Patchwork patch sysfs-implement-sysfs-tagged-directory-support.patch added to gregkh-2.6 tree

login
register
mail settings
Submitter gregkh@suse.de
Date April 29, 2010, 8:29 p.m.
Message ID <12725729473590@kroah.org>
Download mbox | patch
Permalink /patch/51384/
State Not Applicable
Delegated to: David Miller
Headers show

Comments

gregkh@suse.de - April 29, 2010, 8:29 p.m.
This is a note to let you know that I've just added the patch titled

    Subject: sysfs: Implement sysfs tagged directory support.

to my gregkh-2.6 tree.  Its filename is

    sysfs-implement-sysfs-tagged-directory-support.patch

This tree can be found at 
    http://www.kernel.org/pub/linux/kernel/people/gregkh/gregkh-2.6/patches/


From ebiederm@xmission.com  Thu Apr 29 12:46:06 2010
From: "Eric W. Biederman" <ebiederm@xmission.com>
Date: Tue, 30 Mar 2010 11:31:26 -0700
Subject: sysfs: Implement sysfs tagged directory support.
To: Greg Kroah-Hartman <gregkh@suse.de>
Cc: Kay Sievers <kay.sievers@vrfy.org>, linux-kernel@vger.kernel.org, Tejun Heo <tj@kernel.org>, Cornelia Huck <cornelia.huck@de.ibm.com>, linux-fsdevel@vger.kernel.org, Eric Dumazet <eric.dumazet@gmail.com>, Benjamin LaHaise <bcrl@lhnet.ca>, Serge Hallyn <serue@us.ibm.com>, <netdev@vger.kernel.org>, "Eric W. Biederman" <ebiederm@xmission.com>, Benjamin Thery <benjamin.thery@bull.net>
Message-ID: <1269973889-25260-3-git-send-email-ebiederm@xmission.com>


From: Eric W. Biederman <ebiederm@xmission.com>

The problem.  When implementing a network namespace I need to be able
to have multiple network devices with the same name.  Currently this
is a problem for /sys/class/net/*, /sys/devices/virtual/net/*, and
potentially a few other directories of the form /sys/ ... /net/*.

What this patch does is to add an additional tag field to the
sysfs dirent structure.  For directories that should show different
contents depending on the context such as /sys/class/net/, and
/sys/devices/virtual/net/ this tag field is used to specify the
context in which those directories should be visible.  Effectively
this is the same as creating multiple distinct directories with
the same name but internally to sysfs the result is nicer.

I am calling the concept of a single directory that looks like multiple
directories all at the same path in the filesystem tagged directories.

For the networking namespace the set of directories whose contents I need
to filter with tags can depend on the presence or absence of hotplug
hardware or which modules are currently loaded.  Which means I need
a simple race free way to setup those directories as tagged.

To achieve a reace free design all tagged directories are created
and managed by sysfs itself.

Users of this interface:
- define a type in the sysfs_tag_type enumeration.
- call sysfs_register_ns_types with the type and it's operations
- sysfs_exit_ns when an individual tag is no longer valid

- Implement mount_ns() which returns the ns of the calling process
  so we can attach it to a sysfs superblock.
- Implement ktype.namespace() which returns the ns of a syfs kobject.

Everything else is left up to sysfs and the driver layer.

For the network namespace mount_ns and namespace() are essentially
one line functions, and look to remain that.

Tags are currently represented a const void * pointers as that is
both generic, prevides enough information for equality comparisons,
and is trivial to create for current users, as it is just the
existing namespace pointer.

The work needed in sysfs is more extensive.  At each directory
or symlink creating I need to check if the directory it is being
created in is a tagged directory and if so generate the appropriate
tag to place on the sysfs_dirent.  Likewise at each symlink or
directory removal I need to check if the sysfs directory it is
being removed from is a tagged directory and if so figure out
which tag goes along with the name I am deleting.

Currently only directories which hold kobjects, and
symlinks are supported.  There is not enough information
in the current file attribute interfaces to give us anything
to discriminate on which makes it useless, and there are
no potential users which makes it an uninteresting problem
to solve.

Signed-off-by: Eric W. Biederman <ebiederm@xmission.com>
Signed-off-by: Benjamin Thery <benjamin.thery@bull.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

---
 drivers/gpio/gpiolib.c |    2 
 drivers/md/bitmap.c    |    4 -
 drivers/md/md.c        |    6 +-
 fs/sysfs/bin.c         |    2 
 fs/sysfs/dir.c         |  112 ++++++++++++++++++++++++++++++++++++++-----------
 fs/sysfs/file.c        |   17 ++++---
 fs/sysfs/group.c       |    6 +-
 fs/sysfs/inode.c       |    4 -
 fs/sysfs/mount.c       |   33 ++++++++++++++
 fs/sysfs/symlink.c     |   15 +++++-
 fs/sysfs/sysfs.h       |   20 +++++++-
 include/linux/sysfs.h  |   10 ++++
 lib/kobject.c          |    1 
 13 files changed, 181 insertions(+), 51 deletions(-)


--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Tejun Heo - April 30, 2010, 4:18 a.m.
On 04/29/2010 10:29 PM, gregkh@suse.de wrote:
> 
> This is a note to let you know that I've just added the patch titled
> 
>     Subject: sysfs: Implement sysfs tagged directory support.
> 
> to my gregkh-2.6 tree.  Its filename is
> 
>     sysfs-implement-sysfs-tagged-directory-support.patch
> 
> This tree can be found at 
>     http://www.kernel.org/pub/linux/kernel/people/gregkh/gregkh-2.6/patches/

I wish at least more comments are added before it goes mainline.  I
don't really understand the current form.

