Re: [Qemu-devel] [PATCH for-4.2 v10 06/15] virtio-iommu: Endpoint and domains structs and helpers

[Peter Xu <peterx@redhat.com>]

On Tue, Jul 30, 2019 at 07:21:28PM +0200, Eric Auger wrote:
>  static void virtio_iommu_device_realize(DeviceState *dev, Error **errp)
>  {
>      VirtIODevice *vdev = VIRTIO_DEVICE(dev);
> @@ -334,6 +444,8 @@ static void virtio_iommu_device_realize(DeviceState *dev, Error **errp)
>      virtio_add_feature(&s->features, VIRTIO_IOMMU_F_BYPASS);
>      virtio_add_feature(&s->features, VIRTIO_IOMMU_F_MMIO);
>  
> +    qemu_mutex_init(&s->mutex);

It's a bit strange to init a mutex which has already been used in
patch 3. :)

Thanks,

> +
>      memset(s->as_by_bus_num, 0, sizeof(s->as_by_bus_num));
>      s->as_by_busptr = g_hash_table_new(NULL, NULL);
>  
> @@ -342,11 +454,20 @@ static void virtio_iommu_device_realize(DeviceState *dev, Error **errp)
>      } else {
>          error_setg(errp, "VIRTIO-IOMMU is not attached to any PCI bus!");
>      }
> +
> +    s->domains = g_tree_new_full((GCompareDataFunc)int_cmp,
> +                                 NULL, NULL, virtio_iommu_put_domain);
> +    s->endpoints = g_tree_new_full((GCompareDataFunc)int_cmp,
> +                                   NULL, NULL, virtio_iommu_put_endpoint);
>  }

-- 
Peter Xu