Re: [PATCH 2/2] SLUB: Disable debugging if it increases the minimum page order

[Pekka Enberg <penberg@cs.helsinki.fi>]

Hi Christoph,

On Thu, 11 Jun 2009, Pekka Enberg wrote:
> > I think my patch is the simplest solution here: it turns off debugging
> > for those caches where the metadata bumps up the minimum allocation
> > order and I suspect that we disable cache only for 4096 in practice.

On Thu, 2009-06-11 at 11:20 -0400, Christoph Lameter wrote:
> There could be other slab caches that would randomly be affected by this.
> I ran a couple of times into situations where I had to hack slab to do
> debugging of exempted slabs.

Yes, I do understand that and I have hit that too in the past. But
surely a printk() telling the user we disabled debugging can cure that?

My main point is that a lot of _testers_ will probably enable all SLUB
debugging by default because we encourage them to and it's pretty bad
that we end up causing order 1 allocations and oom conditions.

So I still think we need to fix _at minimum_ the kmalloc-4096 case
(assuming Larry won't hit the same problem still). I see you're not
happy with my patch so any suggestions how to handle that?

			Pekka