On Wed 08-07-15 20:32:51, Vladimir Davydov wrote:
> I like the gist of this patch. A few comments below.
>
> On Wed, Jul 08, 2015 at 02:27:51PM +0200, Michal Hocko wrote:
[...]
> > +/**
> > + * mm_inherit_memcg - Initialize mm_struct::memcg from an existing mm_struct
> > + * @newmm: new mm struct
> > + * @oldmm: old mm struct to inherit from
> > + *
> > + * Should be called for each new mm_struct.
> > + */
> > +static inline
> > +void mm_inherit_memcg(struct mm_struct *newmm, struct mm_struct *oldmm)
> > +{
> > + struct mem_cgroup *memcg = oldmm->memcg;
>
> FWIW, if CONFIG_SPARSE_RCU_POINTER is on, this will trigger a compile
> time warning, as well as any unannotated dereference of mm_struct->memcg
> below.
The idea was that this would be a false positive because the
oldmm->memcg should be stable. But now that I am reading your race
scenario below I am not so sure anymore and we may need to use rcu
locking here. More below
>
> > +
> > + __mm_set_memcg(newmm, memcg);
> > +}
> > +
> > +/**
> > + * mm_drop_iter - drop mm_struct::memcg association
>
> s/mm_drop_iter/mm_drop_memcg
Thanks
>
> > + * @mm: mm struct
> > + *
> > + * Should be called after the mm has been removed from all tasks
> > + * and before it is freed (e.g. from mmput)
> > + */
> > +static inline void mm_drop_memcg(struct mm_struct *mm)
> > +{
> > + if (mm->memcg)
> > + css_put(&mm->memcg->css);
> > + mm->memcg = NULL;
> > +}
[...]
> > @@ -474,7 +519,7 @@ static inline void mem_cgroup_count_vm_event(struct mm_struct *mm,
> > return;
> >
> > rcu_read_lock();
> > - memcg = mem_cgroup_from_task(rcu_dereference(mm->owner));
> > + memcg = rcu_dereference(mm->memcg);
> > if (unlikely(!memcg))
> > goto out;
> >
>
> If I'm not mistaken, mm->memcg equals NULL for any task in the root
> memory cgroup
right
> (BTW, it it's true, it's worth mentioning in the comment
> to mm->memcg definition IMO). As a result, we won't account the stats
> for such tasks, will we?
well spotted! This is certainly a bug. There are more places which are
checking for mm->memcg being NULL and falling back to root_mem_cgroup. I
think it would be better to simply use root_mem_cgroup right away. We
can setup init_mm.memcg = root_mem_cgroup during initialization and be
done with it. What do you think? The diff is in the very end of the
email (completely untested yet).
[...]
> > + * No need to take a reference here because the memcg is pinned by the
> > + * mm_struct.
> > + */
>
> But after we drop the reference to the mm below, mc.from can pass away
> and we can get use-after-free in mem_cgroup_move_task, can't we?
Right, the comment is a left over from the previous attempt when I was
holding the reference throughout the migration.
But then I managed to convince myself that...
> AFAIU the real reason why we can skip taking a reference to mc.from, as
> well as to mc.to, is that task migration proceeds under cgroup_mutex,
> which blocks cgroup destruction.
is true. But now that I am thinking about that again I think I just
misled myself. If a task p is moving from A -> B but p->mm->memcg = C
then we are not protected. I will think about this some more.
[...]
> > @@ -4932,14 +4943,26 @@ static void mem_cgroup_move_task(struct cgroup_subsys_state *css,
> > {
> > struct task_struct *p = cgroup_taskset_first(tset);
> > struct mm_struct *mm = get_task_mm(p);
> > + struct mem_cgroup *old_memcg = NULL;
> >
> > if (mm) {
> > + old_memcg = READ_ONCE(mm->memcg);
> > + __mm_set_memcg(mm, mem_cgroup_from_css(css));
> > +
> > if (mc.to)
> > mem_cgroup_move_charge(mm);
> > mmput(mm);
> > }
> > if (mc.to)
> > mem_cgroup_clear_mc();
> > +
> > + /*
> > + * Be careful and drop the reference only after we are done because
> > + * p's task_css memcg might be different from p->memcg and nothing else
> > + * might be pinning the old memcg.
> > + */
> > + if (old_memcg)
> > + css_put(&old_memcg->css);
>
> Please explain why the following race is impossible:
>
> CPU0 CPU1
> ---- ----
> [current = T]
> dup_mm or exec_mmap
> mm_inherit_memcg
> memcg = current->mm->memcg;
> mem_cgroup_move_task
> p = T;
> mm = get_task_mm(p);
> old_memcg = mm->memcg;
> css_put(&old_memcg->css);
> /* old_memcg can be freed now */
> css_get(memcg); /* BUG */
I guess you are right. The window seem to be very small but CPU0 simly
might get preempted by the moving task and so even cgroup pinning
wouldn't help here.
