linux/include/uapi/linux/vfio.h

1307 lines
50 KiB

License cleanup: add SPDX license identifier to uapi header files with a license Many user space API headers have licensing information, which is either incomplete, badly formatted or just a shorthand for referring to the license under which the file is supposed to be. This makes it hard for compliance tools to determine the correct license. Update these files with an SPDX license identifier. The identifier was chosen based on the license information in the file. GPL/LGPL licensed headers get the matching GPL/LGPL SPDX license identifier with the added 'WITH Linux-syscall-note' exception, which is the officially assigned exception identifier for the kernel syscall exception: NOTE! This copyright does *not* cover user programs that use kernel services by normal system calls - this is merely considered normal use of the kernel, and does *not* fall under the heading of "derived work". This exception makes it possible to include GPL headers into non GPL code, without confusing license compliance tools. Headers which have either explicit dual licensing or are just licensed under a non GPL license are updated with the corresponding SPDX identifier and the GPLv2 with syscall exception identifier. The format is: ((GPL-2.0 WITH Linux-syscall-note) OR SPDX-ID-OF-OTHER-LICENSE) SPDX license identifiers are a legally binding shorthand, which can be used instead of the full boiler plate text. The update does not remove existing license information as this has to be done on a case by case basis and the copyright holders might have to be consulted. This will happen in a separate step. This patch is based on work done by Thomas Gleixner and Kate Stewart and Philippe Ombredanne. See the previous patch in this series for the methodology of how this patch was researched. Reviewed-by: Kate Stewart <kstewart@linuxfoundation.org> Reviewed-by: Philippe Ombredanne <pombredanne@nexb.com> Reviewed-by: Thomas Gleixner <tglx@linutronix.de> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2017-11-01 15:09:13 +01:00
vfio/iommu_type1: Multi-IOMMU domain support We currently have a problem that we cannot support advanced features of an IOMMU domain (ex. IOMMU_CACHE), because we have no guarantee that those features will be supported by all of the hardware units involved with the domain over its lifetime. For instance, the Intel VT-d architecture does not require that all DRHDs support snoop control. If we create a domain based on a device behind a DRHD that does support snoop control and enable SNP support via the IOMMU_CACHE mapping option, we cannot then add a device behind a DRHD which does not support snoop control or we'll get reserved bit faults from the SNP bit in the pagetables. To add to the complexity, we can't know the properties of a domain until a device is attached. We could pass this problem off to userspace and require that a separate vfio container be used, but we don't know how to handle page accounting in that case. How do we know that a page pinned in one container is the same page as a different container and avoid double billing the user for the page. The solution is therefore to support multiple IOMMU domains per container. In the majority of cases, only one domain will be required since hardware is typically consistent within a system. However, this provides us the ability to validate compatibility of domains and support mixed environments where page table flags can be different between domains. To do this, our DMA tracking needs to change. We currently try to coalesce user mappings into as few tracking entries as possible. The problem then becomes that we lose granularity of user mappings. We've never guaranteed that a user is able to unmap at a finer granularity than the original mapping, but we must honor the granularity of the original mapping. This coalescing code is therefore removed, allowing only unmaps covering complete maps. The change in accounting is fairly small here, a typical QEMU VM will start out with roughly a dozen entries, so it's arguable if this coalescing was ever needed. We also move IOMMU domain creation to the point where a group is attached to the container. An interesting side-effect of this is that we now have access to the device at the time of domain creation and can probe the devices within the group to determine the bus_type. This finally makes vfio_iommu_type1 completely device/bus agnostic. In fact, each IOMMU domain can host devices on different buses managed by different physical IOMMUs, and present a single DMA mapping interface to the user. When a new domain is created, mappings are replayed to bring the IOMMU pagetables up to the state of the current container. And of course, DMA mapping and unmapping automatically traverse all of the configured IOMMU domains. Signed-off-by: Alex Williamson <alex.williamson@redhat.com> Cc: Varun Sethi <Varun.Sethi@freescale.com>
