On Sat 24-10-15 14:46:58, Ben Hutchings wrote:
[...]
> > --- a/mm/memcontrol.c
> > +++ b/mm/memcontrol.c
> > @@ -3687,6 +3687,7 @@ static int __mem_cgroup_usage_register_e
> > > > ret = page_counter_memparse(args, "-1", &threshold);
> > > > if (ret)
> > > > > return ret;
> > +> > threshold <<= PAGE_SHIFT;
> >
> > > > mutex_lock(&memcg->thresholds_lock);
> >
>
> mem_cgroup_usage() returns a u64 and I think that the types of
> threshold and mem_cgroup_threshold::threshold also need be changed to
> u64 to avoid overflow on large 32-bit systems.
You are absolutely right! I have compltely missed that 3e32cb2e0a12 has
changed the type as well. Should have noticed that during the review.
The whole thing is just way too confusing. All the tracking is done in
page units yet tresholds are in bytes. This only calls for troubles. The
patch below simply turns thresholds to page units as well. I hope I
haven't screwed anything, I didn't get to more than compile test it.
---
>From 58c4824a350e98779ddb9e2e760ec4eef5aed466 Mon Sep 17 00:00:00 2001
From: Michal Hocko <mhocko@xxxxxxxx>
Date: Sun, 25 Oct 2015 17:12:44 +0100
Subject: [PATCH] memcg: Fix thresholds for 32b architectures.
424cdc141380 ("memcg: convert threshold to bytes") has fixed a
regression introduced by 3e32cb2e0a12 ("mm: memcontrol: lockless page
counters") where thresholds were silently converted to use page units
rather than bytes when interpreting the user input.
The fix is not complete, though, as properly pointed out by Ben
Hutchings during stable backport review. The page count is converted
to bytes but unsigned long is used to hold the value which would
be obviously not sufficient for 32b systems with more than 4G
thresholds. The same applies to usage as taken from mem_cgroup_usage
which might overflow.
Let's remove this bytes vs. pages internal tracking differences and
handle thresholds in page units internally. Chage mem_cgroup_usage()
to return the value in page units and revert 424cdc141380 because this
should be sufficient for the consistent handling.
mem_cgroup_read_u64 as the only users of mem_cgroup_usage outside of
the threshold handling code is converted to give the proper in bytes
result. It is doing that already for page_counter output so this is
more consistent as well.
The value presented to the userspace is still in bytes units.
Fixes: 424cdc141380 ("memcg: convert threshold to bytes")
Fixes: 3e32cb2e0a12 ("mm: memcontrol: lockless page counters")
Reported-by: Ben Hutchings <ben@xxxxxxxxxxxxxxx>
Signed-off-by: Michal Hocko <mhocko@xxxxxxxx>
---
mm/memcontrol.c | 9 ++++-----
1 file changed, 4 insertions(+), 5 deletions(-)
diff --git a/mm/memcontrol.c b/mm/memcontrol.c
index f3cc594ffa2d..2823cafc269e 100644
--- a/mm/memcontrol.c
+++ b/mm/memcontrol.c
@@ -2802,7 +2802,7 @@ static unsigned long tree_stat(struct mem_cgroup *memcg,
return val;
}
-static inline u64 mem_cgroup_usage(struct mem_cgroup *memcg, bool swap)
+static inline unsigned long mem_cgroup_usage(struct mem_cgroup *memcg, bool swap)
{
u64 val;
@@ -2817,7 +2817,7 @@ static inline u64 mem_cgroup_usage(struct mem_cgroup *memcg, bool swap)
else
val = page_counter_read(&memcg->memsw);
}
- return val << PAGE_SHIFT;
+ return val;
}
enum {
@@ -2851,9 +2851,9 @@ static u64 mem_cgroup_read_u64(struct cgroup_subsys_state *css,
switch (MEMFILE_ATTR(cft->private)) {
case RES_USAGE:
if (counter == &memcg->memory)
- return mem_cgroup_usage(memcg, false);
+ return (u64)mem_cgroup_usage(memcg, false) * PAGE_SIZE;
if (counter == &memcg->memsw)
- return mem_cgroup_usage(memcg, true);
+ return (u64)mem_cgroup_usage(memcg, true) * PAGE_SIZE;
return (u64)page_counter_read(counter) * PAGE_SIZE;
case RES_LIMIT:
return (u64)counter->limit * PAGE_SIZE;
@@ -3353,7 +3353,6 @@ static int __mem_cgroup_usage_register_event(struct mem_cgroup *memcg,
ret = page_counter_memparse(args, "-1", &threshold);
if (ret)
return ret;
- threshold <<= PAGE_SHIFT;
mutex_lock(&memcg->thresholds_lock);
--
2.6.1
--
Michal Hocko
SUSE Labs
--
To unsubscribe from this list: send the line "unsubscribe stable" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html