1VFIO - "Virtual Function I/O"[1] 2------------------------------------------------------------------------------- 3Many modern system now provide DMA and interrupt remapping facilities 4to help ensure I/O devices behave within the boundaries they've been 5allotted. This includes x86 hardware with AMD-Vi and Intel VT-d, 6POWER systems with Partitionable Endpoints (PEs) and embedded PowerPC 7systems such as Freescale PAMU. The VFIO driver is an IOMMU/device 8agnostic framework for exposing direct device access to userspace, in 9a secure, IOMMU protected environment. In other words, this allows 10safe[2], non-privileged, userspace drivers. 11 12Why do we want that? Virtual machines often make use of direct device 13access ("device assignment") when configured for the highest possible 14I/O performance. From a device and host perspective, this simply 15turns the VM into a userspace driver, with the benefits of 16significantly reduced latency, higher bandwidth, and direct use of 17bare-metal device drivers[3]. 18 19Some applications, particularly in the high performance computing 20field, also benefit from low-overhead, direct device access from 21userspace. Examples include network adapters (often non-TCP/IP based) 22and compute accelerators. Prior to VFIO, these drivers had to either 23go through the full development cycle to become proper upstream 24driver, be maintained out of tree, or make use of the UIO framework, 25which has no notion of IOMMU protection, limited interrupt support, 26and requires root privileges to access things like PCI configuration 27space. 28 29The VFIO driver framework intends to unify these, replacing both the 30KVM PCI specific device assignment code as well as provide a more 31secure, more featureful userspace driver environment than UIO. 32 33Groups, Devices, and IOMMUs 34------------------------------------------------------------------------------- 35 36Devices are the main target of any I/O driver. Devices typically 37create a programming interface made up of I/O access, interrupts, 38and DMA. Without going into the details of each of these, DMA is 39by far the most critical aspect for maintaining a secure environment 40as allowing a device read-write access to system memory imposes the 41greatest risk to the overall system integrity. 42 43To help mitigate this risk, many modern IOMMUs now incorporate 44isolation properties into what was, in many cases, an interface only 45meant for translation (ie. solving the addressing problems of devices 46with limited address spaces). With this, devices can now be isolated 47from each other and from arbitrary memory access, thus allowing 48things like secure direct assignment of devices into virtual machines. 49 50This isolation is not always at the granularity of a single device 51though. Even when an IOMMU is capable of this, properties of devices, 52interconnects, and IOMMU topologies can each reduce this isolation. 53For instance, an individual device may be part of a larger multi- 54function enclosure. While the IOMMU may be able to distinguish 55between devices within the enclosure, the enclosure may not require 56transactions between devices to reach the IOMMU. Examples of this 57could be anything from a multi-function PCI device with backdoors 58between functions to a non-PCI-ACS (Access Control Services) capable 59bridge allowing redirection without reaching the IOMMU. Topology 60can also play a factor in terms of hiding devices. A PCIe-to-PCI 61bridge masks the devices behind it, making transaction appear as if 62from the bridge itself. Obviously IOMMU design plays a major factor 63as well. 64 65Therefore, while for the most part an IOMMU may have device level 66granularity, any system is susceptible to reduced granularity. The 67IOMMU API therefore supports a notion of IOMMU groups. A group is 68a set of devices which is isolatable from all other devices in the 69system. Groups are therefore the unit of ownership used by VFIO. 70 71While the group is the minimum granularity that must be used to 72ensure secure user access, it's not necessarily the preferred 73granularity. In IOMMUs which make use of page tables, it may be 74possible to share a set of page tables between different groups, 75reducing the overhead both to the platform (reduced TLB thrashing, 76reduced duplicate page tables), and to the user (programming only 77a single set of translations). For this reason, VFIO makes use of 78a container class, which may hold one or more groups. A container 79is created by simply opening the /dev/vfio/vfio character device. 