Re: [PATCH 4/5] msix: Don't process table changes while disabled

["Michael S. Tsirkin" <mst@redhat.com>]

On Tue, Oct 18, 2011 at 09:50:53AM +0200, Jan Kiszka wrote:
> As long as MSI-X is disabled, it's incorrect to invoke
> msix_handle_mask_update on per-vector mask changes. That may misguide
> the config notifier callback or spuriously trigger an MSI event.
> 
> Signed-off-by: Jan Kiszka <jan.kiszka@siemens.com>

Same question as on the previous patch. At the moment, the only user
(virtio) seems to check msix_enabled before registering
notifiers. See virtio_pci_query_guest_notifiers.
Is there some way you see to trigger incorrect behaviour?

> ---
>  hw/msix.c |   14 +++++++++-----
>  1 files changed, 9 insertions(+), 5 deletions(-)
> 
> diff --git a/hw/msix.c b/hw/msix.c
> index e351c68..f5c8d08 100644
> --- a/hw/msix.c
> +++ b/hw/msix.c
> @@ -277,12 +277,16 @@ static void msix_mmio_write(void *opaque, target_phys_addr_t addr,
>      bool was_masked = msix_is_masked(dev, vector);
>  
>      pci_set_long(dev->msix_table_page + offset, val);
> -    if (kvm_enabled() && kvm_irqchip_in_kernel()) {
> -        kvm_msix_update(dev, vector, was_masked, msix_is_masked(dev, vector));
> -    }
>  
> -    if (was_masked != msix_is_masked(dev, vector)) {
> -        msix_handle_mask_update(dev, vector);
> +    if (msix_enabled(dev)) {
> +        if (kvm_enabled() && kvm_irqchip_in_kernel()) {
> +            kvm_msix_update(dev, vector, was_masked,
> +                            msix_is_masked(dev, vector));
> +        }
> +
> +        if (was_masked != msix_is_masked(dev, vector)) {
> +            msix_handle_mask_update(dev, vector);
> +        }
>      }
>  }
>  
> -- 
> 1.7.3.4