Thanks.
Tejun Heo - April 30, 2010, 5:37 a.m.
Hello,

On 04/30/2010 07:24 AM, Eric W. Biederman wrote:
>>> I wish at least more comments are added before it goes mainline.  I
>>> don't really understand the current form.
>>
>> Ok, that's fine with me, I'll pull it back out.
> 
> ?????
> 
> Tejun you have offered nothing constructive to the review, except looking
> and saying you don't understand what is going on.

Eric, no need to get too touchy and you're right in part in saying all
I'm saying is basically "I don't understand it" which is the same
reason why I'm not nacking it and explicitly stated that I would be
okay with the series going in if Greg/Kay would be okay with it.
Again, about the same thing with the above comment, I was *wishing*
for more comments *before it goes mainline*.

> Tejun I think for the code to make any sense to you I would need to rip
> out out and/or rewrite the kobject layer, and possible the device
> model code as well.

And yes, in the long run, please do that.

> Tejun I'm sorry you can't understand the code, and I'm sorry the code
> may be over-general.  In part that is because making the code
> over-general is what you asked for when reviewing it the first time.

Please give me some credit.  I mean that the code is difficult to
follow and justify when I say I don't understand it.  Yeah, I tried to
understand it and I think I understand how it *works* in its current
form but I just don't think the design is justified or logical.  You
say it's infeasible to do it in straightforward manner in reasonable
amount of time and that's why I neither acked or nacked the series and
deferred the decision to the subsystem maintainer.

But, at the very least, please add some comments.  Try to explain what
each callbacks are supposed to do and why they're there.  Not everyone
lives in your head.

Thanks.
gregkh@suse.de - April 30, 2010, 3:58 p.m.
On Fri, Apr 30, 2010 at 10:43:21AM -0500, Serge E. Hallyn wrote:
> Quoting Tejun Heo (tj@kernel.org):
> > Hello,
> > 
> > On 04/30/2010 04:29 PM, Serge E. Hallyn wrote:
> > > Hmm, but looking back over the previous thread (Mar 31) I guess you
> > > mean more in-line comments around the callbacks, presumably things
> > > like class_dir_child_ns_type() and struct kobj_ns_type_operations
> > > members?
> > 
> > In-line.  What they're, how they're supposed to be used, which calling
> > context is expected, what can be returned and so on.
> > 
> > > It sounds like what you'd really like is to have any explicit
> > > mention to namespaces pulled out of drivers/base (layering as you
> > > keep saying)?  But will there be a use for this outside of
> > > namespaces?  Does trying to anticipate that fall into the category
> > > of over-abstraction?
> > 
> > I wouldn't mind limited amount of layering exceptions as long as
> > they're clearly documented.  What I'm primarily worried about is not
> > the possibility of other users but more the obfuscation of the whole
> > sysfs-kobject-driver model thing which is already overly abstracted
> > and obfuscated (at least it seems to me that way).
> > 
> > NS needs tagged support in the driver model which in itself is fine
> > and I also understand that from someone who's primarily working on NS,
> > adding a bit on top of the whole thing wouldn't seem like much of a
> > problem.  To me it seems like worsening a problem which is already
> > pretty bad.  I hope you could understand my POV too.
> 
> I do.  I can take a stab monday at pushing a cloned version of Eric's
> tree with comments added, if Eric doesn't have time.  (Or a patch on
> top of Greg's tree)

On top of Greg's tree please :)

thanks,

greg k-h
--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Patch

--- a/drivers/gpio/gpiolib.c
+++ b/drivers/gpio/gpiolib.c
@@ -399,7 +399,7 @@  static int gpio_setup_irq(struct gpio_de
 			goto free_id;
 		}
 
-		pdesc->value_sd = sysfs_get_dirent(dev->kobj.sd, "value");
+		pdesc->value_sd = sysfs_get_dirent(dev->kobj.sd, NULL, "value");
 		if (!pdesc->value_sd) {
 			ret = -ENODEV;
 			goto free_id;
--- a/drivers/md/bitmap.c
+++ b/drivers/md/bitmap.c
@@ -1678,9 +1678,9 @@  int bitmap_create(mddev_t *mddev)
 
 	bitmap->mddev = mddev;
 
