CRED: Inaugurate COW credentials

Inaugurate copy-on-write credentials management.  This uses RCU to manage the
credentials pointer in the task_struct with respect to accesses by other tasks.
A process may only modify its own credentials, and so does not need locking to
access or modify its own credentials.

A mutex (cred_replace_mutex) is added to the task_struct to control the effect
of PTRACE_ATTACHED on credential calculations, particularly with respect to
execve().

With this patch, the contents of an active credentials struct may not be
changed directly; rather a new set of credentials must be prepared, modified
and committed using something like the following sequence of events:

	struct cred *new = prepare_creds();
	int ret = blah(new);
	if (ret < 0) {
		abort_creds(new);
		return ret;
	}
	return commit_creds(new);

There are some exceptions to this rule: the keyrings pointed to by the active
credentials may be instantiated - keyrings violate the COW rule as managing
COW keyrings is tricky, given that it is possible for a task to directly alter
the keys in a keyring in use by another task.

To help enforce this, various pointers to sets of credentials, such as those in
the task_struct, are declared const.  The purpose of this is compile-time
discouragement of altering credentials through those pointers.  Once a set of
credentials has been made public through one of these pointers, it may not be
modified, except under special circumstances:

  (1) Its reference count may incremented and decremented.

  (2) The keyrings to which it points may be modified, but not replaced.

The only safe way to modify anything else is to create a replacement and commit
using the functions described in Documentation/credentials.txt (which will be
added by a later patch).

This patch and the preceding patches have been tested with the LTP SELinux
testsuite.

This patch makes several logical sets of alteration:

 (1) execve().

     This now prepares and commits credentials in various places in the
     security code rather than altering the current creds directly.

 (2) Temporary credential overrides.

     do_coredump() and sys_faccessat() now prepare their own credentials and
     temporarily override the ones currently on the acting thread, whilst
     preventing interference from other threads by holding cred_replace_mutex
     on the thread being dumped.

     This will be replaced in a future patch by something that hands down the
     credentials directly to the functions being called, rather than altering
     the task's objective credentials.

 (3) LSM interface.

     A number of functions have been changed, added or removed:

     (*) security_capset_check(), ->capset_check()
     (*) security_capset_set(), ->capset_set()

     	 Removed in favour of security_capset().

     (*) security_capset(), ->capset()

     	 New.  This is passed a pointer to the new creds, a pointer to the old
     	 creds and the proposed capability sets.  It should fill in the new
     	 creds or return an error.  All pointers, barring the pointer to the
     	 new creds, are now const.

     (*) security_bprm_apply_creds(), ->bprm_apply_creds()

     	 Changed; now returns a value, which will cause the process to be
     	 killed if it's an error.

     (*) security_task_alloc(), ->task_alloc_security()

     	 Removed in favour of security_prepare_creds().

     (*) security_cred_free(), ->cred_free()

     	 New.  Free security data attached to cred->security.

     (*) security_prepare_creds(), ->cred_prepare()

     	 New. Duplicate any security data attached to cred->security.

     (*) security_commit_creds(), ->cred_commit()

     	 New. Apply any security effects for the upcoming installation of new
     	 security by commit_creds().

     (*) security_task_post_setuid(), ->task_post_setuid()

     	 Removed in favour of security_task_fix_setuid().

     (*) security_task_fix_setuid(), ->task_fix_setuid()

     	 Fix up the proposed new credentials for setuid().  This is used by
     	 cap_set_fix_setuid() to implicitly adjust capabilities in line with
     	 setuid() changes.  Changes are made to the new credentials, rather
     	 than the task itself as in security_task_post_setuid().

     (*) security_task_reparent_to_init(), ->task_reparent_to_init()

     	 Removed.  Instead the task being reparented to init is referred
     	 directly to init's credentials.

	 NOTE!  This results in the loss of some state: SELinux's osid no
	 longer records the sid of the thread that forked it.

     (*) security_key_alloc(), ->key_alloc()
     (*) security_key_permission(), ->key_permission()

     	 Changed.  These now take cred pointers rather than task pointers to
     	 refer to the security context.

 (4) sys_capset().

     This has been simplified and uses less locking.  The LSM functions it
     calls have been merged.

 (5) reparent_to_kthreadd().

     This gives the current thread the same credentials as init by simply using
     commit_thread() to point that way.

 (6) __sigqueue_alloc() and switch_uid()

     __sigqueue_alloc() can't stop the target task from changing its creds
     beneath it, so this function gets a reference to the currently applicable
     user_struct which it then passes into the sigqueue struct it returns if
     successful.

     switch_uid() is now called from commit_creds(), and possibly should be
     folded into that.  commit_creds() should take care of protecting
     __sigqueue_alloc().

 (7) [sg]et[ug]id() and co and [sg]et_current_groups.

     The set functions now all use prepare_creds(), commit_creds() and
     abort_creds() to build and check a new set of credentials before applying
     it.

     security_task_set[ug]id() is called inside the prepared section.  This
     guarantees that nothing else will affect the creds until we've finished.

     The calling of set_dumpable() has been moved into commit_creds().

     Much of the functionality of set_user() has been moved into
     commit_creds().

     The get functions all simply access the data directly.

 (8) security_task_prctl() and cap_task_prctl().

     security_task_prctl() has been modified to return -ENOSYS if it doesn't
     want to handle a function, or otherwise return the return value directly
     rather than through an argument.

     Additionally, cap_task_prctl() now prepares a new set of credentials, even
     if it doesn't end up using it.

 (9) Keyrings.

     A number of changes have been made to the keyrings code:

     (a) switch_uid_keyring(), copy_keys(), exit_keys() and suid_keys() have
     	 all been dropped and built in to the credentials functions directly.
     	 They may want separating out again later.

     (b) key_alloc() and search_process_keyrings() now take a cred pointer
     	 rather than a task pointer to specify the security context.

     (c) copy_creds() gives a new thread within the same thread group a new
     	 thread keyring if its parent had one, otherwise it discards the thread
     	 keyring.

     (d) The authorisation key now points directly to the credentials to extend
     	 the search into rather pointing to the task that carries them.

     (e) Installing thread, process or session keyrings causes a new set of
     	 credentials to be created, even though it's not strictly necessary for
     	 process or session keyrings (they're shared).

(10) Usermode helper.

     The usermode helper code now carries a cred struct pointer in its
     subprocess_info struct instead of a new session keyring pointer.  This set
     of credentials is derived from init_cred and installed on the new process
     after it has been cloned.

     call_usermodehelper_setup() allocates the new credentials and
     call_usermodehelper_freeinfo() discards them if they haven't been used.  A
     special cred function (prepare_usermodeinfo_creds()) is provided
     specifically for call_usermodehelper_setup() to call.

     call_usermodehelper_setkeys() adjusts the credentials to sport the
     supplied keyring as the new session keyring.

(11) SELinux.

     SELinux has a number of changes, in addition to those to support the LSM
     interface changes mentioned above:

     (a) selinux_setprocattr() no longer does its check for whether the
     	 current ptracer can access processes with the new SID inside the lock
     	 that covers getting the ptracer's SID.  Whilst this lock ensures that
     	 the check is done with the ptracer pinned, the result is only valid
     	 until the lock is released, so there's no point doing it inside the
     	 lock.

(12) is_single_threaded().

     This function has been extracted from selinux_setprocattr() and put into
     a file of its own in the lib/ directory as join_session_keyring() now
     wants to use it too.

     The code in SELinux just checked to see whether a task shared mm_structs
     with other tasks (CLONE_VM), but that isn't good enough.  We really want
     to know if they're part of the same thread group (CLONE_THREAD).

(13) nfsd.

     The NFS server daemon now has to use the COW credentials to set the
     credentials it is going to use.  It really needs to pass the credentials
     down to the functions it calls, but it can't do that until other patches
     in this series have been applied.

