnt);
+	preempt_enable();
 }
 EXPORT_SYMBOL_GPL(mnt_drop_write);
 
@@ -370,24 +284,41 @@ static int mnt_make_readonly(struct vfsmount *mnt)
 {
 	int ret = 0;
 
-	lock_mnt_writers();
+	spin_lock(&vfsmount_lock);
+	mnt->mnt_flags |= MNT_WRITE_HOLD;
 	/*
-	 * With all the locks held, this value is stable
+	 * After storing MNT_WRITE_HOLD, we'll read the counters. This store
+	 * should be visible before we do.
 	 */
-	if (atomic_read(&mnt->__mnt_writers) > 0) {
-		ret = -EBUSY;
-		goto out;
-	}
+	smp_mb();
+
 	/*
-	 * nobody can do a successful mnt_want_write() with all
-	 * of the counts in MNT_DENIED_WRITE and the locks held.
+	 * With writers on hold, if this value is zero, then there are
+	 * definitely no active writers (although held writers may subsequently
+	 * increment the count, they'll have to wait, and decrement it after
+	 * seeing MNT_READONLY).
+	 *
+	 * It is OK to have counter incremented on one CPU and decremented on
+	 * another: the sum will add up correctly. The danger would be when we
+	 * sum up each counter, if we read a counter before it is incremented,
+	 * but then read another CPU's count which it has been subsequently
+	 * decremented from -- we would see more decrements than we should.
+	 * MNT_WRITE_HOLD protects against this scenario, because
+	 * mnt_want_write first increments count, then smp_mb, then spins on
+	 * MNT_WRITE_HOLD, so it can't be decremented by another CPU while
+	 * we're counting up here.
 	 */
-	spin_lock(&vfsmount_lock);
-	if (!ret)
+	if (count_mnt_writers(mnt) > 0)
+		ret = -EBUSY;
+	else
 		mnt->mnt_flags |= MNT_READONLY;
+	/*
+	 * MNT_READONLY must become visible before ~MNT_WRITE_HOLD, so writers
+	 * that become unheld will see MNT_READONLY.
+	 */
+	smp_wmb();
+	mnt->mnt_flags &= ~MNT_WRITE_HOLD;
 	spin_unlock(&vfsmount_lock);
-out:
-	unlock_mnt_writers();
 	return ret;
 }
 
@@ -410,6 +341,9 @@ void free_vfsmnt(struct vfsmount *mnt)
 {
 	kfree(mnt->mnt_devname);
 	mnt_free_id(mnt);
+#ifdef CONFIG_SMP
+	free_percpu(mnt->mnt_writers);
+#endif
 	kmem_cache_free(mnt_cache, mnt);
 }
 
@@ -604,38 +538,18 @@ static struct vfsmount *clone_mnt(struct vfsmount *old, struct dentry *root,
 
 static inline void __mntput(struct vfsmount *mnt)
 {
-	int cpu;
 	struct super_block *sb = mnt->mnt_sb;
-	/*
-	 * We don't have to hold all of the locks at the
-	 * same time here because we know that we're the
-	 * last reference to mnt and that no new writers
-	 * can come in.
-	 */
-	for_each_possible_cpu(cpu) {
-		struct mnt_writer *cpu_writer = &per_cpu(mnt_writers, cpu);
-		spin_lock(&cpu_writer->lock);
-		if (cpu_writer->mnt != mnt) {
-			spin_unlock(&cpu_writer->lock);
-			continue;
-		}
-		atomic_add(cpu_writer->count, &mnt->__mnt_writers);
-		cpu_writer->count = 0;
-		/*
-		 * Might as well do this so that no one
-		 * ever sees the pointer and expects
-		 * it to be valid.
-		 */
-		cpu_writer->mnt = NULL;
-		spin_unlock(&cpu_writer->lock);
-	}
 	/*
 	 * This probably indicates that somebody messed
 	 * up a mnt_want/drop_write() pair.  If this
 	 * happens, the filesystem was probably unable
 	 * to make r/w->r/o transitions.
 	 */
-	WARN_ON(atomic_read(&mnt->__mnt_writers));
+	/*
+	 * atomic_dec_and_lock() used to deal with ->mnt_count decrements
+	 * provides barriers, so count_mnt_writers() below is safe.  AV
+	 */
+	WARN_ON(count_mnt_writers(mnt));
 	dput(mnt->mnt_root);
 	free_vfsmnt(mnt);
 	deactivate_super(sb);
diff --git a/include/linux/mount.h b/include/linux/mount.h
index 51f55f903af..ac49c1f8e5c 100644
--- a/include/linux/mount.h
+++ b/include/linux/mount.h
@@ -30,7 +30,7 @@ struct mnt_namespace;
 #define MNT_STRICTATIME 0x80
 