-	bm = sysfs_get_dirent(mddev->kobj.sd, "bitmap");
+	bm = sysfs_get_dirent(mddev->kobj.sd, NULL, "bitmap");
 	if (bm) {
-		bitmap->sysfs_can_clear = sysfs_get_dirent(bm, "can_clear");
+		bitmap->sysfs_can_clear = sysfs_get_dirent(bm, NULL, "can_clear");
 		sysfs_put(bm);
 	} else
 		bitmap->sysfs_can_clear = NULL;
--- a/drivers/md/md.c
+++ b/drivers/md/md.c
@@ -1766,7 +1766,7 @@  static int bind_rdev_to_array(mdk_rdev_t
 		kobject_del(&rdev->kobj);
 		goto fail;
 	}
-	rdev->sysfs_state = sysfs_get_dirent(rdev->kobj.sd, "state");
+	rdev->sysfs_state = sysfs_get_dirent(rdev->kobj.sd, NULL, "state");
 
 	list_add_rcu(&rdev->same_set, &mddev->disks);
 	bd_claim_by_disk(rdev->bdev, rdev->bdev->bd_holder, mddev->gendisk);
@@ -4183,7 +4183,7 @@  static int md_alloc(dev_t dev, char *nam
 	mutex_unlock(&disks_mutex);
 	if (!error) {
 		kobject_uevent(&mddev->kobj, KOBJ_ADD);
-		mddev->sysfs_state = sysfs_get_dirent(mddev->kobj.sd, "array_state");
+		mddev->sysfs_state = sysfs_get_dirent(mddev->kobj.sd, NULL, "array_state");
 	}
 	mddev_put(mddev);
 	return error;
@@ -4392,7 +4392,7 @@  static int do_md_run(mddev_t * mddev)
 			printk(KERN_WARNING
 			       "md: cannot register extra attributes for %s\n",
 			       mdname(mddev));
-		mddev->sysfs_action = sysfs_get_dirent(mddev->kobj.sd, "sync_action");
+		mddev->sysfs_action = sysfs_get_dirent(mddev->kobj.sd, NULL, "sync_action");
 	} else if (mddev->ro == 2) /* auto-readonly not meaningful */
 		mddev->ro = 0;
 
--- a/fs/sysfs/bin.c
+++ b/fs/sysfs/bin.c
@@ -501,7 +501,7 @@  int sysfs_create_bin_file(struct kobject
 void sysfs_remove_bin_file(struct kobject *kobj,
 			   const struct bin_attribute *attr)
 {
-	sysfs_hash_and_remove(kobj->sd, attr->attr.name);
+	sysfs_hash_and_remove(kobj->sd, NULL, attr->attr.name);
 }
 
 EXPORT_SYMBOL_GPL(sysfs_create_bin_file);
--- a/fs/sysfs/dir.c
+++ b/fs/sysfs/dir.c
@@ -380,9 +380,15 @@  int __sysfs_add_one(struct sysfs_addrm_c
 {
 	struct sysfs_inode_attrs *ps_iattr;
 
-	if (sysfs_find_dirent(acxt->parent_sd, sd->s_name))
+	if (sysfs_find_dirent(acxt->parent_sd, sd->s_ns, sd->s_name))
 		return -EEXIST;
 
+	if (sysfs_ns_type(acxt->parent_sd) && !sd->s_ns) {
+		WARN(1, KERN_WARNING "sysfs: ns required in '%s' for '%s'\n",
+			acxt->parent_sd->s_name, sd->s_name);
+		return -EINVAL;
+	}
+
 	sd->s_parent = sysfs_get(acxt->parent_sd);
 
 	sysfs_link_sibling(sd);
@@ -533,13 +539,17 @@  void sysfs_addrm_finish(struct sysfs_add
  *	Pointer to sysfs_dirent if found, NULL if not.
  */
 struct sysfs_dirent *sysfs_find_dirent(struct sysfs_dirent *parent_sd,
+				       const void *ns,
 				       const unsigned char *name)
 {
 	struct sysfs_dirent *sd;
 
-	for (sd = parent_sd->s_dir.children; sd; sd = sd->s_sibling)
+	for (sd = parent_sd->s_dir.children; sd; sd = sd->s_sibling) {
+		if (sd->s_ns != ns)
+			continue;
 		if (!strcmp(sd->s_name, name))
 			return sd;
+	}
 	return NULL;
 }
 
@@ -558,12 +568,13 @@  struct sysfs_dirent *sysfs_find_dirent(s
  *	Pointer to sysfs_dirent if found, NULL if not.
  */
 struct sysfs_dirent *sysfs_get_dirent(struct sysfs_dirent *parent_sd,
+				      const void *ns,
 				      const unsigned char *name)
 {
 	struct sysfs_dirent *sd;
 
 	mutex_lock(&sysfs_mutex);
-	sd = sysfs_find_dirent(parent_sd, name);
+	sd = sysfs_find_dirent(parent_sd, ns, name);
 	sysfs_get(sd);
 	mutex_unlock(&sysfs_mutex);
 