2014-02-26 11:38:36 -07:00
vfio: powerpc/spapr: Register memory and define IOMMU v2 The existing implementation accounts the whole DMA window in the locked_vm counter. This is going to be worse with multiple containers and huge DMA windows. Also, real-time accounting would requite additional tracking of accounted pages due to the page size difference - IOMMU uses 4K pages and system uses 4K or 64K pages. Another issue is that actual pages pinning/unpinning happens on every DMA map/unmap request. This does not affect the performance much now as we spend way too much time now on switching context between guest/userspace/host but this will start to matter when we add in-kernel DMA map/unmap acceleration. This introduces a new IOMMU type for SPAPR - VFIO_SPAPR_TCE_v2_IOMMU. New IOMMU deprecates VFIO_IOMMU_ENABLE/VFIO_IOMMU_DISABLE and introduces 2 new ioctls to register/unregister DMA memory - VFIO_IOMMU_SPAPR_REGISTER_MEMORY and VFIO_IOMMU_SPAPR_UNREGISTER_MEMORY - which receive user space address and size of a memory region which needs to be pinned/unpinned and counted in locked_vm. New IOMMU splits physical pages pinning and TCE table update into 2 different operations. It requires: 1) guest pages to be registered first 2) consequent map/unmap requests to work only with pre-registered memory. For the default single window case this means that the entire guest (instead of 2GB) needs to be pinned before using VFIO. When a huge DMA window is added, no additional pinning will be required, otherwise it would be guest RAM + 2GB. The new memory registration ioctls are not supported by VFIO_SPAPR_TCE_IOMMU. Dynamic DMA window and in-kernel acceleration will require memory to be preregistered in order to work. The accounting is done per the user process. This advertises v2 SPAPR TCE IOMMU and restricts what the userspace can do with v1 or v2 IOMMUs. In order to support memory pre-registration, we need a way to track the use of every registered memory region and only allow unregistration if a region is not in use anymore. So we need a way to tell from what region the just cleared TCE was from. This adds a userspace view of the TCE table into iommu_table struct. It contains userspace address, one per TCE entry. The table is only allocated when the ownership over an IOMMU group is taken which means it is only used from outside of the powernv code (such as VFIO). As v2 IOMMU supports IODA2 and pre-IODA2 IOMMUs (which do not support DDW API), this creates a default DMA window for IODA2 for consistency. Signed-off-by: Alexey Kardashevskiy <aik@ozlabs.ru> [aw: for the vfio related changes] Acked-by: Alex Williamson <alex.williamson@redhat.com> Reviewed-by: David Gibson <david@gibson.dropbear.id.au> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
2015-06-05 16:35:25 +10:00
vfio: Define capability chains We have a few cases where we need to extend the data returned from the INFO ioctls in VFIO. For instance we already have devices exposed through vfio-pci where VFIO_DEVICE_GET_REGION_INFO reports the region as mmap-capable, but really only supports sparse mmaps, avoiding the MSI-X table. If we wanted to provide in-kernel emulation or extended functionality for devices, we'd also want the ability to tell the user not to mmap various regions, rather than forcing them to figure it out on their own. Another example is VFIO_IOMMU_GET_INFO. We'd really like to expose the actual IOVA capabilities of the IOMMU rather than letting the user assume the address space they have available to them. We could add IOVA base and size fields to struct vfio_iommu_type1_info, but what if we have multiple IOVA ranges. For instance x86 uses a range of addresses at 0xfee00000 for MSI vectors. These typically are not available for standard DMA IOVA mappings and splits our available IOVA space into two ranges. POWER systems have both an IOVA window below 4G as well as dynamic data window which they can use to remap all of guest memory. Representing variable sized arrays within a fixed structure makes it very difficult to parse, we'd therefore like to put this data beyond fixed fields within the data structures. One way to do this is to emulate capabilities in PCI configuration space. A new flag indciates whether capabilties are supported and a new fixed field reports the offset of the first entry. Users can then walk the chain to find capabilities, adding capabilities does not require additional fields in the fixed structure, and parsing variable sized data becomes trivial. This patch outlines the theory and base header structure, which should be shared by all future users. Signed-off-by: Alex Williamson <alex.williamson@redhat.com>