 #define MNT_SHRINKABLE	0x100
-#define MNT_IMBALANCED_WRITE_COUNT	0x200 /* just for debugging */
+#define MNT_WRITE_HOLD	0x200
 
 #define MNT_SHARED	0x1000	/* if the vfsmount is a shared mount */
 #define MNT_UNBINDABLE	0x2000	/* if the vfsmount is a unbindable mount */
@@ -65,13 +65,22 @@ struct vfsmount {
 	int mnt_expiry_mark;		/* true if marked for expiry */
 	int mnt_pinned;
 	int mnt_ghosts;
-	/*
-	 * This value is not stable unless all of the mnt_writers[] spinlocks
-	 * are held, and all mnt_writer[]s on this mount have 0 as their ->count
-	 */
-	atomic_t __mnt_writers;
+#ifdef CONFIG_SMP
+	int *mnt_writers;
+#else
+	int mnt_writers;
+#endif
 };
 
+static inline int *get_mnt_writers_ptr(struct vfsmount *mnt)
+{
+#ifdef CONFIG_SMP
+	return mnt->mnt_writers;
+#else
+	return &mnt->mnt_writers;
+#endif
+}
+
 static inline struct vfsmount *mntget(struct vfsmount *mnt)
 {
 	if (mnt)
-- 
cgit v1.2.3-70-g09d2


From 96029c4e09ccbd73a6d0ed2b29e80bf2586ad7ef Mon Sep 17 00:00:00 2001
From: "npiggin@suse.de" <npiggin@suse.de>
Date: Sun, 26 Apr 2009 20:25:55 +1000
Subject: fs: introduce mnt_clone_write

This patch speeds up lmbench lat_mmap test by about another 2% after the
first patch.

Before:
 avg = 462.286
 std = 5.46106

After:
 avg = 453.12
 std = 9.58257

(50 runs of each, stddev gives a reasonable confidence)

It does this by introducing mnt_clone_write, which avoids some heavyweight
operations of mnt_want_write if called on a vfsmount which we know already
has a write count; and mnt_want_write_file, which can call mnt_clone_write
if the file is open for write.

After these two patches, mnt_want_write and mnt_drop_write go from 7% on
the profile down to 1.3% (including mnt_clone_write).

[AV: mnt_want_write_file() should take file alone and derive mnt from it;
not only all callers have that form, but that's the only mnt about which
we know that it's already held for write if file is opened for write]

Cc: Dave Hansen <haveblue@us.ibm.com>
Signed-off-by: Nick Piggin <npiggin@suse.de>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
---
 fs/file_table.c       |  2 +-
 fs/inode.c            |  2 +-
 fs/namespace.c        | 40 ++++++++++++++++++++++++++++++++++++++++
 fs/open.c             |  4 ++--
 fs/xattr.c            |  4 ++--
 include/linux/mount.h |  4 ++++
 6 files changed, 50 insertions(+), 6 deletions(-)

