Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > linux.kernel > #1395592 > unrolled thread

[PATCH v2] ksm: fix conflict between mmput and scan_get_next_rmap_item

Back to article view | Back to linux.kernel

Contents

  [PATCH v2] ksm: fix conflict between mmput and scan_get_next_rmap_item Zhou Chengming <zhouchengming1@huawei.com> - 2016-05-06 05:30 +0200
    Re: [PATCH v2] ksm: fix conflict between mmput and  scan_get_next_rmap_item Andrea Arcangeli <aarcange@redhat.com> - 2016-05-06 16:30 +0200
      Re: [PATCH v2] ksm: fix conflict between mmput and scan_get_next_rmap_item zhouchengming <zhouchengming1@huawei.com> - 2016-05-08 08:10 +0200

#1395592 — [PATCH v2] ksm: fix conflict between mmput and scan_get_next_rmap_item

A concurrency issue about KSM in the function scan_get_next_rmap_item.

task A (ksmd):				|task B (the mm's task):
					|
mm = slot->mm;				|
down_read(&mm->mmap_sem);		|
					|
...					|
					|
spin_lock(&ksm_mmlist_lock);		|
					|
ksm_scan.mm_slot go to the next slot;	|
					|
spin_unlock(&ksm_mmlist_lock);		|
					|mmput() ->
					|	ksm_exit():
					|
					|spin_lock(&ksm_mmlist_lock);
					|if (mm_slot && ksm_scan.mm_slot != mm_slot) {
					|	if (!mm_slot->rmap_list) {
					|		easy_to_free = 1;
					|		...
					|
					|if (easy_to_free) {
					|	mmdrop(mm);
					|	...
					|
					|So this mm_struct will be freed successfully.
					|
up_read(&mm->mmap_sem);			|

As we can see above, the ksmd thread may access a mm_struct that already
been freed to the kmem_cache.
Suppose a fork will get this mm_struct from the kmem_cache, the ksmd thread
then call up_read(&mm->mmap_sem), will cause mmap_sem.count to become -1.
From the suggestion of Andrea Arcangeli, unmerge_and_remove_all_rmap_items
has the same SMP race condition, so fix it too. My prev fix in function
scan_get_next_rmap_item will introduce a different SMP race condition,
so just invert the up_read/spin_unlock order as Andrea Arcangeli said.

Signed-off-by: Zhou Chengming <zhouchengming1@huawei.com>
Suggested-by: Andrea Arcangeli <aarcange@redhat.com>
---
 mm/ksm.c |   17 ++++++++++-------
 1 files changed, 10 insertions(+), 7 deletions(-)

diff --git a/mm/ksm.c b/mm/ksm.c
index ca6d2a0..d87bafc 100644
--- a/mm/ksm.c
+++ b/mm/ksm.c
@@ -777,6 +777,7 @@ static int unmerge_and_remove_all_rmap_items(void)
 		}
 
 		remove_trailing_rmap_items(mm_slot, &mm_slot->rmap_list);
+		up_read(&mm->mmap_sem);
 
 		spin_lock(&ksm_mmlist_lock);
 		ksm_scan.mm_slot = list_entry(mm_slot->mm_list.next,
@@ -784,16 +785,12 @@ static int unmerge_and_remove_all_rmap_items(void)
 		if (ksm_test_exit(mm)) {
 			hash_del(&mm_slot->link);
 			list_del(&mm_slot->mm_list);
-			spin_unlock(&ksm_mmlist_lock);
 
 			free_mm_slot(mm_slot);
 			clear_bit(MMF_VM_MERGEABLE, &mm->flags);
-			up_read(&mm->mmap_sem);
 			mmdrop(mm);
-		} else {
-			spin_unlock(&ksm_mmlist_lock);
-			up_read(&mm->mmap_sem);
 		}
+		spin_unlock(&ksm_mmlist_lock);
 	}
 
 	/* Clean up stable nodes, but don't worry if some are still busy */
@@ -1650,16 +1647,22 @@ next_mm:
 		 */
 		hash_del(&slot->link);
 		list_del(&slot->mm_list);
-		spin_unlock(&ksm_mmlist_lock);
 
 		free_mm_slot(slot);
 		clear_bit(MMF_VM_MERGEABLE, &mm->flags);
 		up_read(&mm->mmap_sem);
 		mmdrop(mm);
 	} else {
-		spin_unlock(&ksm_mmlist_lock);
 		up_read(&mm->mmap_sem);
 	}
+	/*
+	 * up_read(&mm->mmap_sem) first because after
+	 * spin_unlock(&ksm_mmlist_lock) run, the "mm" may
+	 * already have been freed under us by __ksm_exit()
+	 * because the "mm_slot" is still hashed and
+	 * ksm_scan.mm_slot doesn't point to it anymore.
+	 */
+	spin_unlock(&ksm_mmlist_lock);
 
 	/* Repeat until we've completed scanning the whole list */
 	slot = ksm_scan.mm_slot;
-- 
1.7.7

[toc] | [next] | [standalone]

#1395893 — Re: [PATCH v2] ksm: fix conflict between mmput and scan_get_next_rmap_item

