From: Madhuparna Bhowmik <[email protected]>
list_for_each_entry_rcu has built-in RCU and lock checking.
Pass cond argument to list_for_each_entry_rcu() to silence
false lockdep warning when CONFIG_PROVE_RCU_LIST is enabled
by default.
Even though the function css_next_child() already checks if
cgroup_mutex or rcu_read_lock() is held using
cgroup_assert_mutex_or_rcu_locked(), there is a need to pass
cond to list_for_each_entry_rcu() to avoid false positive
lockdep warning.
Signed-off-by: Madhuparna Bhowmik <[email protected]>
---
kernel/cgroup/cgroup.c | 3 ++-
1 file changed, 2 insertions(+), 1 deletion(-)
diff --git a/kernel/cgroup/cgroup.c b/kernel/cgroup/cgroup.c
index 735af8f15f95..c2959764ad95 100644
--- a/kernel/cgroup/cgroup.c
+++ b/kernel/cgroup/cgroup.c
@@ -4152,7 +4152,8 @@ struct cgroup_subsys_state *css_next_child(struct cgroup_subsys_state *pos,
} else if (likely(!(pos->flags & CSS_RELEASED))) {
next = list_entry_rcu(pos->sibling.next, struct cgroup_subsys_state, sibling);
} else {
- list_for_each_entry_rcu(next, &parent->children, sibling)
+ list_for_each_entry_rcu(next, &parent->children, sibling,
+ lockdep_is_held(&cgroup_mutex))
if (next->serial_nr > pos->serial_nr)
break;
}
--
2.17.1
Hello.
On Sat, Jan 18, 2020 at 08:40:51AM +0530, [email protected] wrote:
> From: Madhuparna Bhowmik <[email protected]>
>
> list_for_each_entry_rcu has built-in RCU and lock checking.
> Pass cond argument to list_for_each_entry_rcu() to silence
> false lockdep warning when CONFIG_PROVE_RCU_LIST is enabled
> by default.
I assume if you've seen the RCU warning, you haven't seen the warning
from cgroup_assert_mutex_or_rcu_locked() above.
The patch makes sense to me from the consistency POV.
Acked-by: Michal Koutn? <[email protected]>
Michal
On Wed, Jan 29, 2020 at 03:22:55PM +0100, Michal Koutn? wrote:
> Hello.
>
> On Sat, Jan 18, 2020 at 08:40:51AM +0530, [email protected] wrote:
> > From: Madhuparna Bhowmik <[email protected]>
> >
> > list_for_each_entry_rcu has built-in RCU and lock checking.
> > Pass cond argument to list_for_each_entry_rcu() to silence
> > false lockdep warning when CONFIG_PROVE_RCU_LIST is enabled
> > by default.
> I assume if you've seen the RCU warning, you haven't seen the warning
> from cgroup_assert_mutex_or_rcu_locked() above.
>
No, I haven't seen any warning from cgroup_assert_mutex_or_rcu_locked(),
I am just doing the conversions to prevent any false lockdep warnings
because of CONFIG_PROVE_RCU_LIST in the future.
> The patch makes sense to me from the consistency POV.
>
Thank you,
Madhuparna
> Acked-by: Michal Koutn? <[email protected]>
>
> Michal
On Sat, Jan 18, 2020 at 08:40:51AM +0530, [email protected] wrote:
> From: Madhuparna Bhowmik <[email protected]>
>
> list_for_each_entry_rcu has built-in RCU and lock checking.
> Pass cond argument to list_for_each_entry_rcu() to silence
> false lockdep warning when CONFIG_PROVE_RCU_LIST is enabled
> by default.
>
> Even though the function css_next_child() already checks if
> cgroup_mutex or rcu_read_lock() is held using
> cgroup_assert_mutex_or_rcu_locked(), there is a need to pass
> cond to list_for_each_entry_rcu() to avoid false positive
> lockdep warning.
>
> Signed-off-by: Madhuparna Bhowmik <[email protected]>
Applied to cgroup/for-5.7.
Thanks.
--
tejun