(limited to 'fs/namespace.c')

diff --git a/fs/file_table.c b/fs/file_table.c
index 54018fe4884..3d66dbcebef 100644
--- a/fs/file_table.c
+++ b/fs/file_table.c
@@ -214,7 +214,7 @@ int init_file(struct file *file, struct vfsmount *mnt, struct dentry *dentry,
 	 */
 	if ((mode & FMODE_WRITE) && !special_file(dentry->d_inode->i_mode)) {
 		file_take_write(file);
-		error = mnt_want_write(mnt);
+		error = mnt_clone_write(mnt);
 		WARN_ON(error);
 	}
 	return error;
diff --git a/fs/inode.c b/fs/inode.c
index ca337014ae2..a88baebf77c 100644
--- a/fs/inode.c
+++ b/fs/inode.c
@@ -1422,7 +1422,7 @@ void file_update_time(struct file *file)
 	if (IS_NOCMTIME(inode))
 		return;
 
-	err = mnt_want_write(file->f_path.mnt);
+	err = mnt_want_write_file(file);
 	if (err)
 		return;
 
diff --git a/fs/namespace.c b/fs/namespace.c
index 22ae06ad751..120b8a6b99e 100644
--- a/fs/namespace.c
+++ b/fs/namespace.c
@@ -264,6 +264,46 @@ out:
 }
 EXPORT_SYMBOL_GPL(mnt_want_write);
 
+/**
+ * mnt_clone_write - get write access to a mount
+ * @mnt: the mount on which to take a write
+ *
+ * This is effectively like mnt_want_write, except
+ * it must only be used to take an extra write reference
+ * on a mountpoint that we already know has a write reference
+ * on it. This allows some optimisation.
+ *
+ * After finished, mnt_drop_write must be called as usual to
+ * drop the reference.
+ */
+int mnt_clone_write(struct vfsmount *mnt)
+{
+	/* superblock may be r/o */
+	if (__mnt_is_readonly(mnt))
+		return -EROFS;
+	preempt_disable();
+	inc_mnt_writers(mnt);
+	preempt_enable();
+	return 0;
+}
+EXPORT_SYMBOL_GPL(mnt_clone_write);
+
+/**
+ * mnt_want_write_file - get write access to a file's mount
+ * @file: the file who's mount on which to take a write
+ *
+ * This is like mnt_want_write, but it takes a file and can
+ * do some optimisations if the file is open for write already
+ */
+int mnt_want_write_file(struct file *file)
+{
+	if (!(file->f_mode & FMODE_WRITE))
+		return mnt_want_write(file->f_path.mnt);
+	else
+		return mnt_clone_write(file->f_path.mnt);
+}
+EXPORT_SYMBOL_GPL(mnt_want_write_file);
+
 /**
  * mnt_drop_write - give up write access to a mount
  * @mnt: the mount on which to give up write access
diff --git a/fs/open.c b/fs/open.c
index bdfbf03615a..7200e23d925 100644
--- a/fs/open.c
+++ b/fs/open.c
@@ -612,7 +612,7 @@ SYSCALL_DEFINE2(fchmod, unsigned int, fd, mode_t, mode)
 
 	audit_inode(NULL, dentry);
 
-	err = mnt_want_write(file->f_path.mnt);
+	err = mnt_want_write_file(file);
 	if (err)
 		goto out_putf;
 	mutex_lock(&inode->i_mutex);
@@ -761,7 +761,7 @@ SYSCALL_DEFINE3(fchown, unsigned int, fd, uid_t, user, gid_t, group)
 	if (!file)
 		goto out;
 
-	error = mnt_want_write(file->f_path.mnt);
+	error = mnt_want_write_file(file);
 	if (error)
 		goto out_fput;
 	dentry = file->f_path.dentry;
diff --git a/fs/xattr.c b/fs/xattr.c
index d51b8f9db92..1c3d0af59dd 100644
--- a/fs/xattr.c
+++ b/fs/xattr.c
@@ -297,7 +297,7 @@ SYSCALL_DEFINE5(fsetxattr, int, fd, const char __user *, name,
 		return error;
 	dentry = f->f_path.dentry;
 	audit_inode(NULL, dentry);
-	error = mnt_want_write(f->f_path.mnt);
+	error = mnt_want_write_file(f);
 	if (!error) {
 		error = setxattr(dentry, name, value, size, flags);
 		mnt_drop_write(f->f_path.mnt);
@@ -524,7 +524,7 @@ SYSCALL_DEFINE2(fremovexattr, int, fd, const char __user *, name)
 		return error;
 	dentry = f->f_path.dentry;
 	audit_inode(NULL, dentry);
-	error = mnt_want_write(f->f_path.mnt);
+	error = mnt_want_write_file(f);
 	if (!error) {
 		error = removexattr(dentry, name);
 		mnt_drop_write(f->f_path.mnt);
diff --git a/include/linux/mount.h b/include/linux/mount.h
index ac49c1f8e5c..5d527536486 100644
--- a/include/linux/mount.h
+++ b/include/linux/mount.h
@@ -88,7 +88,11 @@ static inline struct vfsmount *mntget(struct vfsmount *mnt)
 	return mnt;
 }
 