Signed-off-by: David Howells <dhowells@redhat.com>
Acked-by: James Morris <jmorris@namei.org>
Signed-off-by: James Morris <jmorris@namei.org>

12 files changed, 652 insertions, 374 deletions

diff --git a/kernel/auditsc.c b/kernel/auditsc.c
index ae8ef88ade3..bc1e2d854bf 100644
--- a/kernel/auditsc.c
+++ b/kernel/auditsc.c
@@ -2546,18 +2546,17 @@ int __audit_signal_info(int sig, struct task_struct *t)
 
 /**
  * __audit_log_bprm_fcaps - store information about a loading bprm and relevant fcaps
- * @bprm pointer to the bprm being processed
- * @caps the caps read from the disk
+ * @bprm: pointer to the bprm being processed
+ * @new: the proposed new credentials
+ * @old: the old credentials
  *
  * Simply check if the proc already has the caps given by the file and if not
  * store the priv escalation info for later auditing at the end of the syscall
  *
- * this can fail and we don't care.  See the note in audit.h for
- * audit_log_bprm_fcaps() for my explaination....
- *
  * -Eric
  */
-void __audit_log_bprm_fcaps(struct linux_binprm *bprm, kernel_cap_t *pP, kernel_cap_t *pE)
+int __audit_log_bprm_fcaps(struct linux_binprm *bprm,
+			   const struct cred *new, const struct cred *old)
 {
 	struct audit_aux_data_bprm_fcaps *ax;
 	struct audit_context *context = current->audit_context;
@@ -2566,7 +2565,7 @@ void __audit_log_bprm_fcaps(struct linux_binprm *bprm, kernel_cap_t *pP, kernel_
 
 	ax = kmalloc(sizeof(*ax), GFP_KERNEL);
 	if (!ax)
-		return;
+		return -ENOMEM;
 
 	ax->d.type = AUDIT_BPRM_FCAPS;
 	ax->d.next = context->aux;
@@ -2581,26 +2580,27 @@ void __audit_log_bprm_fcaps(struct linux_binprm *bprm, kernel_cap_t *pP, kernel_
 	ax->fcap.fE = !!(vcaps.magic_etc & VFS_CAP_FLAGS_EFFECTIVE);
 	ax->fcap_ver = (vcaps.magic_etc & VFS_CAP_REVISION_MASK) >> VFS_CAP_REVISION_SHIFT;
 
-	ax->old_pcap.permitted = *pP;
-	ax->old_pcap.inheritable = current->cred->cap_inheritable;
-	ax->old_pcap.effective = *pE;
+	ax->old_pcap.permitted   = old->cap_permitted;
+	ax->old_pcap.inheritable = old->cap_inheritable;
+	ax->old_pcap.effective   = old->cap_effective;
 
-	ax->new_pcap.permitted = current->cred->cap_permitted;
-	ax->new_pcap.inheritable = current->cred->cap_inheritable;
-	ax->new_pcap.effective = current->cred->cap_effective;
+	ax->new_pcap.permitted   = new->cap_permitted;
+	ax->new_pcap.inheritable = new->cap_inheritable;
+	ax->new_pcap.effective   = new->cap_effective;
+	return 0;
 }
 
 /**
  * __audit_log_capset - store information about the arguments to the capset syscall
- * @pid target pid of the capset call
- * @eff effective cap set
- * @inh inheritible cap set
- * @perm permited cap set
+ * @pid: target pid of the capset call
+ * @new: the new credentials
+ * @old: the old (current) credentials
  *
  * Record the aguments userspace sent to sys_capset for later printing by the
  * audit system if applicable
  */
-int __audit_log_capset(pid_t pid, kernel_cap_t *eff, kernel_cap_t *inh, kernel_cap_t *perm)
+int __audit_log_capset(pid_t pid,
+		       const struct cred *new, const struct cred *old)
 {
 	struct audit_aux_data_capset *ax;
 	struct audit_context *context = current->audit_context;
@@ -2617,9 +2617,9 @@ int __audit_log_capset(pid_t pid, kernel_cap_t *eff, kernel_cap_t *inh, kernel_c
 	context->aux = (void *)ax;
 
 	ax->pid = pid;
-	ax->cap.effective = *eff;
-	ax->cap.inheritable = *eff;
-	ax->cap.permitted = *perm;
+	ax->cap.effective   = new->cap_effective;
+	ax->cap.inheritable = new->cap_effective;
+	ax->cap.permitted   = new->cap_permitted;
 
 	return 0;
 }
diff --git a/kernel/capability.c b/kernel/capability.c
index a404b980b1b..36b4b4daebe 100644
--- a/kernel/capability.c
+++ b/kernel/capability.c
@@ -15,12 +15,7 @@
 #include <linux/syscalls.h>
 #include <linux/pid_namespace.h>
 #include <asm/uaccess.h>
-
-/*
- * This lock protects task->cap_* for all tasks including current.
- * Locking rule: acquire this prior to tasklist_lock.
- */
-static DEFINE_SPINLOCK(task_capability_lock);
+#include "cred-internals.h"
 
 /*
  * Leveraged for setting/resetting capabilities
@@ -128,12 +123,11 @@ static int cap_validate_magic(cap_user_header_t header, unsigned *tocopy)
 }
 
 /*
- * If we have configured with filesystem capability support, then the
- * only thing that can change the capabilities of the current process
- * is the current process. As such, we can't be in this code at the
- * same time as we are in the process of setting capabilities in this
- * process. The net result is that we can limit our use of locks to
- * when we are reading the caps of another process.
+ * The only thing that can change the capabilities of the current
+ * process is the current process. As such, we can't be in this code
+ * at the same time as we are in the process of setting capabilities
+ * in this process. The net result is that we can limit our use of
+ * locks to when we are reading the caps of another process.
  */
 static inline int cap_get_target_pid(pid_t pid, kernel_cap_t *pEp,
 				     kernel_cap_t *pIp, kernel_cap_t *pPp)
@@ -143,7 +137,6 @@ static inline int cap_get_target_pid(pid_t pid, kernel_cap_t *pEp,
 	if (pid && (pid != task_pid_vnr(current))) {
 		struct task_struct *target;
 
-		spin_lock(&task_capability_lock);
 		read_lock(&tasklist_lock);
 
 		target = find_task_by_vpid(pid);
@@ -153,34 +146,12 @@ static inline int cap_get_target_pid(pid_t pid, kernel_cap_t *pEp,
 			ret = security_capget(target, pEp, pIp, pPp);
 
 		read_unlock(&tasklist_lock);
-		spin_unlock(&task_capability_lock);
 	} else
 		ret = security_capget(current, pEp, pIp, pPp);
 
 	return ret;
 }
 
-/*
- * Atomically modify the effective capabilities returning the original
- * value. No permission check is performed here - it is assumed that the
- * caller is permitted to set the desired effective capabilities.
- */
-kernel_cap_t cap_set_effective(const kernel_cap_t pE_new)
-{
-	kernel_cap_t pE_old;
-
-	spin_lock(&task_capability_lock);
-
-	pE_old = current->cred->cap_effective;
-	current->cred->cap_effective = pE_new;
-
-	spin_unlock(&task_capability_lock);
-
-	return pE_old;
-}
-
-EXPORT_SYMBOL(cap_set_effective);
-
 /**
  * sys_capget - get the capabilities of a given process.
  * @header: pointer to struct that contains capability version and
@@ -208,7 +179,6 @@ asmlinkage long sys_capget(cap_user_header_t header, cap_user_data_t dataptr)
 		return -EINVAL;
 
 	ret = cap_get_target_pid(pid, &pE, &pI, &pP);
-
 	if (!ret) {
 		struct __user_cap_data_struct kdata[_KERNEL_CAPABILITY_U32S];
 		unsigned i;
@@ -270,6 +240,7 @@ asmlinkage long sys_capset(cap_user_header_t header, const cap_user_data_t data)
 	struct __user_cap_data_struct kdata[_KERNEL_CAPABILITY_U32S];
 	unsigned i, tocopy;
 	kernel_cap_t inheritable, permitted, effective;
+	struct cred *new;
 	int ret;
 	pid_t pid;
 
@@ -284,8 +255,8 @@ asmlinkage long sys_capset(cap_user_header_t header, const cap_user_data_t data)
 	if (pid != 0 && pid != task_pid_vnr(current))
 		return -EPERM;
 
