with CONFIG_PROVE_RCU=y, I saw this warning, it's because
css_id() is not under rcu_read_lock().
===================================================
[ INFO: suspicious rcu_dereference_check() usage. ]
---------------------------------------------------
kernel/cgroup.c:4438 invoked rcu_dereference_check() without protection!
other info that might help us debug this:
rcu_scheduler_active = 1, debug_locks = 1
1 lock held by kswapd0/31:
#0: (swap_lock){+.+.-.}, at: [<c05058bb>] swap_info_get+0x4b/0xd0
stack backtrace:
Pid: 31, comm: kswapd0 Not tainted 2.6.34-rc5-tip+ #13
Call Trace:
[<c083c5d6>] ? printk+0x1d/0x1f
[<c0480744>] lockdep_rcu_dereference+0x94/0xb0
[<c049d6ed>] css_id+0x5d/0x60
[<c05165a5>] mem_cgroup_uncharge_swapcache+0x45/0xa0
[<c0505e4f>] swapcache_free+0x3f/0x60
[<c04e79e2>] __remove_mapping+0xb2/0xf0
[<c04e7cbb>] shrink_page_list+0x26b/0x490
[<c047f85d>] ? put_lock_stats+0xd/0x30
[<c083fd67>] ? _raw_spin_unlock_irq+0x27/0x50
[<c0482566>] ? trace_hardirqs_on_caller+0xb6/0x220
[<c04e8158>] shrink_inactive_list+0x278/0x620
[<c04729e1>] ? sched_clock_cpu+0x121/0x180
[<c047e9b8>] ? trace_hardirqs_off_caller+0x18/0x130
[<c047eadb>] ? trace_hardirqs_off+0xb/0x10
[<c0843438>] ? sub_preempt_count+0x8/0x90
[<c047f85d>] ? put_lock_stats+0xd/0x30
[<c04e8704>] shrink_zone+0x204/0x3c0
[<c083fcac>] ? _raw_spin_unlock+0x2c/0x50
[<c04e951e>] kswapd+0x61e/0x7c0
[<c04e6ed0>] ? isolate_pages_global+0x0/0x1f0
[<c046bae0>] ? autoremove_wake_function+0x0/0x50
[<c04e8f00>] ? kswapd+0x0/0x7c0
[<c046b5e4>] kthread+0x74/0x80
[<c046b570>] ? kthread+0x0/0x80
[<c04035ba>] kernel_thread_helper+0x6/0x10
> From: KAMEZAWA Hiroyuki <[email protected]>
>
> css_id() should be called under rcu_read_lock().
> Following is a report from Li Zefan.
> ==
> ===================================================
> [ INFO: suspicious rcu_dereference_check() usage. ]
> ---------------------------------------------------
> kernel/cgroup.c:4438 invoked rcu_dereference_check() without protection!
>
> other info that might help us debug this:
>
>
> rcu_scheduler_active = 1, debug_locks = 1
> 1 lock held by kswapd0/31:
> #0: (swap_lock){+.+.-.}, at: [<c05058bb>] swap_info_get+0x4b/0xd0
>
> stack backtrace:
...
>
> And css_is_ancestor() should be called under rcu_read_lock().
>
>
> Reported-by: Li Zefan <[email protected]>
> Cc: Daisuke Nishimura <[email protected]>
> Cc: Balbir Singh <[email protected]>
> Signed-off-by: KAMEZAWA Hiroyuki <[email protected]>
With this patch applied, I did some more test, and no warning was triggered.
Tested-by: Li Zefan <[email protected]>
On Fri, 23 Apr 2010 14:10:32 +0800
Li Zefan <[email protected]> wrote:
> > From: KAMEZAWA Hiroyuki <[email protected]>
> >
> > css_id() should be called under rcu_read_lock().
> > Following is a report from Li Zefan.
> > ==
> > ===================================================
> > [ INFO: suspicious rcu_dereference_check() usage. ]
> > ---------------------------------------------------
> > kernel/cgroup.c:4438 invoked rcu_dereference_check() without protection!
> >
> > other info that might help us debug this:
> >
> >
> > rcu_scheduler_active = 1, debug_locks = 1
> > 1 lock held by kswapd0/31:
> > #0: (swap_lock){+.+.-.}, at: [<c05058bb>] swap_info_get+0x4b/0xd0
> >
> > stack backtrace:
> ...
> >
> > And css_is_ancestor() should be called under rcu_read_lock().
> >
> >
> > Reported-by: Li Zefan <[email protected]>
> > Cc: Daisuke Nishimura <[email protected]>
> > Cc: Balbir Singh <[email protected]>
> > Signed-off-by: KAMEZAWA Hiroyuki <[email protected]>
>
> With this patch applied, I did some more test, and no warning was triggered.
>
> Tested-by: Li Zefan <[email protected]>
>
Thank you!.
-Kame
* KAMEZAWA Hiroyuki <[email protected]> [2010-04-23 13:03:49]:
> On Fri, 23 Apr 2010 12:58:14 +0900
> KAMEZAWA Hiroyuki <[email protected]> wrote:
>
> > On Fri, 23 Apr 2010 11:55:16 +0800
> > Li Zefan <[email protected]> wrote:
> >
> > > Li Zefan wrote:
> > > > KAMEZAWA Hiroyuki wrote:
> > > >> On Fri, 23 Apr 2010 11:00:41 +0800
> > > >> Li Zefan <[email protected]> wrote:
> > > >>
> > > >>> with CONFIG_PROVE_RCU=y, I saw this warning, it's because
> > > >>> css_id() is not under rcu_read_lock().
> > > >>>
> > > >> Ok. Thank you for reporting.
> > > >> This is ok ?
> > > >
> > > > Yes, and I did some more simple tests on memcg, no more warning
> > > > showed up.
> > > >
> > >
> > > oops, after trigging oom, I saw 2 more warnings:
> > >
> >
> > Thank you for good testing.
> v3 here...sorry too rapid posting...
>
Looking at the patch we seem to be protecting the use of only css_*().
I wonder if we should push down the rcu_read_*lock() semnatics to the
css routines or is it just too instrusive to do it that way?
--
Three Cheers,
Balbir
On Fri, 23 Apr 2010 12:30:11 +0530
Balbir Singh <[email protected]> wrote:
> * KAMEZAWA Hiroyuki <[email protected]> [2010-04-23 13:03:49]:
>
> > On Fri, 23 Apr 2010 12:58:14 +0900
> > KAMEZAWA Hiroyuki <[email protected]> wrote:
> >
> > > On Fri, 23 Apr 2010 11:55:16 +0800
> > > Li Zefan <[email protected]> wrote:
> > >
> > > > Li Zefan wrote:
> > > > > KAMEZAWA Hiroyuki wrote:
> > > > >> On Fri, 23 Apr 2010 11:00:41 +0800
> > > > >> Li Zefan <[email protected]> wrote:
> > > > >>
> > > > >>> with CONFIG_PROVE_RCU=y, I saw this warning, it's because
> > > > >>> css_id() is not under rcu_read_lock().
> > > > >>>
> > > > >> Ok. Thank you for reporting.
> > > > >> This is ok ?
> > > > >
> > > > > Yes, and I did some more simple tests on memcg, no more warning
> > > > > showed up.
> > > > >
> > > >
> > > > oops, after trigging oom, I saw 2 more warnings:
> > > >
> > >
> > > Thank you for good testing.
