Re: [Xen-devel] [PATCH v9 05/28] ARM: GICv3: forward pending LPIs to guests

[Julien Grall <julien.grall@xxxxxxx>]

Hi Andre,

On 11/05/17 18:53, Andre Przywara wrote:
> diff --git a/xen/arch/arm/vgic-v3.c b/xen/arch/arm/vgic-v3.c
> index 44d2b50..87f58f6 100644
> --- a/xen/arch/arm/vgic-v3.c
> +++ b/xen/arch/arm/vgic-v3.c
> @@ -1567,12 +1567,30 @@ static struct pending_irq 
> *vgic_v3_lpi_to_pending(struct domain *d,
>      return pirq;
>  }
>
> +/* Retrieve the priority of an LPI from its struct pending_irq. */
> +static int vgic_v3_lpi_get_priority(struct domain *d, uint32_t vlpi)
> +{
> +    struct pending_irq *p = vgic_v3_lpi_to_pending(d, vlpi);
> +
> +    /*
> +     * Cope with the case where this function is called with an invalid LPI.
> +     * It is expected that a caller will bail out handling this LPI at a
> +     * later point in time, but for the sake of this function let us return
> +     * some value here and avoid a NULL pointer dereference.
> +     */
> +    if ( !p )
> +        return 0xff;
I am sorry but I am still against a such as it is. This is only a way to 
workaround a broken design for LPIs.
Looking at the code, I think we don't need to take the rank lock for 
reading the priority as this can be done atomically. In this case, the 
call to get_priority could be moved after the check of p in the caller.
> +
> +    return p->lpi_priority;
> +}
> +
Cheers,

--
Julien Grall

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