Merge git://git.kernel.org/pub/scm/linux/kernel/git/mingo/linux-2.6-sched

* git://git.kernel.org/pub/scm/linux/kernel/git/mingo/linux-2.6-sched: (140 commits)
  sched: sync wakeups preempt too
  sched: affine sync wakeups
  sched: guest CPU accounting: maintain guest state in KVM
  sched: guest CPU accounting: maintain stats in account_system_time()
  sched: guest CPU accounting: add guest-CPU /proc/<pid>/stat fields
  sched: guest CPU accounting: add guest-CPU /proc/stat field
  sched: domain sysctl fixes: add terminator comment
  sched: domain sysctl fixes: do not crash on allocation failure
  sched: domain sysctl fixes: unregister the sysctl table before domains
  sched: domain sysctl fixes: use for_each_online_cpu()
  sched: domain sysctl fixes: use kcalloc()
  Make scheduler debug file operations const
  sched: enable wake-idle on CONFIG_SCHED_MC=y
  sched: reintroduce topology.h tunings
  sched: allow the immediate migration of cache-cold tasks
  sched: debug, improve migration statistics
  sched: debug: increase width of debug line
  sched: activate task_hot() only on fair-scheduled tasks
  sched: reintroduce cache-hot affinity
  sched: speed up context-switches a bit
  ...

1 files changed, 238 insertions, 11 deletions

diff --git a/kernel/user.c b/kernel/user.c
index 9ca2848fc356..f0e561e6d085 100644
--- a/kernel/user.c
+++ b/kernel/user.c
@@ -50,12 +50,16 @@ struct user_struct root_user = {
         .uid_keyring    = &root_user_keyring,
         .session_keyring = &root_session_keyring,
 #endif
+#ifdef CONFIG_FAIR_USER_SCHED
+        .tg             = &init_task_group,
+#endif
 };
 
 /*
  * These routines must be called with the uidhash spinlock held!
  */
-static inline void uid_hash_insert(struct user_struct *up, struct hlist_head *hashent)
+static inline void uid_hash_insert(struct user_struct *up,
+                                                struct hlist_head *hashent)
 {
         hlist_add_head(&up->uidhash_node, hashent);
 }
@@ -65,13 +69,14 @@ static inline void uid_hash_remove(struct user_struct *up)
         hlist_del_init(&up->uidhash_node);
 }
 
-static inline struct user_struct *uid_hash_find(uid_t uid, struct hlist_head *hashent)
+static inline struct user_struct *uid_hash_find(uid_t uid,
+                                                struct hlist_head *hashent)
 {
         struct user_struct *user;
         struct hlist_node *h;
 
         hlist_for_each_entry(user, h, hashent, uidhash_node) {
-                if(user->uid == uid) {
+                if (user->uid == uid) {
                         atomic_inc(&user->__count);
                         return user;
                 }
@@ -80,6 +85,203 @@ static inline struct user_struct *uid_hash_find(uid_t uid, struct hlist_head *ha
         return NULL;
 }
 