+struct file; /* forward dec */
+
 extern int mnt_want_write(struct vfsmount *mnt);
+extern int mnt_want_write_file(struct file *file);
+extern int mnt_clone_write(struct vfsmount *mnt);
 extern void mnt_drop_write(struct vfsmount *mnt);
 extern void mntput_no_expire(struct vfsmount *mnt);
 extern void mnt_pin(struct vfsmount *mnt);
-- 
cgit v1.2.3-70-g09d2


From f3da392e9ff14b9f388e74319e6d195848991c07 Mon Sep 17 00:00:00 2001
From: Alexey Dobriyan <adobriyan@gmail.com>
Date: Mon, 4 May 2009 03:32:03 +0400
Subject: dcache: extrace and use d_unlinked()

d_unlinked() will be used in middle-term to ban checkpointing when opened
but unlinked file is detected, and in long term, to detect such situation
and special case on it.

Signed-off-by: Alexey Dobriyan <adobriyan@gmail.com>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
---
 fs/dcache.c            | 7 +++----
 fs/namespace.c         | 8 ++++----
 include/linux/dcache.h | 5 +++++
 3 files changed, 12 insertions(+), 8 deletions(-)

(limited to 'fs/namespace.c')

diff --git a/fs/dcache.c b/fs/dcache.c
index 75659a6fd1f..9e5cd3c3a6b 100644
--- a/fs/dcache.c
+++ b/fs/dcache.c
@@ -1910,7 +1910,7 @@ char *__d_path(const struct path *path, struct path *root,
 
 	spin_lock(&vfsmount_lock);
 	prepend(&end, &buflen, "\0", 1);
-	if (!IS_ROOT(dentry) && d_unhashed(dentry) &&
+	if (d_unlinked(dentry) &&
 		(prepend(&end, &buflen, " (deleted)", 10) != 0))
 			goto Elong;
 
@@ -2035,7 +2035,7 @@ char *dentry_path(struct dentry *dentry, char *buf, int buflen)
 
 	spin_lock(&dcache_lock);
 	prepend(&end, &buflen, "\0", 1);
-	if (!IS_ROOT(dentry) && d_unhashed(dentry) &&
+	if (d_unlinked(dentry) &&
 		(prepend(&end, &buflen, "//deleted", 9) != 0))
 			goto Elong;
 	if (buflen < 1)
@@ -2097,9 +2097,8 @@ SYSCALL_DEFINE2(getcwd, char __user *, buf, unsigned long, size)
 	read_unlock(&current->fs->lock);
 
 	error = -ENOENT;
-	/* Has the current directory has been unlinked? */
 	spin_lock(&dcache_lock);
-	if (IS_ROOT(pwd.dentry) || !d_unhashed(pwd.dentry)) {
+	if (!d_unlinked(pwd.dentry)) {
 		unsigned long len;
 		struct path tmp = root;
 		char * cwd;
diff --git a/fs/namespace.c b/fs/namespace.c
index 120b8a6b99e..7e537f0393b 100644
--- a/fs/namespace.c
+++ b/fs/namespace.c
@@ -1384,7 +1384,7 @@ static int graft_tree(struct vfsmount *mnt, struct path *path)
 		goto out_unlock;
 
 	err = -ENOENT;
-	if (IS_ROOT(path->dentry) || !d_unhashed(path->dentry))
+	if (!d_unlinked(path->dentry))
 		err = attach_recursive_mnt(mnt, path, NULL);
 out_unlock:
 	mutex_unlock(&path->dentry->d_inode->i_mutex);
@@ -1566,7 +1566,7 @@ static int do_move_mount(struct path *path, char *old_name)
 	if (IS_DEADDIR(path->dentry->d_inode))
 		goto out1;
 