-	if (copy_from_user(&kdata, data, tocopy
-			   * sizeof(struct __user_cap_data_struct)))
+	if (copy_from_user(&kdata, data,
+			   tocopy * sizeof(struct __user_cap_data_struct)))
 		return -EFAULT;
 
 	for (i = 0; i < tocopy; i++) {
@@ -300,24 +271,23 @@ asmlinkage long sys_capset(cap_user_header_t header, const cap_user_data_t data)
 		i++;
 	}
 
-	ret = audit_log_capset(pid, &effective, &inheritable, &permitted);
-	if (ret)
+	new = prepare_creds();
+	if (!new)
+		return -ENOMEM;
+
+	ret = security_capset(new, current_cred(),
+			      &effective, &inheritable, &permitted);
+	if (ret < 0)
+		goto error;
+
+	ret = audit_log_capset(pid, new, current_cred());
+	if (ret < 0)
 		return ret;
 
-	/* This lock is required even when filesystem capability support is
-	 * configured - it protects the sys_capget() call from returning
-	 * incorrect data in the case that the targeted process is not the
-	 * current one.
-	 */
-	spin_lock(&task_capability_lock);
-
-	ret = security_capset_check(&effective, &inheritable, &permitted);
-	/* Having verified that the proposed changes are legal, we now put them
-	 * into effect.
-	 */
-	if (!ret)
-		security_capset_set(&effective, &inheritable, &permitted);
-	spin_unlock(&task_capability_lock);
+	return commit_creds(new);
+
+error:
+	abort_creds(new);
 	return ret;
 }
 
diff --git a/kernel/cred-internals.h b/kernel/cred-internals.h
new file mode 100644
index 00000000000..2dc4fc2d0bf
--- /dev/null
+++ b/kernel/cred-internals.h
@@ -0,0 +1,21 @@
+/* Internal credentials stuff
+ *
+ * Copyright (C) 2008 Red Hat, Inc. All Rights Reserved.
+ * Written by David Howells (dhowells@redhat.com)
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public Licence
+ * as published by the Free Software Foundation; either version
+ * 2 of the Licence, or (at your option) any later version.
+ */
+
+/*
+ * user.c
+ */
+static inline void sched_switch_user(struct task_struct *p)
+{
+#ifdef CONFIG_USER_SCHED
+	sched_move_task(p);
+#endif	/* CONFIG_USER_SCHED */
+}
+
diff --git a/kernel/cred.c b/kernel/cred.c
index ac73e361768..cb6b5eda978 100644
--- a/kernel/cred.c
+++ b/kernel/cred.c
@@ -15,6 +15,10 @@
 #include <linux/keyctl.h>
 #include <linux/init_task.h>
 #include <linux/security.h>
+#include <linux/cn_proc.h>
+#include "cred-internals.h"
+
+static struct kmem_cache *cred_jar;
 
 /*
  * The common credentials for the initial task's thread group
@@ -64,7 +68,7 @@ static void release_tgcred_rcu(struct rcu_head *rcu)
 /*
  * Release a set of thread group credentials.
  */
-static void release_tgcred(struct cred *cred)
+void release_tgcred(struct cred *cred)
 {
 #ifdef CONFIG_KEYS
 	struct thread_group_cred *tgcred = cred->tgcred;
@@ -81,79 +85,322 @@ static void put_cred_rcu(struct rcu_head *rcu)
 {
 	struct cred *cred = container_of(rcu, struct cred, rcu);
 
-	BUG_ON(atomic_read(&cred->usage) != 0);
+	if (atomic_read(&cred->usage) != 0)
+		panic("CRED: put_cred_rcu() sees %p with usage %d\n",
+		      cred, atomic_read(&cred->usage));
 
+	security_cred_free(cred);
 	key_put(cred->thread_keyring);
 	key_put(cred->request_key_auth);
 	release_tgcred(cred);
 	put_group_info(cred->group_info);
 	free_uid(cred->user);
-	security_cred_free(cred);
-	kfree(cred);
+	kmem_cache_free(cred_jar, cred);
 }
 
 /**
  * __put_cred - Destroy a set of credentials
- * @sec: The record to release
+ * @cred: The record to release
  *
  * Destroy a set of credentials on which no references remain.
  */
 void __put_cred(struct cred *cred)
 {
+	BUG_ON(atomic_read(&cred->usage) != 0);
+
 	call_rcu(&cred->rcu, put_cred_rcu);
 }
 EXPORT_SYMBOL(__put_cred);
 
+/**
+ * prepare_creds - Prepare a new set of credentials for modification
+ *
+ * Prepare a new set of task credentials for modification.  A task's creds
+ * shouldn't generally be modified directly, therefore this function is used to
+ * prepare a new copy, which the caller then modifies and then commits by
+ * calling commit_creds().
+ *
+ * Returns a pointer to the new creds-to-be if successful, NULL otherwise.
+ *
+ * Call commit_creds() or abort_creds() to clean up.
+ */
+struct cred *prepare_creds(void)
+{
+	struct task_struct *task = current;
+	const struct cred *old;
+	struct cred *new;
+
+	BUG_ON(atomic_read(&task->cred->usage) < 1);
+
+	new = kmem_cache_alloc(cred_jar, GFP_KERNEL);
+	if (!new)
+		return NULL;
+
+	old = task->cred;
+	memcpy(new, old, sizeof(struct cred));
+
+	atomic_set(&new->usage, 1);
+	get_group_info(new->group_info);
+	get_uid(new->user);
+
+#ifdef CONFIG_KEYS
+	key_get(new->thread_keyring);
+	key_get(new->request_key_auth);
+	atomic_inc(&new->tgcred->usage);
+#endif
+
+#ifdef CONFIG_SECURITY
+	new->security = NULL;
+#endif
+
+	if (security_prepare_creds(new, old, GFP_KERNEL) < 0)
+		goto error;
+	return new;
+
+error:
+	abort_creds(new);
+	return NULL;
+}
+EXPORT_SYMBOL(prepare_creds);
+
+/*
+ * prepare new credentials for the usermode helper dispatcher
+ */
+struct cred *prepare_usermodehelper_creds(void)
+{
+#ifdef CONFIG_KEYS
+	struct thread_group_cred *tgcred = NULL;
+#endif
+	struct cred *new;
+
+#ifdef CONFIG_KEYS
+	tgcred = kzalloc(sizeof(*new->tgcred), GFP_ATOMIC);
+	if (!tgcred)
+		return NULL;
+#endif
+
+	new = kmem_cache_alloc(cred_jar, GFP_ATOMIC);
+	if (!new)
+		return NULL;
+
+	memcpy(new, &init_cred, sizeof(struct cred));
+
+	atomic_set(&new->usage, 1);
+	get_group_info(new->group_info);
+	get_uid(new->user);
+
+#ifdef CONFIG_KEYS
+	new->thread_keyring = NULL;
+	new->request_key_auth = NULL;
+	new->jit_keyring = KEY_REQKEY_DEFL_DEFAULT;
+
+	atomic_set(&tgcred->usage, 1);
+	spin_lock_init(&tgcred->lock);
+	new->tgcred = tgcred;
+#endif
+
+#ifdef CONFIG_SECURITY
+	new->security = NULL;
+#endif
+	if (security_prepare_creds(new, &init_cred, GFP_ATOMIC) < 0)
+		goto error;
+
+	BUG_ON(atomic_read(&new->usage) != 1);
+	return new;
+
+error:
+	put_cred(new);
+	return NULL;
+}
+
 /*
  * Copy credentials for the new process created by fork()
+ *
+ * We share if we can, but under some circumstances we have to generate a new
+ * set.
  */
 int copy_creds(struct task_struct *p, unsigned long clone_flags)
 {
-	struct cred *pcred;
-	int ret;
+#ifdef CONFIG_KEYS
+	struct thread_group_cred *tgcred;
+#endif
+	struct cred *new;
+
+	mutex_init(&p->cred_exec_mutex);
 
-	pcred = kmemdup(p->cred, sizeof(*p->cred), GFP_KERNEL);
-	if (!pcred)
+	if (
+#ifdef CONFIG_KEYS
+		!p->cred->thread_keyring &&
+#endif
+		clone_flags & CLONE_THREAD
+	    ) {
+		get_cred(p->cred);
+		atomic_inc(&p->cred->user->processes);
+		return 0;
+	}
+
+	new = prepare_creds();
+	if (!new)
 		return -ENOMEM;
 