@@ -572,7 +583,8 @@  struct sysfs_dirent *sysfs_get_dirent(st
 EXPORT_SYMBOL_GPL(sysfs_get_dirent);
 
 static int create_dir(struct kobject *kobj, struct sysfs_dirent *parent_sd,
-		      const char *name, struct sysfs_dirent **p_sd)
+	enum kobj_ns_type type, const void *ns, const char *name,
+	struct sysfs_dirent **p_sd)
 {
 	umode_t mode = S_IFDIR| S_IRWXU | S_IRUGO | S_IXUGO;
 	struct sysfs_addrm_cxt acxt;
@@ -583,6 +595,9 @@  static int create_dir(struct kobject *ko
 	sd = sysfs_new_dirent(name, mode, SYSFS_DIR);
 	if (!sd)
 		return -ENOMEM;
+
+	sd->s_flags |= (type << SYSFS_NS_TYPE_SHIFT);
+	sd->s_ns = ns;
 	sd->s_dir.kobj = kobj;
 
 	/* link in */
@@ -601,7 +616,25 @@  static int create_dir(struct kobject *ko
 int sysfs_create_subdir(struct kobject *kobj, const char *name,
 			struct sysfs_dirent **p_sd)
 {
-	return create_dir(kobj, kobj->sd, name, p_sd);
+	return create_dir(kobj, kobj->sd,
+			  KOBJ_NS_TYPE_NONE, NULL, name, p_sd);
+}
+
+static enum kobj_ns_type sysfs_read_ns_type(struct kobject *kobj)
+{
+	const struct kobj_ns_type_operations *ops;
+	enum kobj_ns_type type;
+
+	ops = kobj_child_ns_ops(kobj);
+	if (!ops)
+		return KOBJ_NS_TYPE_NONE;
+
+	type = ops->type;
+	BUG_ON(type <= KOBJ_NS_TYPE_NONE);
+	BUG_ON(type >= KOBJ_NS_TYPES);
+	BUG_ON(!kobj_ns_type_registered(type));
+
+	return type;
 }
 
 /**
@@ -610,7 +643,9 @@  int sysfs_create_subdir(struct kobject *
  */
 int sysfs_create_dir(struct kobject * kobj)
 {
+	enum kobj_ns_type type;
 	struct sysfs_dirent *parent_sd, *sd;
+	const void *ns = NULL;
 	int error = 0;
 
 	BUG_ON(!kobj);
@@ -620,7 +655,11 @@  int sysfs_create_dir(struct kobject * ko
 	else
 		parent_sd = &sysfs_root;
 
-	error = create_dir(kobj, parent_sd, kobject_name(kobj), &sd);
+	if (sysfs_ns_type(parent_sd))
+		ns = kobj->ktype->namespace(kobj);
+	type = sysfs_read_ns_type(kobj);
+
+	error = create_dir(kobj, parent_sd, type, ns, kobject_name(kobj), &sd);
 	if (!error)
 		kobj->sd = sd;
 	return error;
@@ -630,13 +669,19 @@  static struct dentry * sysfs_lookup(stru
 				struct nameidata *nd)
 {
 	struct dentry *ret = NULL;
-	struct sysfs_dirent *parent_sd = dentry->d_parent->d_fsdata;
+	struct dentry *parent = dentry->d_parent;
+	struct sysfs_dirent *parent_sd = parent->d_fsdata;
 	struct sysfs_dirent *sd;
 	struct inode *inode;
+	enum kobj_ns_type type;
+	const void *ns;
 
 	mutex_lock(&sysfs_mutex);
 
-	sd = sysfs_find_dirent(parent_sd, dentry->d_name.name);
+	type = sysfs_ns_type(parent_sd);
+	ns = sysfs_info(dir->i_sb)->ns[type];
+
+	sd = sysfs_find_dirent(parent_sd, ns, dentry->d_name.name);
 
 	/* no such entry */
 	if (!sd) {
@@ -735,7 +780,8 @@  void sysfs_remove_dir(struct kobject * k
 }
 
 int sysfs_rename(struct sysfs_dirent *sd,
-	struct sysfs_dirent *new_parent_sd, const char *new_name)
+	struct sysfs_dirent *new_parent_sd, const void *new_ns,
+	const char *new_name)
 {
 	const char *dup_name = NULL;
 	int error;
@@ -743,12 +789,12 @@  int sysfs_rename(struct sysfs_dirent *sd
 	mutex_lock(&sysfs_mutex);
 
 	error = 0;
-	if ((sd->s_parent == new_parent_sd) &&
+	if ((sd->s_parent == new_parent_sd) && (sd->s_ns == new_ns) &&
 	    (strcmp(sd->s_name, new_name) == 0))
 		goto out;	/* nothing to rename */
 
 	error = -EEXIST;
-	if (sysfs_find_dirent(new_parent_sd, new_name))
+	if (sysfs_find_dirent(new_parent_sd, new_ns, new_name))
 		goto out;
 