> > v3 here...sorry too rapid posting...
> >
>
> Looking at the patch we seem to be protecting the use of only css_*().
> I wonder if we should push down the rcu_read_*lock() semnatics to the
> css routines or is it just too instrusive to do it that way?
>
Maybe worth to consider for future patches for clean up.
Thanks,
-Kame
On Fri, 23 Apr 2010 11:00:41 +0800
Li Zefan <[email protected]> wrote:
> with CONFIG_PROVE_RCU=y, I saw this warning, it's because
> css_id() is not under rcu_read_lock().
>
Ok. Thank you for reporting.
This is ok ?
==
From: KAMEZAWA Hiroyuki <[email protected]>
css_id() should be called under rcu_read_lock().
Following is a report from Li Zefan.
==
===================================================
[ INFO: suspicious rcu_dereference_check() usage. ]
---------------------------------------------------
kernel/cgroup.c:4438 invoked rcu_dereference_check() without protection!
other info that might help us debug this:
rcu_scheduler_active = 1, debug_locks = 1
1 lock held by kswapd0/31:
#0: (swap_lock){+.+.-.}, at: [<c05058bb>] swap_info_get+0x4b/0xd0
stack backtrace:
Pid: 31, comm: kswapd0 Not tainted 2.6.34-rc5-tip+ #13
Call Trace:
[<c083c5d6>] ? printk+0x1d/0x1f
[<c0480744>] lockdep_rcu_dereference+0x94/0xb0
[<c049d6ed>] css_id+0x5d/0x60
[<c05165a5>] mem_cgroup_uncharge_swapcache+0x45/0xa0
[<c0505e4f>] swapcache_free+0x3f/0x60
[<c04e79e2>] __remove_mapping+0xb2/0xf0
[<c04e7cbb>] shrink_page_list+0x26b/0x490
[<c047f85d>] ? put_lock_stats+0xd/0x30
[<c083fd67>] ? _raw_spin_unlock_irq+0x27/0x50
[<c0482566>] ? trace_hardirqs_on_caller+0xb6/0x220
[<c04e8158>] shrink_inactive_list+0x278/0x620
[<c04729e1>] ? sched_clock_cpu+0x121/0x180
[<c047e9b8>] ? trace_hardirqs_off_caller+0x18/0x130
[<c047eadb>] ? trace_hardirqs_off+0xb/0x10
[<c0843438>] ? sub_preempt_count+0x8/0x90
[<c047f85d>] ? put_lock_stats+0xd/0x30
[<c04e8704>] shrink_zone+0x204/0x3c0
[<c083fcac>] ? _raw_spin_unlock+0x2c/0x50
[<c04e951e>] kswapd+0x61e/0x7c0
[<c04e6ed0>] ? isolate_pages_global+0x0/0x1f0
[<c046bae0>] ? autoremove_wake_function+0x0/0x50
[<c04e8f00>] ? kswapd+0x0/0x7c0
[<c046b5e4>] kthread+0x74/0x80
[<c046b570>] ? kthread+0x0/0x80
[<c04035ba>] kernel_thread_helper+0x6/0x10
Reported-by: Li Zefan <[email protected]>
Cc: Daisuke Nishimura <[email protected]>
Cc: Balbir Singh <[email protected]>
Signed-off-by: KAMEZAWA Hiroyuki <[email protected]>
---
mm/memcontrol.c | 2 ++
1 file changed, 2 insertions(+)
Index: linux-2.6.34-rc5-mm1/mm/memcontrol.c
===================================================================
--- linux-2.6.34-rc5-mm1.orig/mm/memcontrol.c
+++ linux-2.6.34-rc5-mm1/mm/memcontrol.c
@@ -2401,7 +2401,9 @@ mem_cgroup_uncharge_swapcache(struct pag
/* record memcg information */
if (do_swap_account && swapout && memcg) {
+ rcu_read_lock();
swap_cgroup_record(ent, css_id(&memcg->css));
+ rcu_read_unlock();
mem_cgroup_get(memcg);
}
if (swapout && memcg)
On Fri, 23 Apr 2010 11:55:16 +0800
Li Zefan <[email protected]> wrote:
> Li Zefan wrote:
> > KAMEZAWA Hiroyuki wrote:
> >> On Fri, 23 Apr 2010 11:00:41 +0800
> >> Li Zefan <[email protected]> wrote:
> >>
> >>> with CONFIG_PROVE_RCU=y, I saw this warning, it's because
> >>> css_id() is not under rcu_read_lock().
> >>>
> >> Ok. Thank you for reporting.
> >> This is ok ?
> >
> > Yes, and I did some more simple tests on memcg, no more warning
> > showed up.
> >
>
> oops, after trigging oom, I saw 2 more warnings:
>
Thank you for good testing.
=
From: KAMEZAWA Hiroyuki <[email protected]>
css_id() should be called under rcu_read_lock().
And css_is_ancestor() should be called under rcu_read_lock().
Following is a report from Li Zefan.
==
===================================================
[ INFO: suspicious rcu_dereference_check() usage. ]
---------------------------------------------------
kernel/cgroup.c:4438 invoked rcu_dereference_check() without protection!
other info that might help us debug this:
rcu_scheduler_active = 1, debug_locks = 1
1 lock held by kswapd0/31:
#0: (swap_lock){+.+.-.}, at: [<c05058bb>] swap_info_get+0x4b/0xd0
stack backtrace:
Pid: 31, comm: kswapd0 Not tainted 2.6.34-rc5-tip+ #13
Call Trace:
[<c083c5d6>] ? printk+0x1d/0x1f
[<c0480744>] lockdep_rcu_dereference+0x94/0xb0
[<c049d6ed>] css_id+0x5d/0x60
[<c05165a5>] mem_cgroup_uncharge_swapcache+0x45/0xa0
[<c0505e4f>] swapcache_free+0x3f/0x60
[<c04e79e2>] __remove_mapping+0xb2/0xf0
[<c04e7cbb>] shrink_page_list+0x26b/0x490
[<c047f85d>] ? put_lock_stats+0xd/0x30
[<c083fd67>] ? _raw_spin_unlock_irq+0x27/0x50
[<c0482566>] ? trace_hardirqs_on_caller+0xb6/0x220
[<c04e8158>] shrink_inactive_list+0x278/0x620
[<c04729e1>] ? sched_clock_cpu+0x121/0x180
[<c047e9b8>] ? trace_hardirqs_off_caller+0x18/0x130
[<c047eadb>] ? trace_hardirqs_off+0xb/0x10
[<c0843438>] ? sub_preempt_count+0x8/0x90
[<c047f85d>] ? put_lock_stats+0xd/0x30
[<c04e8704>] shrink_zone+0x204/0x3c0
[<c083fcac>] ? _raw_spin_unlock+0x2c/0x50
[<c04e951e>] kswapd+0x61e/0x7c0
[<c04e6ed0>] ? isolate_pages_global+0x0/0x1f0
[<c046bae0>] ? autoremove_wake_function+0x0/0x50
[<c04e8f00>] ? kswapd+0x0/0x7c0
[<c046b5e4>] kthread+0x74/0x80
[<c046b570>] ? kthread+0x0/0x80
[<c04035ba>] kernel_thread_helper+0x6/0x10
Reported-by: Li Zefan <[email protected]>
Cc: Daisuke Nishimura <[email protected]>
Cc: Balbir Singh <[email protected]>
Signed-off-by: KAMEZAWA Hiroyuki <[email protected]>
---
mm/memcontrol.c | 10 +++++++++-
1 file changed, 9 insertions(+), 1 deletion(-)
Index: linux-2.6.34-rc5-mm1/mm/memcontrol.c
===================================================================
--- linux-2.6.34-rc5-mm1.orig/mm/memcontrol.c
+++ linux-2.6.34-rc5-mm1/mm/memcontrol.c
@@ -838,10 +838,12 @@ int task_in_mem_cgroup(struct task_struc
* enabled in "curr" and "curr" is a child of "mem" in *cgroup*
* hierarchy(even if use_hierarchy is disabled in "mem").
