Re: [RFC PATCH 2/2] kernel/sched.c: VLA in middle of struct

[Ingo Molnar <mingo@elte.hu>]

* Jeff Garzik <jeff@garzik.org> wrote:

> The semantics for variable-length arrays __in the middle of structs__
> are quite muddy, and a case in sched.c presents an interesting case,
> as the preceding code comment indicates:
> 
> 	/*
> 	 * The cpus mask in sched_group and sched_domain hangs off
> 	 the end.  * FIXME: use cpumask_var_t or dynamic percpu alloc
> 	 to avoid * wasting space for nr_cpu_ids < CONFIG_NR_CPUS.  */
> 	struct static_sched_group {
> 		struct sched_group sg; DECLARE_BITMAP(cpus,
> 		CONFIG_NR_CPUS);
> 	};
> 
> 	struct static_sched_domain {
> 		struct sched_domain sd; DECLARE_BITMAP(span,
> 		CONFIG_NR_CPUS);
> 	};
> 
> Both sched_group and sched_domain have the following trailing struct
> member:
> 
> 	unsigned long cpumask[];
> 
> So this change is intended largely to spawn a discussion, because
> I'm not sure this VLA-in-middle-of-struct behavior is guaranteed to
> always behave as expected?
> 
> Maybe a C expert can say whether cpumask[0] is better than cpumask[],
> or have other comments?

This bit of the code is a bit temporary, with a half-done cpumask 
conversion ...

I'm not convinced at all that this code wants to be converted to 
cpuvar_t. These are all either statically percpu or dynamically 
allocated structures, so we wont allocate NR_CPUS of them. So i'd 
lean towards embedding a full struct cpumask and that's it.

That cpumask[] should probably be cpumask[0], to document the 
aliasing to ->span and ->cpus properly.

Rusty, what do you think?

	Ingo