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

[KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>]

On Sun, 9 May 2010 18:32:32 -0700 (PDT)
Linus Torvalds <torvalds@linux-foundation.org> wrote:

> 
> 
> On Sun, 9 May 2010, Linus Torvalds wrote:
> > 
> > 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".
> 
> Btw, I also think that Mel's patch could be made a lot _less_ magic by 
> just marking that initial stack vma with a VM_STACK_INCOMPLETE_SETUP bit, 
> instead of doing that "maybe_stack" thing. We could easily make that 
> initial vma setup very explicit indeed, and then just clear that bit when 
> we've moved the stack to its final position.
> 

Hmm. vm_flags is still 32bit..(I think it should be long long)

Using combination of existing flags...

#define VM_STACK_INCOMPLETE_SETUP (VM_RAND_READ | VM_SEC_READ)

Can be used instead of checking mapcount, I think.

Thanks,
-Kame

--
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>