Re: [Intel-gfx] [PATCH v1 1/5] iommufd: Create access in vfio_iommufd_emulated_bind()

[Jason Gunthorpe <jgg@nvidia.com>]

On Wed, Mar 08, 2023 at 05:13:36AM -0800, Yi Liu wrote:

> +int iommufd_access_set_ioas(struct iommufd_access *access, u32 ioas_id)
> +{
> +	struct iommufd_ioas *new_ioas = NULL, *cur_ioas;
> +	struct iommufd_ctx *ictx = access->ictx;
> +	struct iommufd_object *obj;
> +	int rc = 0;
> +
> +	if (ioas_id) {
> +		obj = iommufd_get_object(ictx, ioas_id, IOMMUFD_OBJ_IOAS);
> +		if (IS_ERR(obj))
> +			return PTR_ERR(obj);
> +		new_ioas = container_of(obj, struct iommufd_ioas, obj);
> +	}
> +
> +	mutex_lock(&access->ioas_lock);
> +	cur_ioas = access->ioas;
> +	if (cur_ioas == new_ioas)
> +		goto out_unlock;
> +
> +	if (new_ioas) {
> +		rc = iopt_add_access(&new_ioas->iopt, access);
> +		if (rc)
> +			goto out_unlock;
> +		iommufd_ref_to_users(obj);
> +	}
> +
> +	if (cur_ioas) {
> +		iopt_remove_access(&cur_ioas->iopt, access);
> +		refcount_dec(&cur_ioas->obj.users);
> +	}

This should match the physical side with an add/remove/replace
API. Especially since remove is implicit in destroy this series only
needs the add API

And the locking shouldn't come in another patch that brings the
replace/remove since with just split add we don't need it.

That will make this patch alot smaller

Jason