user namespaces: introduce user_struct->user_namespace relationship

PATCH In review Serge E. Hallyn 2 participants 5 messages Aug 26, 2008 – Aug 28, 2008

Subsystems (5)

MM core Scheduler Security (LSM) exec and binfmt +1 more subsystems

lore

Share & tools

lore

Serge E. Hallyn

author

2008-08-26 18:53 UTC #0 lore raw

PATCH 1/3 user namespaces: introduce user_struct->user_namespace relationship

The following 3 patches are a small start to the long task of fleshing out user namespace support. These three patches just clear up the relationship of struct user_struct to struct user_namespace (patch 1), fix up the refcounting (patch 2), and complete the action of switching users when cloning a new user_namespace (patch 3).

Patch 1:

When a task does clone(CLONE_NEWNS), the task's user is the 'creator' of the new user_namespace, and the user_namespace is tacked onto a list of those created by this user.

Changelog:
Aug 25: make free_user not inlined as it's not trivial. (Eric
Biederman suggestion)
Aug 1: renamed user->user_namespace to user_ns, as the next
patch did anyway.
Aug 1: move put_user_ns call in one free_user() definition
to move it outside the lock in free_user. put_user_ns calls free_user on the user_ns->creator, which in turn would grab the lock again.

Signed-off-by Serge Hallyn <[email protected]>

---
 include/linux/sched.h          |    1 +
 include/linux/user_namespace.h |    1 +
 kernel/user.c                  |   11 +++++++++--
 kernel/user_namespace.c        |   20 +++++++++++---------
 4 files changed, 22 insertions(+), 11 deletions(-)

include/linux/sched.h +1 -0

index cfb0d87..9bebf95 100644
@@ -601,6 +601,7 @@ struct user_struct {
 	/* Hash table maintenance information */
 	struct hlist_node uidhash_node;
 	uid_t uid;
+	struct user_namespace *user_ns;
 
 #ifdef CONFIG_USER_SCHED
 	struct task_group *tg;

include/linux/user_namespace.h +1 -0

index b5f41d4..f9477c3 100644
@@ -13,6 +13,7 @@ struct user_namespace {
 	struct kref		kref;
 	struct hlist_head	uidhash_table[UIDHASH_SZ];
 	struct user_struct	*root_user;
+	struct user_struct	*creator;
 };
 
 extern struct user_namespace init_user_ns;

kernel/user.c +9 -2

index 865ecf5..ee841c7 100644
@@ -22,6 +22,7 @@ struct user_namespace init_user_ns = {
 		.refcount	= ATOMIC_INIT(2),
 	},
 	.root_user = &root_user,
+	.creator = &root_user,
 };
 EXPORT_SYMBOL_GPL(init_user_ns);

@@ -53,6 +54,7 @@ struct user_struct root_user = {
 	.files		= ATOMIC_INIT(0),
 	.sigpending	= ATOMIC_INIT(0),
 	.locked_shm     = 0,
+	.user_ns	= &init_user_ns,
 #ifdef CONFIG_USER_SCHED
 	.tg		= &init_task_group,
 #endif

@@ -319,12 +321,13 @@ done:
  * IRQ state (as stored in flags) is restored and uidhash_lock released
  * upon function exit.
  */
-static inline void free_user(struct user_struct *up, unsigned long flags)
+static void free_user(struct user_struct *up, unsigned long flags)
 {
 	/* restore back the count */
 	atomic_inc(&up->__count);
 	spin_unlock_irqrestore(&uidhash_lock, flags);
 
+	put_user_ns(up->user_ns);
 	INIT_WORK(&up->work, remove_user_sysfs_dir);
 	schedule_work(&up->work);
 }

