Re: [2/3] mm: fix up some user-visible effects of the stack guard page

[Linus Torvalds]

On Fri, Aug 20, 2010 at 8:54 AM, Linus Torvalds
<torvalds@xxxxxxxxxxxxxxxxxxxx> wrote:
> On Fri, Aug 20, 2010 at 5:54 AM, Ian Campbell <ijc@xxxxxxxxxxxxxx> wrote:
>>
>> Since we have split the original VMA into 3, shouldn't only the bottom
>> one still have VM_GROWSDOWN set? (how can the top two grow down with the
>> bottom one in the way?) Certainly it seems wrong to enforce a guard page
>> on anything but the bottom VMA (which is what appears to be happening).
>
> Yes, it does seem like we should teach vma splitting to remove
> VM_GROWSDOWN on all but the lowest mapping.

Actually, thinking some more about it, that may not be a good idea.
Why? Simply because we may want to merge the vma's back together if
you do munlock. And it won't (and mustn't) merge if the vm_flags
differ in VM_GROWSDOWN.

So I do think we want to keep VM_GROWSDOWN (and VM_GROWSUP on PA-RISC)
even across a vma split.

Of course, we could set a flag whether the vma really does have a
guard page or not.

That said, it does strike me as rather odd to do VM ops on partial
stacks. What are you doing, exactly, to hit this?

                         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/