I guess we need
---
diff --git a/include/linux/memcontrol.h b/include/linux/memcontrol.h
index b3e7e30b5a74..6fbd33273b6d 100644
--- a/include/linux/memcontrol.h
+++ b/include/linux/memcontrol.h
@@ -300,9 +300,17 @@ void __mm_set_memcg(struct mm_struct *mm, struct mem_cgroup *memcg)
static inline
void mm_inherit_memcg(struct mm_struct *newmm, struct mm_struct *oldmm)
{
- struct mem_cgroup *memcg = oldmm->memcg;
+ struct mem_cgroup *memcg;
+ /*
+ * oldmm might be under move and just replacing its memcg (see
+ * mem_cgroup_move_task) so we have to protect from its memcg
+ * going away between we dereference and take a reference.
+ */
+ rcu_read_lock();
+ memcg = rcu_dereference(oldmm->memcg);
__mm_set_memcg(newmm, memcg);
+ rcu_read_unlock();
}
/**
Make sure that all tasks have non NULL memcg.
---
diff --git a/include/linux/memcontrol.h b/include/linux/memcontrol.h
index f23e29f3d4fa..b3e7e30b5a74 100644
--- a/include/linux/memcontrol.h
+++ b/include/linux/memcontrol.h
@@ -286,8 +286,7 @@ extern struct cgroup_subsys_state *mem_cgroup_root_css;
static inline
void __mm_set_memcg(struct mm_struct *mm, struct mem_cgroup *memcg)
{
- if (memcg)
- css_get(&memcg->css);
+ css_get(&memcg->css);
rcu_assign_pointer(mm->memcg, memcg);
}
@@ -307,7 +306,7 @@ void mm_inherit_memcg(struct mm_struct *newmm, struct mm_struct *oldmm)
}
/**
- * mm_drop_iter - drop mm_struct::memcg association
+ * mm_drop_memcg - drop mm_struct::memcg association
* @mm: mm struct
*
* Should be called after the mm has been removed from all tasks
@@ -315,8 +314,7 @@ void mm_inherit_memcg(struct mm_struct *newmm, struct mm_struct *oldmm)
*/
static inline void mm_drop_memcg(struct mm_struct *mm)
{
- if (mm->memcg)
- css_put(&mm->memcg->css);
+ css_put(&mm->memcg->css);
mm->memcg = NULL;
}
@@ -382,8 +380,7 @@ static inline bool mm_match_cgroup(struct mm_struct *mm,
rcu_read_lock();
task_memcg = rcu_dereference(mm->memcg);
- if (task_memcg)
- match = mem_cgroup_is_descendant(task_memcg, memcg);
+ match = mem_cgroup_is_descendant(task_memcg, memcg);
rcu_read_unlock();
return match;
}
@@ -526,8 +523,6 @@ static inline void mem_cgroup_count_vm_event(struct mm_struct *mm,
rcu_read_lock();
memcg = rcu_dereference(mm->memcg);
- if (unlikely(!memcg))
- goto out;
switch (idx) {
case PGFAULT:
@@ -539,7 +534,6 @@ static inline void mem_cgroup_count_vm_event(struct mm_struct *mm,
default:
BUG();
}
-out:
rcu_read_unlock();
}
#ifdef CONFIG_TRANSPARENT_HUGEPAGE
diff --git a/mm/memcontrol.c b/mm/memcontrol.c
index e7169a9f7a47..23ee92c396e2 100644
--- a/mm/memcontrol.c
+++ b/mm/memcontrol.c
@@ -801,11 +801,8 @@ static struct mem_cgroup *get_mem_cgroup_from_mm(struct mm_struct *mm)
*/
if (unlikely(!mm))
memcg = root_mem_cgroup;
- else {
+ else
memcg = rcu_dereference(mm->memcg);
- if (unlikely(!memcg))
- memcg = root_mem_cgroup;
- }
} while (!css_tryget_online(&memcg->css));
rcu_read_unlock();
return memcg;
@@ -4176,6 +4173,11 @@ mem_cgroup_css_alloc(struct cgroup_subsys_state *parent_css)
memcg->soft_limit = PAGE_COUNTER_MAX;
page_counter_init(&memcg->memsw, NULL);
page_counter_init(&memcg->kmem, NULL);
+ /*
+ * Make sure all tasks will inherit root_mem_cgroup
+ * implicitly.
+ */
+ __mm_set_memcg(&init_mm, root_mem_cgroup);
}
memcg->last_scanned_node = MAX_NUMNODES;
@@ -4787,8 +4789,6 @@ static int mem_cgroup_can_attach(struct cgroup_subsys_state *css,
* mm_struct.
*/
from = READ_ONCE(mm->memcg);
- if (!from)
- from = root_mem_cgroup;
if (from == to)
goto out;
@@ -4979,8 +4979,7 @@ static void mem_cgroup_move_task(struct cgroup_subsys_state *css,
* p's task_css memcg might be different from p->memcg and nothing else
* might be pinning the old memcg.
*/
- if (old_memcg)
- css_put(&old_memcg->css);
+ css_put(&old_memcg->css);
}
#else /* !CONFIG_MMU */
static int mem_cgroup_can_attach(struct cgroup_subsys_state *css,
--
Michal Hocko
SUSE Labs
--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@xxxxxxxxx. For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@xxxxxxxxx";> email@xxxxxxxxx </a>