@@ -340,13 +343,14 @@ static inline void uids_mutex_unlock(void) { }
  * IRQ state (as stored in flags) is restored and uidhash_lock released
  * upon function exit.
  */
-static inline void free_user(struct user_struct *up, unsigned long flags)
+static void free_user(struct user_struct *up, unsigned long flags)
 {
 	uid_hash_remove(up);
 	spin_unlock_irqrestore(&uidhash_lock, flags);
 	sched_destroy_user(up);
 	key_put(up->uid_keyring);
 	key_put(up->session_keyring);
+	put_user_ns(up->user_ns);
 	kmem_cache_free(uid_cachep, up);
 }

@@ -409,6 +413,8 @@ struct user_struct *alloc_uid(struct user_namespace *ns, uid_t uid)
 		if (sched_create_user(new) < 0)
 			goto out_free_user;
 
+		new->user_ns = get_user_ns(ns);
+
 		if (uids_user_create(new))
 			goto out_destoy_sched;

@@ -441,6 +447,7 @@ struct user_struct *alloc_uid(struct user_namespace *ns, uid_t uid)
 
 out_destoy_sched:
 	sched_destroy_user(new);
+	put_user_ns(new->user_ns);
 out_free_user:
 	kmem_cache_free(uid_cachep, new);
 out_unlock:

kernel/user_namespace.c +11 -9

index 532858f..f9f7ad7 100644
@@ -18,7 +18,6 @@
 static struct user_namespace *clone_user_ns(struct user_namespace *old_ns)
 {
 	struct user_namespace *ns;
-	struct user_struct *new_user;
 	int n;
 
 	ns = kmalloc(sizeof(struct user_namespace), GFP_KERNEL);

@@ -37,15 +36,17 @@ static struct user_namespace *clone_user_ns(struct user_namespace *old_ns)
 		return ERR_PTR(-ENOMEM);
 	}
 
-	/* Reset current->user with a new one */
-	new_user = alloc_uid(ns, current->uid);
-	if (!new_user) {
-		free_uid(ns->root_user);
-		kfree(ns);
-		return ERR_PTR(-ENOMEM);
-	}
+	/* pin the creating user */
+	ns->creator = current->user;
+	atomic_inc(&ns->creator->__count);
+
+	/*
+	 * The alloc_uid() incremented the userns refcount,
+	 * so drop it again
+	 */
+	put_user_ns(ns);
 
-	switch_uid(new_user);
+	switch_uid(ns->root_user);
 	return ns;
 }

@@ -71,6 +72,7 @@ void free_user_ns(struct kref *kref)
 
 	ns = container_of(kref, struct user_namespace, kref);
 	release_uids(ns);
+	free_uid(ns->creator);
 	kfree(ns);
 }
 EXPORT_SYMBOL(free_user_ns);

Signature · 1 line

1.5.4.3

Serge E. Hallyn

author

2008-08-26 18:56 UTC #1 lore raw

PATCH 2/3 user namespaces: move user_ns from nsproxy into user struct

When we get the sysfs support needed to support fair user scheduling along with user namespaces, then we will need to be able to get the user namespace from the user struct.

So we need the user_ns to be a part of struct user. Once we can access it from tsk->user, we no longer have a use for tsk->nsproxy->user_ns.

When a user_namespace is created, the user which created it is marked as its 'creator'. The user_namespace pins the creator. Each userid in a user_ns pins the user_ns. This keeps refcounting nice and simple.

At the same time, this patch simplifies the refcounting. The current user and userns locking works as follows:

The task pins the user struct.

The task pins the nsproxy, the nsproxy pins the user_ns.

When a new user_ns is created, it creates a root user for it, and pins it. When the nsproxy releases the user_ns, the userns tries to release all its user structs.

So you see that the refcounting "works" for now, but only because the nsproxy (and therefore usr_ns) and user structs will be freed at the same time - when the last task using them is released.

Now we need to put the user_ns in the struct user. You can see that will mess up the refcounting.

Fortunately, once the user_ns is available from tsk->user, we don't need it in nsproxy.

So here is how the refcounting *should* be done:

The task pins the user struct.

The user struct pins its user namespace.

The user namespace pins the struct user which created it.