 	/* rename sysfs_dirent */
@@ -770,6 +816,7 @@  int sysfs_rename(struct sysfs_dirent *sd
 		sd->s_parent = new_parent_sd;
 		sysfs_link_sibling(sd);
 	}
+	sd->s_ns = new_ns;
 
 	error = 0;
  out:
@@ -780,19 +827,28 @@  int sysfs_rename(struct sysfs_dirent *sd
 
 int sysfs_rename_dir(struct kobject *kobj, const char *new_name)
 {
-	return sysfs_rename(kobj->sd, kobj->sd->s_parent, new_name);
+	struct sysfs_dirent *parent_sd = kobj->sd->s_parent;
+	const void *new_ns = NULL;
+
+	if (sysfs_ns_type(parent_sd))
+		new_ns = kobj->ktype->namespace(kobj);
+
+	return sysfs_rename(kobj->sd, parent_sd, new_ns, new_name);
 }
 
 int sysfs_move_dir(struct kobject *kobj, struct kobject *new_parent_kobj)
 {
 	struct sysfs_dirent *sd = kobj->sd;
 	struct sysfs_dirent *new_parent_sd;
+	const void *new_ns = NULL;
 
 	BUG_ON(!sd->s_parent);
+	if (sysfs_ns_type(sd->s_parent))
+		new_ns = kobj->ktype->namespace(kobj);
 	new_parent_sd = new_parent_kobj && new_parent_kobj->sd ?
 		new_parent_kobj->sd : &sysfs_root;
 
-	return sysfs_rename(sd, new_parent_sd, sd->s_name);
+	return sysfs_rename(sd, new_parent_sd, new_ns, sd->s_name);
 }
 
 /* Relationship between s_mode and the DT_xxx types */
@@ -807,32 +863,35 @@  static int sysfs_dir_release(struct inod
 	return 0;
 }
 
-static struct sysfs_dirent *sysfs_dir_pos(struct sysfs_dirent *parent_sd,
-	ino_t ino, struct sysfs_dirent *pos)
+static struct sysfs_dirent *sysfs_dir_pos(const void *ns,
+	struct sysfs_dirent *parent_sd,	ino_t ino, struct sysfs_dirent *pos)
 {
 	if (pos) {
 		int valid = !(pos->s_flags & SYSFS_FLAG_REMOVED) &&
 			pos->s_parent == parent_sd &&
 			ino == pos->s_ino;
 		sysfs_put(pos);
-		if (valid)
-			return pos;
+		if (!valid)
+			pos = NULL;
 	}
-	pos = NULL;
-	if ((ino > 1) && (ino < INT_MAX)) {
+	if (!pos && (ino > 1) && (ino < INT_MAX)) {
 		pos = parent_sd->s_dir.children;
 		while (pos && (ino > pos->s_ino))
 			pos = pos->s_sibling;
 	}
+	while (pos && pos->s_ns != ns)
+		pos = pos->s_sibling;
 	return pos;
 }
 
-static struct sysfs_dirent *sysfs_dir_next_pos(struct sysfs_dirent *parent_sd,
-	ino_t ino, struct sysfs_dirent *pos)
+static struct sysfs_dirent *sysfs_dir_next_pos(const void *ns,
+	struct sysfs_dirent *parent_sd,	ino_t ino, struct sysfs_dirent *pos)
 {
-	pos = sysfs_dir_pos(parent_sd, ino, pos);
+	pos = sysfs_dir_pos(ns, parent_sd, ino, pos);
 	if (pos)
 		pos = pos->s_sibling;
+	while (pos && pos->s_ns != ns)
+		pos = pos->s_sibling;
 	return pos;
 }
 
@@ -841,8 +900,13 @@  static int sysfs_readdir(struct file * f
 	struct dentry *dentry = filp->f_path.dentry;
 	struct sysfs_dirent * parent_sd = dentry->d_fsdata;
 	struct sysfs_dirent *pos = filp->private_data;
+	enum kobj_ns_type type;
+	const void *ns;
 	ino_t ino;
 
+	type = sysfs_ns_type(parent_sd);
+	ns = sysfs_info(dentry->d_sb)->ns[type];
+
 	if (filp->f_pos == 0) {
 		ino = parent_sd->s_ino;
 		if (filldir(dirent, ".", 1, filp->f_pos, ino, DT_DIR) == 0)
@@ -857,9 +921,9 @@  static int sysfs_readdir(struct file * f
 			filp->f_pos++;
 	}
 	mutex_lock(&sysfs_mutex);
-	for (pos = sysfs_dir_pos(parent_sd, filp->f_pos, pos);
+	for (pos = sysfs_dir_pos(ns, parent_sd, filp->f_pos, pos);
 	     pos;
-	     pos = sysfs_dir_next_pos(parent_sd, filp->f_pos, pos)) {
+	     pos = sysfs_dir_next_pos(ns, parent_sd, filp->f_pos, pos)) {
 		const char * name;
 		unsigned int type;
 		int len, ret;
--- a/fs/sysfs/file.c
+++ b/fs/sysfs/file.c
@@ -478,9 +478,12 @@  void sysfs_notify(struct kobject *k, con
 	mutex_lock(&sysfs_mutex);
 
 	if (sd && dir)
-		sd = sysfs_find_dirent(sd, dir);
+		/* Only directories are tagged, so no need to pass
+		 * a tag explicitly.
+		 */
+		sd = sysfs_find_dirent(sd, NULL, dir);
 	if (sd && attr)
-		sd = sysfs_find_dirent(sd, attr);
+		sd = sysfs_find_dirent(sd, NULL, attr);
 	if (sd)
 		sysfs_notify_dirent(sd);
 