*/
+ rcu_read_lock();
if (mem->use_hierarchy)
ret = css_is_ancestor(&curr->css, &mem->css);
else
ret = (curr == mem);
+ rcu_read_unlock();
css_put(&curr->css);
return ret;
}
@@ -1360,9 +1362,13 @@ static int memcg_oom_wake_function(wait_
* Both of oom_wait_info->mem and wake_mem are stable under us.
* Then we can use css_is_ancestor without taking care of RCU.
*/
+ rcu_read_lock();
if (!css_is_ancestor(&oom_wait_info->mem->css, &wake_mem->css) &&
- !css_is_ancestor(&wake_mem->css, &oom_wait_info->mem->css))
+ !css_is_ancestor(&wake_mem->css, &oom_wait_info->mem->css)) {
+ rcu_read_unlock();
return 0;
+ }
+ rcu_read_unlock();
wakeup:
return autoremove_wake_function(wait, mode, sync, arg);
@@ -2401,7 +2407,9 @@ mem_cgroup_uncharge_swapcache(struct pag
/* record memcg information */
if (do_swap_account && swapout && memcg) {
+ rcu_read_lock();
swap_cgroup_record(ent, css_id(&memcg->css));
+ rcu_read_unlock();
mem_cgroup_get(memcg);
}
if (swapout && memcg)
On Fri, 23 Apr 2010 11:55:16 +0800
Li Zefan <[email protected]> wrote:
> Li Zefan wrote:
> > KAMEZAWA Hiroyuki wrote:
> >> On Fri, 23 Apr 2010 11:00:41 +0800
> >> Li Zefan <[email protected]> wrote:
> >>
> >>> with CONFIG_PROVE_RCU=y, I saw this warning, it's because
> >>> css_id() is not under rcu_read_lock().
> >>>
> >> Ok. Thank you for reporting.
> >> This is ok ?
> >
> > Yes, and I did some more simple tests on memcg, no more warning
> > showed up.
> >
>
> oops, after trigging oom, I saw 2 more warnings:
>
ok, I will update. thank you.
On Fri, 23 Apr 2010 13:03:49 +0900, KAMEZAWA Hiroyuki <[email protected]> wrote:
> On Fri, 23 Apr 2010 12:58:14 +0900
> KAMEZAWA Hiroyuki <[email protected]> wrote:
>
> > On Fri, 23 Apr 2010 11:55:16 +0800
> > Li Zefan <[email protected]> wrote:
> >
> > > Li Zefan wrote:
> > > > KAMEZAWA Hiroyuki wrote:
> > > >> On Fri, 23 Apr 2010 11:00:41 +0800
> > > >> Li Zefan <[email protected]> wrote:
> > > >>
> > > >>> with CONFIG_PROVE_RCU=y, I saw this warning, it's because
> > > >>> css_id() is not under rcu_read_lock().
> > > >>>
> > > >> Ok. Thank you for reporting.
> > > >> This is ok ?
> > > >
> > > > Yes, and I did some more simple tests on memcg, no more warning
> > > > showed up.
> > > >
> > >
> > > oops, after trigging oom, I saw 2 more warnings:
> > >
> >
> > Thank you for good testing.
> v3 here...sorry too rapid posting...
>
Thank you for your report & patch.
(and I'm sorry that I've not been active these days ;( )
This patch looks good to me and, IIUC, would be enough to fix this bug.
Reviewed-by: Daisuke Nishimura <[email protected]>
BTW, it wouldn't cause any problem, I think former rcu_read_lock()/unlock()
in task_in_mem_cgroup() is unnecessary, because try_get_mem_cgroup_from_mm()
calls them for itself.
Thanks,
Daisuke Nishimura.
> ==
> From: KAMEZAWA Hiroyuki <[email protected]>
>
> css_id() should be called under rcu_read_lock().
> Following is a report from Li Zefan.
> ==
> ===================================================
> [ INFO: suspicious rcu_dereference_check() usage. ]
> ---------------------------------------------------
> kernel/cgroup.c:4438 invoked rcu_dereference_check() without protection!
>
> other info that might help us debug this:
>
>
> rcu_scheduler_active = 1, debug_locks = 1
> 1 lock held by kswapd0/31:
> #0: (swap_lock){+.+.-.}, at: [<c05058bb>] swap_info_get+0x4b/0xd0
>
> stack backtrace:
> Pid: 31, comm: kswapd0 Not tainted 2.6.34-rc5-tip+ #13
> Call Trace:
> [<c083c5d6>] ? printk+0x1d/0x1f
> [<c0480744>] lockdep_rcu_dereference+0x94/0xb0
> [<c049d6ed>] css_id+0x5d/0x60
> [<c05165a5>] mem_cgroup_uncharge_swapcache+0x45/0xa0
> [<c0505e4f>] swapcache_free+0x3f/0x60
> [<c04e79e2>] __remove_mapping+0xb2/0xf0
> [<c04e7cbb>] shrink_page_list+0x26b/0x490
> [<c047f85d>] ? put_lock_stats+0xd/0x30
> [<c083fd67>] ? _raw_spin_unlock_irq+0x27/0x50
> [<c0482566>] ? trace_hardirqs_on_caller+0xb6/0x220
> [<c04e8158>] shrink_inactive_list+0x278/0x620
> [<c04729e1>] ? sched_clock_cpu+0x121/0x180
> [<c047e9b8>] ? trace_hardirqs_off_caller+0x18/0x130
> [<c047eadb>] ? trace_hardirqs_off+0xb/0x10
> [<c0843438>] ? sub_preempt_count+0x8/0x90
> [<c047f85d>] ? put_lock_stats+0xd/0x30
> [<c04e8704>] shrink_zone+0x204/0x3c0
> [<c083fcac>] ? _raw_spin_unlock+0x2c/0x50
> [<c04e951e>] kswapd+0x61e/0x7c0
> [<c04e6ed0>] ? isolate_pages_global+0x0/0x1f0
> [<c046bae0>] ? autoremove_wake_function+0x0/0x50
> [<c04e8f00>] ? kswapd+0x0/0x7c0
> [<c046b5e4>] kthread+0x74/0x80
> [<c046b570>] ? kthread+0x0/0x80
> [<c04035ba>] kernel_thread_helper+0x6/0x10
>
> And css_is_ancestor() should be called under rcu_read_lock().
