Merge branch 'linus' into perfcounters/core

Conflicts:
	arch/x86/include/asm/kmap_types.h
	include/linux/mm.h

	include/asm-generic/kmap_types.h

Merge reason: We crossed changes with kmap_types.h cleanups in mainline.

Signed-off-by: Ingo Molnar <mingo@elte.hu>

31 files changed, 804 insertions, 588 deletions

diff --git a/kernel/Makefile b/kernel/Makefile
index 90b53f6dc22..9df4501cb92 100644
--- a/kernel/Makefile
+++ b/kernel/Makefile
@@ -11,6 +11,7 @@ obj-y     = sched.o fork.o exec_domain.o panic.o printk.o \
             hrtimer.o rwsem.o nsproxy.o srcu.o semaphore.o \
             notifier.o ksysfs.o pm_qos_params.o sched_clock.o cred.o \
             async.o
+obj-y += groups.o
 
 ifdef CONFIG_FUNCTION_TRACER
 # Do not trace debug files and internal ftrace files
diff --git a/kernel/cpuset.c b/kernel/cpuset.c
index d5a7e17474e..7e75a41bd50 100644
--- a/kernel/cpuset.c
+++ b/kernel/cpuset.c
@@ -97,12 +97,6 @@ struct cpuset {
 
         struct cpuset *parent;          /* my parent */
 
-        /*
-         * Copy of global cpuset_mems_generation as of the most
-         * recent time this cpuset changed its mems_allowed.
-         */
-        int mems_generation;
-
         struct fmeter fmeter;           /* memory_pressure filter */
 
         /* partition number for rebuild_sched_domains() */
@@ -176,27 +170,6 @@ static inline int is_spread_slab(const struct cpuset *cs)
         return test_bit(CS_SPREAD_SLAB, &cs->flags);
 }
 
-/*
- * Increment this integer everytime any cpuset changes its
- * mems_allowed value.  Users of cpusets can track this generation
- * number, and avoid having to lock and reload mems_allowed unless
- * the cpuset they're using changes generation.
- *
- * A single, global generation is needed because cpuset_attach_task() could
- * reattach a task to a different cpuset, which must not have its
- * generation numbers aliased with those of that tasks previous cpuset.
- *
- * Generations are needed for mems_allowed because one task cannot
- * modify another's memory placement.  So we must enable every task,
- * on every visit to __alloc_pages(), to efficiently check whether
- * its current->cpuset->mems_allowed has changed, requiring an update
- * of its current->mems_allowed.
- *
- * Since writes to cpuset_mems_generation are guarded by the cgroup lock
- * there is no need to mark it atomic.
- */
-static int cpuset_mems_generation;
-
 static struct cpuset top_cpuset = {
         .flags = ((1 << CS_CPU_EXCLUSIVE) | (1 << CS_MEM_EXCLUSIVE)),
 };
@@ -228,8 +201,9 @@ static struct cpuset top_cpuset = {
  * If a task is only holding callback_mutex, then it has read-only
  * access to cpusets.
  *
- * The task_struct fields mems_allowed and mems_generation may only
- * be accessed in the context of that task, so require no locks.
+ * Now, the task_struct fields mems_allowed and mempolicy may be changed
+ * by other task, we use alloc_lock in the task_struct fields to protect
+ * them.
  *
  * The cpuset_common_file_read() handlers only hold callback_mutex across
  * small pieces of code, such as when reading out possibly multi-word
@@ -331,75 +305,22 @@ static void guarantee_online_mems(const struct cpuset *cs, nodemask_t *pmask)
         BUG_ON(!nodes_intersects(*pmask, node_states[N_HIGH_MEMORY]));
 }
 
-/**
- * cpuset_update_task_memory_state - update task memory placement
- *
- * If the current tasks cpusets mems_allowed changed behind our
- * backs, update current->mems_allowed, mems_generation and task NUMA
- * mempolicy to the new value.
- *
- * Task mempolicy is updated by rebinding it relative to the
- * current->cpuset if a task has its memory placement changed.
- * Do not call this routine if in_interrupt().
- *
- * Call without callback_mutex or task_lock() held.  May be
- * called with or without cgroup_mutex held.  Thanks in part to
- * 'the_top_cpuset_hack', the task's cpuset pointer will never
- * be NULL.  This routine also might acquire callback_mutex during
- * call.
- *
- * Reading current->cpuset->mems_generation doesn't need task_lock
- * to guard the current->cpuset derefence, because it is guarded
- * from concurrent freeing of current->cpuset using RCU.
- *
- * The rcu_dereference() is technically probably not needed,
- * as I don't actually mind if I see a new cpuset pointer but
- * an old value of mems_generation.  However this really only
- * matters on alpha systems using cpusets heavily.  If I dropped
- * that rcu_dereference(), it would save them a memory barrier.
- * For all other arch's, rcu_dereference is a no-op anyway, and for
- * alpha systems not using cpusets, another planned optimization,
- * avoiding the rcu critical section for tasks in the root cpuset
- * which is statically allocated, so can't vanish, will make this
- * irrelevant.  Better to use RCU as intended, than to engage in
- * some cute trick to save a memory barrier that is impossible to
- * test, for alpha systems using cpusets heavily, which might not
- * even exist.
- *
- * This routine is needed to update the per-task mems_allowed data,
- * within the tasks context, when it is trying to allocate memory
- * (in various mm/mempolicy.c routines) and notices that some other
- * task has been modifying its cpuset.
+/*
+ * update task's spread flag if cpuset's page/slab spread flag is set
+ *
+ * Called with callback_mutex/cgroup_mutex held
  */
-
-void cpuset_update_task_memory_state(void)
+static void cpuset_update_task_spread_flag(struct cpuset *cs,
+                                        struct task_struct *tsk)
 {
-        int my_cpusets_mem_gen;
-        struct task_struct *tsk = current;
-        struct cpuset *cs;
-
-        rcu_read_lock();
-        my_cpusets_mem_gen = task_cs(tsk)->mems_generation;
-        rcu_read_unlock();
-
-        if (my_cpusets_mem_gen != tsk->cpuset_mems_generation) {
-                mutex_lock(&callback_mutex);
-                task_lock(tsk);
-                cs = task_cs(tsk); /* Maybe changed when task not locked */
-                guarantee_online_mems(cs, &tsk->mems_allowed);
-                tsk->cpuset_mems_generation = cs->mems_generation;
-                if (is_spread_page(cs))
-                        tsk->flags |= PF_SPREAD_PAGE;
-                else
-                        tsk->flags &= ~PF_SPREAD_PAGE;
-                if (is_spread_slab(cs))
-                        tsk->flags |= PF_SPREAD_SLAB;
-                else
-                        tsk->flags &= ~PF_SPREAD_SLAB;
-                task_unlock(tsk);
-                mutex_unlock(&callback_mutex);
-                mpol_rebind_task(tsk, &tsk->mems_allowed);
-        }
+        if (is_spread_page(cs))
+                tsk->flags |= PF_SPREAD_PAGE;
+        else
+                tsk->flags &= ~PF_SPREAD_PAGE;
+        if (is_spread_slab(cs))
+                tsk->flags |= PF_SPREAD_SLAB;
+        else
+                tsk->flags &= ~PF_SPREAD_SLAB;
 }
 
 /*
@@ -1007,14 +928,6 @@ static int update_cpumask(struct cpuset *cs, struct cpuset *trialcs,
  *    other task, the task_struct mems_allowed that we are hacking