2016-02-22 16:02:32 -07:00
2018-09-25 19:16:20 -04:00
vfio_pci: Add NVIDIA GV100GL [Tesla V100 SXM2] subdriver POWER9 Witherspoon machines come with 4 or 6 V100 GPUs which are not pluggable PCIe devices but still have PCIe links which are used for config space and MMIO. In addition to that the GPUs have 6 NVLinks which are connected to other GPUs and the POWER9 CPU. POWER9 chips have a special unit on a die called an NPU which is an NVLink2 host bus adapter with p2p connections to 2 to 3 GPUs, 3 or 2 NVLinks to each. These systems also support ATS (address translation services) which is a part of the NVLink2 protocol. Such GPUs also share on-board RAM (16GB or 32GB) to the system via the same NVLink2 so a CPU has cache-coherent access to a GPU RAM. This exports GPU RAM to the userspace as a new VFIO device region. This preregisters the new memory as device memory as it might be used for DMA. This inserts pfns from the fault handler as the GPU memory is not onlined until the vendor driver is loaded and trained the NVLinks so doing this earlier causes low level errors which we fence in the firmware so it does not hurt the host system but still better be avoided; for the same reason this does not map GPU RAM into the host kernel (usual thing for emulated access otherwise). This exports an ATSD (Address Translation Shootdown) register of NPU which allows TLB invalidations inside GPU for an operating system. The register conveniently occupies a single 64k page. It is also presented to the userspace as a new VFIO device region. One NPU has 8 ATSD registers, each of them can be used for TLB invalidation in a GPU linked to this NPU. This allocates one ATSD register per an NVLink bridge allowing passing up to 6 registers. Due to the host firmware bug (just recently fixed), only 1 ATSD register per NPU was actually advertised to the host system so this passes that alone register via the first NVLink bridge device in the group which is still enough as QEMU collects them all back and presents to the guest via vPHB to mimic the emulated NPU PHB on the host. In order to provide the userspace with the information about GPU-to-NVLink connections, this exports an additional capability called "tgt" (which is an abbreviated host system bus address). The "tgt" property tells the GPU its own system address and allows the guest driver to conglomerate the routing information so each GPU knows how to get directly to the other GPUs. For ATS to work, the nest MMU (an NVIDIA block in a P9 CPU) needs to know LPID (a logical partition ID or a KVM guest hardware ID in other words) and PID (a memory context ID of a userspace process, not to be confused with a linux pid). This assigns a GPU to LPID in the NPU and this is why this adds a listener for KVM on an IOMMU group. A PID comes via NVLink from a GPU and NPU uses a PID wildcard to pass it through. This requires coherent memory and ATSD to be available on the host as the GPU vendor only supports configurations with both features enabled and other configurations are known not to work. Because of this and because of the ways the features are advertised to the host system (which is a device tree with very platform specific properties), this requires enabled POWERNV platform. The V100 GPUs do not advertise any of these capabilities via the config space and there are more than just one device ID so this relies on the platform to tell whether these GPUs have special abilities such as NVLinks. Signed-off-by: Alexey Kardashevskiy <aik@ozlabs.ru> Acked-by: Alex Williamson <alex.williamson@redhat.com> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
2018-12-20 12:10:36 +11:00
vfio: ABI for mdev display dma-buf operation Add VFIO_DEVICE_QUERY_GFX_PLANE ioctl command to let user query and get a plane and its information. So far, two types of buffers are supported: buffers based on dma-buf and buffers based on region. This ioctl can be invoked with: 1) Either DMABUF or REGION flag. Vendor driver returns a plane_info successfully only when the specific kind of buffer is supported. 