 #ifdef CONFIG_KEYS
-	if (clone_flags & CLONE_THREAD) {
-		atomic_inc(&pcred->tgcred->usage);
-	} else {
-		pcred->tgcred = kmalloc(sizeof(struct cred), GFP_KERNEL);
-		if (!pcred->tgcred) {
-			kfree(pcred);
+	/* new threads get their own thread keyrings if their parent already
+	 * had one */
+	if (new->thread_keyring) {
+		key_put(new->thread_keyring);
+		new->thread_keyring = NULL;
+		if (clone_flags & CLONE_THREAD)
+			install_thread_keyring_to_cred(new);
+	}
+
+	/* we share the process and session keyrings between all the threads in
+	 * a process - this is slightly icky as we violate COW credentials a
+	 * bit */
+	if (!(clone_flags & CLONE_THREAD)) {
+		tgcred = kmalloc(sizeof(*tgcred), GFP_KERNEL);
+		if (!tgcred) {
+			put_cred(new);
 			return -ENOMEM;
 		}
-		atomic_set(&pcred->tgcred->usage, 1);
-		spin_lock_init(&pcred->tgcred->lock);
-		pcred->tgcred->process_keyring = NULL;
-		pcred->tgcred->session_keyring =
-			key_get(p->cred->tgcred->session_keyring);
+		atomic_set(&tgcred->usage, 1);
+		spin_lock_init(&tgcred->lock);
+		tgcred->process_keyring = NULL;
+		tgcred->session_keyring = key_get(new->tgcred->session_keyring);
+
+		release_tgcred(new);
+		new->tgcred = tgcred;
 	}
 #endif
 
-#ifdef CONFIG_SECURITY
-	pcred->security = NULL;
-#endif
+	atomic_inc(&new->user->processes);
+	p->cred = new;
+	return 0;
+}
 
-	ret = security_cred_alloc(pcred);
-	if (ret < 0) {
-		release_tgcred(pcred);
-		kfree(pcred);
-		return ret;
+/**
+ * commit_creds - Install new credentials upon the current task
+ * @new: The credentials to be assigned
+ *
+ * Install a new set of credentials to the current task, using RCU to replace
+ * the old set.
+ *
+ * This function eats the caller's reference to the new credentials.
+ *
+ * Always returns 0 thus allowing this function to be tail-called at the end
+ * of, say, sys_setgid().
+ */
+int commit_creds(struct cred *new)
+{
+	struct task_struct *task = current;
+	const struct cred *old;
+
+	BUG_ON(atomic_read(&new->usage) < 1);
+	BUG_ON(atomic_read(&task->cred->usage) < 1);
+
+	old = task->cred;
+	security_commit_creds(new, old);
+
+	/* dumpability changes */
+	if (old->euid != new->euid ||
+	    old->egid != new->egid ||
+	    old->fsuid != new->fsuid ||
+	    old->fsgid != new->fsgid ||
+	    !cap_issubset(new->cap_permitted, old->cap_permitted)) {
+		set_dumpable(task->mm, suid_dumpable);
+		task->pdeath_signal = 0;
+		smp_wmb();
 	}
 
-	atomic_set(&pcred->usage, 1);
-	get_group_info(pcred->group_info);
-	get_uid(pcred->user);
-	key_get(pcred->thread_keyring);
-	key_get(pcred->request_key_auth);
+	/* alter the thread keyring */
+	if (new->fsuid != old->fsuid)
+		key_fsuid_changed(task);
+	if (new->fsgid != old->fsgid)
+		key_fsgid_changed(task);
+
+	/* do it
+	 * - What if a process setreuid()'s and this brings the
+	 *   new uid over his NPROC rlimit?  We can check this now
+	 *   cheaply with the new uid cache, so if it matters
+	 *   we should be checking for it.  -DaveM
+	 */
+	if (new->user != old->user)
+		atomic_inc(&new->user->processes);
+	rcu_assign_pointer(task->cred, new);
+	if (new->user != old->user)
+		atomic_dec(&old->user->processes);
+
+	sched_switch_user(task);
+
+	/* send notifications */
+	if (new->uid   != old->uid  ||
+	    new->euid  != old->euid ||
+	    new->suid  != old->suid ||
+	    new->fsuid != old->fsuid)
+		proc_id_connector(task, PROC_EVENT_UID);
 
-	atomic_inc(&pcred->user->processes);
+	if (new->gid   != old->gid  ||
+	    new->egid  != old->egid ||
+	    new->sgid  != old->sgid ||
+	    new->fsgid != old->fsgid)
+		proc_id_connector(task, PROC_EVENT_GID);
 
-	/* RCU assignment is unneeded here as no-one can have accessed this
-	 * pointer yet, barring us */
-	p->cred = pcred;
+	put_cred(old);
 	return 0;
 }
+EXPORT_SYMBOL(commit_creds);
+
+/**
+ * abort_creds - Discard a set of credentials and unlock the current task
+ * @new: The credentials that were going to be applied
+ *
+ * Discard a set of credentials that were under construction and unlock the
+ * current task.
+ */
+void abort_creds(struct cred *new)
+{
+	BUG_ON(atomic_read(&new->usage) < 1);
+	put_cred(new);
+}
+EXPORT_SYMBOL(abort_creds);
+
+/**
+ * override_creds - Temporarily override the current process's credentials
+ * @new: The credentials to be assigned
+ *
+ * Install a set of temporary override credentials on the current process,
+ * returning the old set for later reversion.
+ */
+const struct cred *override_creds(const struct cred *new)
+{
+	const struct cred *old = current->cred;
+
+	rcu_assign_pointer(current->cred, get_cred(new));
+	return old;
+}
+EXPORT_SYMBOL(override_creds);
+
+/**
+ * revert_creds - Revert a temporary credentials override
+ * @old: The credentials to be restored
+ *
+ * Revert a temporary set of override credentials to an old set, discarding the
+ * override set.
+ */
+void revert_creds(const struct cred *old)
+{
+	const struct cred *override = current->cred;
+
+	rcu_assign_pointer(current->cred, old);
+	put_cred(override);
+}
+EXPORT_SYMBOL(revert_creds);
+
+/*
+ * initialise the credentials stuff
+ */
+void __init cred_init(void)
+{
+	/* allocate a slab in which we can store credentials */
+	cred_jar = kmem_cache_create("cred_jar", sizeof(struct cred),
+				     0, SLAB_HWCACHE_ALIGN|SLAB_PANIC, NULL);
+}
diff --git a/kernel/exit.c b/kernel/exit.c
index bbc22530f2c..c0711da1548 100644
--- a/kernel/exit.c
+++ b/kernel/exit.c
@@ -47,12 +47,14 @@
 #include <linux/blkdev.h>
 #include <linux/task_io_accounting_ops.h>
 #include <linux/tracehook.h>
+#include <linux/init_task.h>
 #include <trace/sched.h>
 
 #include <asm/uaccess.h>
 #include <asm/unistd.h>
 #include <asm/pgtable.h>
 #include <asm/mmu_context.h>
+#include "cred-internals.h"
 
 static void exit_mm(struct task_struct * tsk);
 