>
>
> Reported-by: Li Zefan <[email protected]>
> Cc: Daisuke Nishimura <[email protected]>
> Cc: Balbir Singh <[email protected]>
> Signed-off-by: KAMEZAWA Hiroyuki <[email protected]>
> ---
> mm/memcontrol.c | 18 ++++++++++++++++--
> 1 file changed, 16 insertions(+), 2 deletions(-)
>
> Index: linux-2.6.34-rc5-mm1/mm/memcontrol.c
> ===================================================================
> --- linux-2.6.34-rc5-mm1.orig/mm/memcontrol.c
> +++ linux-2.6.34-rc5-mm1/mm/memcontrol.c
> @@ -838,10 +838,12 @@ int task_in_mem_cgroup(struct task_struc
> * enabled in "curr" and "curr" is a child of "mem" in *cgroup*
> * hierarchy(even if use_hierarchy is disabled in "mem").
> */
> + rcu_read_lock();
> if (mem->use_hierarchy)
> ret = css_is_ancestor(&curr->css, &mem->css);
> else
> ret = (curr == mem);
> + rcu_read_unlock();
> css_put(&curr->css);
> return ret;
> }
> @@ -1360,9 +1362,13 @@ static int memcg_oom_wake_function(wait_
> * Both of oom_wait_info->mem and wake_mem are stable under us.
> * Then we can use css_is_ancestor without taking care of RCU.
> */
> + rcu_read_lock();
> if (!css_is_ancestor(&oom_wait_info->mem->css, &wake_mem->css) &&
> - !css_is_ancestor(&wake_mem->css, &oom_wait_info->mem->css))
> + !css_is_ancestor(&wake_mem->css, &oom_wait_info->mem->css)) {
> + rcu_read_unlock();
> return 0;
> + }
> + rcu_read_unlock();
>
> wakeup:
> return autoremove_wake_function(wait, mode, sync, arg);
> @@ -2401,7 +2407,9 @@ mem_cgroup_uncharge_swapcache(struct pag
>
> /* record memcg information */
> if (do_swap_account && swapout && memcg) {
> + rcu_read_lock();
> swap_cgroup_record(ent, css_id(&memcg->css));
> + rcu_read_unlock();
> mem_cgroup_get(memcg);
> }
> if (swapout && memcg)
> @@ -2458,8 +2466,10 @@ static int mem_cgroup_move_swap_account(
> {
> unsigned short old_id, new_id;
>
> + rcu_read_lock();
> old_id = css_id(&from->css);
> new_id = css_id(&to->css);
> + rcu_read_unlock();
>
> if (swap_cgroup_cmpxchg(entry, old_id, new_id) == old_id) {
> mem_cgroup_swap_statistics(from, false);
> @@ -4303,7 +4313,11 @@ static int is_target_pte_for_mc(struct v
> }
> /* Threre is a swap entry and a page doesn't exist or isn't charged */
> if (ent.val && !ret) {
> - if (css_id(&mc.from->css) == lookup_swap_cgroup(ent)) {
> + unsigned short id;
> + rcu_read_lock();
> + id = css_id(&mc.from->css);
> + rcu_read_unlock();
> + if (id == lookup_swap_cgroup(ent)) {
> ret = MC_TARGET_SWAP;
> if (target)
> target->ent = ent;
>
>
KAMEZAWA Hiroyuki wrote:
> On Fri, 23 Apr 2010 11:00:41 +0800
> Li Zefan <[email protected]> wrote:
>
>> with CONFIG_PROVE_RCU=y, I saw this warning, it's because
>> css_id() is not under rcu_read_lock().
>>
>
> Ok. Thank you for reporting.
> This is ok ?
Yes, and I did some more simple tests on memcg, no more warning
showed up.
> ==
> From: KAMEZAWA Hiroyuki <[email protected]>
>
KAMEZAWA Hiroyuki wrote:
> On Fri, 23 Apr 2010 11:55:16 +0800
> Li Zefan <[email protected]> wrote:
>
>> Li Zefan wrote:
>>> KAMEZAWA Hiroyuki wrote:
>>>> On Fri, 23 Apr 2010 11:00:41 +0800
>>>> Li Zefan <[email protected]> wrote:
>>>>
>>>>> with CONFIG_PROVE_RCU=y, I saw this warning, it's because
>>>>> css_id() is not under rcu_read_lock().
>>>>>
>>>> Ok. Thank you for reporting.
>>>> This is ok ?
>>> Yes, and I did some more simple tests on memcg, no more warning
>>> showed up.
>>>
>> oops, after trigging oom, I saw 2 more warnings:
>>
>
> ok, I will update. thank you.
>
one more:
===================================================
[ INFO: suspicious rcu_dereference_check() usage. ]
---------------------------------------------------
kernel/cgroup.c:4438 invoked rcu_dereference_check() without protection!
other info that might help us debug this:
rcu_scheduler_active = 1, debug_locks = 1
3 locks held by bash/2270:
#0: (cgroup_mutex){+.+.+.}, at: [<c049ab37>] cgroup_lock_live_group+0x17/0x30
#1: (&mm->mmap_sem){++++++}, at: [<c0517302>] mem_cgroup_can_attach+0xb2/0x130
#2: (&(&mm->page_table_lock)->rlock){+.+.-.}, at: [<c0513c23>] mem_cgroup_count_precharge_pte_range+0x93/0x130
stack backtrace:
Pid: 2270, comm: bash Not tainted 2.6.34-rc5-tip+ #14
Call Trace:
[<c083c636>] ? printk+0x1d/0x1f
[<c0480744>] lockdep_rcu_dereference+0x94/0xb0
[<c049d6ed>] css_id+0x5d/0x60
[<c051373f>] is_target_pte_for_mc+0x16f/0x1c0
[<c083f46b>] ? _raw_spin_lock+0x6b/0x80
[<c0513c4d>] mem_cgroup_count_precharge_pte_range+0xbd/0x130
[<c0513b90>] ? mem_cgroup_count_precharge_pte_range+0x0/0x130
[<c05030bd>] walk_page_range+0x25d/0x3f0
[<c0517344>] mem_cgroup_can_attach+0xf4/0x130
[<c0513b90>] ? mem_cgroup_count_precharge_pte_range+0x0/0x130
[<c0517250>] ? mem_cgroup_can_attach+0x0/0x130
[<c049e000>] cgroup_attach_task+0x70/0x280
[<c049e633>] cgroup_tasks_write+0x63/0x1c0
[<c049e660>] ? cgroup_tasks_write+0x90/0x1c0
[<c049d515>] cgroup_file_write+0x1f5/0x230
[<c0842f90>] ? do_page_fault+0x0/0x500
[<c047107b>] ? up_read+0x1b/0x30
[<c0843195>] ? do_page_fault+0x205/0x500
[<c051a8c4>] vfs_write+0xa4/0x1a0
[<c049d320>] ? cgroup_file_write+0x0/0x230
[<c051b3f6>] sys_write+0x46/0x70
[<c0403090>] sysenter_do_call+0x12/0x36
Li Zefan wrote:
> KAMEZAWA Hiroyuki wrote:
>> On Fri, 23 Apr 2010 11:00:41 +0800
>> Li Zefan <[email protected]> wrote:
>>
>>> with CONFIG_PROVE_RCU=y, I saw this warning, it's because
>>> css_id() is not under rcu_read_lock().
>>>
>> Ok. Thank you for reporting.
>> This is ok ?
>
> Yes, and I did some more simple tests on memcg, no more warning
> showed up.