On Fri, May 06, 2016 at 11:27:36AM +0800, Zhou Chengming wrote:
> @@ -1650,16 +1647,22 @@ next_mm:
>  		 */
>  		hash_del(&slot->link);
>  		list_del(&slot->mm_list);
> -		spin_unlock(&ksm_mmlist_lock);
>  
>  		free_mm_slot(slot);
>  		clear_bit(MMF_VM_MERGEABLE, &mm->flags);
>  		up_read(&mm->mmap_sem);
>  		mmdrop(mm);
>  	} else {
> -		spin_unlock(&ksm_mmlist_lock);
>  		up_read(&mm->mmap_sem);
>  	}
> +	/*
> +	 * up_read(&mm->mmap_sem) first because after
> +	 * spin_unlock(&ksm_mmlist_lock) run, the "mm" may
> +	 * already have been freed under us by __ksm_exit()
> +	 * because the "mm_slot" is still hashed and
> +	 * ksm_scan.mm_slot doesn't point to it anymore.
> +	 */
> +	spin_unlock(&ksm_mmlist_lock);
>  
>  	/* Repeat until we've completed scanning the whole list */
>  	slot = ksm_scan.mm_slot;

Reviewed-by: Andrea Arcangeli <aarcange@redhat.com>

While the above patch is correct, I would however prefer if you could
update it to keep releasing the ksm_mmlist_lock as before (I'm talking
only about the quoted part, not the other one not quoted), because
it's "strictier" and it better documents that it's only needed up
until:

  		hash_del(&slot->link);
  		list_del(&slot->mm_list);

It should be also a bit more scalable but to me this is just about
keeping implicit documentation on the locking by keeping it strict.

The fact up_read happens exactly after clear_bit also avoided me to
overlook that it was really needed, same thing with the
ksm_mmlist_lock after list_del, I'd like to keep it there and just
invert the order of spin_unlock; up_read in the else branch.

That should be enough because after hash_del get_mm_slot will return
NULL so the mmdrop will not happen anymore in __ksm_exit, this is
further explicit by the code doing mmdrop itself just after
up_read.

The SMP race condition is fixed by just the two liner that reverse the
order of spin_unlock; up_read without increasing the size of the
spinlock critical section for the ksm_scan.address == 0 case. This is
also why it wasn't reproducible because it's about 1 instruction window.

Thanks!
Andrea

[toc] | [prev] | [next] | [standalone]

#1396382

On 2016/5/6 22:24, Andrea Arcangeli wrote:
> On Fri, May 06, 2016 at 11:27:36AM +0800, Zhou Chengming wrote:
>> @@ -1650,16 +1647,22 @@ next_mm:
>>   		 */
>>   		hash_del(&slot->link);
>>   		list_del(&slot->mm_list);
>> -		spin_unlock(&ksm_mmlist_lock);
>>
>>   		free_mm_slot(slot);
>>   		clear_bit(MMF_VM_MERGEABLE,&mm->flags);
>>   		up_read(&mm->mmap_sem);
>>   		mmdrop(mm);
>>   	} else {
>> -		spin_unlock(&ksm_mmlist_lock);
>>   		up_read(&mm->mmap_sem);
>>   	}
>> +	/*
>> +	 * up_read(&mm->mmap_sem) first because after
>> +	 * spin_unlock(&ksm_mmlist_lock) run, the "mm" may
>> +	 * already have been freed under us by __ksm_exit()
>> +	 * because the "mm_slot" is still hashed and
>> +	 * ksm_scan.mm_slot doesn't point to it anymore.
>> +	 */
>> +	spin_unlock(&ksm_mmlist_lock);
>>
>>   	/* Repeat until we've completed scanning the whole list */
>>   	slot = ksm_scan.mm_slot;
>
> Reviewed-by: Andrea Arcangeli<aarcange@redhat.com>
>
> While the above patch is correct, I would however prefer if you could
> update it to keep releasing the ksm_mmlist_lock as before (I'm talking
> only about the quoted part, not the other one not quoted), because
> it's "strictier" and it better documents that it's only needed up
> until:
>
>    		hash_del(&slot->link);
>    		list_del(&slot->mm_list);
>
> It should be also a bit more scalable but to me this is just about
> keeping implicit documentation on the locking by keeping it strict.
>
> The fact up_read happens exactly after clear_bit also avoided me to
> overlook that it was really needed, same thing with the
> ksm_mmlist_lock after list_del, I'd like to keep it there and just
> invert the order of spin_unlock; up_read in the else branch.

Thanks a lot for your review and comment. It's my fault to misunderstand
your last reply. Yes it's better and more scalable to just invert the 
order of spin_unlock/up_read in the else branch. And it's also enough.

Thanks!
>
> That should be enough because after hash_del get_mm_slot will return
> NULL so the mmdrop will not happen anymore in __ksm_exit, this is
> further explicit by the code doing mmdrop itself just after
> up_read.
>
> The SMP race condition is fixed by just the two liner that reverse the
> order of spin_unlock; up_read without increasing the size of the
> spinlock critical section for the ksm_scan.address == 0 case. This is
> also why it wasn't reproducible because it's about 1 instruction window.
>
> Thanks!
> Andrea
>
> .
>

[toc] | [prev] | [standalone]

Back to top | Article view | linux.kernel

csiph-web