-	if (!IS_ROOT(path->dentry) && d_unhashed(path->dentry))
+	if (d_unlinked(path->dentry))
 		goto out1;
 
 	err = -EINVAL;
@@ -2129,9 +2129,9 @@ SYSCALL_DEFINE2(pivot_root, const char __user *, new_root,
 	error = -ENOENT;
 	if (IS_DEADDIR(new.dentry->d_inode))
 		goto out2;
-	if (d_unhashed(new.dentry) && !IS_ROOT(new.dentry))
+	if (d_unlinked(new.dentry))
 		goto out2;
-	if (d_unhashed(old.dentry) && !IS_ROOT(old.dentry))
+	if (d_unlinked(old.dentry))
 		goto out2;
 	error = -EBUSY;
 	if (new.mnt == root.mnt ||
diff --git a/include/linux/dcache.h b/include/linux/dcache.h
index 72ce2ae8859..30b93b2a01a 100644
--- a/include/linux/dcache.h
+++ b/include/linux/dcache.h
@@ -353,6 +353,11 @@ static inline int d_unhashed(struct dentry *dentry)
 	return (dentry->d_flags & DCACHE_UNHASHED);
 }
 
+static inline int d_unlinked(struct dentry *dentry)
+{
+	return d_unhashed(dentry) && !IS_ROOT(dentry);
+}
+
 static inline struct dentry *dget_parent(struct dentry *dentry)
 {
 	struct dentry *ret;
-- 
cgit v1.2.3-70-g09d2


From 6fac98dd218653c6aff8a0f56305c424930cea2a Mon Sep 17 00:00:00 2001
From: Al Viro <viro@zeniv.linux.org.uk>
Date: Fri, 8 May 2009 13:31:17 -0400
Subject: Push BKL into do_mount()

Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
---
 arch/alpha/kernel/osf_sys.c | 3 ---
 fs/compat.c                 | 2 --
 fs/namespace.c              | 4 ++--
 3 files changed, 2 insertions(+), 7 deletions(-)

(limited to 'fs/namespace.c')

diff --git a/arch/alpha/kernel/osf_sys.c b/arch/alpha/kernel/osf_sys.c
index 42ee05981e7..9a3334ae282 100644
--- a/arch/alpha/kernel/osf_sys.c
+++ b/arch/alpha/kernel/osf_sys.c
@@ -371,8 +371,6 @@ SYSCALL_DEFINE4(osf_mount, unsigned long, typenr, char __user *, path,
 	int retval = -EINVAL;
 	char *name;
 
-	lock_kernel();
-
 	name = getname(path);
 	retval = PTR_ERR(name);
 	if (IS_ERR(name))
@@ -392,7 +390,6 @@ SYSCALL_DEFINE4(osf_mount, unsigned long, typenr, char __user *, path,
 	}
 	putname(name);
  out:
-	unlock_kernel();
 	return retval;
 }
 
diff --git a/fs/compat.c b/fs/compat.c
index bb2a9b2e817..6aefb776dfe 100644
--- a/fs/compat.c
+++ b/fs/compat.c
@@ -812,10 +812,8 @@ asmlinkage long compat_sys_mount(char __user * dev_name, char __user * dir_name,
 		}
 	}
 
-	lock_kernel();
 	retval = do_mount((char*)dev_page, dir_page, (char*)type_page,
 			flags, (void*)data_page);
-	unlock_kernel();
 
  out4:
 	free_page(data_page);
diff --git a/fs/namespace.c b/fs/namespace.c
index 7e537f0393b..4740f7bdb55 100644
--- a/fs/namespace.c
+++ b/fs/namespace.c
@@ -1921,6 +1921,7 @@ long do_mount(char *dev_name, char *dir_name, char *type_page,
 	if (retval)
 		goto dput_out;
 
