Merge tag 'for-linus-iommufd' of git://git.kernel.org/pub/scm/linux/kernel/git/jgg/iommufd

Pull iommufd updates from Jason Gunthorpe: - The iova_bitmap logic for efficiently reporting dirty pages back to userspace has a few more tricky corner case bugs that have been resolved and backed with new tests. The revised version has simpler logic. - Shared branch with iommu for handle support when doing domain attach. Handles allow the domain owner to include additional private data on a per-device basis. - IO Page Fault Reporting to userspace via iommufd. Page faults can be generated on fault capable HWPTs when a translation is not present. Routing them to userspace would allow a VMM to be able to virtualize them into an emulated vIOMMU. This is the next step to fully enabling vSVA support. * tag 'for-linus-iommufd' of git://git.kernel.org/pub/scm/linux/kernel/git/jgg/iommufd: (26 commits) iommufd: Put constants for all the uAPI enums iommufd: Fix error pointer checking iommufd: Add check on user response code iommufd: Remove IOMMUFD_PAGE_RESP_FAILURE iommufd: Require drivers to supply the cache_invalidate_user ops iommufd/selftest: Add coverage for IOPF test iommufd/selftest: Add IOPF support for mock device iommufd: Associate fault object with iommufd_hw_pgtable iommufd: Fault-capable hwpt attach/detach/replace iommufd: Add iommufd fault object iommufd: Add fault and response message definitions iommu: Extend domain attach group with handle support iommu: Add attach handle to struct iopf_group iommu: Remove sva handle list iommu: Introduce domain attachment handle iommufd/iova_bitmap: Remove iterator logic iommufd/iova_bitmap: Dynamic pinning on iova_bitmap_set() iommufd/iova_bitmap: Consolidate iova_bitmap_set exit conditionals iommufd/iova_bitmap: Move initial pinning to iova_bitmap_for_each() iommufd/iova_bitmap: Cache mapped length in iova_bitmap_map struct ...