+#ifdef CONFIG_FAIR_USER_SCHED
+
+static struct kobject uids_kobject; /* represents /sys/kernel/uids directory */
+static DEFINE_MUTEX(uids_mutex);
+
+static void sched_destroy_user(struct user_struct *up)
+{
+        sched_destroy_group(up->tg);
+}
+
+static int sched_create_user(struct user_struct *up)
+{
+        int rc = 0;
+
+        up->tg = sched_create_group();
+        if (IS_ERR(up->tg))
+                rc = -ENOMEM;
+
+        return rc;
+}
+
+static void sched_switch_user(struct task_struct *p)
+{
+        sched_move_task(p);
+}
+
+static inline void uids_mutex_lock(void)
+{
+        mutex_lock(&uids_mutex);
+}
+
+static inline void uids_mutex_unlock(void)
+{
+        mutex_unlock(&uids_mutex);
+}
+
+/* return cpu shares held by the user */
+ssize_t cpu_shares_show(struct kset *kset, char *buffer)
+{
+        struct user_struct *up = container_of(kset, struct user_struct, kset);
+
+        return sprintf(buffer, "%lu\n", sched_group_shares(up->tg));
+}
+
+/* modify cpu shares held by the user */
+ssize_t cpu_shares_store(struct kset *kset, const char *buffer, size_t size)
+{
+        struct user_struct *up = container_of(kset, struct user_struct, kset);
+        unsigned long shares;
+        int rc;
+
+        sscanf(buffer, "%lu", &shares);
+
+        rc = sched_group_set_shares(up->tg, shares);
+
+        return (rc ? rc : size);
+}
+
+static void user_attr_init(struct subsys_attribute *sa, char *name, int mode)
+{
+        sa->attr.name = name;
+        sa->attr.mode = mode;
+        sa->show = cpu_shares_show;
+        sa->store = cpu_shares_store;
+}
+
+/* Create "/sys/kernel/uids/<uid>" directory and
+ *  "/sys/kernel/uids/<uid>/cpu_share" file for this user.
+ */
+static int user_kobject_create(struct user_struct *up)
+{
+        struct kset *kset = &up->kset;
+        struct kobject *kobj = &kset->kobj;
+        int error;
+
+        memset(kset, 0, sizeof(struct kset));
+        kobj->parent = &uids_kobject;   /* create under /sys/kernel/uids dir */
+        kobject_set_name(kobj, "%d", up->uid);
+        kset_init(kset);
+        user_attr_init(&up->user_attr, "cpu_share", 0644);
+
+        error = kobject_add(kobj);
+        if (error)
+                goto done;
+
+        error = sysfs_create_file(kobj, &up->user_attr.attr);
+        if (error)
+                kobject_del(kobj);
+
+        kobject_uevent(kobj, KOBJ_ADD);
+
+done:
+        return error;
+}
+
+/* create these in sysfs filesystem:
+ *      "/sys/kernel/uids" directory
+ *      "/sys/kernel/uids/0" directory (for root user)
+ *      "/sys/kernel/uids/0/cpu_share" file (for root user)
+ */
+int __init uids_kobject_init(void)
+{
+        int error;
+
+        /* create under /sys/kernel dir */
+        uids_kobject.parent = &kernel_subsys.kobj;
+        uids_kobject.kset = &kernel_subsys;
+        kobject_set_name(&uids_kobject, "uids");
+        kobject_init(&uids_kobject);
+
+        error = kobject_add(&uids_kobject);
+        if (!error)
+                error = user_kobject_create(&root_user);
+
+        return error;
+}
+
+/* work function to remove sysfs directory for a user and free up
+ * corresponding structures.
+ */
+static void remove_user_sysfs_dir(struct work_struct *w)
+{
+        struct user_struct *up = container_of(w, struct user_struct, work);
+        struct kobject *kobj = &up->kset.kobj;
+        unsigned long flags;
+        int remove_user = 0;
+
+        /* Make uid_hash_remove() + sysfs_remove_file() + kobject_del()
+         * atomic.
+         */
+        uids_mutex_lock();
+
+        local_irq_save(flags);
+
+        if (atomic_dec_and_lock(&up->__count, &uidhash_lock)) {
+                uid_hash_remove(up);
+                remove_user = 1;
+                spin_unlock_irqrestore(&uidhash_lock, flags);
+        } else {
+                local_irq_restore(flags);
+        }
+
+        if (!remove_user)
+                goto done;
+
+        sysfs_remove_file(kobj, &up->user_attr.attr);
+        kobject_uevent(kobj, KOBJ_REMOVE);
+        kobject_del(kobj);
+
+        sched_destroy_user(up);
+        key_put(up->uid_keyring);
+        key_put(up->session_keyring);
+        kmem_cache_free(uid_cachep, up);
+
+done:
+        uids_mutex_unlock();
+}
+
+/* IRQs are disabled and uidhash_lock is held upon function entry.
+ * 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)
+{
+        /* restore back the count */
+        atomic_inc(&up->__count);
+        spin_unlock_irqrestore(&uidhash_lock, flags);
+
+        INIT_WORK(&up->work, remove_user_sysfs_dir);
+        schedule_work(&up->work);
+}
+
+#else   /* CONFIG_FAIR_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) { }
+static inline int user_kobject_create(struct user_struct *up) { return 0; }
+static inline void uids_mutex_lock(void) { }
+static inline void uids_mutex_unlock(void) { }
+
+/* IRQs are disabled and uidhash_lock is held upon function entry.
+ * 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)
+{
+        uid_hash_remove(up);
+        spin_unlock_irqrestore(&uidhash_lock, flags);
+        sched_destroy_user(up);
+        key_put(up->uid_keyring);
+        key_put(up->session_keyring);
+        kmem_cache_free(uid_cachep, up);
+}
+
+#endif  /* CONFIG_FAIR_USER_SCHED */
+
 /*
  * Locate the user_struct for the passed UID.  If found, take a ref on it.  The
  * caller must undo that ref with free_uid().
@@ -106,15 +308,10 @@ void free_uid(struct user_struct *up)
                 return;
 