80 81On its own, the container provides little functionality, with all 82but a couple version and extension query interfaces locked away. 83The user needs to add a group into the container for the next level 84of functionality. To do this, the user first needs to identify the 85group associated with the desired device. This can be done using 86the sysfs links described in the example below. By unbinding the 87device from the host driver and binding it to a VFIO driver, a new 88VFIO group will appear for the group as /dev/vfio/$GROUP, where 89$GROUP is the IOMMU group number of which the device is a member. 90If the IOMMU group contains multiple devices, each will need to 91be bound to a VFIO driver before operations on the VFIO group 92are allowed (it's also sufficient to only unbind the device from 93host drivers if a VFIO driver is unavailable; this will make the 94group available, but not that particular device). TBD - interface 95for disabling driver probing/locking a device. 96 97Once the group is ready, it may be added to the container by opening 98the VFIO group character device (/dev/vfio/$GROUP) and using the 99VFIO_GROUP_SET_CONTAINER ioctl, passing the file descriptor of the 100previously opened container file. If desired and if the IOMMU driver 101supports sharing the IOMMU context between groups, multiple groups may 102be set to the same container. If a group fails to set to a container 103with existing groups, a new empty container will need to be used 104instead. 105 106With a group (or groups) attached to a container, the remaining 107ioctls become available, enabling access to the VFIO IOMMU interfaces. 108Additionally, it now becomes possible to get file descriptors for each 109device within a group using an ioctl on the VFIO group file descriptor. 110 111The VFIO device API includes ioctls for describing the device, the I/O 112regions and their read/write/mmap offsets on the device descriptor, as 113well as mechanisms for describing and registering interrupt 114notifications. 115 116VFIO Usage Example 117------------------------------------------------------------------------------- 118 119Assume user wants to access PCI device 0000:06:0d.0 120 121$ readlink /sys/bus/pci/devices/0000:06:0d.0/iommu_group 122../../../../kernel/iommu_groups/26 123 124This device is therefore in IOMMU group 26. This device is on the 125pci bus, therefore the user will make use of vfio-pci to manage the 126group: 127 128# modprobe vfio-pci 129 130Binding this device to the vfio-pci driver creates the VFIO group 131character devices for this group: 132 133$ lspci -n -s 0000:06:0d.0 13406:0d.0 0401: 1102:0002 (rev 08) 135# echo 0000:06:0d.0 > /sys/bus/pci/devices/0000:06:0d.0/driver/unbind 136# echo 1102 0002 > /sys/bus/pci/drivers/vfio-pci/new_id 137 138Now we need to look at what other devices are in the group to free 139it for use by VFIO: 140 141$ ls -l /sys/bus/pci/devices/0000:06:0d.0/iommu_group/devices 142total 0 143lrwxrwxrwx. 1 root root 0 Apr 23 16:13 0000:00:1e.0 -> 144 ../../../../devices/pci0000:00/0000:00:1e.0 145lrwxrwxrwx. 1 root root 0 Apr 23 16:13 0000:06:0d.0 -> 146 ../../../../devices/pci0000:00/0000:00:1e.0/0000:06:0d.0 147lrwxrwxrwx. 1 root root 0 Apr 23 16:13 0000:06:0d.1 -> 148 ../../../../devices/pci0000:00/0000:00:1e.0/0000:06:0d.1 149 150This device is behind a PCIe-to-PCI bridge[4], therefore we also 151need to add device 0000:06:0d.1 to the group following the same 152procedure as above. Device 0000:00:1e.0 is a bridge that does 153not currently have a host driver, therefore it's not required to 154bind this device to the vfio-pci driver (vfio-pci does not currently 155support PCI bridges). 