[PATCH v3] drm/amdkfd: Prevent user space using both svm and kfd api to register same user buffer

[Felix Kuehling]

Am 2023-02-08 um 18:26 schrieb Xiaogang.Chen:
> From: Xiaogang Chen <xiaogang.chen at amd.com>
>
> When xnack is on user space can use svm page restore to set a vm range without
> setup it first, then use regular api to register. Currently kfd api and svm are
> not interoperable. We already have check on that, but for user buffer the mapping
> address is not same as buffer cpu virtual address. Add checking on that to
> avoid error propagate to hmm.
>
> Signed-off-by: Xiaogang Chen <xiaogang.chen at amd.com>

Reviewed-by: Felix Kuehling <Felix.Kuehling at amd.com>


> ---
>   drivers/gpu/drm/amd/amdkfd/kfd_chardev.c | 14 ++++++++++++++
>   1 file changed, 14 insertions(+)
>
> diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_chardev.c b/drivers/gpu/drm/amd/amdkfd/kfd_chardev.c
> index f79b8e964140..072fa4fbd27f 100644
> --- a/drivers/gpu/drm/amd/amdkfd/kfd_chardev.c
> +++ b/drivers/gpu/drm/amd/amdkfd/kfd_chardev.c
> @@ -1065,6 +1065,20 @@ static int kfd_ioctl_alloc_memory_of_gpu(struct file *filep,
>   		mutex_unlock(&p->svms.lock);
>   		return -EADDRINUSE;
>   	}
> +
> +	/* When register user buffer check if it has been registered by svm by
> +	 * buffer cpu virtual address.
> +	 */
> +	if ((flags & KFD_IOC_ALLOC_MEM_FLAGS_USERPTR) &&
> +	    interval_tree_iter_first(&p->svms.objects,
> +				     args->mmap_offset >> PAGE_SHIFT,
> +				     (args->mmap_offset  + args->size - 1) >> PAGE_SHIFT)) {
> +		pr_err("User Buffer Address: 0x%llx already allocated by SVM\n",
> +			args->mmap_offset);
> +		mutex_unlock(&p->svms.lock);
> +		return -EADDRINUSE;
> +	}
> +
>   	mutex_unlock(&p->svms.lock);
>   #endif
>   	mutex_lock(&p->mutex);