@@ -338,12 +340,12 @@ static void reparent_to_kthreadd(void)
 	/* cpus_allowed? */
 	/* rt_priority? */
 	/* signals? */
-	security_task_reparent_to_init(current);
 	memcpy(current->signal->rlim, init_task.signal->rlim,
 	       sizeof(current->signal->rlim));
-	atomic_inc(&(INIT_USER->__count));
+
+	atomic_inc(&init_cred.usage);
+	commit_creds(&init_cred);
 	write_unlock_irq(&tasklist_lock);
-	switch_uid(INIT_USER);
 }
 
 void __set_special_pids(struct pid *pid)
@@ -1085,7 +1087,6 @@ NORET_TYPE void do_exit(long code)
 	check_stack_usage();
 	exit_thread();
 	cgroup_exit(tsk, 1);
-	exit_keys(tsk);
 
 	if (group_dead && tsk->signal->leader)
 		disassociate_ctty(1);
diff --git a/kernel/fork.c b/kernel/fork.c
index ded1972672a..82a7948a664 100644
--- a/kernel/fork.c
+++ b/kernel/fork.c
@@ -1084,10 +1084,8 @@ static struct task_struct *copy_process(unsigned long clone_flags,
 		goto bad_fork_cleanup_sighand;
 	if ((retval = copy_mm(clone_flags, p)))
 		goto bad_fork_cleanup_signal;
-	if ((retval = copy_keys(clone_flags, p)))
-		goto bad_fork_cleanup_mm;
 	if ((retval = copy_namespaces(clone_flags, p)))
-		goto bad_fork_cleanup_keys;
+		goto bad_fork_cleanup_mm;
 	if ((retval = copy_io(clone_flags, p)))
 		goto bad_fork_cleanup_namespaces;
 	retval = copy_thread(0, clone_flags, stack_start, stack_size, p, regs);
@@ -1252,8 +1250,6 @@ bad_fork_cleanup_io:
 	put_io_context(p->io_context);
 bad_fork_cleanup_namespaces:
 	exit_task_namespaces(p);
-bad_fork_cleanup_keys:
-	exit_keys(p);
 bad_fork_cleanup_mm:
 	if (p->mm)
 		mmput(p->mm);
@@ -1281,6 +1277,7 @@ bad_fork_cleanup_cgroup:
 bad_fork_cleanup_put_domain:
 	module_put(task_thread_info(p)->exec_domain->module);
 bad_fork_cleanup_count:
+	atomic_dec(&p->cred->user->processes);
 	put_cred(p->cred);
 bad_fork_free:
 	free_task(p);
diff --git a/kernel/kmod.c b/kernel/kmod.c
index f044f8f5770..b46dbb90866 100644
--- a/kernel/kmod.c
+++ b/kernel/kmod.c
@@ -118,10 +118,10 @@ EXPORT_SYMBOL(request_module);
 struct subprocess_info {
 	struct work_struct work;
 	struct completion *complete;
+	struct cred *cred;
 	char *path;
 	char **argv;
 	char **envp;
-	struct key *ring;
 	enum umh_wait wait;
 	int retval;
 	struct file *stdin;
@@ -134,19 +134,20 @@ struct subprocess_info {
 static int ____call_usermodehelper(void *data)
 {
 	struct subprocess_info *sub_info = data;
-	struct key *new_session, *old_session;
 	int retval;
 
-	/* Unblock all signals and set the session keyring. */
-	new_session = key_get(sub_info->ring);
+	BUG_ON(atomic_read(&sub_info->cred->usage) != 1);
+
+	/* Unblock all signals */
 	spin_lock_irq(&current->sighand->siglock);
-	old_session = __install_session_keyring(new_session);
 	flush_signal_handlers(current, 1);
 	sigemptyset(&current->blocked);
 	recalc_sigpending();
 	spin_unlock_irq(&current->sighand->siglock);
 
-	key_put(old_session);
+	/* Install the credentials */
+	commit_creds(sub_info->cred);
+	sub_info->cred = NULL;
 
 	/* Install input pipe when needed */
 	if (sub_info->stdin) {
@@ -185,6 +186,8 @@ void call_usermodehelper_freeinfo(struct subprocess_info *info)
 {
 	if (info->cleanup)
 		(*info->cleanup)(info->argv, info->envp);
+	if (info->cred)
+		put_cred(info->cred);
 	kfree(info);
 }
 EXPORT_SYMBOL(call_usermodehelper_freeinfo);
@@ -240,6 +243,8 @@ static void __call_usermodehelper(struct work_struct *work)
 	pid_t pid;
 	enum umh_wait wait = sub_info->wait;
 
+	BUG_ON(atomic_read(&sub_info->cred->usage) != 1);
+
 	/* CLONE_VFORK: wait until the usermode helper has execve'd
 	 * successfully We need the data structures to stay around
 	 * until that is done.  */
@@ -362,6 +367,9 @@ struct subprocess_info *call_usermodehelper_setup(char *path, char **argv,
 	sub_info->path = path;
 	sub_info->argv = argv;
 	sub_info->envp = envp;
+	sub_info->cred = prepare_usermodehelper_creds();
+	if (!sub_info->cred)
+		return NULL;
 
   out:
 	return sub_info;
@@ -376,7 +384,13 @@ EXPORT_SYMBOL(call_usermodehelper_setup);
 void call_usermodehelper_setkeys(struct subprocess_info *info,
 				 struct key *session_keyring)
 {
-	info->ring = session_keyring;
+#ifdef CONFIG_KEYS
+	struct thread_group_cred *tgcred = info->cred->tgcred;
+	key_put(tgcred->session_keyring);
+	tgcred->session_keyring = key_get(session_keyring);
+#else
+	BUG();
+#endif
 }
 EXPORT_SYMBOL(call_usermodehelper_setkeys);
 
@@ -444,6 +458,8 @@ int call_usermodehelper_exec(struct subprocess_info *sub_info,
 	DECLARE_COMPLETION_ONSTACK(done);
 	int retval = 0;
 
+	BUG_ON(atomic_read(&sub_info->cred->usage) != 1);
+
 	helper_lock();
 	if (sub_info->path[0] == '\0')
 		goto out;
diff --git a/kernel/ptrace.c b/kernel/ptrace.c
index b9d5f4e4f6a..f764b880695 100644
--- a/kernel/ptrace.c
+++ b/kernel/ptrace.c
@@ -171,6 +171,14 @@ int ptrace_attach(struct task_struct *task)
 	if (same_thread_group(task, current))
 		goto out;
 
+	/* Protect exec's credential calculations against our interference;
+	 * SUID, SGID and LSM creds get determined differently under ptrace.
+	 */
+	retval = mutex_lock_interruptible(&current->cred_exec_mutex);
+	if (retval  < 0)
+		goto out;
+
+	retval = -EPERM;
 repeat:
 	/*
 	 * Nasty, nasty.
@@ -210,6 +218,7 @@ repeat:
 bad:
 	write_unlock_irqrestore(&tasklist_lock, flags);
 	task_unlock(task);
+	mutex_unlock(&current->cred_exec_mutex);
 out:
 	return retval;
 }
diff --git a/kernel/signal.c b/kernel/signal.c
index 84989124baf..2a64304ed54 100644
--- a/kernel/signal.c
+++ b/kernel/signal.c
@@ -180,7 +180,7 @@ int next_signal(struct sigpending *pending, sigset_t *mask)
 /*
  * allocate a new signal queue record
  * - this may be called without locks if and only if t == current, otherwise an
- *   appopriate lock must be held to protect t's user_struct
+ *   appopriate lock must be held to stop the target task from exiting
  */
 static struct sigqueue *__sigqueue_alloc(struct task_struct *t, gfp_t flags,
 					 int override_rlimit)
@@ -194,7 +194,7 @@ static struct sigqueue *__sigqueue_alloc(struct task_struct *t, gfp_t flags,
 	 * caller must be holding the RCU readlock (by way of a spinlock) and
 	 * we use RCU protection here
 	 */
-	user = __task_cred(t)->user;
+	user = get_uid(__task_cred(t)->user);
 	atomic_inc(&user->sigpending);
 	if (override_rlimit ||
 	    atomic_read(&user->sigpending) <=
@@ -202,12 +202,14 @@ static struct sigqueue *__sigqueue_alloc(struct task_struct *t, gfp_t flags,
 		q = kmem_cache_alloc(sigqueue_cachep, flags);
 	if (unlikely(q == NULL)) {
 		atomic_dec(&user->sigpending);
+		free_uid(user);
 	} else {
 		INIT_LIST_HEAD(&q->list);
 		q->flags = 0;
-		q->user = get_uid(user);
+		q->user = user;
 	}
-	return(q);
+
+	return q;
 }
 
 static void __sigqueue_free(struct sigqueue *q)
diff --git a/kernel/sys.c b/kernel/sys.c
index ccc9eb736d3..ab735040468 100644
--- a/kernel/sys.c
+++ b/kernel/sys.c
@@ -180,7 +180,7 @@ asmlinkage long sys_setpriority(int which, int who, int niceval)
 			} while_each_pid_thread(pgrp, PIDTYPE_PGID, p);
 			break;
 		case PRIO_USER:
-			user = cred->user;
+			user = (struct user_struct *) cred->user;
 			if (!who)
 				who = cred->uid;
 			else if ((who != cred->uid) &&
@@ -479,47 +479,48 @@ void ctrl_alt_del(void)
  */
 asmlinkage long sys_setregid(gid_t rgid, gid_t egid)
 {
-	struct cred *cred = current->cred;
-	int old_rgid = cred->gid;
-	int old_egid = cred->egid;
-	int new_rgid = old_rgid;
-	int new_egid = old_egid;
+	const struct cred *old;
+	struct cred *new;
 	int retval;
 
