Re: [PATCH 01/16] xen: sched: fix locking when allocating an RTDS pCPU

[George Dunlap <george.dunlap@citrix.com>]

On 18/03/16 19:04, Dario Faggioli wrote:
> as doing that include changing the scheduler lock
> mapping for the pCPU itself, and the correct way
> of doing that is:
>  - take the lock that the pCPU is using right now
>    (which may be the lock of another scheduler);
>  - change the mapping of the lock to the RTDS one;
>  - release the lock (the one that has actually been
>    taken!)
> 
> Signed-off-by: Dario Faggioli <dario.faggioli@citrix.com>

Reviewed-by: George Dunlap <george.dunlap@citrix.com>

> ---
> Cc: Meng Xu <mengxu@cis.upenn.edu>
> Cc: George Dunlap <george.dunlap@eu.citrix.com>
> Cc: Tianyang Chen <tiche@seas.upenn.edu>
> ---
>  xen/common/sched_rt.c |    9 +++++++--
>  1 file changed, 7 insertions(+), 2 deletions(-)
> 
> diff --git a/xen/common/sched_rt.c b/xen/common/sched_rt.c
> index c896a6f..d98bfb6 100644
> --- a/xen/common/sched_rt.c
> +++ b/xen/common/sched_rt.c
> @@ -653,11 +653,16 @@ static void *
>  rt_alloc_pdata(const struct scheduler *ops, int cpu)
>  {
>      struct rt_private *prv = rt_priv(ops);
> +    spinlock_t *old_lock;
>      unsigned long flags;
>  
> -    spin_lock_irqsave(&prv->lock, flags);
> +    /* Move the scheduler lock to our global runqueue lock.  */
> +    old_lock = pcpu_schedule_lock_irqsave(cpu, &flags);
> +
>      per_cpu(schedule_data, cpu).schedule_lock = &prv->lock;
> -    spin_unlock_irqrestore(&prv->lock, flags);
> +
> +    /* _Not_ pcpu_schedule_unlock(): per_cpu().schedule_lock changed! */
> +    spin_unlock_irqrestore(old_lock, flags);
>  
>      if ( !alloc_cpumask_var(&_cpumask_scratch[cpu]) )
>          return NULL;
> 


_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xen.org
http://lists.xen.org/xen-devel