Re: [PATCH 3/3 -mm] generic-ipi: fix the race between generic_smp_call_function_*() and hotplug_cfd()

[Peter Zijlstra <peterz@infradead.org>]

On Wed, 2009-07-29 at 16:31 -0700, Andrew Morton wrote:

> -void generic_smp_call_function_interrupt(void)
> +static void
> +__generic_smp_call_function_interrupt(int cpu, int run_callbacks)
>  {
>  	struct call_function_data *data;
> -	int cpu = smp_processor_id();
>  
>  	/*
>  	 * Ensure entry is visible on call_function_queue after we have
> @@ -169,12 +165,18 @@ void generic_smp_call_function_interrupt

> +static void
> +__generic_smp_call_function_single_interrupt(int cpu, int run_callbacks)
>  {
> -	struct call_single_queue *q = &__get_cpu_var(call_single_queue);
> +	struct call_single_queue *q = &per_cpu(call_single_queue, cpu);
>  	unsigned int data_flags;
>  	LIST_HEAD(list);
>  

It introduces this run_callbacks thing to two functions, but nothing
actually uses that... makes me suspicious there's something missing.


> case CPU_DEAD_FROZEN:
> +               local_irq_save(flags);
> +               __generic_smp_call_function_interrupt(cpu, 0);
> +               __generic_smp_call_function_single_interrupt(cpu, 0);
> +               local_irq_restore(flags);

Doing the callbacks from a different cpu than they were queued on seems
like a fine way to mess things up.