+	new = prepare_creds();
+	if (!new)
+		return -ENOMEM;
+	old = current_cred();
+
 	retval = security_task_setgid(rgid, egid, (gid_t)-1, LSM_SETID_RE);
 	if (retval)
-		return retval;
+		goto error;
 
+	retval = -EPERM;
 	if (rgid != (gid_t) -1) {
-		if ((old_rgid == rgid) ||
-		    (cred->egid == rgid) ||
+		if (old->gid == rgid ||
+		    old->egid == rgid ||
 		    capable(CAP_SETGID))
-			new_rgid = rgid;
+			new->gid = rgid;
 		else
-			return -EPERM;
+			goto error;
 	}
 	if (egid != (gid_t) -1) {
-		if ((old_rgid == egid) ||
-		    (cred->egid == egid) ||
-		    (cred->sgid == egid) ||
+		if (old->gid == egid ||
+		    old->egid == egid ||
+		    old->sgid == egid ||
 		    capable(CAP_SETGID))
-			new_egid = egid;
+			new->egid = egid;
 		else
-			return -EPERM;
-	}
-	if (new_egid != old_egid) {
-		set_dumpable(current->mm, suid_dumpable);
-		smp_wmb();
+			goto error;
 	}
+
 	if (rgid != (gid_t) -1 ||
-	    (egid != (gid_t) -1 && egid != old_rgid))
-		cred->sgid = new_egid;
-	cred->fsgid = new_egid;
-	cred->egid = new_egid;
-	cred->gid = new_rgid;
-	key_fsgid_changed(current);
-	proc_id_connector(current, PROC_EVENT_GID);
-	return 0;
+	    (egid != (gid_t) -1 && egid != old->gid))
+		new->sgid = new->egid;
+	new->fsgid = new->egid;
+
+	return commit_creds(new);
+
+error:
+	abort_creds(new);
+	return retval;
 }
 
 /*
@@ -529,40 +530,42 @@ asmlinkage long sys_setregid(gid_t rgid, gid_t egid)
  */
 asmlinkage long sys_setgid(gid_t gid)
 {
-	struct cred *cred = current->cred;
-	int old_egid = cred->egid;
+	const struct cred *old;
+	struct cred *new;
 	int retval;
 
+	new = prepare_creds();
+	if (!new)
+		return -ENOMEM;
+	old = current_cred();
+
 	retval = security_task_setgid(gid, (gid_t)-1, (gid_t)-1, LSM_SETID_ID);
 	if (retval)
-		return retval;
+		goto error;
 
-	if (capable(CAP_SETGID)) {
-		if (old_egid != gid) {
-			set_dumpable(current->mm, suid_dumpable);
-			smp_wmb();
-		}
-		cred->gid = cred->egid = cred->sgid = cred->fsgid = gid;
-	} else if ((gid == cred->gid) || (gid == cred->sgid)) {
-		if (old_egid != gid) {
-			set_dumpable(current->mm, suid_dumpable);
-			smp_wmb();
-		}
-		cred->egid = cred->fsgid = gid;
-	}
+	retval = -EPERM;
+	if (capable(CAP_SETGID))
+		new->gid = new->egid = new->sgid = new->fsgid = gid;
+	else if (gid == old->gid || gid == old->sgid)
+		new->egid = new->fsgid = gid;
 	else
-		return -EPERM;
+		goto error;
 
-	key_fsgid_changed(current);
-	proc_id_connector(current, PROC_EVENT_GID);
-	return 0;
+	return commit_creds(new);
+
+error:
+	abort_creds(new);
+	return retval;
 }
   
-static int set_user(uid_t new_ruid, int dumpclear)
+/*
+ * change the user struct in a credentials set to match the new UID
+ */
+static int set_user(struct cred *new)
 {
 	struct user_struct *new_user;
 
-	new_user = alloc_uid(current->nsproxy->user_ns, new_ruid);
+	new_user = alloc_uid(current->nsproxy->user_ns, new->uid);
 	if (!new_user)
 		return -EAGAIN;
 
@@ -573,13 +576,8 @@ static int set_user(uid_t new_ruid, int dumpclear)
 		return -EAGAIN;
 	}
 
-	switch_uid(new_user);
-
-	if (dumpclear) {
-		set_dumpable(current->mm, suid_dumpable);
-		smp_wmb();
-	}
-	current->cred->uid = new_ruid;
+	free_uid(new->user);
+	new->user = new_user;
 	return 0;
 }
 
@@ -600,55 +598,56 @@ static int set_user(uid_t new_ruid, int dumpclear)
  */
 asmlinkage long sys_setreuid(uid_t ruid, uid_t euid)
 {
-	struct cred *cred = current->cred;
-	int old_ruid, old_euid, old_suid, new_ruid, new_euid;
+	const struct cred *old;
+	struct cred *new;
 	int retval;
 
+	new = prepare_creds();
+	if (!new)
+		return -ENOMEM;
+	old = current_cred();
+
 	retval = security_task_setuid(ruid, euid, (uid_t)-1, LSM_SETID_RE);
 	if (retval)
-		return retval;
-
-	new_ruid = old_ruid = cred->uid;
-	new_euid = old_euid = cred->euid;
-	old_suid = cred->suid;
+		goto error;
 
+	retval = -EPERM;
 	if (ruid != (uid_t) -1) {
-		new_ruid = ruid;
-		if ((old_ruid != ruid) &&
-		    (cred->euid != ruid) &&
+		new->uid = ruid;
+		if (old->uid != ruid &&
+		    old->euid != ruid &&
 		    !capable(CAP_SETUID))
-			return -EPERM;
+			goto error;
 	}
 
 	if (euid != (uid_t) -1) {
-		new_euid = euid;
-		if ((old_ruid != euid) &&
-		    (cred->euid != euid) &&
-		    (cred->suid != euid) &&
+		new->euid = euid;
+		if (old->uid != euid &&
+		    old->euid != euid &&
+		    old->suid != euid &&
 		    !capable(CAP_SETUID))
-			return -EPERM;
+			goto error;
 	}
 
-	if (new_ruid != old_ruid && set_user(new_ruid, new_euid != old_euid) < 0)
-		return -EAGAIN;
+	retval = -EAGAIN;
+	if (new->uid != old->uid && set_user(new) < 0)
+		goto error;
 
-	if (new_euid != old_euid) {
-		set_dumpable(current->mm, suid_dumpable);
-		smp_wmb();
-	}
-	cred->fsuid = cred->euid = new_euid;
 	if (ruid != (uid_t) -1 ||
-	    (euid != (uid_t) -1 && euid != old_ruid))
-		cred->suid = cred->euid;
-	cred->fsuid = cred->euid;
-
-	key_fsuid_changed(current);
-	proc_id_connector(current, PROC_EVENT_UID);
+	    (euid != (uid_t) -1 && euid != old->uid))
+		new->suid = new->euid;
+	new->fsuid = new->euid;
 
-	return security_task_post_setuid(old_ruid, old_euid, old_suid, LSM_SETID_RE);
-}
+	retval = security_task_fix_setuid(new, old, LSM_SETID_RE);
+	if (retval < 0)
+		goto error;
 
+	return commit_creds(new);
 
+error:
+	abort_creds(new);
+	return retval;
+}
 		
 /*
  * setuid() is implemented like SysV with SAVED_IDS 
@@ -663,37 +662,41 @@ asmlinkage long sys_setreuid(uid_t ruid, uid_t euid)
  */
 asmlinkage long sys_setuid(uid_t uid)
 {
-	struct cred *cred = current->cred;
-	int old_euid = cred->euid;
-	int old_ruid, old_suid, new_suid;
+	const struct cred *old;
+	struct cred *new;
 	int retval;
 
