Re: [PATCH 05/14] vfio: refactor noiommu group creation

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



On Wed, 11 Aug 2021 17:14:51 +0200
Christoph Hellwig <hch@xxxxxx> wrote:
> @@ -833,14 +789,61 @@ void vfio_uninit_group_dev(struct vfio_device *device)
>  }
>  EXPORT_SYMBOL_GPL(vfio_uninit_group_dev);
>  
> -struct vfio_group *vfio_group_find_or_alloc(struct device *dev)
> +#ifdef CONFIG_VFIO_NOIOMMU
> +static struct vfio_group *vfio_noiommu_group_alloc(struct device *dev)
>  {
>  	struct iommu_group *iommu_group;
>  	struct vfio_group *group;
> +	int ret;
>  
> -	iommu_group = vfio_iommu_group_get(dev);
> -	if (!iommu_group)
> +	iommu_group = iommu_group_alloc();
> +	if (IS_ERR(iommu_group))
> +		return ERR_CAST(iommu_group);
> +
> +	iommu_group_set_name(iommu_group, "vfio-noiommu");
> +	iommu_group_set_iommudata(iommu_group, &noiommu, NULL);
> +	ret = iommu_group_add_device(iommu_group, dev);
> +	if (ret)
> +		goto out_put_group;
> +
> +	group = vfio_create_group(iommu_group);
> +	if (IS_ERR(group)) {
> +		ret = PTR_ERR(group);
> +		goto out_remove_device;
> +	}
> +
> +	return group;
> +
> +out_remove_device:
> +	iommu_group_remove_device(dev);
> +out_put_group:
> +	iommu_group_put(iommu_group);
> +	return ERR_PTR(ret);
> +}
> +#endif
> +
> +static struct vfio_group *vfio_group_find_or_alloc(struct device *dev)
> +{
> +	struct iommu_group *iommu_group;
> +	struct vfio_group *group;
> +
> +	iommu_group = iommu_group_get(dev);
> +	if (!iommu_group) {
> +#ifdef CONFIG_VFIO_NOIOMMU
> +		/*
> +		 * With noiommu enabled, create an IOMMU group for devices that
> +		 * don't already have one and don't have an iommu_ops on their
> +		 * bus.  Taint the kernel because we're about to give a DMA
> +		 * capable device to a user without IOMMU protection.
> +		 */
> +		if (noiommu && !iommu_present(dev->bus)) {
> +			add_taint(TAINT_USER, LOCKDEP_STILL_OK);
> +			dev_warn(dev, "Adding kernel taint for vfio-noiommu group on device\n");
> +			return vfio_noiommu_group_alloc(dev);

Nit, we taint regardless of the success of this function, should we
move the tainting back into the function (using the flags to skip for
mdev in subsequent patches) or swap the order to check the return value
before tainting?  Thanks,

Alex




[Index of Archives]     [KVM ARM]     [KVM ia64]     [KVM ppc]     [Virtualization Tools]     [Spice Development]     [Libvirt]     [Libvirt Users]     [Linux USB Devel]     [Linux Audio Users]     [Yosemite Questions]     [Linux Kernel]     [Linux SCSI]     [XFree86]

  Powered by Linux