A user namespace now doesn't need to release its userids, because it is only released when its last user disappears.

This patch makes those changes.

Signed-off-by Serge Hallyn <[email protected]>

---
 include/linux/init_task.h      |    1 -
 include/linux/key.h            |    3 ++
 include/linux/nsproxy.h        |    1 -
 include/linux/sched.h          |    1 +
 include/linux/user_namespace.h |   10 +++-----
 kernel/fork.c                  |    3 +-
 kernel/nsproxy.c               |   10 +-------
 kernel/sys.c                   |    4 +-
 kernel/user.c                  |   42 +++++++++------------------------------
 kernel/user_namespace.c        |   36 +++++++++------------------------
 security/keys/process_keys.c   |    7 +++++-
 11 files changed, 39 insertions(+), 79 deletions(-)

include/linux/init_task.h +0 -1

index 021d8e7..550058b 100644
@@ -57,7 +57,6 @@ extern struct nsproxy init_nsproxy;
 	.mnt_ns		= NULL,						\
 	INIT_NET_NS(net_ns)                                             \
 	INIT_IPC_NS(ipc_ns)						\
-	.user_ns	= &init_user_ns,				\
 }
 
 #define INIT_SIGHAND(sighand) {						\

include/linux/key.h +3 -0

index c45c962..ba53aef 100644
@@ -277,6 +277,8 @@ extern ctl_table key_sysctls[];
  * the userspace interface
  */
 extern void switch_uid_keyring(struct user_struct *new_user);
+extern void switch_uid_keyring_task(struct task_struct *tsk,
+				struct user_struct *new_user);
 extern int copy_keys(unsigned long clone_flags, struct task_struct *tsk);
 extern int copy_thread_group_keys(struct task_struct *tsk);
 extern void exit_keys(struct task_struct *tsk);

@@ -305,6 +307,7 @@ extern void key_init(void);
 #define key_ref_to_ptr(k)		({ NULL; })
 #define is_key_possessed(k)		0
 #define switch_uid_keyring(u)		do { } while(0)
+#define switch_uid_keyring_task(t,u)	do { } while(0)
 #define __install_session_keyring(t, k)	({ NULL; })
 #define copy_keys(f,t)			0
 #define copy_thread_group_keys(t)	0

include/linux/nsproxy.h +0 -1

index c8a768e..afad7de 100644
@@ -27,7 +27,6 @@ struct nsproxy {
 	struct ipc_namespace *ipc_ns;
 	struct mnt_namespace *mnt_ns;
 	struct pid_namespace *pid_ns;
-	struct user_namespace *user_ns;
 	struct net 	     *net_ns;
 };
 extern struct nsproxy init_nsproxy;

include/linux/sched.h +1 -0

index 9bebf95..47147fb 100644
@@ -1716,6 +1716,7 @@ static inline struct user_struct *get_uid(struct user_struct *u)
 }
 extern void free_uid(struct user_struct *);
 extern void switch_uid(struct user_struct *);
+extern void task_switch_uid(struct task_struct *tsk, struct user_struct *);
 extern void release_uids(struct user_namespace *ns);
 
 #include <asm/current.h>

include/linux/user_namespace.h +4 -6

index f9477c3..1b4959d 100644
@@ -27,8 +27,7 @@ static inline struct user_namespace *get_user_ns(struct user_namespace *ns)
 	return ns;
 }
 
-extern struct user_namespace *copy_user_ns(int flags,
-					   struct user_namespace *old_ns);
+extern int create_new_userns(int flags, struct task_struct *tsk);
 extern void free_user_ns(struct kref *kref);
 
 static inline void put_user_ns(struct user_namespace *ns)

@@ -44,13 +43,12 @@ static inline struct user_namespace *get_user_ns(struct user_namespace *ns)
 	return &init_user_ns;
 }
 