+	new = prepare_creds();
+	if (!new)
+		return -ENOMEM;
+	old = current_cred();
+
 	retval = security_task_setuid(uid, (uid_t)-1, (uid_t)-1, LSM_SETID_ID);
 	if (retval)
-		return retval;
+		goto error;
 
-	old_ruid = cred->uid;
-	old_suid = cred->suid;
-	new_suid = old_suid;
-	
+	retval = -EPERM;
 	if (capable(CAP_SETUID)) {
-		if (uid != old_ruid && set_user(uid, old_euid != uid) < 0)
-			return -EAGAIN;
-		new_suid = uid;
-	} else if ((uid != cred->uid) && (uid != new_suid))
-		return -EPERM;
-
-	if (old_euid != uid) {
-		set_dumpable(current->mm, suid_dumpable);
-		smp_wmb();
+		new->suid = new->uid = uid;
+		if (uid != old->uid && set_user(new) < 0) {
+			retval = -EAGAIN;
+			goto error;
+		}
+	} else if (uid != old->uid && uid != new->suid) {
+		goto error;
 	}
-	cred->fsuid = cred->euid = uid;
-	cred->suid = new_suid;
 
-	key_fsuid_changed(current);
-	proc_id_connector(current, PROC_EVENT_UID);
+	new->fsuid = new->euid = uid;
+
+	retval = security_task_fix_setuid(new, old, LSM_SETID_ID);
+	if (retval < 0)
+		goto error;
+
+	return commit_creds(new);
 
-	return security_task_post_setuid(old_ruid, old_euid, old_suid, LSM_SETID_ID);
+error:
+	abort_creds(new);
+	return retval;
 }
 
 
@@ -703,47 +706,53 @@ asmlinkage long sys_setuid(uid_t uid)
  */
 asmlinkage long sys_setresuid(uid_t ruid, uid_t euid, uid_t suid)
 {
-	struct cred *cred = current->cred;
-	int old_ruid = cred->uid;
-	int old_euid = cred->euid;
-	int old_suid = cred->suid;
+	const struct cred *old;
+	struct cred *new;
 	int retval;
 
+	new = prepare_creds();
+	if (!new)
+		return -ENOMEM;
+
 	retval = security_task_setuid(ruid, euid, suid, LSM_SETID_RES);
 	if (retval)
-		return retval;
+		goto error;
+	old = current_cred();
 
+	retval = -EPERM;
 	if (!capable(CAP_SETUID)) {
-		if ((ruid != (uid_t) -1) && (ruid != cred->uid) &&
-		    (ruid != cred->euid) && (ruid != cred->suid))
-			return -EPERM;
-		if ((euid != (uid_t) -1) && (euid != cred->uid) &&
-		    (euid != cred->euid) && (euid != cred->suid))
-			return -EPERM;
-		if ((suid != (uid_t) -1) && (suid != cred->uid) &&
-		    (suid != cred->euid) && (suid != cred->suid))
-			return -EPERM;
+		if (ruid != (uid_t) -1 && ruid != old->uid &&
+		    ruid != old->euid  && ruid != old->suid)
+			goto error;
+		if (euid != (uid_t) -1 && euid != old->uid &&
+		    euid != old->euid  && euid != old->suid)
+			goto error;
+		if (suid != (uid_t) -1 && suid != old->uid &&
+		    suid != old->euid  && suid != old->suid)
+			goto error;
 	}
+
+	retval = -EAGAIN;
 	if (ruid != (uid_t) -1) {
-		if (ruid != cred->uid &&
-		    set_user(ruid, euid != cred->euid) < 0)
-			return -EAGAIN;
+		new->uid = ruid;
+		if (ruid != old->uid && set_user(new) < 0)
+			goto error;
 	}
-	if (euid != (uid_t) -1) {
-		if (euid != cred->euid) {
-			set_dumpable(current->mm, suid_dumpable);
-			smp_wmb();
-		}
-		cred->euid = euid;
-	}
-	cred->fsuid = cred->euid;
+	if (euid != (uid_t) -1)
+		new->euid = euid;
 	if (suid != (uid_t) -1)
-		cred->suid = suid;
+		new->suid = suid;
+	new->fsuid = new->euid;
 
-	key_fsuid_changed(current);
-	proc_id_connector(current, PROC_EVENT_UID);
+	retval = security_task_fix_setuid(new, old, LSM_SETID_RES);
+	if (retval < 0)
+		goto error;
 
-	return security_task_post_setuid(old_ruid, old_euid, old_suid, LSM_SETID_RES);
+	return commit_creds(new);
+
+error:
+	abort_creds(new);
+	return retval;
 }
 
 asmlinkage long sys_getresuid(uid_t __user *ruid, uid_t __user *euid, uid_t __user *suid)
@@ -763,40 +772,45 @@ asmlinkage long sys_getresuid(uid_t __user *ruid, uid_t __user *euid, uid_t __us
  */
 asmlinkage long sys_setresgid(gid_t rgid, gid_t egid, gid_t sgid)
 {
-	struct cred *cred = current->cred;
+	const struct cred *old;
+	struct cred *new;
 	int retval;
 
+	new = prepare_creds();
+	if (!new)
+		return -ENOMEM;
+	old = current_cred();
+
 	retval = security_task_setgid(rgid, egid, sgid, LSM_SETID_RES);
 	if (retval)
-		return retval;
+		goto error;
 
+	retval = -EPERM;
 	if (!capable(CAP_SETGID)) {
-		if ((rgid != (gid_t) -1) && (rgid != cred->gid) &&
-		    (rgid != cred->egid) && (rgid != cred->sgid))
-			return -EPERM;
-		if ((egid != (gid_t) -1) && (egid != cred->gid) &&
-		    (egid != cred->egid) && (egid != cred->sgid))
-			return -EPERM;
-		if ((sgid != (gid_t) -1) && (sgid != cred->gid) &&
-		    (sgid != cred->egid) && (sgid != cred->sgid))
-			return -EPERM;
+		if (rgid != (gid_t) -1 && rgid != old->gid &&
+		    rgid != old->egid  && rgid != old->sgid)
+			goto error;
+		if (egid != (gid_t) -1 && egid != old->gid &&
+		    egid != old->egid  && egid != old->sgid)
+			goto error;
+		if (sgid != (gid_t) -1 && sgid != old->gid &&
+		    sgid != old->egid  && sgid != old->sgid)
+			goto error;
 	}
-	if (egid != (gid_t) -1) {
-		if (egid != cred->egid) {
-			set_dumpable(current->mm, suid_dumpable);
-			smp_wmb();
-		}
-		cred->egid = egid;
-	}
-	cred->fsgid = cred->egid;
+
 	if (rgid != (gid_t) -1)
-		cred->gid = rgid;
+		new->gid = rgid;
+	if (egid != (gid_t) -1)
+		new->egid = egid;
 	if (sgid != (gid_t) -1)
-		cred->sgid = sgid;
+		new->sgid = sgid;
+	new->fsgid = new->egid;
 
-	key_fsgid_changed(current);
-	proc_id_connector(current, PROC_EVENT_GID);
-	return 0;
+	return commit_creds(new);
+
+error:
+	abort_creds(new);
+	return retval;
 }
 
 asmlinkage long sys_getresgid(gid_t __user *rgid, gid_t __user *egid, gid_t __user *sgid)
@@ -820,28 +834,35 @@ asmlinkage long sys_getresgid(gid_t __user *rgid, gid_t __user *egid, gid_t __us
  */
 asmlinkage long sys_setfsuid(uid_t uid)
 {
-	struct cred *cred = current->cred;
-	int old_fsuid;
+	const struct cred *old;
+	struct cred *new;
+	uid_t old_fsuid;
+
+	new = prepare_creds();
+	if (!new)
+		return current_fsuid();
+	old = current_cred();
+	old_fsuid = old->fsuid;
 
-	old_fsuid = cred->fsuid;
-	if (security_task_setuid(uid, (uid_t)-1, (uid_t)-1, LSM_SETID_FS))
-		return old_fsuid;
+	if (security_task_setuid(uid, (uid_t)-1, (uid_t)-1, LSM_SETID_FS) < 0)
+		goto error;
 
