Re: [PATCH] IRQ fix incorrect logic in __clear_irq_vector (v2)

[Keir Fraser <keir.xen@gmail.com>]

On 12/08/2011 14:54, "Andrew Cooper" <andrew.cooper3@citrix.com> wrote:

> In the old code, tmp_mask is the cpu_and of cfg->cpu_mask and
> cpu_online_map.  However, in the usual case of moving an IRQ from one
> PCPU to another because the scheduler decides its a good idea,
> cfg->cpu_mask and cfg->old_cpu_mask do not intersect.  This causes the
> old cpu vector_irq table to keep the irq reference when it shouldn't.
> 
> This leads to a resource leak if a domain is shut down wile an irq has
> a move pending, which results in Xen's create_irq() eventually failing
> with -ENOSPC when all vector_irq tables are full of stale references.
> 
> v2: reuse tmp_mask to take account of online cpus

Nasty bug, nice fix!

The extra field in irq_cfg sounds plausible to me -- I don't mind adding it
if it's a nice cleanup.

 Thanks,
 Keir

> Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com>
> 
> diff -r 1f08b380d438 -r bd106cc2aa65 xen/arch/x86/irq.c
> --- a/xen/arch/x86/irq.c Wed Aug 10 14:43:34 2011 +0100
> +++ b/xen/arch/x86/irq.c Fri Aug 12 14:54:11 2011 +0100
> @@ -216,6 +216,7 @@ static void __clear_irq_vector(int irq)
>  
>      if (likely(!cfg->move_in_progress))
>          return;
> +    cpus_and(tmp_mask, cfg->old_cpu_mask, cpu_online_map);
>      for_each_cpu_mask(cpu, tmp_mask) {
>          for (vector = FIRST_DYNAMIC_VECTOR; vector <= LAST_DYNAMIC_VECTOR;
>                                  vector++) {
> 
> _______________________________________________
> Xen-devel mailing list
> Xen-devel@lists.xensource.com
> http://lists.xensource.com/xen-devel