[PATCH 1/2] uprobes: fix the memcg accounting

From	Oleg Nesterov <oleg@redhat.com>
Newsgroups	linux.kernel
Subject	[PATCH 1/2] uprobes: fix the memcg accounting
Date	2016-08-17 17:40 +0200
Message-ID	<s7b6F-2UT-19@gated-at.bofh.it> (permalink)
References	(1 earlier) <s6NnH-49C-1@gated-at.bofh.it> <s6Nxn-4cZ-1@gated-at.bofh.it> <s6NH3-4gh-13@gated-at.bofh.it> <s6PSy-5xO-15@gated-at.bofh.it> <s7b6F-2UT-21@gated-at.bofh.it>
Organization	linux.* mail to news gateway

Show all headers | View raw

__replace_page() wronlgy calls mem_cgroup_cancel_charge() in "success" path,
it should only do this if page_check_address() fails.

This means that every enable/disable leads to unbalanced mem_cgroup_uncharge()
from put_page(old_page), it is trivial to underflow the page_counter->count
and trigger OOM.

Reported-and-tested-by: Brenden Blanco <bblanco@plumgrid.com>
Signed-off-by: Oleg Nesterov <oleg@redhat.com>
Acked-by: Michal Hocko <mhocko@kernel.org>
Reviewed-by: Johannes Weiner <hannes@cmpxchg.org>
Fixes: 00501b531c47 ("mm: memcontrol: rewrite charge API")
Cc: stable@vger.kernel.org # 3.17+
---
 kernel/events/uprobes.c | 5 +++--
 1 file changed, 3 insertions(+), 2 deletions(-)

diff --git a/kernel/events/uprobes.c b/kernel/events/uprobes.c
index b7a525a..8c50276 100644
--- a/kernel/events/uprobes.c
+++ b/kernel/events/uprobes.c
@@ -172,8 +172,10 @@ static int __replace_page(struct vm_area_struct *vma, unsigned long addr,
 	mmu_notifier_invalidate_range_start(mm, mmun_start, mmun_end);
 	err = -EAGAIN;
 	ptep = page_check_address(page, mm, addr, &ptl, 0);
-	if (!ptep)
+	if (!ptep) {
+		mem_cgroup_cancel_charge(kpage, memcg, false);
 		goto unlock;
+	}
 
 	get_page(kpage);
 	page_add_new_anon_rmap(kpage, vma, addr, false);
@@ -200,7 +202,6 @@ static int __replace_page(struct vm_area_struct *vma, unsigned long addr,
 
 	err = 0;
  unlock:
-	mem_cgroup_cancel_charge(kpage, memcg, false);
 	mmu_notifier_invalidate_range_end(mm, mmun_start, mmun_end);
 	unlock_page(page);
 	return err;
-- 
2.5.0

Thread

uprobes: memory leak in enable/disable loop Brenden Blanco <bblanco@plumgrid.com> - 2016-08-15 23:00 +0200
  Re: uprobes: memory leak in enable/disable loop Oleg Nesterov <oleg@redhat.com> - 2016-08-16 16:20 +0200
    Re: uprobes: memory leak in enable/disable loop Oleg Nesterov <oleg@redhat.com> - 2016-08-16 16:30 +0200
      Re: uprobes: memory leak in enable/disable loop Michal Hocko <mhocko@kernel.org> - 2016-08-16 16:40 +0200
      Re: uprobes: memory leak in enable/disable loop Oleg Nesterov <oleg@redhat.com> - 2016-08-16 16:40 +0200
        Re: uprobes: memory leak in enable/disable loop Brenden Blanco <bblanco@plumgrid.com> - 2016-08-16 19:00 +0200
          [PATCH 0/2] uprobes: memory leak in enable/disable loop Oleg Nesterov <oleg@redhat.com> - 2016-08-17 17:40 +0200
            [PATCH 1/2] uprobes: fix the memcg accounting Oleg Nesterov <oleg@redhat.com> - 2016-08-17 17:40 +0200
              [tip:perf/urgent] uprobes: Fix the memcg accounting tip-bot for Oleg Nesterov <tipbot@zytor.com> - 2016-08-18 10:20 +0200
            [PATCH 2/2] uprobes: rename the "struct page *" args of  __replace_page() Oleg Nesterov <oleg@redhat.com> - 2016-08-17 17:50 +0200
              [tip:perf/core] uprobes: Rename the "struct page *" args of  __replace_page() tip-bot for Oleg Nesterov <tipbot@zytor.com> - 2016-08-18 13:00 +0200
        Re: uprobes: memory leak in enable/disable loop Johannes Weiner <hannes@cmpxchg.org> - 2016-08-16 19:50 +0200

csiph-web