-	if (uid == cred->uid || uid == cred->euid ||
-	    uid == cred->suid || uid == cred->fsuid ||
+	if (uid == old->uid  || uid == old->euid  ||
+	    uid == old->suid || uid == old->fsuid ||
 	    capable(CAP_SETUID)) {
 		if (uid != old_fsuid) {
-			set_dumpable(current->mm, suid_dumpable);
-			smp_wmb();
+			new->fsuid = uid;
+			if (security_task_fix_setuid(new, old, LSM_SETID_FS) == 0)
+				goto change_okay;
 		}
-		cred->fsuid = uid;
 	}
 
-	key_fsuid_changed(current);
-	proc_id_connector(current, PROC_EVENT_UID);
-
-	security_task_post_setuid(old_fsuid, (uid_t)-1, (uid_t)-1, LSM_SETID_FS);
+error:
+	abort_creds(new);
+	return old_fsuid;
 
+change_okay:
+	commit_creds(new);
 	return old_fsuid;
 }
 
@@ -850,24 +871,34 @@ asmlinkage long sys_setfsuid(uid_t uid)
  */
 asmlinkage long sys_setfsgid(gid_t gid)
 {
-	struct cred *cred = current->cred;
-	int old_fsgid;
+	const struct cred *old;
+	struct cred *new;
+	gid_t old_fsgid;
+
+	new = prepare_creds();
+	if (!new)
+		return current_fsgid();
+	old = current_cred();
+	old_fsgid = old->fsgid;
 
-	old_fsgid = cred->fsgid;
 	if (security_task_setgid(gid, (gid_t)-1, (gid_t)-1, LSM_SETID_FS))
-		return old_fsgid;
+		goto error;
 
-	if (gid == cred->gid || gid == cred->egid ||
-	    gid == cred->sgid || gid == cred->fsgid ||
+	if (gid == old->gid  || gid == old->egid  ||
+	    gid == old->sgid || gid == old->fsgid ||
 	    capable(CAP_SETGID)) {
 		if (gid != old_fsgid) {
-			set_dumpable(current->mm, suid_dumpable);
-			smp_wmb();
+			new->fsgid = gid;
+			goto change_okay;
 		}
-		cred->fsgid = gid;
-		key_fsgid_changed(current);
-		proc_id_connector(current, PROC_EVENT_GID);
 	}
+
+error:
+	abort_creds(new);
+	return old_fsgid;
+
+change_okay:
+	commit_creds(new);
 	return old_fsgid;
 }
 
@@ -1136,7 +1167,7 @@ EXPORT_SYMBOL(groups_free);
 
 /* export the group_info to a user-space array */
 static int groups_to_user(gid_t __user *grouplist,
-    struct group_info *group_info)
+			  const struct group_info *group_info)
 {
 	int i;
 	unsigned int count = group_info->ngroups;
@@ -1227,31 +1258,25 @@ int groups_search(const struct group_info *group_info, gid_t grp)
 }
 
 /**
- * set_groups - Change a group subscription in a security record
- * @sec: The security record to alter
- * @group_info: The group list to impose
+ * set_groups - Change a group subscription in a set of credentials
+ * @new: The newly prepared set of credentials to alter
+ * @group_info: The group list to install
  *
- * Validate a group subscription and, if valid, impose it upon a task security
- * record.
+ * Validate a group subscription and, if valid, insert it into a set
+ * of credentials.
  */
-int set_groups(struct cred *cred, struct group_info *group_info)
+int set_groups(struct cred *new, struct group_info *group_info)
 {
 	int retval;
-	struct group_info *old_info;
 
 	retval = security_task_setgroups(group_info);
 	if (retval)
 		return retval;
 
+	put_group_info(new->group_info);
 	groups_sort(group_info);
 	get_group_info(group_info);
-
-	spin_lock(&cred->lock);
-	old_info = cred->group_info;
-	cred->group_info = group_info;
-	spin_unlock(&cred->lock);
-
-	put_group_info(old_info);
+	new->group_info = group_info;
 	return 0;
 }
 
@@ -1266,7 +1291,20 @@ EXPORT_SYMBOL(set_groups);
  */
 int set_current_groups(struct group_info *group_info)
 {
-	return set_groups(current->cred, group_info);
+	struct cred *new;
+	int ret;
+
+	new = prepare_creds();
+	if (!new)
+		return -ENOMEM;
+
+	ret = set_groups(new, group_info);
+	if (ret < 0) {
+		abort_creds(new);
+		return ret;
+	}
+
+	return commit_creds(new);
 }
 
 EXPORT_SYMBOL(set_current_groups);
@@ -1666,9 +1704,11 @@ asmlinkage long sys_prctl(int option, unsigned long arg2, unsigned long arg3,
 	unsigned char comm[sizeof(me->comm)];
 	long error;
 
-	if (security_task_prctl(option, arg2, arg3, arg4, arg5, &error))
+	error = security_task_prctl(option, arg2, arg3, arg4, arg5);
+	if (error != -ENOSYS)
 		return error;
 
+	error = 0;
 	switch (option) {
 		case PR_SET_PDEATHSIG:
 			if (!valid_signal(arg2)) {
diff --git a/kernel/user.c b/kernel/user.c
index 104d22ac84d..d476307dd4b 100644
--- a/kernel/user.c
+++ b/kernel/user.c
@@ -16,6 +16,7 @@
 #include <linux/interrupt.h>
 #include <linux/module.h>
 #include <linux/user_namespace.h>
+#include "cred-internals.h"
 
 struct user_namespace init_user_ns = {
 	.kref = {
@@ -104,16 +105,10 @@ static int sched_create_user(struct user_struct *up)
 	return rc;
 }
 
-static void sched_switch_user(struct task_struct *p)
-{
-	sched_move_task(p);
-}
-
 #else	/* CONFIG_USER_SCHED */
 
 static void sched_destroy_user(struct user_struct *up) { }
 static int sched_create_user(struct user_struct *up) { return 0; }
-static void sched_switch_user(struct task_struct *p) { }
 
 #endif	/* CONFIG_USER_SCHED */
 
@@ -448,36 +443,6 @@ out_unlock:
 	return NULL;
 }
 
-void switch_uid(struct user_struct *new_user)
-{
-	struct user_struct *old_user;
-
-	/* What if a process setreuid()'s and this brings the
-	 * new uid over his NPROC rlimit?  We can check this now
-	 * cheaply with the new uid cache, so if it matters
-	 * we should be checking for it.  -DaveM
-	 */
-	old_user = current->cred->user;
-	atomic_inc(&new_user->processes);
-	atomic_dec(&old_user->processes);
-	switch_uid_keyring(new_user);
-	current->cred->user = new_user;
-	sched_switch_user(current);
-
-	/*
-	 * We need to synchronize with __sigqueue_alloc()
-	 * doing a get_uid(p->user).. If that saw the old
-	 * user value, we need to wait until it has exited
-	 * its critical region before we can free the old
-	 * structure.
-	 */
-	smp_mb();
-	spin_unlock_wait(&current->sighand->siglock);
-
-	free_uid(old_user);
-	suid_keys(current);
-}
-
 #ifdef CONFIG_USER_NS
 void release_uids(struct user_namespace *ns)
 {
diff --git a/kernel/user_namespace.c b/kernel/user_namespace.c
index f82730adea0..0d9c51d6733 100644
--- a/kernel/user_namespace.c
+++ b/kernel/user_namespace.c
@@ -19,6 +19,7 @@ static struct user_namespace *clone_user_ns(struct user_namespace *old_ns)
 {
 	struct user_namespace *ns;
 	struct user_struct *new_user;
+	struct cred *new;
 	int n;
 
 	ns = kmalloc(sizeof(struct user_namespace), GFP_KERNEL);
@@ -45,7 +46,16 @@ static struct user_namespace *clone_user_ns(struct user_namespace *old_ns)
 		return ERR_PTR(-ENOMEM);
 	}
 
-	switch_uid(new_user);
+	/* Install the new user */
+	new = prepare_creds();
+	if (!new) {
+		free_uid(new_user);
+		free_uid(ns->root_user);
+		kfree(ns);
+	}
+	free_uid(new->user);
+	new->user = new_user;
+	commit_creds(new);
 	return ns;
 }