>
oops, after trigging oom, I saw 2 more warnings:
===================================================
[ INFO: suspicious rcu_dereference_check() usage. ]
---------------------------------------------------
kernel/cgroup.c:4459 invoked rcu_dereference_check() without protection!
other info that might help us debug this:
rcu_scheduler_active = 1, debug_locks = 1
2 locks held by firefox/2258:
#0: (&mm->mmap_sem){++++++}, at: [<c0843090>] do_page_fault+0x100/0x500
#1: (tasklist_lock){.?.?.-}, at: [<c04df1ac>] mem_cgroup_out_of_memory+0x2c/0x90
stack backtrace:
Pid: 2258, comm: firefox Not tainted 2.6.34-rc5-tip+ #14
Call Trace:
[<c083c636>] ? printk+0x1d/0x1f
[<c0480744>] lockdep_rcu_dereference+0x94/0xb0
[<c049d61e>] css_is_ancestor+0xce/0xe0
[<c0517c41>] task_in_mem_cgroup+0xd1/0xf0
[<c0517b70>] ? task_in_mem_cgroup+0x0/0xf0
[<c04def10>] select_bad_process+0x70/0xe0
[<c04df1c1>] mem_cgroup_out_of_memory+0x41/0x90
[<c04826db>] ? trace_hardirqs_on+0xb/0x10
[<c05159e3>] mem_cgroup_handle_oom+0xf3/0x130
[<c046bae0>] ? autoremove_wake_function+0x0/0x50
[<c0516e01>] __mem_cgroup_try_charge+0x391/0x3d0
[<c047eadb>] ? trace_hardirqs_off+0xb/0x10
[<c05174c0>] mem_cgroup_charge_common+0x40/0x70
[<c0517620>] mem_cgroup_cache_charge+0x130/0x150
[<c04db6e7>] add_to_page_cache_locked+0x37/0x130
[<c04e5719>] ? __lru_cache_add+0x69/0xb0
[<c04db811>] add_to_page_cache_lru+0x31/0x80
[<c0549084>] mpage_readpages+0x84/0xf0
[<c057e4d0>] ? ext3_get_block+0x0/0x110
[<c057c760>] ? ext3_readpages+0x0/0x20
[<c057c77e>] ext3_readpages+0x1e/0x20
[<c057e4d0>] ? ext3_get_block+0x0/0x110
[<c04e4889>] __do_page_cache_readahead+0x219/0x2b0
[<c04e4748>] ? __do_page_cache_readahead+0xd8/0x2b0
[<c04e4946>] ra_submit+0x26/0x30
[<c04dcf86>] filemap_fault+0x436/0x470
[<c04f6a95>] __do_fault+0x55/0x550
[<c04f7afb>] handle_mm_fault+0x17b/0xad0
[<c0843090>] ? do_page_fault+0x100/0x500
[<c0842f90>] ? do_page_fault+0x0/0x500
[<c0843109>] do_page_fault+0x179/0x500
[<c04532b1>] ? __do_softirq+0x111/0x260
[<c045344f>] ? do_softirq+0x4f/0x70
[<c047ea65>] ? trace_hardirqs_off_caller+0xc5/0x130
[<c0840b0f>] ? error_code+0x67/0x70
[<c047ea14>] ? trace_hardirqs_off_caller+0x74/0x130
[<c0842f90>] ? do_page_fault+0x0/0x500
[<c0840b13>] error_code+0x6b/0x70
[<c0840000>] ? _raw_read_trylock+0x40/0x90
[<c0842f90>] ? do_page_fault+0x0/0x500
===================================================
[ INFO: suspicious rcu_dereference_check() usage. ]
---------------------------------------------------
kernel/cgroup.c:4460 invoked rcu_dereference_check() without protection!
other info that might help us debug this:
rcu_scheduler_active = 1, debug_locks = 1
2 locks held by firefox/2258:
#0: (&mm->mmap_sem){++++++}, at: [<c0843090>] do_page_fault+0x100/0x500
#1: (tasklist_lock){.?.?.-}, at: [<c04df1ac>] mem_cgroup_out_of_memory+0x2c/0x90
stack backtrace:
Pid: 2258, comm: firefox Not tainted 2.6.34-rc5-tip+ #14
Call Trace:
[<c083c636>] ? printk+0x1d/0x1f
[<c0480744>] lockdep_rcu_dereference+0x94/0xb0
[<c049d5e6>] css_is_ancestor+0x96/0xe0
[<c0517c41>] task_in_mem_cgroup+0xd1/0xf0
[<c0517b70>] ? task_in_mem_cgroup+0x0/0xf0
[<c04def10>] select_bad_process+0x70/0xe0
[<c04df1c1>] mem_cgroup_out_of_memory+0x41/0x90
[<c04826db>] ? trace_hardirqs_on+0xb/0x10
[<c05159e3>] mem_cgroup_handle_oom+0xf3/0x130
[<c046bae0>] ? autoremove_wake_function+0x0/0x50
[<c0516e01>] __mem_cgroup_try_charge+0x391/0x3d0
[<c047eadb>] ? trace_hardirqs_off+0xb/0x10
[<c05174c0>] mem_cgroup_charge_common+0x40/0x70
[<c0517620>] mem_cgroup_cache_charge+0x130/0x150
[<c04db6e7>] add_to_page_cache_locked+0x37/0x130
[<c04e5719>] ? __lru_cache_add+0x69/0xb0
[<c04db811>] add_to_page_cache_lru+0x31/0x80
[<c0549084>] mpage_readpages+0x84/0xf0
[<c057e4d0>] ? ext3_get_block+0x0/0x110
[<c057c760>] ? ext3_readpages+0x0/0x20
[<c057c77e>] ext3_readpages+0x1e/0x20
[<c057e4d0>] ? ext3_get_block+0x0/0x110
[<c04e4889>] __do_page_cache_readahead+0x219/0x2b0
[<c04e4748>] ? __do_page_cache_readahead+0xd8/0x2b0
[<c04e4946>] ra_submit+0x26/0x30
[<c04dcf86>] filemap_fault+0x436/0x470
[<c04f6a95>] __do_fault+0x55/0x550
[<c04f7afb>] handle_mm_fault+0x17b/0xad0
[<c0843090>] ? do_page_fault+0x100/0x500
[<c0842f90>] ? do_page_fault+0x0/0x500
[<c0843109>] do_page_fault+0x179/0x500
[<c04532b1>] ? __do_softirq+0x111/0x260
[<c045344f>] ? do_softirq+0x4f/0x70
[<c047ea65>] ? trace_hardirqs_off_caller+0xc5/0x130
[<c0840b0f>] ? error_code+0x67/0x70
[<c047ea14>] ? trace_hardirqs_off_caller+0x74/0x130
[<c0842f90>] ? do_page_fault+0x0/0x500
[<c0840b13>] error_code+0x6b/0x70
[<c0840000>] ? _raw_read_trylock+0x40/0x90
[<c0842f90>] ? do_page_fault+0x0/0x500
On Fri, 23 Apr 2010 12:58:14 +0900
KAMEZAWA Hiroyuki <[email protected]> wrote:
> On Fri, 23 Apr 2010 11:55:16 +0800
> Li Zefan <[email protected]> wrote:
>
> > Li Zefan wrote:
> > > KAMEZAWA Hiroyuki wrote:
> > >> On Fri, 23 Apr 2010 11:00:41 +0800
> > >> Li Zefan <[email protected]> wrote:
> > >>
> > >>> with CONFIG_PROVE_RCU=y, I saw this warning, it's because
> > >>> css_id() is not under rcu_read_lock().
