Re: linux-next: Tree for June 5

[Mike Travis <travis@sgi.com>]

Ingo Molnar wrote:
> * Vegard Nossum <vegard.nossum@gmail.com> wrote:
> 
>>> Thanks, yes I had that some after thought.  It should check the node 
>>> index if CONFIG_DEBUG_PER_CPU_MAPS is enabled.  One gotcha is that 
>>> nr_node_ids is intialized to MAX_NUMNODES until 
>>> setup_node_to_cpumask_map() sets it to the correct value.  So uses 
>>> before that should be caught by the earlier check.
>> I think it should always check the node index. The code in 
>> kernel/sched.c (see above) calls node_to_cpumask(i) on nodes 0 < i < 
>> MAX_NUMNODES and it WILL use invalid pointers. Or should 
>> kernel/sched.c be changed to use nr_node_ids instead of MAX_NUMNODES? 
>> I believe there are more places that do this than just sched.c.
>>
>> I have attached two patches. The sched one fixes Andrew's boot 
>> problem. The x86 one is untested, but I believe it is better to BUG 
>> than silently corrupt some arbitrary memory. (Then the callers can be 
>> found easily and fixed at least.)
> 
> nice fixes! I have applied both of them to -tip, this one to 
> tip/sched-devel:
> 
>> Subject: [PATCH] sched: don't call node_to_cpumask() on nodes > nr_node_ids
> 
> AFAICS this is not yet required for v2.6.26, as the requirement to never 
> iterate to MAX_NUMNODES and call nr_cpus_node() with the index only got 
> introduced by Mike's patch.
> 
> and this one to tip/x86/numa:
> 
>> Subject: [PATCH] x86: don't return invalid pointers from node_to_cpumask()
> 
> and i've undone the revert of "x86: remove the static 256k 
> node_to_cpumask_map" as well.
> 
> agreed?
> 
> 	Ingo

Hi Ingo,

My -tip branch has:

	a953e4597abd51b74c99e0e3b7074532a60fd031

	sched: replace MAX_NUMNODES with nr_node_ids in kernel/sched.c
	committed: 2008-05-23 09:22:17

The check for node > nr_node_ids however should be included (at least
when CONFIG_DEBUG_PER_CPU_MAPS is enabled.)

Thanks,
Mike