include/linux/vfio.h

8  * it under the terms of the GNU General Public License version 2 as
43  * supports no ioctls outside of VFIO_CHECK_EXTENSION.  Use of VFIO's No-IOMMU
55  * It's *always* the caller's responsibility to indicate the size of
63  * For extension of INFO ioctls, VFIO makes use of a capability chain
66  * structure defines the offset of the first capability in the chain.
68  * bitmap is set.  This offset field is relative to the start of the
77 	__u32	next;		/* Offset of next capability */
81  * Callers of INFO ioctls passing insufficiently sized buffers will see
94  * Report the version of the VFIO API.  This allows us to bump the entire
149  * groups.  Only when a container is set are all of the interfaces
150  * of the VFIO file descriptor and the VFIO group file descriptor
161  * opposite of the SET_CONTAINER call and returns the group to
177  * the devices subdirectory of the IOMMU group sysfs entry.  The
210  * of the following corresponding to device flags in vfio_device_info structure.
224  * Implementation of region mapping is bus driver specific.  This is
238 	__u32	cap_offset;	/* Offset within info struct of first cap */
240 	__u64	offset;		/* Region offset from start of device fd */
245  * The sparse mmap capability allows finer granularity of specifying areas
247  * mmap the offset ranges specified by the areas array.  mmaps outside of the
251  * The structures below define version 1 of this capability.
256 	__u64	offset;	/* Offset of mmap'able area within region */
257 	__u64	size;	/* Size of mmap'able area */
269  * device or class of devices to be exposed.  This helps solve the problem for
270  * vfio bus drivers of defining which region indexes correspond to which region
277  * address space associated with it due to implementation of device file
280  * The current implementation is now part of the user ABI, so we can't use this
285  * The structure below defines version 1 of this capability.
309  * Implementation of IRQ mapping is bus driver specific.  Indexes
328  * and MSI-X where the driver may only use a subset of the available
329  * indexes, but VFIO needs to enable a specific number of vectors
330  * upfront.  In the case of MSI-X, where the user can enable MSI-X and
332  * whether to allocate the maximum supported number of vectors or tear
343 	__u32	count;		/* Number of IRQs within this index */
350  * Set signaling, masking, and unmasking of interrupts.  Caller provides
352  * the range of subindexes being specified.
354  * The DATA flags specify the type of data provided.  If DATA_NONE, the
359  * DATA_BOOL allows sparse support for the same on arrays of interrupts.
365  * A value of -1 can be used to either de-assign interrupts if already
377  * Setting of an event triggering mechanism to userspace for ACTION_TRIGGER
415  * The VFIO-PCI bus driver makes use of the following fixed region and
416  * IRQ index mapping.  Unimplemented regions return a size of zero.
417  * Unimplemented IRQ types return a count of zero.
452  * The vfio-ccw bus driver makes use of the following fixed region and
453  * IRQ index mapping. Unimplemented regions return a size of zero.
454  * Unimplemented IRQ types return a count of zero.
519 	__u64	iova_pgsizes;		/* Bitmap of supported page sizes */
537 	__u64	size;				/* Size of mapping (bytes) */
548  * field.  No guarantee is made to the user that arbitrary unmaps of iova
556 	__u64	size;				/* Size of mapping (bytes) */
572  * the details of Dynamic DMA window capability.
575  * @max_dynamic_windows_supported tells the maximum number of windows
577  * @levels tells the maximum number of levels in multi-level IOMMU tables;
579  * the amount of physically contiguous memory required for the table.
582 	__u64 pgsizes;			/* Bitmap of supported page sizes */
667 	__u64	size;				/* Size of mapping (bytes) */
685  * size and number of levels in the TCE table being created.
687  * It allocates and returns an offset on a PCI bus of the new DMA window.