-static inline struct user_namespace *copy_user_ns(int flags,
-						  struct user_namespace *old_ns)
+static inline int create_new_userns(int flags, struct task_struct *tsk)
 {
 	if (flags & CLONE_NEWUSER)
-		return ERR_PTR(-EINVAL);
+		return -EINVAL;
 
-	return old_ns;
+	return 0;
 }
 
 static inline void put_user_ns(struct user_namespace *ns)

kernel/fork.c +1 -2

index 7ce2ebe..b5d4dc5 100644
@@ -941,8 +941,7 @@ static struct task_struct *copy_process(unsigned long clone_flags,
 	retval = -EAGAIN;
 	if (atomic_read(&p->user->processes) >=
 			p->signal->rlim[RLIMIT_NPROC].rlim_cur) {
-		if (!capable(CAP_SYS_ADMIN) && !capable(CAP_SYS_RESOURCE) &&
-		    p->user != current->nsproxy->user_ns->root_user)
+		if (!capable(CAP_SYS_ADMIN) && !capable(CAP_SYS_RESOURCE))
 			goto bad_fork_free;
 	}

kernel/nsproxy.c +2 -8

index 1d3ef29..766c672 100644
@@ -80,11 +80,9 @@ static struct nsproxy *create_new_namespaces(unsigned long flags,
 		goto out_pid;
 	}
 
-	new_nsp->user_ns = copy_user_ns(flags, tsk->nsproxy->user_ns);
-	if (IS_ERR(new_nsp->user_ns)) {
-		err = PTR_ERR(new_nsp->user_ns);
+	err = create_new_userns(flags, tsk);
+	if (err)
 		goto out_user;
-	}
 
 	new_nsp->net_ns = copy_net_ns(flags, tsk->nsproxy->net_ns);
 	if (IS_ERR(new_nsp->net_ns)) {

@@ -95,8 +93,6 @@ static struct nsproxy *create_new_namespaces(unsigned long flags,
 	return new_nsp;
 
 out_net:
-	if (new_nsp->user_ns)
-		put_user_ns(new_nsp->user_ns);
 out_user:
 	if (new_nsp->pid_ns)
 		put_pid_ns(new_nsp->pid_ns);

@@ -173,8 +169,6 @@ void free_nsproxy(struct nsproxy *ns)
 		put_ipc_ns(ns->ipc_ns);
 	if (ns->pid_ns)
 		put_pid_ns(ns->pid_ns);
-	if (ns->user_ns)
-		put_user_ns(ns->user_ns);
 	put_net(ns->net_ns);
 	kmem_cache_free(nsproxy_cachep, ns);
 }

kernel/sys.c +2 -2

index 038a7bc..d89ddd9 100644
@@ -553,13 +553,13 @@ static int set_user(uid_t new_ruid, int dumpclear)
 {
 	struct user_struct *new_user;
 
-	new_user = alloc_uid(current->nsproxy->user_ns, new_ruid);
+	new_user = alloc_uid(current->user->user_ns, new_ruid);
 	if (!new_user)
 		return -EAGAIN;
 
 	if (atomic_read(&new_user->processes) >=
 				current->signal->rlim[RLIMIT_NPROC].rlim_cur &&
-			new_user != current->nsproxy->user_ns->root_user) {
+			!capable(CAP_SYS_RESOURCE) && !capable(CAP_SYS_ADMIN)) {
 		free_uid(new_user);
 		return -EAGAIN;
 	}

kernel/user.c +10 ~~-32~~

74 lines not on this page · raw message

kernel/user_namespace.c +10 ~~-26~~

index f9f7ad7..d59f193 100644
@@ -15,14 +15,17 @@
  * @old_ns: namespace to clone
  * Return NULL on error (failure to kmalloc), new ns otherwise
  */
-static struct user_namespace *clone_user_ns(struct user_namespace *old_ns)
+int create_new_userns(int flags, struct task_struct *tsk)
 {
 	struct user_namespace *ns;
 	int n;
 
