From: Lu Baolu <[email protected]> Introduce a new domain type for a user I/O page table, which is nested on top of another user space address represented by a UNMANAGED domain. The mappings of a nested domain are managed by user space software, so it is not necessary to have map/unmap callbacks. But the updates of the PTEs in the nested page table will be propagated to the hardware caches on both IOMMU (IOTLB) and devices (DevTLB/ATC).
A nested domain is allocated by the domain_alloc_user op, and attached to a device through the existing iommu_attach_device/group() interfaces. Add a new domain op cache_invalidate_user for the userspace to flush the hardware caches for a nested domain through iommufd. No wrapper for it, as it's only supposed to be used by iommufd. Pass in invalidation requests to the cache_invalidate_user op, in form of a user data array that conatins a number of invalidation entries. Add an iommu_user_data_array struct and an iommu_copy_user_data_from_array helper for iommu drivers to walk through the invalidation request array and fetch the data entry inside. Signed-off-by: Lu Baolu <[email protected]> Reviewed-by: Kevin Tian <[email protected]> Co-developed-by: Nicolin Chen <[email protected]> Signed-off-by: Nicolin Chen <[email protected]> Signed-off-by: Yi Liu <[email protected]> --- include/linux/iommu.h | 59 +++++++++++++++++++++++++++++++++++++++++++ 1 file changed, 59 insertions(+) diff --git a/include/linux/iommu.h b/include/linux/iommu.h index 12e12e5563e6..439e295c91a3 100644 --- a/include/linux/iommu.h +++ b/include/linux/iommu.h @@ -66,6 +66,9 @@ struct iommu_domain_geometry { #define __IOMMU_DOMAIN_SVA (1U << 4) /* Shared process address space */ +#define __IOMMU_DOMAIN_NESTED (1U << 5) /* User-managed address space nested + on a stage-2 translation */ + #define IOMMU_DOMAIN_ALLOC_FLAGS ~__IOMMU_DOMAIN_DMA_FQ /* * This are the possible domain-types @@ -92,6 +95,7 @@ struct iommu_domain_geometry { __IOMMU_DOMAIN_DMA_API | \ __IOMMU_DOMAIN_DMA_FQ) #define IOMMU_DOMAIN_SVA (__IOMMU_DOMAIN_SVA) +#define IOMMU_DOMAIN_NESTED (__IOMMU_DOMAIN_NESTED) struct iommu_domain { unsigned type; @@ -241,6 +245,21 @@ struct iommu_user_data { size_t len; }; +/** + * struct iommu_user_data_array - iommu driver specific user space data array + * @uptr: Pointer to the user buffer array for copy_from_user() + * @entry_len: The fixed-width length of a entry in the array, in bytes + * @entry_num: The number of total entries in the array + * + * A array having a @entry_num number of @entry_len sized entries, each entry is + * user space data, i.e. an uAPI that is defined in include/uapi/linux/iommufd.h + */ +struct iommu_user_data_array { + void __user *uptr; + size_t entry_len; + int entry_num; +}; + /** * iommu_copy_user_data - Copy iommu driver specific user space data * @dst_data: Pointer to an iommu driver specific user data that is defined in @@ -263,6 +282,34 @@ static inline int iommu_copy_user_data(void *dst_data, src_data->uptr, src_data->len); } +/** + * iommu_copy_user_data_from_array - Copy iommu driver specific user space data + * from an iommu_user_data_array input + * @dst_data: Pointer to an iommu driver specific user data that is defined in + * include/uapi/linux/iommufd.h + * @src_data: Pointer to a struct iommu_user_data_array for user space data array + * @index: Index to offset the location in the array to copy user data from + * @data_len: Length of current user data structure, i.e. sizeof(struct _dst) + * @min_len: Initial length of user data structure for backward compatibility. + * This should be offsetofend using the last member in the user data + * struct that was initially added to include/uapi/linux/iommufd.h + */ +static inline int +iommu_copy_user_data_from_array(void *dst_data, + const struct iommu_user_data_array *src_array, + int index, size_t data_len, size_t min_len) +{ + struct iommu_user_data src_data; + + if (WARN_ON(!src_array || index >= src_array->entry_num)) + return -EINVAL; + if (!src_array->entry_num) + return -EINVAL; + src_data.uptr = src_array->uptr + src_array->entry_len * index; + src_data.len = src_array->entry_len; + return iommu_copy_user_data(dst_data, &src_data, data_len, min_len); +} + /** * struct iommu_ops - iommu ops and capabilities * @capable: check capability @@ -374,6 +421,15 @@ struct iommu_ops { * @iotlb_sync_map: Sync mappings created recently using @map to the hardware * @iotlb_sync: Flush all queued ranges from the hardware TLBs and empty flush * queue + * @cache_invalidate_user: Flush hardware cache for user space IO page table. + * The @domain must be IOMMU_DOMAIN_NESTED. The @array + * passes in the cache invalidation requests, in form + * of a driver data structure. The driver must update + * array->entry_num to report the number of handled + * invalidation requests. The 32-bit @error_code can + * forward a driver specific error code to user space. + * Both the driver data structure and the error code + * must be defined in include/uapi/linux/iommufd.h * @iova_to_phys: translate iova to physical address * @enforce_cache_coherency: Prevent any kind of DMA from bypassing IOMMU_CACHE, * including no-snoop TLPs on PCIe or other platform @@ -403,6 +459,9 @@ struct iommu_domain_ops { size_t size); void (*iotlb_sync)(struct iommu_domain *domain, struct iommu_iotlb_gather *iotlb_gather); + int (*cache_invalidate_user)(struct iommu_domain *domain, + struct iommu_user_data_array *array, + u32 *error_code); phys_addr_t (*iova_to_phys)(struct iommu_domain *domain, dma_addr_t iova); -- 2.34.1