Linus Torvalds committed 2 years ago

Commit ef7c8f22 parents 07e773d 136a80619 files · +1206 −239Browse files

Showing the changes to include/uapi (1 of 19 files). Show all

Unified · Split

include/uapi/linux/iommufd.h+123−18

@@ -37,19 +37,20 @@ · 36 unmodified lines
37	37	enum {
38	38	IOMMUFD_CMD_BASE = 0x80,
39	39	IOMMUFD_CMD_DESTROY = IOMMUFD_CMD_BASE,
40		− IOMMUFD_CMD_IOAS_ALLOC,
41		− IOMMUFD_CMD_IOAS_ALLOW_IOVAS,
42		− IOMMUFD_CMD_IOAS_COPY,
43		− IOMMUFD_CMD_IOAS_IOVA_RANGES,
44		− IOMMUFD_CMD_IOAS_MAP,
45		− IOMMUFD_CMD_IOAS_UNMAP,
46		− IOMMUFD_CMD_OPTION,
47		− IOMMUFD_CMD_VFIO_IOAS,
48		− IOMMUFD_CMD_HWPT_ALLOC,
49		− IOMMUFD_CMD_GET_HW_INFO,
50		− IOMMUFD_CMD_HWPT_SET_DIRTY_TRACKING,
51		− IOMMUFD_CMD_HWPT_GET_DIRTY_BITMAP,
52		− IOMMUFD_CMD_HWPT_INVALIDATE,
	40	+ IOMMUFD_CMD_IOAS_ALLOC = 0x81,
	41	+ IOMMUFD_CMD_IOAS_ALLOW_IOVAS = 0x82,
	42	+ IOMMUFD_CMD_IOAS_COPY = 0x83,
	43	+ IOMMUFD_CMD_IOAS_IOVA_RANGES = 0x84,
	44	+ IOMMUFD_CMD_IOAS_MAP = 0x85,
	45	+ IOMMUFD_CMD_IOAS_UNMAP = 0x86,
	46	+ IOMMUFD_CMD_OPTION = 0x87,
	47	+ IOMMUFD_CMD_VFIO_IOAS = 0x88,
	48	+ IOMMUFD_CMD_HWPT_ALLOC = 0x89,
	49	+ IOMMUFD_CMD_GET_HW_INFO = 0x8a,
	50	+ IOMMUFD_CMD_HWPT_SET_DIRTY_TRACKING = 0x8b,
	51	+ IOMMUFD_CMD_HWPT_GET_DIRTY_BITMAP = 0x8c,
	52	+ IOMMUFD_CMD_HWPT_INVALIDATE = 0x8d,
	53	+ IOMMUFD_CMD_FAULT_QUEUE_ALLOC = 0x8e,
53	54	};
54	55
55	56	/**
@@ -356,10 +357,13 @@ struct iommu_vfio_ioas { · 300 unmodified lines
356	357	* the parent HWPT in a nesting configuration.
357	358	* @IOMMU_HWPT_ALLOC_DIRTY_TRACKING: Dirty tracking support for device IOMMU is
358	359	* enforced on device attachment
	360	+ * @IOMMU_HWPT_FAULT_ID_VALID: The fault_id field of hwpt allocation data is
	361	+ * valid.
359	362	*/
360	363	enum iommufd_hwpt_alloc_flags {
361	364	IOMMU_HWPT_ALLOC_NEST_PARENT = 1 << 0,
362	365	IOMMU_HWPT_ALLOC_DIRTY_TRACKING = 1 << 1,
	366	+ IOMMU_HWPT_FAULT_ID_VALID = 1 << 2,
363	367	};
364	368
365	369	/**
@@ -396,8 +400,8 @@ struct iommu_hwpt_vtd_s1 { · 30 unmodified lines
396	400	* @IOMMU_HWPT_DATA_VTD_S1: Intel VT-d stage-1 page table
397	401	*/
398	402	enum iommu_hwpt_data_type {
399		− IOMMU_HWPT_DATA_NONE,
400		− IOMMU_HWPT_DATA_VTD_S1,
	403	+ IOMMU_HWPT_DATA_NONE = 0,
	404	+ IOMMU_HWPT_DATA_VTD_S1 = 1,
401	405	};
402	406
403	407	/**
@@ -411,6 +415,9 @@ enum iommu_hwpt_data_type { · 7 unmodified lines
411	415	* @data_type: One of enum iommu_hwpt_data_type
412	416	* @data_len: Length of the type specific data
413	417	* @data_uptr: User pointer to the type specific data
	418	+ * @fault_id: The ID of IOMMUFD_FAULT object. Valid only if flags field of
	419	+ * IOMMU_HWPT_FAULT_ID_VALID is set.
	420	+ * @__reserved2: Padding to 64-bit alignment. Must be 0.
414	421	*
415	422	* Explicitly allocate a hardware page table object. This is the same object
416	423	* type that is returned by iommufd_device_attach() and represents the
@@ -441,6 +448,8 @@ struct iommu_hwpt_alloc { · 24 unmodified lines
441	448	__u32 data_type;
442	449	__u32 data_len;
443	450	__aligned_u64 data_uptr;
	451	+ __u32 fault_id;
	452	+ __u32 __reserved2;
444	453	};
445	454	#define IOMMU_HWPT_ALLOC _IO(IOMMUFD_TYPE, IOMMUFD_CMD_HWPT_ALLOC)
446	455
@@ -482,8 +491,8 @@ struct iommu_hw_info_vtd { · 35 unmodified lines
482	491	* @IOMMU_HW_INFO_TYPE_INTEL_VTD: Intel VT-d iommu info type
483	492	*/
484	493	enum iommu_hw_info_type {
485		− IOMMU_HW_INFO_TYPE_NONE,
486		− IOMMU_HW_INFO_TYPE_INTEL_VTD,
	494	+ IOMMU_HW_INFO_TYPE_NONE = 0,
	495	+ IOMMU_HW_INFO_TYPE_INTEL_VTD = 1,
487	496	};
488	497
489	498	/**
@@ -620,7 +629,7 @@ struct iommu_hwpt_get_dirty_bitmap { · 130 unmodified lines
620	629	* @IOMMU_HWPT_INVALIDATE_DATA_VTD_S1: Invalidation data for VTD_S1
621	630	*/
622	631	enum iommu_hwpt_invalidate_data_type {
623		− IOMMU_HWPT_INVALIDATE_DATA_VTD_S1,
	632	+ IOMMU_HWPT_INVALIDATE_DATA_VTD_S1 = 0,
624	633	};
625	634
626	635	/**