> > >>>
> > >> Ok. Thank you for reporting.
> > >> This is ok ?
> > >
> > > Yes, and I did some more simple tests on memcg, no more warning
> > > showed up.
> > >
> >
> > oops, after trigging oom, I saw 2 more warnings:
> >
>
> Thank you for good testing.
v3 here...sorry too rapid posting...
==
From: KAMEZAWA Hiroyuki <[email protected]>
css_id() should be called under rcu_read_lock().
Following is a report from Li Zefan.
==
===================================================
[ INFO: suspicious rcu_dereference_check() usage. ]
---------------------------------------------------
kernel/cgroup.c:4438 invoked rcu_dereference_check() without protection!
other info that might help us debug this:
rcu_scheduler_active = 1, debug_locks = 1
1 lock held by kswapd0/31:
#0: (swap_lock){+.+.-.}, at: [<c05058bb>] swap_info_get+0x4b/0xd0
stack backtrace:
Pid: 31, comm: kswapd0 Not tainted 2.6.34-rc5-tip+ #13
Call Trace:
[<c083c5d6>] ? printk+0x1d/0x1f
[<c0480744>] lockdep_rcu_dereference+0x94/0xb0
[<c049d6ed>] css_id+0x5d/0x60
[<c05165a5>] mem_cgroup_uncharge_swapcache+0x45/0xa0
[<c0505e4f>] swapcache_free+0x3f/0x60
[<c04e79e2>] __remove_mapping+0xb2/0xf0
[<c04e7cbb>] shrink_page_list+0x26b/0x490
[<c047f85d>] ? put_lock_stats+0xd/0x30
[<c083fd67>] ? _raw_spin_unlock_irq+0x27/0x50
[<c0482566>] ? trace_hardirqs_on_caller+0xb6/0x220
[<c04e8158>] shrink_inactive_list+0x278/0x620
[<c04729e1>] ? sched_clock_cpu+0x121/0x180
[<c047e9b8>] ? trace_hardirqs_off_caller+0x18/0x130
[<c047eadb>] ? trace_hardirqs_off+0xb/0x10
[<c0843438>] ? sub_preempt_count+0x8/0x90
[<c047f85d>] ? put_lock_stats+0xd/0x30
[<c04e8704>] shrink_zone+0x204/0x3c0
[<c083fcac>] ? _raw_spin_unlock+0x2c/0x50
[<c04e951e>] kswapd+0x61e/0x7c0
[<c04e6ed0>] ? isolate_pages_global+0x0/0x1f0
[<c046bae0>] ? autoremove_wake_function+0x0/0x50
[<c04e8f00>] ? kswapd+0x0/0x7c0
[<c046b5e4>] kthread+0x74/0x80
[<c046b570>] ? kthread+0x0/0x80
[<c04035ba>] kernel_thread_helper+0x6/0x10
And css_is_ancestor() should be called under rcu_read_lock().
Reported-by: Li Zefan <[email protected]>
Cc: Daisuke Nishimura <[email protected]>
Cc: Balbir Singh <[email protected]>
Signed-off-by: KAMEZAWA Hiroyuki <[email protected]>
---
mm/memcontrol.c | 18 ++++++++++++++++--
1 file changed, 16 insertions(+), 2 deletions(-)
Index: linux-2.6.34-rc5-mm1/mm/memcontrol.c
===================================================================
--- linux-2.6.34-rc5-mm1.orig/mm/memcontrol.c
+++ linux-2.6.34-rc5-mm1/mm/memcontrol.c
@@ -838,10 +838,12 @@ int task_in_mem_cgroup(struct task_struc
* enabled in "curr" and "curr" is a child of "mem" in *cgroup*
* hierarchy(even if use_hierarchy is disabled in "mem").
*/
+ rcu_read_lock();
if (mem->use_hierarchy)
ret = css_is_ancestor(&curr->css, &mem->css);
else
ret = (curr == mem);
+ rcu_read_unlock();
css_put(&curr->css);
return ret;
}
@@ -1360,9 +1362,13 @@ static int memcg_oom_wake_function(wait_
* Both of oom_wait_info->mem and wake_mem are stable under us.
* Then we can use css_is_ancestor without taking care of RCU.
*/
+ rcu_read_lock();
if (!css_is_ancestor(&oom_wait_info->mem->css, &wake_mem->css) &&
- !css_is_ancestor(&wake_mem->css, &oom_wait_info->mem->css))
+ !css_is_ancestor(&wake_mem->css, &oom_wait_info->mem->css)) {
+ rcu_read_unlock();
return 0;
+ }
+ rcu_read_unlock();
wakeup:
return autoremove_wake_function(wait, mode, sync, arg);
@@ -2401,7 +2407,9 @@ mem_cgroup_uncharge_swapcache(struct pag
/* record memcg information */
if (do_swap_account && swapout && memcg) {
+ rcu_read_lock();
swap_cgroup_record(ent, css_id(&memcg->css));
+ rcu_read_unlock();
mem_cgroup_get(memcg);
}
if (swapout && memcg)
@@ -2458,8 +2466,10 @@ static int mem_cgroup_move_swap_account(
{
unsigned short old_id, new_id;
+ rcu_read_lock();
old_id = css_id(&from->css);
new_id = css_id(&to->css);
+ rcu_read_unlock();
if (swap_cgroup_cmpxchg(entry, old_id, new_id) == old_id) {
mem_cgroup_swap_statistics(from, false);
@@ -4303,7 +4313,11 @@ static int is_target_pte_for_mc(struct v
}
/* Threre is a swap entry and a page doesn't exist or isn't charged */
if (ent.val && !ret) {
- if (css_id(&mc.from->css) == lookup_swap_cgroup(ent)) {
+ unsigned short id;
+ rcu_read_lock();
+ id = css_id(&mc.from->css);
+ rcu_read_unlock();
+ if (id == lookup_swap_cgroup(ent)) {
ret = MC_TARGET_SWAP;
if (target)
target->ent = ent;
* KAMEZAWA Hiroyuki <[email protected]> [2010-04-23 12:14:24]:
> On Fri, 23 Apr 2010 11:00:41 +0800
> Li Zefan <[email protected]> wrote:
>
> > with CONFIG_PROVE_RCU=y, I saw this warning, it's because
> > css_id() is not under rcu_read_lock().
> >
>
> Ok. Thank you for reporting.
> This is ok ?
> ==
> From: KAMEZAWA Hiroyuki <[email protected]>
>
> css_id() should be called under rcu_read_lock().
> Following is a report from Li Zefan.
> ==
> ===================================================
> [ INFO: suspicious rcu_dereference_check() usage. ]
> ---------------------------------------------------
> kernel/cgroup.c:4438 invoked rcu_dereference_check() without protection!