@@ -569,7 +572,7 @@  int sysfs_add_file_to_group(struct kobje
 	int error;
 
 	if (group)
-		dir_sd = sysfs_get_dirent(kobj->sd, group);
+		dir_sd = sysfs_get_dirent(kobj->sd, NULL, group);
 	else
 		dir_sd = sysfs_get(kobj->sd);
 
@@ -599,7 +602,7 @@  int sysfs_chmod_file(struct kobject *kob
 	mutex_lock(&sysfs_mutex);
 
 	rc = -ENOENT;
-	sd = sysfs_find_dirent(kobj->sd, attr->name);
+	sd = sysfs_find_dirent(kobj->sd, NULL, attr->name);
 	if (!sd)
 		goto out;
 
@@ -624,7 +627,7 @@  EXPORT_SYMBOL_GPL(sysfs_chmod_file);
 
 void sysfs_remove_file(struct kobject * kobj, const struct attribute * attr)
 {
-	sysfs_hash_and_remove(kobj->sd, attr->name);
+	sysfs_hash_and_remove(kobj->sd, NULL, attr->name);
 }
 
 void sysfs_remove_files(struct kobject * kobj, const struct attribute **ptr)
@@ -646,11 +649,11 @@  void sysfs_remove_file_from_group(struct
 	struct sysfs_dirent *dir_sd;
 
 	if (group)
-		dir_sd = sysfs_get_dirent(kobj->sd, group);
+		dir_sd = sysfs_get_dirent(kobj->sd, NULL, group);
 	else
 		dir_sd = sysfs_get(kobj->sd);
 	if (dir_sd) {
-		sysfs_hash_and_remove(dir_sd, attr->name);
+		sysfs_hash_and_remove(dir_sd, NULL, attr->name);
 		sysfs_put(dir_sd);
 	}
 }
--- a/fs/sysfs/group.c
+++ b/fs/sysfs/group.c
@@ -23,7 +23,7 @@  static void remove_files(struct sysfs_di
 	int i;
 
 	for (i = 0, attr = grp->attrs; *attr; i++, attr++)
-		sysfs_hash_and_remove(dir_sd, (*attr)->name);
+		sysfs_hash_and_remove(dir_sd, NULL, (*attr)->name);
 }
 
 static int create_files(struct sysfs_dirent *dir_sd, struct kobject *kobj,
@@ -39,7 +39,7 @@  static int create_files(struct sysfs_dir
 		 * visibility.  Do this by first removing then
 		 * re-adding (if required) the file */
 		if (update)
-			sysfs_hash_and_remove(dir_sd, (*attr)->name);
+			sysfs_hash_and_remove(dir_sd, NULL, (*attr)->name);
 		if (grp->is_visible) {
 			mode = grp->is_visible(kobj, *attr, i);
 			if (!mode)
@@ -132,7 +132,7 @@  void sysfs_remove_group(struct kobject *
 	struct sysfs_dirent *sd;
 
 	if (grp->name) {
-		sd = sysfs_get_dirent(dir_sd, grp->name);
+		sd = sysfs_get_dirent(dir_sd, NULL, grp->name);
 		if (!sd) {
 			WARN(!sd, KERN_WARNING "sysfs group %p not found for "
 				"kobject '%s'\n", grp, kobject_name(kobj));
--- a/fs/sysfs/inode.c
+++ b/fs/sysfs/inode.c
@@ -324,7 +324,7 @@  void sysfs_delete_inode(struct inode *in
 	sysfs_put(sd);
 }
 
-int sysfs_hash_and_remove(struct sysfs_dirent *dir_sd, const char *name)
+int sysfs_hash_and_remove(struct sysfs_dirent *dir_sd, const void *ns, const char *name)
 {
 	struct sysfs_addrm_cxt acxt;
 	struct sysfs_dirent *sd;
@@ -334,7 +334,7 @@  int sysfs_hash_and_remove(struct sysfs_d
 
 	sysfs_addrm_start(&acxt, dir_sd);
 
-	sd = sysfs_find_dirent(dir_sd, name);
+	sd = sysfs_find_dirent(dir_sd, ns, name);
 	if (sd)
 		sysfs_remove_one(&acxt, sd);
 
--- a/fs/sysfs/mount.c
+++ b/fs/sysfs/mount.c
@@ -35,7 +35,7 @@  static const struct super_operations sys
 struct sysfs_dirent sysfs_root = {
 	.s_name		= "",
 	.s_count	= ATOMIC_INIT(1),
-	.s_flags	= SYSFS_DIR,
+	.s_flags	= SYSFS_DIR | (KOBJ_NS_TYPE_NONE << SYSFS_NS_TYPE_SHIFT),
 	.s_mode		= S_IFDIR | S_IRWXU | S_IRUGO | S_IXUGO,
 	.s_ino		= 1,
 };
@@ -76,7 +76,13 @@  static int sysfs_test_super(struct super
 {
 	struct sysfs_super_info *sb_info = sysfs_info(sb);
 	struct sysfs_super_info *info = data;
+	enum kobj_ns_type type;
 	int found = 1;
+
+	for (type = KOBJ_NS_TYPE_NONE; type < KOBJ_NS_TYPES; type++) {
+		if (sb_info->ns[type] != info->ns[type])
+			found = 0;
+	}
 	return found;
 }
 