         local_irq_save(flags);
-        if (atomic_dec_and_lock(&up->__count, &uidhash_lock)) {
-                uid_hash_remove(up);
-                spin_unlock_irqrestore(&uidhash_lock, flags);
-                key_put(up->uid_keyring);
-                key_put(up->session_keyring);
-                kmem_cache_free(uid_cachep, up);
-        } else {
+        if (atomic_dec_and_lock(&up->__count, &uidhash_lock))
+                free_user(up, flags);
+        else
                 local_irq_restore(flags);
-        }
 }
 
 struct user_struct * alloc_uid(struct user_namespace *ns, uid_t uid)
@@ -122,6 +319,11 @@ struct user_struct * alloc_uid(struct user_namespace *ns, uid_t uid)
         struct hlist_head *hashent = uidhashentry(ns, uid);
         struct user_struct *up;
 
+        /* Make uid_hash_find() + user_kobject_create() + uid_hash_insert()
+         * atomic.
+         */
+        uids_mutex_lock();
+
         spin_lock_irq(&uidhash_lock);
         up = uid_hash_find(uid, hashent);
         spin_unlock_irq(&uidhash_lock);
@@ -150,6 +352,22 @@ struct user_struct * alloc_uid(struct user_namespace *ns, uid_t uid)
                         return NULL;
                 }
 
+                if (sched_create_user(new) < 0) {
+                        key_put(new->uid_keyring);
+                        key_put(new->session_keyring);
+                        kmem_cache_free(uid_cachep, new);
+                        return NULL;
+                }
+
+                if (user_kobject_create(new)) {
+                        sched_destroy_user(new);
+                        key_put(new->uid_keyring);
+                        key_put(new->session_keyring);
+                        kmem_cache_free(uid_cachep, new);
+                        uids_mutex_unlock();
+                        return NULL;
+                }
+
                 /*
                  * Before adding this, check whether we raced
                  * on adding the same user already..
@@ -157,6 +375,11 @@ struct user_struct * alloc_uid(struct user_namespace *ns, uid_t uid)
                 spin_lock_irq(&uidhash_lock);
                 up = uid_hash_find(uid, hashent);
                 if (up) {
+                        /* This case is not possible when CONFIG_FAIR_USER_SCHED
+                         * is defined, since we serialize alloc_uid() using
+                         * uids_mutex. Hence no need to call
+                         * sched_destroy_user() or remove_user_sysfs_dir().
+                         */
                         key_put(new->uid_keyring);
                         key_put(new->session_keyring);
                         kmem_cache_free(uid_cachep, new);
@@ -167,6 +390,9 @@ struct user_struct * alloc_uid(struct user_namespace *ns, uid_t uid)
                 spin_unlock_irq(&uidhash_lock);
 
         }
+
+        uids_mutex_unlock();
+
         return up;
 }
 
@@ -184,6 +410,7 @@ void switch_uid(struct user_struct *new_user)
         atomic_dec(&old_user->processes);
         switch_uid_keyring(new_user);
         current->user = new_user;
+        sched_switch_user(current);
 
         /*
          * We need to synchronize with __sigqueue_alloc()