+	if (!(flags & CLONE_NEWUSER))
+		return 0;
+
 	ns = kmalloc(sizeof(struct user_namespace), GFP_KERNEL);
 	if (!ns)
-		return ERR_PTR(-ENOMEM);
+		return -ENOMEM;
 
 	kref_init(&ns->kref);

@@ -33,37 +36,19 @@ static struct user_namespace *clone_user_ns(struct user_namespace *old_ns)
 	ns->root_user = alloc_uid(ns, 0);
 	if (!ns->root_user) {
 		kfree(ns);
-		return ERR_PTR(-ENOMEM);
+		return -ENOMEM;
 	}
 
 	/* pin the creating user */
-	ns->creator = current->user;
+	ns->creator = tsk->user;
 	atomic_inc(&ns->creator->__count);
 
-	/*
-	 * The alloc_uid() incremented the userns refcount,
-	 * so drop it again
-	 */
+	/* alloc_uid() incremented the userns refcount.  drop it again */
 	put_user_ns(ns);
 
-	switch_uid(ns->root_user);
-	return ns;
-}
-
-struct user_namespace * copy_user_ns(int flags, struct user_namespace *old_ns)
-{
-	struct user_namespace *new_ns;
-
-	BUG_ON(!old_ns);
-	get_user_ns(old_ns);
-
-	if (!(flags & CLONE_NEWUSER))
-		return old_ns;
-
-	new_ns = clone_user_ns(old_ns);
+	task_switch_uid(tsk, ns->root_user);
 
-	put_user_ns(old_ns);
-	return new_ns;
+	return 0;
 }
 
 void free_user_ns(struct kref *kref)

@@ -71,7 +56,6 @@ void free_user_ns(struct kref *kref)
 	struct user_namespace *ns;
 
 	ns = container_of(kref, struct user_namespace, kref);
-	release_uids(ns);
 	free_uid(ns->creator);
 	kfree(ns);
 }

security/keys/process_keys.c +6 -1

index 5be6d01..8d1cfb2 100644
@@ -119,7 +119,8 @@ error:
 /*
  * deal with the UID changing
  */
-void switch_uid_keyring(struct user_struct *new_user)
+void switch_uid_keyring_task(struct task_struct *task,
+				struct user_struct *new_user)
 {
 #if 0 /* do nothing for now */
 	struct key *old;

@@ -142,6 +143,10 @@ void switch_uid_keyring(struct user_struct *new_user)
 
 } /* end switch_uid_keyring() */
 
+void switch_uid_keyring(struct user_struct *new_user)
+{
+	switch_uid_keyring_task(current, new_user);
+}
 /*****************************************************************************/
 /*
  * install a fresh thread keyring, discarding the old one

Signature · 1 line

1.5.4.3

Serge E. Hallyn

author

2008-08-26 18:56 UTC #2 lore raw

PATCH 3/3 user namespaces: reset task's credentials on CLONE_NEWUSER

In reply to Serge E. Hallyn

Currently, creating a new user namespace does not reset the task's uid or gid. Since generally that is done as root because it requires CAP_SYS_ADMIN, and since the first uid in the new namespace is 0, one usually doesn't notice. However, if one does

capset cap_sys_admin=ep ns_exec
su - hallyn
ns_exec -U /bin/sh
id

then one will see hallyn's userid, and all preexisting groups.

With this patch, cloning a new user namespace will set the task's uid and gid to 0, and reset the group_info to the empty set assigned to init.