@@ -93,6 +99,7 @@  static int sysfs_get_sb(struct file_syst
 	int flags, const char *dev_name, void *data, struct vfsmount *mnt)
 {
 	struct sysfs_super_info *info;
+	enum kobj_ns_type type;
 	struct super_block *sb;
 	int error;
 
@@ -100,6 +107,10 @@  static int sysfs_get_sb(struct file_syst
 	info = kzalloc(sizeof(*info), GFP_KERNEL);
 	if (!info)
 		goto out;
+
+	for (type = KOBJ_NS_TYPE_NONE; type < KOBJ_NS_TYPES; type++)
+		info->ns[type] = kobj_ns_current(type);
+
 	sb = sget(fs_type, sysfs_test_super, sysfs_set_super, info);
 	if (IS_ERR(sb) || sb->s_fs_info != info)
 		kfree(info);
@@ -137,6 +148,26 @@  static struct file_system_type sysfs_fs_
 	.kill_sb	= sysfs_kill_sb,
 };
 
+void sysfs_exit_ns(enum kobj_ns_type type, const void *ns)
+{
+	struct super_block *sb;
+
+	mutex_lock(&sysfs_mutex);
+	spin_lock(&sb_lock);
+	list_for_each_entry(sb, &sysfs_fs_type.fs_supers, s_instances) {
+		struct sysfs_super_info *info = sysfs_info(sb);
+		/* Ignore superblocks that are in the process of unmounting */
+		if (sb->s_count <= S_BIAS)
+			continue;
+		/* Ignore superblocks with the wrong ns */
+		if (info->ns[type] != ns)
+			continue;
+		info->ns[type] = NULL;
+	}
+	spin_unlock(&sb_lock);
+	mutex_unlock(&sysfs_mutex);
+}
+
 int __init sysfs_init(void)
 {
 	int err = -ENOMEM;
--- a/fs/sysfs/symlink.c
+++ b/fs/sysfs/symlink.c
@@ -58,6 +58,8 @@  static int sysfs_do_create_link(struct k
 	if (!sd)
 		goto out_put;
 
+	if (sysfs_ns_type(parent_sd))
+		sd->s_ns = target->ktype->namespace(target);
 	sd->s_symlink.target_sd = target_sd;
 	target_sd = NULL;	/* reference is now owned by the symlink */
 
@@ -121,7 +123,7 @@  void sysfs_remove_link(struct kobject *
 	else
 		parent_sd = kobj->sd;
 
-	sysfs_hash_and_remove(parent_sd, name);
+	sysfs_hash_and_remove(parent_sd, NULL, name);
 }
 
 /**
@@ -137,6 +139,7 @@  int sysfs_rename_link(struct kobject *ko
 			const char *old, const char *new)
 {
 	struct sysfs_dirent *parent_sd, *sd = NULL;
+	const void *old_ns = NULL, *new_ns = NULL;
 	int result;
 
 	if (!kobj)
@@ -144,8 +147,11 @@  int sysfs_rename_link(struct kobject *ko
 	else
 		parent_sd = kobj->sd;
 
+	if (targ->sd)
+		old_ns = targ->sd->s_ns;
+
 	result = -ENOENT;
-	sd = sysfs_get_dirent(parent_sd, old);
+	sd = sysfs_get_dirent(parent_sd, old_ns, old);
 	if (!sd)
 		goto out;
 
@@ -155,7 +161,10 @@  int sysfs_rename_link(struct kobject *ko
 	if (sd->s_symlink.target_sd->s_dir.kobj != targ)
 		goto out;
 
-	result = sysfs_rename(sd, parent_sd, new);
+	if (sysfs_ns_type(parent_sd))
+		new_ns = targ->ktype->namespace(targ);
+
+	result = sysfs_rename(sd, parent_sd, new_ns, new);
 
 out:
 	sysfs_put(sd);
--- a/fs/sysfs/sysfs.h
+++ b/fs/sysfs/sysfs.h
@@ -58,6 +58,7 @@  struct sysfs_dirent {
 	struct sysfs_dirent	*s_sibling;
 	const char		*s_name;
 
+	const void		*s_ns;
 	union {
 		struct sysfs_elem_dir		s_dir;
 		struct sysfs_elem_symlink	s_symlink;
@@ -81,14 +82,22 @@  struct sysfs_dirent {
 #define SYSFS_COPY_NAME			(SYSFS_DIR | SYSFS_KOBJ_LINK)
 #define SYSFS_ACTIVE_REF		(SYSFS_KOBJ_ATTR | SYSFS_KOBJ_BIN_ATTR)
 
