Re: [Xen-devel] [PATCH v3 33/39] ARM: new VGIC: Add preliminary stub implementation

[Stefano Stabellini <sstabellini@xxxxxxxxxx>]

On Wed, 21 Mar 2018, Andre Przywara wrote:
> The ARM arch code requires an interrupt controller emulation to implement
> vgic_clear_pending_irqs(), although it is suspected that it is actually
> not necessary. Go with a stub for now to make the linker happy.
> 
> Signed-off-by: Andre Przywara <andre.przywara@xxxxxxxxxx>
> Reviewed-by: Julien Grall <julien.grall@xxxxxxx>
> ---
>  xen/arch/arm/vgic/vgic.c | 8 ++++++++
>  1 file changed, 8 insertions(+)
> 
> diff --git a/xen/arch/arm/vgic/vgic.c b/xen/arch/arm/vgic/vgic.c
> index 23b8abfc5e..b70fdaaecb 100644
> --- a/xen/arch/arm/vgic/vgic.c
> +++ b/xen/arch/arm/vgic/vgic.c
> @@ -791,6 +791,14 @@ void gic_dump_vgic_info(struct vcpu *v)
>      spin_unlock_irqrestore(&v->arch.vgic.ap_list_lock, flags);
>  }
>  
> +void vgic_clear_pending_irqs(struct vcpu *v)
> +{
> +    /*
> +     * TODO: It is unclear whether we really need this, so we might instead
> +     * remove it on the caller site.
> +     */
> +}

This is OK for now.

However, thinking about this issue, is it possible for a vcpu to send an
interrupt to an offline vcpu, maybe an SGI? What would happen in that
case? It looks like that vgic_mmio_write_sgir would allow it. Otherwise,
a vcpu could cause the generation of a physical interrupt, an SPI,
targeting an offline vcpu.

Maybe we should WARN in case ap_list is not empty?


>  /**
>   * arch_move_irqs() - migrate the physical affinity of hardware mapped vIRQs
>   * @v:  the vCPU, already assigned to the new pCPU
> -- 
> 2.14.1
> 

_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxxxxxxxxx
https://lists.xenproject.org/mailman/listinfo/xen-devel