Signed-off-by Serge E. Hallyn <[email protected]>

---
 kernel/user_namespace.c |   14 ++++++++++++++
 1 files changed, 14 insertions(+), 0 deletions(-)

kernel/user_namespace.c +14 -0

index d59f193..16e6296 100644
@@ -10,6 +10,9 @@
 #include <linux/slab.h>
 #include <linux/user_namespace.h>
 
+/* defined in kernel/sys.c */
+extern struct group_info init_groups;
+
 /*
  * Clone a new ns copying an original user ns, setting refcount to 1
  * @old_ns: namespace to clone

@@ -47,6 +50,17 @@ int create_new_userns(int flags, struct task_struct *tsk)
 	put_user_ns(ns);
 
 	task_switch_uid(tsk, ns->root_user);
+	tsk->uid = tsk->euid = tsk->suid = tsk->fsuid = 0;
+	tsk->gid = tsk->egid = tsk->sgid = tsk->fsgid = 0;
+
+	/* this can't be safe for unshare, can it?  it's safe
+	 * for fork, though.  I'm tempted to limit clone_newuser to
+	 * fork only */
+	task_lock(tsk);
+	put_group_info(tsk->group_info);
+	tsk->group_info = &init_groups;
+	get_group_info(tsk->group_info);
+	task_unlock(tsk);
 
 	return 0;
 }

Signature · 1 line

1.5.4.3

Andrew Morton

maintainer

2008-08-28 00:12 UTC #3 lore raw

In reply to Serge E. Hallyn

Quoted 14 lines · On Tue, 26 Aug 2008 13:53:41 -0500 "Serge E. Hallyn" <[email protected]> wrote:

When a task does clone(CLONE_NEWNS), the task's user is the 'creator' of the new user_namespace, and the user_namespace is tacked onto a list of those created by this user.
Changelog:
Aug 25: make free_user not inlined as it's not trivial. (Eric
Biederman suggestion)
Aug 1: renamed user->user_namespace to user_ns, as the next
patch did anyway.
Aug 1: move put_user_ns call in one free_user() definition
to move it outside the lock in free_user. put_user_ns calls free_user on the user_ns->creator, which in turn would grab the lock again.

Quoted 6 lines

Signed-off-by Serge Hallyn <[email protected]>

---
 include/linux/sched.h          |    1 +
 include/linux/user_namespace.h |    1 +
 kernel/user.c                  |   11 +++++++++--
 kernel/user_namespace.c        |   20 +++++++++++---------

The credentials code in linux-next is changing the same code which you're changing, in more-than-trivially-textual ways.

I'd suggest a dhowells cc on these changes, as he's also working in this area, and as you touch the keyring code a bit.

And, of course, please remove that almost-always-wrong extern-declaration-in-C which checkpatch told you about. init_groups is already declared in include/linux/init_task.h anyway...

Serge E. Hallyn

author

2008-08-28 00:52 UTC #4 lore raw

Quoted 24 lines · Andrew Morton

Quoted 14 lines · On Tue, 26 Aug 2008 13:53:41 -0500 "Serge E. Hallyn" <[email protected]> wrote:
When a task does clone(CLONE_NEWNS), the task's user is the 'creator' of the new user_namespace, and the user_namespace is tacked onto a list of those created by this user.
Changelog:
Aug 25: make free_user not inlined as it's not trivial. (Eric
Biederman suggestion)
Aug 1: renamed user->user_namespace to user_ns, as the next
patch did anyway.
Aug 1: move put_user_ns call in one free_user() definition
to move it outside the lock in free_user. put_user_ns calls free_user on the user_ns->creator, which in turn would grab the lock again.
Quoted 6 lines
Signed-off-by Serge Hallyn <[email protected]>
---
 include/linux/sched.h          |    1 +
 include/linux/user_namespace.h |    1 +
 kernel/user.c                  |   11 +++++++++--
 kernel/user_namespace.c        |   20 +++++++++++---------

Quoted 2 lines

The credentials code in linux-next is changing the same code which you're changing, in more-than-trivially-textual ways.

Ok, I knew that was coming eventually. I'll rebase.

Quoted 2 lines

I'd suggest a dhowells cc on these changes, as he's also working in this area, and as you touch the keyring code a bit.

Ok.

Quoted 3 lines

And, of course, please remove that almost-always-wrong extern-declaration-in-C which checkpatch told you about. init_groups is already declared in include/linux/init_task.h anyway...

Sorry, will do with the rebase.

Thanks, Andrew.

-serge