Re: [PATCH v9 12/13] iommu/amd: Clear out the GV flag when handle deferred domain attach

[Joerg Roedel]

On Tue, Aug 01, 2017 at 07:37:28PM +0800, Baoquan He wrote:
> @@ -2466,11 +2472,21 @@ static struct protection_domain *get_domain(struct device *dev)
>  		return ERR_PTR(-EINVAL);
>  
>  	domain = get_dev_data(dev)->domain;
> -	if (domain == NULL && get_dev_data(dev)->defer_attach) {
> +	if (domain == NULL && dev_data->defer_attach) {
> +		u16 alias = amd_iommu_alias_table[dev_data->devid];
>  		get_dev_data(dev)->defer_attach = false;
>  		io_domain = iommu_get_domain_for_dev(dev);
>  		domain = to_pdomain(io_domain);
>  		attach_device(dev, domain);
> +		/*
> +		 * If the deferred attached domain is not v2, should clear out
> +		 * the old GV flag.
> +		 */
> +		if (!(domain->flags & PD_IOMMUV2_MASK)) {
> +			clear_dte_flag_gv(dev_data->devid);
> +			if (alias != dev_data->devid)
> +				clear_dte_flag_gv(dev_data->devid);

Hmm, thinking more about it, I am not sure what the IOMMU responds to
PRI/PASID prefixes if the GV flag is 0.

But until we know it causes problems we should just disable the bit
while doing the copy in the previous patch and avoid any special
handling like done here.