Li Zefan wrote:
> Balbir Singh wrote:
>> YAMAMOTO Takashi wrote:
>>>> Li Zefan wrote:
>>>>> No need for VM_BUG_ON(pc), since 'pc' is the list entry. This should
>>>>> be VM_BUG_ON(page).
>>>>>
>>>>> Signed-off-by: Li Zefan <lizf@cn.fujitsu.com>
>>>>> Acked-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
>>>> pc is of type page_cgroup and we use list_for_each_entry_safe_reverse. Not sure
>>>> why we can't bug on pc.
>>> pc is dereferenced before this VM_BUG_ON.
>>>
>>> YAMAMOTO Takashi
>>>
>> OK, so the VM_BUG_ON needs to move to an earlier location. Agreed.
>>
> 
> No, 'pc' has been dereferenced in list_for_each_entry_safe_reverse().
> 
> 
> #define list_for_each_entry_safe_reverse(pos, n, head, member)		\
> 	for (pos = list_entry((head)->prev, typeof(*pos), member),	\
> 		n = list_entry(pos->member.prev, typeof(*pos), member);	\
>                                ^^^^^^^^^^^
> 	     &pos->member != (head); 					\
>               ^^^^^^^^^^^
> 	     pos = n, n = list_entry(n->member.prev, typeof(*n), member))
> 

Hmm.. We used to have a for loop with !list_empty() as a termination condition
and VM_BUG_ON(!pc) is a spill over. With the new loop, VM_BUG_ON(!pc) does not
make sense.


-- 
	Warm Regards,
	Balbir Singh
	Linux Technology Center
	IBM, ISTL
_______________________________________________
Containers mailing list
Containers@lists.linux-foundation.org
https://lists.linux-foundation.org/mailman/listinfo/containers

Balbir Singh wrote:
> Li Zefan wrote:
>> Balbir Singh wrote:
>>> YAMAMOTO Takashi wrote:
>>>>> Li Zefan wrote:
>>>>>> No need for VM_BUG_ON(pc), since 'pc' is the list entry. This should
>>>>>> be VM_BUG_ON(page).
>>>>>>
>>>>>> Signed-off-by: Li Zefan <lizf@cn.fujitsu.com>
>>>>>> Acked-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
>>>>> pc is of type page_cgroup and we use list_for_each_entry_safe_reverse. Not sure
>>>>> why we can't bug on pc.
>>>> pc is dereferenced before this VM_BUG_ON.
>>>>
>>>> YAMAMOTO Takashi
>>>>
>>> OK, so the VM_BUG_ON needs to move to an earlier location. Agreed.
>>>
>> No, 'pc' has been dereferenced in list_for_each_entry_safe_reverse().
>>
>>
>> #define list_for_each_entry_safe_reverse(pos, n, head, member)		\
>> 	for (pos = list_entry((head)->prev, typeof(*pos), member),	\
>> 		n = list_entry(pos->member.prev, typeof(*pos), member);	\
>>                                ^^^^^^^^^^^
>> 	     &pos->member != (head); 					\
>>               ^^^^^^^^^^^
>> 	     pos = n, n = list_entry(n->member.prev, typeof(*n), member))
>>
> 
> Hmm.. We used to have a for loop with !list_empty() as a termination condition
> and VM_BUG_ON(!pc) is a spill over. With the new loop, VM_BUG_ON(!pc) does not
> make sense.
> 
> 

I see, and I'll post a new patch to just remove it.
_______________________________________________
Containers mailing list
Containers@lists.linux-foundation.org
https://lists.linux-foundation.org/mailman/listinfo/containers

Remove this VM_BUG_ON(), as Balbir stated:

We used to have a for loop with !list_empty() as a termination condition
and VM_BUG_ON(!pc) is a spill over. With the new loop, VM_BUG_ON(!pc) does not
make sense.

Signed-off-by: Li Zefan <lizf@cn.fujitsu.com>
---
 mm/memcontrol.c |    1 -
 1 files changed, 0 insertions(+), 1 deletions(-)

diff --git a/mm/memcontrol.c b/mm/memcontrol.c
index 718acf8..631002d 100644
--- a/mm/memcontrol.c
+++ b/mm/memcontrol.c
@@ -534,7 +534,6 @@ unsigned long mem_cgroup_isolate_pages(unsigned long nr_to_scan,
 		if (scan >= nr_to_scan)
 			break;
 		page = pc->page;
-		VM_BUG_ON(!pc);
 
 		if (unlikely(!PageLRU(page)))
 			continue;
-- 
1.5.4.rc3
_______________________________________________
Containers mailing list
Containers@lists.linux-foundation.org
https://lists.linux-foundation.org/mailman/listinfo/containers