156 157The final step is to provide the user with access to the group if 158unprivileged operation is desired (note that /dev/vfio/vfio provides 159no capabilities on its own and is therefore expected to be set to 160mode 0666 by the system). 161 162# chown user:user /dev/vfio/26 163 164The user now has full access to all the devices and the iommu for this 165group and can access them as follows: 166 167 int container, group, device, i; 168 struct vfio_group_status group_status = 169 { .argsz = sizeof(group_status) }; 170 struct vfio_iommu_type1_info iommu_info = { .argsz = sizeof(iommu_info) }; 171 struct vfio_iommu_type1_dma_map dma_map = { .argsz = sizeof(dma_map) }; 172 struct vfio_device_info device_info = { .argsz = sizeof(device_info) }; 173 174 /* Create a new container */ 175 container = open("/dev/vfio/vfio", O_RDWR); 176 177 if (ioctl(container, VFIO_GET_API_VERSION) != VFIO_API_VERSION) 178 /* Unknown API version */ 179 180 if (!ioctl(container, VFIO_CHECK_EXTENSION, VFIO_TYPE1_IOMMU)) 181 /* Doesn't support the IOMMU driver we want. */ 182 183 /* Open the group */ 184 group = open("/dev/vfio/26", O_RDWR); 185 186 /* Test the group is viable and available */ 187 ioctl(group, VFIO_GROUP_GET_STATUS, &group_status); 188 189 if (!(group_status.flags & VFIO_GROUP_FLAGS_VIABLE)) 190 /* Group is not viable (ie, not all devices bound for vfio) */ 191 192 /* Add the group to the container */ 193 ioctl(group, VFIO_GROUP_SET_CONTAINER, &container); 194 195 /* Enable the IOMMU model we want */ 196 ioctl(container, VFIO_SET_IOMMU, VFIO_TYPE1_IOMMU); 197 198 /* Get addition IOMMU info */ 199 ioctl(container, VFIO_IOMMU_GET_INFO, &iommu_info); 200 201 /* Allocate some space and setup a DMA mapping */ 202 dma_map.vaddr = mmap(0, 1024 * 1024, PROT_READ | PROT_WRITE, 203 MAP_PRIVATE | MAP_ANONYMOUS, 0, 0); 204 dma_map.size = 1024 * 1024; 205 dma_map.iova = 0; /* 1MB starting at 0x0 from device view */ 206 dma_map.flags = VFIO_DMA_MAP_FLAG_READ | VFIO_DMA_MAP_FLAG_WRITE; 207 208 ioctl(container, VFIO_IOMMU_MAP_DMA, &dma_map); 209 210 /* Get a file descriptor for the device */ 211 device = ioctl(group, VFIO_GROUP_GET_DEVICE_FD, "0000:06:0d.0"); 212 213 /* Test and setup the device */ 214 ioctl(device, VFIO_DEVICE_GET_INFO, &device_info); 215 216 for (i = 0; i < device_info.num_regions; i++) { 217 struct vfio_region_info reg = { .argsz = sizeof(reg) }; 218 219 reg.index = i; 220 221 ioctl(device, VFIO_DEVICE_GET_REGION_INFO, ®); 222 223 /* Setup mappings... read/write offsets, mmaps 224 * For PCI devices, config space is a region */ 225 } 226 227 for (i = 0; i < device_info.num_irqs; i++) { 228 struct vfio_irq_info irq = { .argsz = sizeof(irq) }; 229 230 irq.index = i; 231 232 ioctl(device, VFIO_DEVICE_GET_IRQ_INFO, &irq); 233 234 /* Setup IRQs... eventfds, VFIO_DEVICE_SET_IRQS */ 235 } 236 237 /* Gratuitous device reset and go... */ 238 ioctl(device, VFIO_DEVICE_RESET); 239 240VFIO User API 241------------------------------------------------------------------------------- 242 243Please see include/linux/vfio.h for complete API documentation. 244 245VFIO bus driver API 246------------------------------------------------------------------------------- 247 248VFIO bus drivers, such as vfio-pci make use of only a few interfaces 249into VFIO core. When devices are bound and unbound to the driver, 250the driver should call vfio_add_group_dev() and vfio_del_group_dev() 251respectively: 252 253extern int vfio_add_group_dev(struct iommu_group *iommu_group, 254 struct device *dev, 255 const struct vfio_device_ops *ops, 256 void *device_data); 257 258extern void *vfio_del_group_dev(struct device *dev); 259 260vfio_add_group_dev() indicates to the core to begin tracking the 261specified iommu_group and register the specified dev as owned by 262a VFIO bus