lkml.org 
[lkml]   [2011]   [Dec]   [13]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
    /
    Date
    From
    SubjectRe: [PATCH] memcg: clean up soft_limit_tree properly new
    On Tue, 13 Dec 2011 17:00:12 -0800
    Andrew Morton <akpm@linux-foundation.org> wrote:

    > On Mon, 12 Dec 2011 15:09:35 +0100
    > Michal Hocko <mhocko@suse.cz> wrote:
    >
    > > And a follow up patch for the proper clean up:
    > > ---
    > > >From 4b9f5a1e88496af9f336d1ef37cfdf3754a3ba48 Mon Sep 17 00:00:00 2001
    > > From: Michal Hocko <mhocko@suse.cz>
    > > Date: Mon, 12 Dec 2011 15:04:18 +0100
    > > Subject: [PATCH] memcg: clean up soft_limit_tree properly
    > >
    > > If we are not able to allocate tree nodes for all NUMA nodes then we
    > > should better clean up those that were allocated otherwise we will leak
    > > a memory.
    > >
    > > Signed-off-by: Michal Hocko <mhocko@suse.cz>
    > > ---
    > > mm/memcontrol.c | 12 +++++++++++-
    > > 1 files changed, 11 insertions(+), 1 deletions(-)
    > >
    > > diff --git a/mm/memcontrol.c b/mm/memcontrol.c
    > > index 6aff93c..838d812 100644
    > > --- a/mm/memcontrol.c
    > > +++ b/mm/memcontrol.c
    > > @@ -4874,7 +4874,7 @@ static int mem_cgroup_soft_limit_tree_init(void)
    > > tmp = -1;
    > > rtpn = kzalloc_node(sizeof(*rtpn), GFP_KERNEL, tmp);
    > > if (!rtpn)
    > > - return 1;
    > > + goto err_cleanup;
    > >
    > > soft_limit_tree.rb_tree_per_node[node] = rtpn;
    > >
    > > @@ -4885,6 +4885,16 @@ static int mem_cgroup_soft_limit_tree_init(void)
    > > }
    > > }
    > > return 0;
    > > +
    > > +err_cleanup:
    > > + for_each_node_state(node, N_POSSIBLE) {
    > > + if (!soft_limit_tree.rb_tree_per_node[node])
    > > + break;
    > > + kfree(soft_limit_tree.rb_tree_per_node[node]);
    > > + soft_limit_tree.rb_tree_per_node[node] = NULL;
    > > + }
    > > + return 1;
    > > +
    > > }
    >
    > afacit the kernel never frees the soft_limit_tree.rb_tree_per_node[]
    > entries on the mem_cgroup_destroy() path. Bug?
    >

    soft_limit_tree.rb_tree_per_node[] is a global object and allocated once
    at creating root cgroup.

    Nodes of rb_tree for a memcg are contained in struct mem_cgroup_per_zone
    and it's freed at mem_cgroup_destroy().

    Thanks,
    -Kame






    \
     
     \ /
      Last update: 2011-12-14 02:21    [W:0.026 / U:29.864 seconds]
    ©2003-2016 Jasper Spaans. hosted at Digital OceanAdvertise on this site