sched: add fair-user scheduler

Enable user-id based fair group scheduling. This is useful for anyone
who wants to test the group scheduler w/o having to enable
CONFIG_CGROUPS.

A separate scheduling group (i.e struct task_grp) is automatically created for 
every new user added to the system. Upon uid change for a task, it is made to 
move to the corresponding scheduling group.

A /proc tunable (/proc/root_user_share) is also provided to tune root
user's quota of cpu bandwidth.

Signed-off-by: Srivatsa Vaddagiri <vatsa@linux.vnet.ibm.com>
Signed-off-by: Dhaval Giani <dhaval@linux.vnet.ibm.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Reviewed-by: Thomas Gleixner <tglx@linutronix.de>

3 files changed, 104 insertions, 0 deletions

diff --git a/kernel/sched.c b/kernel/sched.c
index e10c403..f33608e 100644
--- a/kernel/sched.c
+++ b/kernel/sched.c
@@ -200,7 +200,12 @@ struct task_grp init_task_grp =  {
 				.cfs_rq = init_cfs_rq_p,
 				 };
 
+#ifdef CONFIG_FAIR_USER_SCHED
+#define INIT_TASK_GRP_LOAD	2*NICE_0_LOAD
+#else
 #define INIT_TASK_GRP_LOAD	NICE_0_LOAD
+#endif
+
 static int init_task_grp_load = INIT_TASK_GRP_LOAD;
 
 /* return group to which a task belongs */
@@ -208,7 +213,11 @@ static inline struct task_grp *task_grp(struct task_struct *p)
 {
 	struct task_grp *tg;
 
+#ifdef CONFIG_FAIR_USER_SCHED
+	tg = p->user->tg;
+#else
 	tg  = &init_task_grp;
+#endif
 
 	return tg;
 }
diff --git a/kernel/sched_debug.c b/kernel/sched_debug.c
index 3e47e87..57ee9d5 100644
--- a/kernel/sched_debug.c
+++ b/kernel/sched_debug.c
@@ -212,6 +212,49 @@ static void sysrq_sched_debug_show(void)
 	sched_debug_show(NULL, NULL);
 }
 
+#ifdef CONFIG_FAIR_USER_SCHED
+
+static DEFINE_MUTEX(root_user_share_mutex);
+
+static int
+root_user_share_read_proc(char *page, char **start, off_t off, int count,
+				 int *eof, void *data)
+{
+	int len;
+
+	len = sprintf(page, "%d\n", init_task_grp_load);
+
+	return len;
+}
+
+static int
+root_user_share_write_proc(struct file *file, const char __user *buffer,
+				 unsigned long count, void *data)
+{
+	unsigned long shares;
+	char kbuf[sizeof(unsigned long)+1];
+	int rc = 0;
+
+	if (copy_from_user(kbuf, buffer, sizeof(kbuf)))
+		return -EFAULT;
+
+	shares = simple_strtoul(kbuf, NULL, 0);
+
+	if (!shares)
+		shares = NICE_0_LOAD;
+
+	mutex_lock(&root_user_share_mutex);
+
+	init_task_grp_load = shares;
+	rc = sched_group_set_shares(&init_task_grp, shares);
+
+	mutex_unlock(&root_user_share_mutex);
+
+	return (rc < 0 ? rc : count);
+}
+
+#endif	/* CONFIG_FAIR_USER_SCHED */
+
 static int sched_debug_open(struct inode *inode, struct file *filp)
 {
 	return single_open(filp, sched_debug_show, NULL);
@@ -234,6 +277,15 @@ static int __init init_sched_debug_procfs(void)
 
 	pe->proc_fops = &sched_debug_fops;
 
+#ifdef CONFIG_FAIR_USER_SCHED
+	pe = create_proc_entry("root_user_share", 0644, NULL);
+	if (!pe)
+		return -ENOMEM;
+
+	pe->read_proc = root_user_share_read_proc;
+	pe->write_proc = root_user_share_write_proc;
+#endif
+
 	return 0;
 }
 
diff --git a/kernel/user.c b/kernel/user.c
index 9ca2848..c6387fa 100644
--- a/kernel/user.c
+++ b/kernel/user.c
@@ -50,8 +50,41 @@ struct user_struct root_user = {
 	.uid_keyring	= &root_user_keyring,
 	.session_keyring = &root_session_keyring,
 #endif
+#ifdef CONFIG_FAIR_USER_SCHED
+	.tg		= &init_task_grp,
+#endif
 };
 
+#ifdef CONFIG_FAIR_USER_SCHED
+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);
+}
+
+#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) { }
+
+#endif	/* CONFIG_FAIR_USER_SCHED */
+
 /*
  * These routines must be called with the uidhash spinlock held!
  */
@@ -109,6 +142,7 @@ void free_uid(struct user_struct *up)
 	if (atomic_dec_and_lock(&up->__count, &uidhash_lock)) {
 		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);
@@ -150,6 +184,13 @@ 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;
+		}
+
 		/*
 		 * Before adding this, check whether we raced
 		 * on adding the same user already..
@@ -157,6 +198,7 @@ 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) {
+			sched_destroy_user(new);
 			key_put(new->uid_keyring);
 			key_put(new->session_keyring);
 			kmem_cache_free(uid_cachep, new);
@@ -184,6 +226,7 @@ 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()