2) Flag PROBE. And at the same time either DMABUF or REGION must be set, so that vendor driver returns success only when the specific kind of buffer is supported. Add VFIO_DEVICE_GET_GFX_DMABUF ioctl command to let user get a specific dma-buf fd of an exposed MDEV buffer provided by dmabuf_id which was returned in VFIO_DEVICE_QUERY_GFX_PLANE ioctl command. The life cycle of an exposed MDEV buffer is handled by userspace and tracked by kernel space. The returned dmabuf_id in struct vfio_device_ query_gfx_plane can be a new id of a new exposed buffer or an old id of a re-exported buffer. Host user can check the value of dmabuf_id to see if it needs to create new resources according to the new exposed buffer or just re-use the existing resource related to the old buffer. v18: - update comments for VFIO_DEVICE_GET_GFX_DMABUF. (Alex) v17: - modify VFIO_DEVICE_GET_GFX_DMABUF interface. (Alex) v16: - add x_hot and y_hot fields. (Gerd) - add comments for VFIO_DEVICE_GET_GFX_DMABUF. (Alex) - rebase to 4.14.0-rc6. v15: - add a ioctl to get a dmabuf for a given dmabuf id. (Gerd) v14: - add PROBE, DMABUF and REGION flags. (Alex) v12: - add drm_format_mod back. (Gerd and Zhenyu) - add region_index. (Gerd) v11: - rename plane_type to drm_plane_type. (Gerd) - move fields of vfio_device_query_gfx_plane to vfio_device_gfx_plane_info. (Gerd) - remove drm_format_mod, start fields. (Daniel) - remove plane_id. v10: - refine the ABI API VFIO_DEVICE_QUERY_GFX_PLANE. (Alex) (Gerd) v3: - add a field gvt_plane_info in the drm_i915_gem_obj structure to save the decoded plane information to avoid look up while need the plane info. (Gerd) Signed-off-by: Tina Zhang <tina.zhang@intel.com> Reviewed-by: Gerd Hoffmann <kraxel@redhat.com> Reviewed-by: Kirti Wankhede <kwankhede@nvidia.com> Acked-by: Alex Williamson <alex.williamson@redhat.com> Cc: Daniel Vetter <daniel.vetter@ffwll.ch> Signed-off-by: Zhenyu Wang <zhenyuw@linux.intel.com>
2017-11-23 16:26:35 +08:00
vfio: hugepage support for vfio_iommu_type1 We currently send all mappings to the iommu in PAGE_SIZE chunks, which prevents the iommu from enabling support for larger page sizes. We still need to pin pages, which means we step through them in PAGE_SIZE chunks, but we can batch up contiguous physical memory chunks to allow the iommu the opportunity to use larger pages. The approach here is a bit different that the one currently used for legacy KVM device assignment. Rather than looking at the vma page size and using that as the maximum size to pass to the iommu, we instead simply look at whether the next page is physically contiguous. This means we might ask the iommu to map a 4MB region, while legacy KVM might limit itself to a maximum of 2MB. Splitting our mapping path also allows us to be smarter about locked memory because we can more easily unwind if the user attempts to exceed the limit. Therefore, rather than assuming that a mapping will result in locked memory, we test each page as it is pinned to determine whether it locks RAM vs an mmap'd MMIO region. This should result in better locking granularity and less locked page fudge factors in userspace. The unmap path uses the same algorithm as legacy KVM. We don't want to track the pfn for each mapping ourselves, but we need the pfn in order to unpin pages. We therefore ask the iommu for the iova to physical address translation, ask it to unpin a page, and see how many pages were actually unpinned. iommus supporting large pages will often return something bigger than a page here, which we know will be physically contiguous and we can unpin a batch of pfns. iommus not supporting large mappings won't see an improvement in batching here as they only unmap a page at a time. With this change, we also make a clarification to the API for mapping and unmapping DMA. We can only guarantee unmaps at the same granularity as used for the original mapping. In other words, unmapping a subregion of a previous mapping is not guaranteed and may result in a larger or smaller unmapping than requested. The size field in the unmapping structure is updated to reflect this. Previously this was unmodified on mapping, always returning the the requested unmap size. This is now updated to return the actual unmap size on success, allowing userspace to appropriately track mappings. Signed-off-by: Alex Williamson <alex.williamson@redhat.com>
2013-06-21 09:38:02 -06:00