driver. The driver provides an ops structure for callbacks 263similar to a file operations structure: 264 265struct vfio_device_ops { 266 int (*open)(void *device_data); 267 void (*release)(void *device_data); 268 ssize_t (*read)(void *device_data, char __user *buf, 269 size_t count, loff_t *ppos); 270 ssize_t (*write)(void *device_data, const char __user *buf, 271 size_t size, loff_t *ppos); 272 long (*ioctl)(void *device_data, unsigned int cmd, 273 unsigned long arg); 274 int (*mmap)(void *device_data, struct vm_area_struct *vma); 275}; 276 277Each function is passed the device_data that was originally registered 278in the vfio_add_group_dev() call above. This allows the bus driver 279an easy place to store its opaque, private data. The open/release 280callbacks are issued when a new file descriptor is created for a 281device (via VFIO_GROUP_GET_DEVICE_FD). The ioctl interface provides 282a direct pass through for VFIO_DEVICE_* ioctls. The read/write/mmap 283interfaces implement the device region access defined by the device's 284own VFIO_DEVICE_GET_REGION_INFO ioctl. 285 286 287PPC64 sPAPR implementation note 288------------------------------------------------------------------------------- 289 290This implementation has some specifics: 291 2921) Only one IOMMU group per container is supported as an IOMMU group 293represents the minimal entity which isolation can be guaranteed for and 294groups are allocated statically, one per a Partitionable Endpoint (PE) 295(PE is often a PCI domain but not always). 296 2972) The hardware supports so called DMA windows - the PCI address range 298within which DMA transfer is allowed, any attempt to access address space 299out of the window leads to the whole PE isolation. 300 3013) PPC64 guests are paravirtualized but not fully emulated. There is an API 302to map/unmap pages for DMA, and it normally maps 1..32 pages per call and 303currently there is no way to reduce the number of calls. In order to make things 304faster, the map/unmap handling has been implemented in real mode which provides 305an excellent performance which has limitations such as inability to do 306locked pages accounting in real time. 307 3084) According to sPAPR specification, A Partitionable Endpoint (PE) is an I/O 309subtree that can be treated as a unit for the purposes of partitioning and 310error recovery. A PE may be a single or multi-function IOA (IO Adapter), a 311function of a multi-function IOA, or multiple IOAs (possibly including switch 312and bridge structures above the multiple IOAs). PPC64 guests detect PCI errors 313and recover from them via EEH RTAS services, which works on the basis of 314additional ioctl commands. 315 316So 4 additional ioctls have been added: 317 318 VFIO_IOMMU_SPAPR_TCE_GET_INFO - returns the size and the start 319 of the DMA window on the PCI bus. 320 321 VFIO_IOMMU_ENABLE - enables the container. The locked pages accounting 322 is done at this point. This lets user first to know what 323 the DMA window is and adjust rlimit before doing any real job. 324 325 VFIO_IOMMU_DISABLE - disables the container. 326 327 VFIO_EEH_PE_OP - provides an API for EEH setup, error detection and recovery. 328 329The code flow from the example above should be slightly changed: 330 331 struct vfio_eeh_pe_op pe_op = { .argsz = sizeof(pe_op), .flags = 0 }; 332 333 ..... 334 /* Add the group to the container */ 335 ioctl(group, VFIO_GROUP_SET_CONTAINER, &container); 336 337 /* Enable the IOMMU model we want */ 338 ioctl(container, VFIO_SET_IOMMU, VFIO_SPAPR_TCE_IOMMU) 339 340 /* Get addition sPAPR IOMMU info */ 341 vfio_iommu_spapr_tce_info spapr_iommu_info; 342 ioctl(container, VFIO_IOMMU_SPAPR_TCE_GET_INFO, &spapr_iommu_info); 343 344 if (ioctl(container, VFIO_IOMMU_ENABLE)) 345 /* Cannot enable container, may be low rlimit */ 346 347 /* Allocate some space and setup a DMA mapping */ 348 dma_map.vaddr = mmap(0, 1024 * 1024, PROT_READ | PROT_WRITE, 349 MAP_PRIVATE | MAP_ANONYMOUS, 0, 0); 350 351 dma_map.size = 1024 * 1024; 352 dma_map.iova = 0; /* 1MB starting at 0x0 from device view */ 353 dma_map.flags = VFIO_DMA_MAP_FLAG_READ | VFIO_DMA_MAP_FLAG_WRITE; 354 355 /* Check here is .iova/.size are within DMA window from spapr_iommu_info */ 356 ioctl(container, VFIO_IOMMU_MAP_DMA, &dma_map); 357 358 /* Get a file descriptor for the device */ 359 device = ioctl(group, VFIO_GROUP_GET_DEVICE_FD, "0000:06:0d.0"); 360 361 .... 