Re: [PATCH 1/2] mm: Take all anon_vma locks in anon_vma_lock

[Rik van Riel]

On 05/03/2010 12:41 PM, Linus Torvalds wrote:
> On Mon, 3 May 2010, Rik van Riel wrote:
> > Both the page migration code and the transparent hugepage patches expect
> > 100% reliable rmap lookups and use page_lock_anon_vma(page) to prevent
> > races with mmap, munmap, expand_stack, etc.
>
> Pretty much same comments as for the other one. Why are we pandering to
> the case that is/should be unusual?

In this case, because the fix from the migration side is
difficult and fragile, while fixing things from the mmap
side is straightforward.

I believe the overhead of patch 1/2 should be minimal
as well, because the locks we take are the _depth_ of
the process tree (truncated every exec), not the width.


As for patch 2/2, Mel has an alternative approach for that:

http://lkml.org/lkml/2010/4/30/198

Does Mel's patch seem more reasonable to you?

--
All rights reversed
--
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/