+	lock_kernel();
 	if (flags & MS_REMOUNT)
 		retval = do_remount(&path, flags & ~MS_REMOUNT, mnt_flags,
 				    data_page);
@@ -1933,6 +1934,7 @@ long do_mount(char *dev_name, char *dir_name, char *type_page,
 	else
 		retval = do_new_mount(&path, type_page, flags, mnt_flags,
 				      dev_name, data_page);
+	unlock_kernel();
 dput_out:
 	path_put(&path);
 	return retval;
@@ -2046,10 +2048,8 @@ SYSCALL_DEFINE5(mount, char __user *, dev_name, char __user *, dir_name,
 	if (retval < 0)
 		goto out3;
 
-	lock_kernel();
 	retval = do_mount((char *)dev_page, dir_page, (char *)type_page,
 			  flags, (void *)data_page);
-	unlock_kernel();
 	free_page(data_page);
 
 out3:
-- 
cgit v1.2.3-70-g09d2


From 7f78d4cd4c5d01864943c22b79df1b6bde923129 Mon Sep 17 00:00:00 2001
From: Al Viro <viro@zeniv.linux.org.uk>
Date: Fri, 8 May 2009 13:34:06 -0400
Subject: Push BKL down beyond VFS-only parts of do_mount()

Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
---
 fs/namespace.c | 9 ++++++---
 1 file changed, 6 insertions(+), 3 deletions(-)

(limited to 'fs/namespace.c')

diff --git a/fs/namespace.c b/fs/namespace.c
index 4740f7bdb55..b94325f00c5 100644
--- a/fs/namespace.c
+++ b/fs/namespace.c
@@ -1515,8 +1515,11 @@ static int do_remount(struct path *path, int flags, int mnt_flags,
 	down_write(&sb->s_umount);
 	if (flags & MS_BIND)
 		err = change_mount_flags(path->mnt, flags);
-	else
+	else {
+		lock_kernel();
 		err = do_remount_sb(sb, flags, data, 0);
+		unlock_kernel();
+	}
 	if (!err)
 		path->mnt->mnt_flags = mnt_flags;
 	up_write(&sb->s_umount);
@@ -1630,7 +1633,9 @@ static int do_new_mount(struct path *path, char *type, int flags,
 	if (!capable(CAP_SYS_ADMIN))
 		return -EPERM;
 
+	lock_kernel();
 	mnt = do_kern_mount(type, flags, name, data);
+	unlock_kernel();
 	if (IS_ERR(mnt))
 		return PTR_ERR(mnt);
 
@@ -1921,7 +1926,6 @@ long do_mount(char *dev_name, char *dir_name, char *type_page,
 	if (retval)
 		goto dput_out;
 
-	lock_kernel();
 	if (flags & MS_REMOUNT)
 		retval = do_remount(&path, flags & ~MS_REMOUNT, mnt_flags,
 				    data_page);
@@ -1934,7 +1938,6 @@ long do_mount(char *dev_name, char *dir_name, char *type_page,
 	else
 		retval = do_new_mount(&path, type_page, flags, mnt_flags,
 				      dev_name, data_page);
-	unlock_kernel();
 dput_out:
 	path_put(&path);
 	return retval;
-- 
cgit v1.2.3-70-g09d2


From 4aa98cf768b6f2ea4b204620d949a665959214f6 Mon Sep 17 00:00:00 2001
From: Al Viro <viro@zeniv.linux.org.uk>
Date: Fri, 8 May 2009 13:36:58 -0400
Subject: Push BKL down into do_remount_sb()

[folded fix from Jiri Slaby]

Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
---
 fs/namespace.c | 10 ++--------
 fs/super.c     | 16 +++++++++++-----
 2 files changed, 13 insertions(+), 13 deletions(-)