362 363 /* Gratuitous device reset and go... */ 364 ioctl(device, VFIO_DEVICE_RESET); 365 366 /* Make sure EEH is supported */ 367 ioctl(container, VFIO_CHECK_EXTENSION, VFIO_EEH); 368 369 /* Enable the EEH functionality on the device */ 370 pe_op.op = VFIO_EEH_PE_ENABLE; 371 ioctl(container, VFIO_EEH_PE_OP, &pe_op); 372 373 /* You're suggested to create additional data struct to represent 374 * PE, and put child devices belonging to same IOMMU group to the 375 * PE instance for later reference. 376 */ 377 378 /* Check the PE's state and make sure it's in functional state */ 379 pe_op.op = VFIO_EEH_PE_GET_STATE; 380 ioctl(container, VFIO_EEH_PE_OP, &pe_op); 381 382 /* Save device state using pci_save_state(). 383 * EEH should be enabled on the specified device. 384 */ 385 386 .... 387 388 /* When 0xFF's returned from reading PCI config space or IO BARs 389 * of the PCI device. Check the PE's state to see if that has been 390 * frozen. 391 */ 392 ioctl(container, VFIO_EEH_PE_OP, &pe_op); 393 394 /* Waiting for pending PCI transactions to be completed and don't 395 * produce any more PCI traffic from/to the affected PE until 396 * recovery is finished. 397 */ 398 399 /* Enable IO for the affected PE and collect logs. Usually, the 400 * standard part of PCI config space, AER registers are dumped 401 * as logs for further analysis. 402 */ 403 pe_op.op = VFIO_EEH_PE_UNFREEZE_IO; 404 ioctl(container, VFIO_EEH_PE_OP, &pe_op); 405 406 /* 407 * Issue PE reset: hot or fundamental reset. Usually, hot reset 408 * is enough. However, the firmware of some PCI adapters would 409 * require fundamental reset. 410 */ 411 pe_op.op = VFIO_EEH_PE_RESET_HOT; 412 ioctl(container, VFIO_EEH_PE_OP, &pe_op); 413 pe_op.op = VFIO_EEH_PE_RESET_DEACTIVATE; 414 ioctl(container, VFIO_EEH_PE_OP, &pe_op); 415 416 /* Configure the PCI bridges for the affected PE */ 417 pe_op.op = VFIO_EEH_PE_CONFIGURE; 418 ioctl(container, VFIO_EEH_PE_OP, &pe_op); 419 420 /* Restored state we saved at initialization time. pci_restore_state() 421 * is good enough as an example. 422 */ 423 424 /* Hopefully, error is recovered successfully. Now, you can resume to 425 * start PCI traffic to/from the affected PE. 426 */ 427 428 .... 429 430------------------------------------------------------------------------------- 431 432[1] VFIO was originally an acronym for "Virtual Function I/O" in its 433initial implementation by Tom Lyon while as Cisco. We've since 434outgrown the acronym, but it's catchy. 435 436[2] "safe" also depends upon a device being "well behaved". It's 437possible for multi-function devices to have backdoors between 438functions and even for single function devices to have alternative 439access to things like PCI config space through MMIO registers. To 440guard against the former we can include additional precautions in the 441IOMMU driver to group multi-function PCI devices together 442(iommu=group_mf). The latter we can't prevent, but the IOMMU should 443still provide isolation. For PCI, SR-IOV Virtual Functions are the 444best indicator of "well behaved", as these are designed for 445virtualization usage models. 446 447[3] As always there are trade-offs to virtual machine device 448assignment that are beyond the scope of VFIO. It's expected that 449future IOMMU technologies will reduce some, but maybe not all, of 450these trade-offs. 451 452[4] In this case the device is below a PCI bridge, so transactions 453from either function of the device are indistinguishable to the iommu: 454 455-[0000:00]-+-1e.0-[06]--+-0d.0 456 \-0d.1 457 45800:1e.0 PCI bridge: Intel Corporation 82801 PCI Bridge (rev 90) 459