>
> other info that might help us debug this:
>
>
> rcu_scheduler_active = 1, debug_locks = 1
> 1 lock held by kswapd0/31:
> #0: (swap_lock){+.+.-.}, at: [<c05058bb>] swap_info_get+0x4b/0xd0
>
> stack backtrace:
> Pid: 31, comm: kswapd0 Not tainted 2.6.34-rc5-tip+ #13
> Call Trace:
> [<c083c5d6>] ? printk+0x1d/0x1f
> [<c0480744>] lockdep_rcu_dereference+0x94/0xb0
> [<c049d6ed>] css_id+0x5d/0x60
> [<c05165a5>] mem_cgroup_uncharge_swapcache+0x45/0xa0
> [<c0505e4f>] swapcache_free+0x3f/0x60
> [<c04e79e2>] __remove_mapping+0xb2/0xf0
> [<c04e7cbb>] shrink_page_list+0x26b/0x490
> [<c047f85d>] ? put_lock_stats+0xd/0x30
> [<c083fd67>] ? _raw_spin_unlock_irq+0x27/0x50
> [<c0482566>] ? trace_hardirqs_on_caller+0xb6/0x220
> [<c04e8158>] shrink_inactive_list+0x278/0x620
> [<c04729e1>] ? sched_clock_cpu+0x121/0x180
> [<c047e9b8>] ? trace_hardirqs_off_caller+0x18/0x130
> [<c047eadb>] ? trace_hardirqs_off+0xb/0x10
> [<c0843438>] ? sub_preempt_count+0x8/0x90
> [<c047f85d>] ? put_lock_stats+0xd/0x30
> [<c04e8704>] shrink_zone+0x204/0x3c0
> [<c083fcac>] ? _raw_spin_unlock+0x2c/0x50
> [<c04e951e>] kswapd+0x61e/0x7c0
> [<c04e6ed0>] ? isolate_pages_global+0x0/0x1f0
> [<c046bae0>] ? autoremove_wake_function+0x0/0x50
> [<c04e8f00>] ? kswapd+0x0/0x7c0
> [<c046b5e4>] kthread+0x74/0x80
> [<c046b570>] ? kthread+0x0/0x80
> [<c04035ba>] kernel_thread_helper+0x6/0x10
>
> Reported-by: Li Zefan <[email protected]>
> Cc: Daisuke Nishimura <[email protected]>
> Cc: Balbir Singh <[email protected]>
> Signed-off-by: KAMEZAWA Hiroyuki <[email protected]>
> ---
> mm/memcontrol.c | 2 ++
> 1 file changed, 2 insertions(+)
>
> Index: linux-2.6.34-rc5-mm1/mm/memcontrol.c
> ===================================================================
> --- linux-2.6.34-rc5-mm1.orig/mm/memcontrol.c
> +++ linux-2.6.34-rc5-mm1/mm/memcontrol.c
> @@ -2401,7 +2401,9 @@ mem_cgroup_uncharge_swapcache(struct pag
>
> /* record memcg information */
> if (do_swap_account && swapout && memcg) {
> + rcu_read_lock();
> swap_cgroup_record(ent, css_id(&memcg->css));
> + rcu_read_unlock();
> mem_cgroup_get(memcg);
Excellent Catch!
Reviewed-by: Balbir Singh <[email protected]>
--
Three Cheers,
Balbir
On Fri, Apr 23, 2010 at 01:03:49PM +0900, KAMEZAWA Hiroyuki wrote:
> On Fri, 23 Apr 2010 12:58:14 +0900
> KAMEZAWA Hiroyuki <[email protected]> wrote:
>
> > On Fri, 23 Apr 2010 11:55:16 +0800
> > Li Zefan <[email protected]> wrote:
> >
> > > Li Zefan wrote:
> > > > KAMEZAWA Hiroyuki wrote:
> > > >> On Fri, 23 Apr 2010 11:00:41 +0800
> > > >> Li Zefan <[email protected]> wrote:
> > > >>
> > > >>> with CONFIG_PROVE_RCU=y, I saw this warning, it's because
> > > >>> css_id() is not under rcu_read_lock().
> > > >>>
> > > >> Ok. Thank you for reporting.
> > > >> This is ok ?
> > > >
> > > > Yes, and I did some more simple tests on memcg, no more warning
> > > > showed up.
> > > >
> > >
> > > oops, after trigging oom, I saw 2 more warnings:
> > >
> >
> > Thank you for good testing.
> v3 here...sorry too rapid posting...
>
> ==
> From: KAMEZAWA Hiroyuki <[email protected]>
I have queued this, thank you all!
However, memcg_oom_wake_function() does not yet exist in the tree
I am using, and is_target_pte_for_mc() has changed. I omitted the
hunk for memcg_oom_wake_function() and edited the hunk for
is_target_pte_for_mc().
I have queued this for others' testing, but if you would rather carry
this patch up the memcg path, please let me know and I will drop it.
Thanx, Paul
> css_id() should be called under rcu_read_lock().
> Following is a report from Li Zefan.
> ==
> ===================================================
> [ INFO: suspicious rcu_dereference_check() usage. ]
> ---------------------------------------------------
> kernel/cgroup.c:4438 invoked rcu_dereference_check() without protection!
>
> other info that might help us debug this:
>
>
> rcu_scheduler_active = 1, debug_locks = 1
> 1 lock held by kswapd0/31:
> #0: (swap_lock){+.+.-.}, at: [<c05058bb>] swap_info_get+0x4b/0xd0
>
> stack backtrace:
> Pid: 31, comm: kswapd0 Not tainted 2.6.34-rc5-tip+ #13
> Call Trace:
> [<c083c5d6>] ? printk+0x1d/0x1f
> [<c0480744>] lockdep_rcu_dereference+0x94/0xb0
> [<c049d6ed>] css_id+0x5d/0x60
> [<c05165a5>] mem_cgroup_uncharge_swapcache+0x45/0xa0
> [<c0505e4f>] swapcache_free+0x3f/0x60
> [<c04e79e2>] __remove_mapping+0xb2/0xf0
> [<c04e7cbb>] shrink_page_list+0x26b/0x490
> [<c047f85d>] ? put_lock_stats+0xd/0x30
> [<c083fd67>] ? _raw_spin_unlock_irq+0x27/0x50
> [<c0482566>] ? trace_hardirqs_on_caller+0xb6/0x220
> [<c04e8158>] shrink_inactive_list+0x278/0x620
> [<c04729e1>] ? sched_clock_cpu+0x121/0x180
> [<c047e9b8>] ? trace_hardirqs_off_caller+0x18/0x130
> [<c047eadb>] ? trace_hardirqs_off+0xb/0x10
> [<c0843438>] ? sub_preempt_count+0x8/0x90
> [<c047f85d>] ? put_lock_stats+0xd/0x30
> [<c04e8704>] shrink_zone+0x204/0x3c0
> [<c083fcac>] ? _raw_spin_unlock+0x2c/0x50
> [<c04e951e>] kswapd+0x61e/0x7c0
> [<c04e6ed0>] ? isolate_pages_global+0x0/0x1f0
> [<c046bae0>] ? autoremove_wake_function+0x0/0x50
> [<c04e8f00>] ? kswapd+0x0/0x7c0
> [<c046b5e4>] kthread+0x74/0x80
> [<c046b570>] ? kthread+0x0/0x80
> [<c04035ba>] kernel_thread_helper+0x6/0x10
>
> And css_is_ancestor() should be called under rcu_read_lock().