@@ -692,4 +701,100 @@ struct iommu_hwpt_invalidate { · 65 unmodified lines
692	701	__u32 __reserved;
693	702	};
694	703	#define IOMMU_HWPT_INVALIDATE _IO(IOMMUFD_TYPE, IOMMUFD_CMD_HWPT_INVALIDATE)
	704	+
	705	+/**
	706	+ * enum iommu_hwpt_pgfault_flags - flags for struct iommu_hwpt_pgfault
	707	+ * @IOMMU_PGFAULT_FLAGS_PASID_VALID: The pasid field of the fault data is
	708	+ * valid.
	709	+ * @IOMMU_PGFAULT_FLAGS_LAST_PAGE: It's the last fault of a fault group.
	710	+ */
	711	+enum iommu_hwpt_pgfault_flags {
	712	+ IOMMU_PGFAULT_FLAGS_PASID_VALID = (1 << 0),
	713	+ IOMMU_PGFAULT_FLAGS_LAST_PAGE = (1 << 1),
	714	+};
	715	+
	716	+/**
	717	+ * enum iommu_hwpt_pgfault_perm - perm bits for struct iommu_hwpt_pgfault
	718	+ * @IOMMU_PGFAULT_PERM_READ: request for read permission
	719	+ * @IOMMU_PGFAULT_PERM_WRITE: request for write permission
	720	+ * @IOMMU_PGFAULT_PERM_EXEC: (PCIE 10.4.1) request with a PASID that has the
	721	+ * Execute Requested bit set in PASID TLP Prefix.
	722	+ * @IOMMU_PGFAULT_PERM_PRIV: (PCIE 10.4.1) request with a PASID that has the
	723	+ * Privileged Mode Requested bit set in PASID TLP
	724	+ * Prefix.
	725	+ */
	726	+enum iommu_hwpt_pgfault_perm {
	727	+ IOMMU_PGFAULT_PERM_READ = (1 << 0),
	728	+ IOMMU_PGFAULT_PERM_WRITE = (1 << 1),
	729	+ IOMMU_PGFAULT_PERM_EXEC = (1 << 2),
	730	+ IOMMU_PGFAULT_PERM_PRIV = (1 << 3),
	731	+};
	732	+
	733	+/**
	734	+ * struct iommu_hwpt_pgfault - iommu page fault data
	735	+ * @flags: Combination of enum iommu_hwpt_pgfault_flags
	736	+ * @dev_id: id of the originated device
	737	+ * @pasid: Process Address Space ID
	738	+ * @grpid: Page Request Group Index
	739	+ * @perm: Combination of enum iommu_hwpt_pgfault_perm
	740	+ * @addr: Fault address
	741	+ * @length: a hint of how much data the requestor is expecting to fetch. For
	742	+ * example, if the PRI initiator knows it is going to do a 10MB
	743	+ * transfer, it could fill in 10MB and the OS could pre-fault in
	744	+ * 10MB of IOVA. It's default to 0 if there's no such hint.
	745	+ * @cookie: kernel-managed cookie identifying a group of fault messages. The
	746	+ * cookie number encoded in the last page fault of the group should
	747	+ * be echoed back in the response message.
	748	+ */
	749	+struct iommu_hwpt_pgfault {
	750	+ __u32 flags;
	751	+ __u32 dev_id;
	752	+ __u32 pasid;
	753	+ __u32 grpid;
	754	+ __u32 perm;
	755	+ __u64 addr;
	756	+ __u32 length;
	757	+ __u32 cookie;
	758	+};
	759	+
	760	+/**
	761	+ * enum iommufd_page_response_code - Return status of fault handlers
	762	+ * @IOMMUFD_PAGE_RESP_SUCCESS: Fault has been handled and the page tables
	763	+ * populated, retry the access. This is the
	764	+ * "Success" defined in PCI 10.4.2.1.
	765	+ * @IOMMUFD_PAGE_RESP_INVALID: Could not handle this fault, don't retry the
	766	+ * access. This is the "Invalid Request" in PCI
	767	+ * 10.4.2.1.
	768	+ */
	769	+enum iommufd_page_response_code {
	770	+ IOMMUFD_PAGE_RESP_SUCCESS = 0,
	771	+ IOMMUFD_PAGE_RESP_INVALID = 1,
	772	+};
	773	+
	774	+/**
	775	+ * struct iommu_hwpt_page_response - IOMMU page fault response
	776	+ * @cookie: The kernel-managed cookie reported in the fault message.
	777	+ * @code: One of response code in enum iommufd_page_response_code.
	778	+ */
	779	+struct iommu_hwpt_page_response {
	780	+ __u32 cookie;
	781	+ __u32 code;
	782	+};
	783	+
	784	+/**
	785	+ * struct iommu_fault_alloc - ioctl(IOMMU_FAULT_QUEUE_ALLOC)
	786	+ * @size: sizeof(struct iommu_fault_alloc)
	787	+ * @flags: Must be 0
	788	+ * @out_fault_id: The ID of the new FAULT
	789	+ * @out_fault_fd: The fd of the new FAULT
	790	+ *
	791	+ * Explicitly allocate a fault handling object.
	792	+ */
	793	+struct iommu_fault_alloc {
	794	+ __u32 size;
	795	+ __u32 flags;
	796	+ __u32 out_fault_id;
	797	+ __u32 out_fault_fd;
	798	+};
	799	+#define IOMMU_FAULT_QUEUE_ALLOC _IO(IOMMUFD_TYPE, IOMMUFD_CMD_FAULT_QUEUE_ALLOC)
695	800	#endif