(limited to 'fs/namespace.c')

diff --git a/fs/namespace.c b/fs/namespace.c
index b94325f00c5..2dd333b0fe7 100644
--- a/fs/namespace.c
+++ b/fs/namespace.c
@@ -1060,11 +1060,8 @@ static int do_umount(struct vfsmount *mnt, int flags)
 		 * we just try to remount it readonly.
 		 */
 		down_write(&sb->s_umount);
-		if (!(sb->s_flags & MS_RDONLY)) {
-			lock_kernel();
+		if (!(sb->s_flags & MS_RDONLY))
 			retval = do_remount_sb(sb, MS_RDONLY, NULL, 0);
-			unlock_kernel();
-		}
 		up_write(&sb->s_umount);
 		return retval;
 	}
@@ -1515,11 +1512,8 @@ static int do_remount(struct path *path, int flags, int mnt_flags,
 	down_write(&sb->s_umount);
 	if (flags & MS_BIND)
 		err = change_mount_flags(path->mnt, flags);
-	else {
-		lock_kernel();
+	else
 		err = do_remount_sb(sb, flags, data, 0);
-		unlock_kernel();
-	}
 	if (!err)
 		path->mnt->mnt_flags = mnt_flags;
 	up_write(&sb->s_umount);
diff --git a/fs/super.c b/fs/super.c
index 2a49fed7745..a64f3620879 100644
--- a/fs/super.c
+++ b/fs/super.c
@@ -542,25 +542,33 @@ int do_remount_sb(struct super_block *sb, int flags, void *data, int force)
 	shrink_dcache_sb(sb);
 	sync_filesystem(sb);
 
+	lock_kernel();
 	/* If we are remounting RDONLY and current sb is read/write,
 	   make sure there are no rw files opened */
 	if ((flags & MS_RDONLY) && !(sb->s_flags & MS_RDONLY)) {
 		if (force)
 			mark_files_ro(sb);
-		else if (!fs_may_remount_ro(sb))
+		else if (!fs_may_remount_ro(sb)) {
+			unlock_kernel();
 			return -EBUSY;
+		}
 		retval = vfs_dq_off(sb, 1);
-		if (retval < 0 && retval != -ENOSYS)
+		if (retval < 0 && retval != -ENOSYS) {
+			unlock_kernel();
 			return -EBUSY;
+		}
 	}
 	remount_rw = !(flags & MS_RDONLY) && (sb->s_flags & MS_RDONLY);
 
 	if (sb->s_op->remount_fs) {
 		retval = sb->s_op->remount_fs(sb, &flags, data);
-		if (retval)
+		if (retval) {
+			unlock_kernel();
 			return retval;
+		}
 	}
 	sb->s_flags = (sb->s_flags & ~MS_RMT_MASK) | (flags & MS_RMT_MASK);
+	unlock_kernel();
 	if (remount_rw)
 		vfs_dq_quota_on_remount(sb);
 	return 0;
@@ -581,9 +589,7 @@ static void do_emergency_remount(struct work_struct *work)
 			 *
 			 * What lock protects sb->s_flags??
 			 */
-			lock_kernel();
 			do_remount_sb(sb, MS_RDONLY, NULL, 1);
-			unlock_kernel();
 		}
 		up_write(&sb->s_umount);
 		put_super(sb);
-- 
cgit v1.2.3-70-g09d2


From 616511d039af402670de8500d0e24495113a9cab Mon Sep 17 00:00:00 2001
From: Trond Myklebust <Trond.Myklebust@netapp.com>
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 <Trond.Myklebust@netapp.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
---
 fs/namespace.c                | 8 ++++++--
 include/linux/mnt_namespace.h | 9 +--------
 2 files changed, 7 insertions(+), 10 deletions(-)

(limited to 'fs/namespace.c')

diff --git a/fs/namespace.c b/fs/namespace.c
index 2dd333b0fe7..6645846f205 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);
diff --git a/include/linux/mnt_namespace.h b/include/linux/mnt_namespace.h
index 3a059298cc1..299d11af5f7 100644
--- a/include/linux/mnt_namespace.h
+++ b/include/linux/mnt_namespace.h
@@ -26,14 +26,7 @@ struct fs_struct;
 
 extern struct mnt_namespace *copy_mnt_ns(unsigned long, struct mnt_namespace *,
 		struct fs_struct *);
-extern void __put_mnt_ns(struct mnt_namespace *ns);
-
-static inline void put_mnt_ns(struct mnt_namespace *ns)
-{
-	if (atomic_dec_and_lock(&ns->count, &vfsmount_lock))
-		/* releases vfsmount_lock */
-		__put_mnt_ns(ns);
-}
+extern void put_mnt_ns(struct mnt_namespace *ns);
 
 static inline void exit_mnt_ns(struct task_struct *p)
 {
-- 
cgit v1.2.3-70-g09d2


