Re: [PATCH 2/2] mm,migration: Fix race between shift_arg_pages andrmap_walk by guaranteeing rmap_walk finds PTEs created within the temporarystack

[Linus Torvalds]

On Mon, 10 May 2010, KAMEZAWA Hiroyuki wrote:
> 
> But, move_page_tables()'s failure is not a big problem.

Well, yes and no.

It's not a problem because it fails, but because it does the allocation. 
Which means that we can't protect the thing with the (natural) anon_vma 
locking.

> Considering cost, as Mel shows, "don't migrate apges in exec's stack" seems
> reasonable. But, I still doubt this check.

Well, I actually always liked Mel's patch, the one he calls "magic". I 
think it's _less_ magic than the crazy "let's create another vma and 
anon_vma chain just because migration has it's thumb up its ass".

So I never disliked that patch. I'm perfectly happy with a "don't migrate 
these pages at all, because they are in a half-way state in the middle of 
execve stack magic".

		Linus
--
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/