Re: [RFC] mlock/stack guard interaction fixup

[Ian Campbell <ijc@hellion.org.uk>]

[-- Attachment #1: Type: text/plain, Size: 1383 bytes --]

On Fri, 2010-08-20 at 16:59 -0700, Linus Torvalds wrote:
> Ian (and others),
>  here's a three-patch series that uses the doubly linked list to do
> your mlock() case hopefully correctly.

Thanks Linus.

> NOTE! It's untested. The first patch (which is the slightly scary one)
> is tested to some degree, the two other ones I checked that they
> compile, but that's it.

I applied on top of 2.6.35.3 and it fixes the simple test case I posted
yesterday as well as the original issue I was seeing with the Xen
toolstack on 2.6.32.20.

I don't know that they are particularly good tests for this change but I
also ran allmodconfig kernel build and ltp on 2.6.35.3+fixes without
issue. Are there any good mlock heavy workloads?

Out of interest, why is there no guard page for the VM_GROWSUP stack
case? Is it just that the memory layout on PA-RISC makes the stack grows
into the heap scenario impossible?

> I'm not going to apply them to my main tree unless they get testing
> and acks.

Tested-by: Ian Campbell <ijc@hellion.org.uk>

>  And as mentioned, I've not done any of the changes that
> having a vm_prev pointer can allow in other places.
> 
> Comments? Fixes? Braindamage?
> 
>                                Linus

-- 
Ian Campbell

Providence, New Jersey, is one of the few cities where Velveeta cheese
appears on the gourmet shelf.

[-- Attachment #2: This is a digitally signed message part --]
[-- Type: application/pgp-signature, Size: 198 bytes --]