-#define SYSFS_FLAG_MASK			~SYSFS_TYPE_MASK
-#define SYSFS_FLAG_REMOVED		0x0200
+#define SYSFS_NS_TYPE_MASK		0xff00
+#define SYSFS_NS_TYPE_SHIFT		8
+
+#define SYSFS_FLAG_MASK			~(SYSFS_NS_TYPE_MASK|SYSFS_TYPE_MASK)
+#define SYSFS_FLAG_REMOVED		0x020000
 
 static inline unsigned int sysfs_type(struct sysfs_dirent *sd)
 {
 	return sd->s_flags & SYSFS_TYPE_MASK;
 }
 
+static inline enum kobj_ns_type sysfs_ns_type(struct sysfs_dirent *sd)
+{
+	return (sd->s_flags & SYSFS_NS_TYPE_MASK) >> SYSFS_NS_TYPE_SHIFT;
+}
+
 #ifdef CONFIG_DEBUG_LOCK_ALLOC
 #define sysfs_dirent_init_lockdep(sd)				\
 do {								\
@@ -115,6 +124,7 @@  struct sysfs_addrm_cxt {
  * mount.c
  */
 struct sysfs_super_info {
+	const void *ns[KOBJ_NS_TYPES];
 };
 #define sysfs_info(SB) ((struct sysfs_super_info *)(SB->s_fs_info))
 extern struct sysfs_dirent sysfs_root;
@@ -140,8 +150,10 @@  void sysfs_remove_one(struct sysfs_addrm
 void sysfs_addrm_finish(struct sysfs_addrm_cxt *acxt);
 
 struct sysfs_dirent *sysfs_find_dirent(struct sysfs_dirent *parent_sd,
+				       const void *ns,
 				       const unsigned char *name);
 struct sysfs_dirent *sysfs_get_dirent(struct sysfs_dirent *parent_sd,
+				      const void *ns,
 				      const unsigned char *name);
 struct sysfs_dirent *sysfs_new_dirent(const char *name, umode_t mode, int type);
 
@@ -152,7 +164,7 @@  int sysfs_create_subdir(struct kobject *
 void sysfs_remove_subdir(struct sysfs_dirent *sd);
 
 int sysfs_rename(struct sysfs_dirent *sd,
-	struct sysfs_dirent *new_parent_sd, const char *new_name);
+	struct sysfs_dirent *new_parent_sd, const void *ns, const char *new_name);
 
 static inline struct sysfs_dirent *__sysfs_get(struct sysfs_dirent *sd)
 {
@@ -182,7 +194,7 @@  int sysfs_setattr(struct dentry *dentry,
 int sysfs_getattr(struct vfsmount *mnt, struct dentry *dentry, struct kstat *stat);
 int sysfs_setxattr(struct dentry *dentry, const char *name, const void *value,
 		size_t size, int flags);
-int sysfs_hash_and_remove(struct sysfs_dirent *dir_sd, const char *name);
+int sysfs_hash_and_remove(struct sysfs_dirent *dir_sd, const void *ns, const char *name);
 int sysfs_inode_init(void);
 
 /*
--- a/include/linux/sysfs.h
+++ b/include/linux/sysfs.h
@@ -20,6 +20,7 @@ 
 
 struct kobject;
 struct module;
+enum kobj_ns_type;
 
 /* FIXME
  * The *owner field is no longer used.
@@ -168,10 +169,14 @@  void sysfs_remove_file_from_group(struct
 void sysfs_notify(struct kobject *kobj, const char *dir, const char *attr);
 void sysfs_notify_dirent(struct sysfs_dirent *sd);
 struct sysfs_dirent *sysfs_get_dirent(struct sysfs_dirent *parent_sd,
+				      const void *ns,
 				      const unsigned char *name);
 struct sysfs_dirent *sysfs_get(struct sysfs_dirent *sd);
 void sysfs_put(struct sysfs_dirent *sd);
 void sysfs_printk_last_file(void);
+
+void sysfs_exit_ns(enum kobj_ns_type type, const void *tag);
+
 int __must_check sysfs_init(void);
 
 #else /* CONFIG_SYSFS */
@@ -301,6 +306,7 @@  static inline void sysfs_notify_dirent(s
 }
 static inline
 struct sysfs_dirent *sysfs_get_dirent(struct sysfs_dirent *parent_sd,
+				      const void *ns,
 				      const unsigned char *name)
 {
 	return NULL;
@@ -313,6 +319,10 @@  static inline void sysfs_put(struct sysf
 {
 }
 
+static inline void sysfs_exit_ns(enum kobj_ns_type type, const void *tag)
+{
+}
+
 static inline int __must_check sysfs_init(void)
 {
 	return 0;
--- a/lib/kobject.c
+++ b/lib/kobject.c
@@ -950,6 +950,7 @@  const void *kobj_ns_initial(enum kobj_ns
 
 void kobj_ns_exit(enum kobj_ns_type type, const void *ns)
 {
+	sysfs_exit_ns(type, ns);
 }