[PATCH] mm: memcontrol: remove BUG_ON in uncharge_list

[PATCH] mm: memcontrol: remove BUG_ON in uncharge_list

[roy.qing.li-Re5JQEeQqe8AvxtiuMwx3w]

From: Li RongQing <roy.qing.li-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org>

when call uncharge_list, if a page is transparent huge, and not need to
BUG_ON about non-transparent huge, since nobody should be be seeing the
page at this stage and this page cannot be raced with a THP split up

Signed-off-by: Li RongQing <roy.qing.li-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org>
---
 mm/memcontrol.c | 1 -
 1 file changed, 1 deletion(-)

diff --git a/mm/memcontrol.c b/mm/memcontrol.c
index 4d9a215..d7a56f1 100644
--- a/mm/memcontrol.c
+++ b/mm/memcontrol.c
@@ -5457,7 +5457,6 @@ static void uncharge_list(struct list_head *page_list)
 
 		if (PageTransHuge(page)) {
 			nr_pages <<= compound_order(page);
-			VM_BUG_ON_PAGE(!PageTransHuge(page), page);
 			nr_huge += nr_pages;
 		}
 
-- 
2.1.4

Re: [PATCH] mm: memcontrol: remove BUG_ON in uncharge_list

[Michal Hocko]

On Wed 08-06-16 15:00:48, roy.qing.li-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org wrote:
> From: Li RongQing <roy.qing.li-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org>
> 
> when call uncharge_list, if a page is transparent huge, and not need to
> BUG_ON about non-transparent huge, since nobody should be be seeing the
> page at this stage and this page cannot be raced with a THP split up

Johannes do you remember why you have kept this bug on even after
0a31bc97c80c ("mm: memcontrol: rewrite uncharge API")?

> Signed-off-by: Li RongQing <roy.qing.li-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org>
> ---
>  mm/memcontrol.c | 1 -
>  1 file changed, 1 deletion(-)
> 
> diff --git a/mm/memcontrol.c b/mm/memcontrol.c
> index 4d9a215..d7a56f1 100644
> --- a/mm/memcontrol.c
> +++ b/mm/memcontrol.c
> @@ -5457,7 +5457,6 @@ static void uncharge_list(struct list_head *page_list)
>  
>  		if (PageTransHuge(page)) {
>  			nr_pages <<= compound_order(page);
> -			VM_BUG_ON_PAGE(!PageTransHuge(page), page);
>  			nr_huge += nr_pages;
>  		}
>  
> -- 
> 2.1.4

-- 
Michal Hocko
SUSE Labs

Re: [PATCH] mm: memcontrol: remove BUG_ON in uncharge_list

[Johannes Weiner]

On Wed, Jun 08, 2016 at 09:25:54AM +0200, Michal Hocko wrote:
> On Wed 08-06-16 15:00:48, roy.qing.li@gmail.com wrote:
> > From: Li RongQing <roy.qing.li@gmail.com>
> > 
> > when call uncharge_list, if a page is transparent huge, and not need to
> > BUG_ON about non-transparent huge, since nobody should be be seeing the
> > page at this stage and this page cannot be raced with a THP split up
> 
> Johannes do you remember why you have kept this bug on even after
> 0a31bc97c80c ("mm: memcontrol: rewrite uncharge API")?

No, I just copied it over without thinking too much about it. That
patch was pretty drastic, so why not keep the sanity checks in case
one of the many assumptions it made were flawed...

But it's probably okay to drop it at this point.

> > Signed-off-by: Li RongQing <roy.qing.li@gmail.com>

Acked-by: Johannes Weiner <hannes@cmpxchg.org>

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>

Re: [PATCH] mm: memcontrol: remove BUG_ON in uncharge_list

[Michal Hocko]

On Wed 08-06-16 10:58:33, Johannes Weiner wrote:
> On Wed, Jun 08, 2016 at 09:25:54AM +0200, Michal Hocko wrote:
> > On Wed 08-06-16 15:00:48, roy.qing.li-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org wrote:
> > > From: Li RongQing <roy.qing.li-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org>
> > > 
> > > when call uncharge_list, if a page is transparent huge, and not need to
> > > BUG_ON about non-transparent huge, since nobody should be be seeing the
> > > page at this stage and this page cannot be raced with a THP split up
> > 
> > Johannes do you remember why you have kept this bug on even after
> > 0a31bc97c80c ("mm: memcontrol: rewrite uncharge API")?
> 
> No, I just copied it over without thinking too much about it. That
> patch was pretty drastic, so why not keep the sanity checks in case
> one of the many assumptions it made were flawed...

Yeah the VM_BUG_ON made sense before because the charge could have gone
away while the page could be visible, thus split.
 
> But it's probably okay to drop it at this point.

Thanks for double checking. I was just not sure I didn't miss anything.
I am always a bit nervous when bug_ons a removed.

> > > Signed-off-by: Li RongQing <roy.qing.li-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org>
> 
> Acked-by: Johannes Weiner <hannes-druUgvl0LCNAfugRpC6u6w@public.gmane.org>

With the reference to the rewrite which made it pointless
Acked-by: Michal Hocko <mhocko-IBi9RG/b67k@public.gmane.org>
-- 
Michal Hocko
SUSE Labs