From cf8d2c11cb77f129675478792122f50827e5b0ae Mon Sep 17 00:00:00 2001
From: Trond Myklebust <Trond.Myklebust@netapp.com>
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 <Trond.Myklebust@netapp.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
---
 fs/namespace.c                | 45 +++++++++++++++++++++++++++++++++++--------
 include/linux/mnt_namespace.h |  1 +
 2 files changed, 38 insertions(+), 8 deletions(-)

(limited to 'fs/namespace.c')

diff --git a/fs/namespace.c b/fs/namespace.c
index 6645846f205..a7bea8c8bd4 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);
diff --git a/include/linux/mnt_namespace.h b/include/linux/mnt_namespace.h
index 299d11af5f7..3beb2592b03 100644
--- a/include/linux/mnt_namespace.h
+++ b/include/linux/mnt_namespace.h
@@ -24,6 +24,7 @@ struct proc_mounts {
 
 struct fs_struct;
 
+extern struct mnt_namespace *create_mnt_ns(struct vfsmount *mnt);
 extern struct mnt_namespace *copy_mnt_ns(unsigned long, struct mnt_namespace *,
 		struct fs_struct *);
 extern void put_mnt_ns(struct mnt_namespace *ns);
-- 
cgit v1.2.3-70-g09d2


From 3b22edc5730b87d360ee7dd7143397ba09b73a47 Mon Sep 17 00:00:00 2001
From: Trond Myklebust <Trond.Myklebust@netapp.com>
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 <Trond.Myklebust@netapp.com>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
---
 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 a7bea8c8bd4..4a86b859516 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 v1.2.3-70-g09d2


From f21f62208a6f60e2e05440b2e438d9541822dc4d Mon Sep 17 00:00:00 2001
From: Al Viro <viro@zeniv.linux.org.uk>
Date: Wed, 24 Jun 2009 03:12:00 -0400
Subject: ... and the same for vfsmount id/mount group id

Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
---
 fs/namespace.c | 26 ++++++++++++++++++++++----
 fs/super.c     |  6 ++++--
 2 files changed, 26 insertions(+), 6 deletions(-)

(limited to 'fs/namespace.c')

diff --git a/fs/namespace.c b/fs/namespace.c
index 4a86b859516..3dc283fd471 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;
 }
 
diff --git a/fs/super.c b/fs/super.c
index 808ffd59e01..2761d3e22ed 100644
--- a/fs/super.c
+++ b/fs/super.c
@@ -620,7 +620,8 @@ int set_anon_super(struct super_block *s, void *data)
 		return -ENOMEM;
 	spin_lock(&unnamed_dev_lock);
 	error = ida_get_new_above(&unnamed_dev_ida, unnamed_dev_start, &dev);
-	unnamed_dev_start = dev + 1;
+	if (!error)
+		unnamed_dev_start = dev + 1;
 	spin_unlock(&unnamed_dev_lock);
 	if (error == -EAGAIN)
 		/* We raced and lost with another CPU. */
@@ -631,7 +632,8 @@ int set_anon_super(struct super_block *s, void *data)
 	if ((dev & MAX_ID_MASK) == (1 << MINORBITS)) {
 		spin_lock(&unnamed_dev_lock);
 		ida_remove(&unnamed_dev_ida, dev);
-		unnamed_dev_start = dev;
+		if (unnamed_dev_start > dev)
+			unnamed_dev_start = dev;
 		spin_unlock(&unnamed_dev_lock);
 		return -EMFILE;
 	}
-- 
cgit v1.2.3-70-g09d2