Re: [PATCH 2.6.24-mm1] Mempolicy: silently restrict nodemask to allowed nodes V3

[David Rientjes <rientjes@google.com>]

On Tue, 12 Feb 2008, KOSAKI Motohiro wrote:

> Hmmmmmm
> sorry, I don't understand yet.
> 
> My test result was
> 
> RHEL5(initrd-2.6.18 + rhel patch)	EINVAL
> 2.6.24					EINVAL
> 2.6.24 + lee-patch			EINVAL
> 
> 
> I don't know current behavior good or wrong.
> but I think it is not regression.
> 

Yes, it's not a regression, but I'm asking why we can't allow this:

	nodemask_t nodes = NODE_MASK_NONE;

	node_set(1, &nodes);
	set_mempolicy(MPOL_DEFAULT, &nodes, MAX_NUMNODES);

It seems like that should not return -EINVAL and that it should just 
effect the system default of a MPOL_DEFAULT policy.

It's not a problem that I'm complaining about specifically in this patch, 
I'm just raising the concern that returning -EINVAL here is really 
unnecessary since mpol_new() will readily accept it.

So you can add my

	Acked-by: David Rientjes <rientjes@google.com>

to this patch, but I would like some counter-arguments presented that show 
why we shouldn't allow the above code to work later on.

		David

--
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>