Re: [PATCH 1/2] mm,migration: Prevent rmap_walk_[anon|ksm] seeing the wrong VMA information

[Minchan Kim]

On Thu, May 6, 2010 at 11:06 PM, Linus Torvalds
<torvalds@xxxxxxxxxxxxxxxxxxxx> wrote:
>
>
> On Thu, 6 May 2010, Minchan Kim wrote:
>> > + Â Â Â Â*/
>> > + Â Â Â avc = list_first_entry(&anon_vma->head, struct anon_vma_chain, same_anon_vma);
>>
>> Dumb question.
>>
>> I can't understand why we should use list_first_entry.
>
> It's not that we "should" use list_entry_first. It's that we want to find
> _any_ entry on the list, and the most natural one is the first one.
>
> So we could take absolutely any 'avc' entry that is reachable from the
> anon_vma, and use that to look up _any_ 'vma' that is associated with that
> anon_vma. And then, from _any_ of those vma's, we know how to get to the
> "root anon_vma" - the one that they are all associated with.
>
> So no, there's absolutely nothing special about the first entry. It's
> just a random easily found one.
>
> Â Â Â Â Â Â Â ÂLinus
>

Thanks, Linus and Mel.
You understood my question correctly. :)

My concern was following case.

Child process does mmap new VMA but anon_vma is reused nearer child's
VMA which is linked parent's VMA by fork.
In that case, anon_vma_prepare calls list_add not list_add_tail.
ex) list_add(&avc->same_anon_vma, &anon_vma->head);

It means list_first_entry is the new VMA not old VMA and new VMA's
root_avc isn't linked at parent's one. It means we are locking each
other locks. That's why I have a question.

But I carefully looked at the reusable_anon_vma and found
list_is_singular. I remember Linus changed it to make problem simple.
So in my scenario, new VMA can't share old VMA's anon_vma.

So my story is broken.
If I miss something, please, correct me. :)

-- 
Kind regards,
Minchan Kim
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/