Re: Race in new page migration code?

[Andi Kleen <ak@suse.de>]

On Monday 16 January 2006 17:28, Hugh Dickins wrote:
> On Mon, 16 Jan 2006, Christoph Lameter wrote:
> > On Mon, 16 Jan 2006, Hugh Dickins wrote:
> > 
> > > > It also applies to the policy compliance check.
> > > 
> > > Good point, I missed that: you've inadventently changed the behaviour
> > > of sys_mbind when it encounters a zero page from a disallowed node.
> > > Another reason to remove your PageReserved test.
> > 
> > The zero page always come from node zero on IA64. I think this is more the 
> > inadvertent fixing of a bug. The policy compliance check currently fails 
> > if an address range contains a zero page but node zero is not contained in 
> > the nodelist.
> 
> To me it sounds more like you introduced a bug than fixed one.
> If MPOL_MF_STRICT and the zero page is found but not in the nodelist
> demanded, then it's right to refuse, I'd say.  If Andi shares your
> view that the zero pages should be ignored, I won't argue; but we
> shouldn't change behaviour by mistake, without review or comment.

I agree with Christoph that the zero page should be ignored - old behaviour
was really a bug.

-Andi

--
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>