>
>
> Reported-by: Li Zefan <[email protected]>
> Cc: Daisuke Nishimura <[email protected]>
> Cc: Balbir Singh <[email protected]>
> Signed-off-by: KAMEZAWA Hiroyuki <[email protected]>
> ---
> mm/memcontrol.c | 18 ++++++++++++++++--
> 1 file changed, 16 insertions(+), 2 deletions(-)
>
> Index: linux-2.6.34-rc5-mm1/mm/memcontrol.c
> ===================================================================
> --- linux-2.6.34-rc5-mm1.orig/mm/memcontrol.c
> +++ linux-2.6.34-rc5-mm1/mm/memcontrol.c
> @@ -838,10 +838,12 @@ int task_in_mem_cgroup(struct task_struc
> * enabled in "curr" and "curr" is a child of "mem" in *cgroup*
> * hierarchy(even if use_hierarchy is disabled in "mem").
> */
> + rcu_read_lock();
> if (mem->use_hierarchy)
> ret = css_is_ancestor(&curr->css, &mem->css);
> else
> ret = (curr == mem);
> + rcu_read_unlock();
> css_put(&curr->css);
> return ret;
> }
> @@ -1360,9 +1362,13 @@ static int memcg_oom_wake_function(wait_
> * Both of oom_wait_info->mem and wake_mem are stable under us.
> * Then we can use css_is_ancestor without taking care of RCU.
> */
> + rcu_read_lock();
> if (!css_is_ancestor(&oom_wait_info->mem->css, &wake_mem->css) &&
> - !css_is_ancestor(&wake_mem->css, &oom_wait_info->mem->css))
> + !css_is_ancestor(&wake_mem->css, &oom_wait_info->mem->css)) {
> + rcu_read_unlock();
> return 0;
> + }
> + rcu_read_unlock();
>
> wakeup:
> return autoremove_wake_function(wait, mode, sync, arg);
> @@ -2401,7 +2407,9 @@ mem_cgroup_uncharge_swapcache(struct pag
>
> /* record memcg information */
> if (do_swap_account && swapout && memcg) {
> + rcu_read_lock();
> swap_cgroup_record(ent, css_id(&memcg->css));
> + rcu_read_unlock();
> mem_cgroup_get(memcg);
> }
> if (swapout && memcg)
> @@ -2458,8 +2466,10 @@ static int mem_cgroup_move_swap_account(
> {
> unsigned short old_id, new_id;
>
> + rcu_read_lock();
> old_id = css_id(&from->css);
> new_id = css_id(&to->css);
> + rcu_read_unlock();
>
> if (swap_cgroup_cmpxchg(entry, old_id, new_id) == old_id) {
> mem_cgroup_swap_statistics(from, false);
> @@ -4303,7 +4313,11 @@ static int is_target_pte_for_mc(struct v
> }
> /* Threre is a swap entry and a page doesn't exist or isn't charged */
> if (ent.val && !ret) {
> - if (css_id(&mc.from->css) == lookup_swap_cgroup(ent)) {
> + unsigned short id;
> + rcu_read_lock();
> + id = css_id(&mc.from->css);
> + rcu_read_unlock();
> + if (id == lookup_swap_cgroup(ent)) {
> ret = MC_TARGET_SWAP;
> if (target)
> target->ent = ent;
>
>
On Fri, 23 Apr 2010 12:34:06 -0700
"Paul E. McKenney" <[email protected]> wrote:
> On Fri, Apr 23, 2010 at 01:03:49PM +0900, KAMEZAWA Hiroyuki wrote:
> > On Fri, 23 Apr 2010 12:58:14 +0900
> > KAMEZAWA Hiroyuki <[email protected]> wrote:
> >
> > > On Fri, 23 Apr 2010 11:55:16 +0800
> > > Li Zefan <[email protected]> wrote:
> > >
> > > > Li Zefan wrote:
> > > > > KAMEZAWA Hiroyuki wrote:
> > > > >> On Fri, 23 Apr 2010 11:00:41 +0800
> > > > >> Li Zefan <[email protected]> wrote:
> > > > >>
> > > > >>> with CONFIG_PROVE_RCU=y, I saw this warning, it's because
> > > > >>> css_id() is not under rcu_read_lock().
> > > > >>>
> > > > >> Ok. Thank you for reporting.
> > > > >> This is ok ?
> > > > >
> > > > > Yes, and I did some more simple tests on memcg, no more warning
> > > > > showed up.
> > > > >
> > > >
> > > > oops, after trigging oom, I saw 2 more warnings:
> > > >
> > >
> > > Thank you for good testing.
> > v3 here...sorry too rapid posting...
> >
> > ==
> > From: KAMEZAWA Hiroyuki <[email protected]>
>
> I have queued this, thank you all!
>
> However, memcg_oom_wake_function() does not yet exist in the tree
> I am using, and is_target_pte_for_mc() has changed. I omitted the
> hunk for memcg_oom_wake_function() and edited the hunk for
> is_target_pte_for_mc().
>
Ok, memcg_oom_wake_function is for -mm. I'll prepare another patch for -mm.
> I have queued this for others' testing, but if you would rather carry
> this patch up the memcg path, please let me know and I will drop it.
>
I think it's ok to be fixed by your tree. I'll look at memcg later and
fix remaining things.
Thanks,
-Kame
On Sat, Apr 24, 2010 at 11:08:05AM +0900, KAMEZAWA Hiroyuki wrote:
> On Fri, 23 Apr 2010 12:34:06 -0700
> "Paul E. McKenney" <[email protected]> wrote:
>
> > On Fri, Apr 23, 2010 at 01:03:49PM +0900, KAMEZAWA Hiroyuki wrote:
> > > On Fri, 23 Apr 2010 12:58:14 +0900
> > > KAMEZAWA Hiroyuki <[email protected]> wrote:
> > >
> > > > On Fri, 23 Apr 2010 11:55:16 +0800
> > > > Li Zefan <[email protected]> wrote:
> > > >
> > > > > Li Zefan wrote:
> > > > > > KAMEZAWA Hiroyuki wrote:
> > > > > >> On Fri, 23 Apr 2010 11:00:41 +0800
> > > > > >> Li Zefan <[email protected]> wrote:
> > > > > >>
> > > > > >>> with CONFIG_PROVE_RCU=y, I saw this warning, it's because
> > > > > >>> css_id() is not under rcu_read_lock().
> > > > > >>>
> > > > > >> Ok. Thank you for reporting.
> > > > > >> This is ok ?
> > > > > >
> > > > > > Yes, and I did some more simple tests on memcg, no more warning
> > > > > > showed up.
> > > > > >
> > > > >
> > > > > oops, after trigging oom, I saw 2 more warnings:
> > > > >
> > > >
> > > > Thank you for good testing.
> > > v3 here...sorry too rapid posting...
> > >
> > > ==
> > > From: KAMEZAWA Hiroyuki <[email protected]>
> >
> > I have queued this, thank you all!
> >
> > However, memcg_oom_wake_function() does not yet exist in the tree
> > I am using, and is_target_pte_for_mc() has changed. I omitted the
> > hunk for memcg_oom_wake_function() and edited the hunk for
> > is_target_pte_for_mc().
> >
> Ok, memcg_oom_wake_function is for -mm. I'll prepare another patch for -mm.
>
>
> > I have queued this for others' testing, but if you would rather carry
> > this patch up the memcg path, please let me know and I will drop it.
> >
> I think it's ok to be fixed by your tree. I'll look at memcg later and
> fix remaining things.
Sounds good!
Thanx, Paul