vfio: hugepage support for vfio_iommu_type1 We currently send all mappings to the iommu in PAGE_SIZE chunks, which prevents the iommu from enabling support for larger page sizes. We still need to pin pages, which means we step through them in PAGE_SIZE chunks, but we can batch up contiguous physical memory chunks to allow the iommu the opportunity to use larger pages. The approach here is a bit different that the one currently used for legacy KVM device assignment. Rather than looking at the vma page size and using that as the maximum size to pass to the iommu, we instead simply look at whether the next page is physically contiguous. This means we might ask the iommu to map a 4MB region, while legacy KVM might limit itself to a maximum of 2MB. Splitting our mapping path also allows us to be smarter about locked memory because we can more easily unwind if the user attempts to exceed the limit. Therefore, rather than assuming that a mapping will result in locked memory, we test each page as it is pinned to determine whether it locks RAM vs an mmap'd MMIO region. This should result in better locking granularity and less locked page fudge factors in userspace. The unmap path uses the same algorithm as legacy KVM. We don't want to track the pfn for each mapping ourselves, but we need the pfn in order to unpin pages. We therefore ask the iommu for the iova to physical address translation, ask it to unpin a page, and see how many pages were actually unpinned. iommus supporting large pages will often return something bigger than a page here, which we know will be physically contiguous and we can unpin a batch of pfns. iommus not supporting large mappings won't see an improvement in batching here as they only unmap a page at a time. With this change, we also make a clarification to the API for mapping and unmapping DMA. We can only guarantee unmaps at the same granularity as used for the original mapping. In other words, unmapping a subregion of a previous mapping is not guaranteed and may result in a larger or smaller unmapping than requested. The size field in the unmapping structure is updated to reflect this. Previously this was unmodified on mapping, always returning the the requested unmap size. This is now updated to return the actual unmap size on success, allowing userspace to appropriately track mappings. Signed-off-by: Alex Williamson <alex.williamson@redhat.com>
2013-06-21 09:38:02 -06:00
vfio: powerpc/spapr: Register memory and define IOMMU v2 The existing implementation accounts the whole DMA window in the locked_vm counter. This is going to be worse with multiple containers and huge DMA windows. Also, real-time accounting would requite additional tracking of accounted pages due to the page size difference - IOMMU uses 4K pages and system uses 4K or 64K pages. Another issue is that actual pages pinning/unpinning happens on every DMA map/unmap request. This does not affect the performance much now as we spend way too much time now on switching context between guest/userspace/host but this will start to matter when we add in-kernel DMA map/unmap acceleration. This introduces a new IOMMU type for SPAPR - VFIO_SPAPR_TCE_v2_IOMMU. New IOMMU deprecates VFIO_IOMMU_ENABLE/VFIO_IOMMU_DISABLE and introduces 2 new ioctls to register/unregister DMA memory - VFIO_IOMMU_SPAPR_REGISTER_MEMORY and VFIO_IOMMU_SPAPR_UNREGISTER_MEMORY - which receive user space address and size of a memory region which needs to be pinned/unpinned and counted in locked_vm. New IOMMU splits physical pages pinning and TCE table update into 2 different operations. It requires: 1) guest pages to be registered first 2) consequent map/unmap requests to work only with pre-registered memory. For the default single window case this means that the entire guest (instead of 2GB) needs to be pinned before using VFIO. When a huge DMA window is added, no additional pinning will be required, otherwise it would be guest RAM + 2GB. The new memory registration ioctls are not supported by VFIO_SPAPR_TCE_IOMMU. Dynamic DMA window and in-kernel acceleration will require memory to be preregistered in order to work. The accounting is done per the user process. This advertises v2 SPAPR TCE IOMMU and restricts what the userspace can do with v1 or v2 IOMMUs. In order to support memory pre-registration, we need a way to track the use of every registered memory region and only allow unregistration if a region is not in use anymore. So we need a way to tell from what region the just cleared TCE was from. This adds a userspace view of the TCE table into iommu_table struct. It contains userspace address, one per TCE entry. The table is only allocated when the ownership over an IOMMU group is taken which means it is only used from outside of the powernv code (such as VFIO). As v2 IOMMU supports IODA2 and pre-IODA2 IOMMUs (which do not support DDW API), this creates a default DMA window for IODA2 for consistency. Signed-off-by: Alexey Kardashevskiy <aik@ozlabs.ru> [aw: for the vfio related changes] Acked-by: Alex Williamson <alex.williamson@redhat.com> Reviewed-by: David Gibson <david@gibson.dropbear.id.au> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
2015-06-05 16:35:25 +10:00