1.. SPDX-License-Identifier: GPL-2.0 2 3=================================================================== 4The Definitive KVM (Kernel-based Virtual Machine) API Documentation 5=================================================================== 6 71. General description 8====================== 9 10The kvm API is a set of ioctls that are issued to control various aspects 11of a virtual machine. The ioctls belong to the following classes: 12 13 - System ioctls: These query and set global attributes which affect the 14 whole kvm subsystem. In addition a system ioctl is used to create 15 virtual machines. 16 17 - VM ioctls: These query and set attributes that affect an entire virtual 18 machine, for example memory layout. In addition a VM ioctl is used to 19 create virtual cpus (vcpus) and devices. 20 21 VM ioctls must be issued from the same process (address space) that was 22 used to create the VM. 23 24 - vcpu ioctls: These query and set attributes that control the operation 25 of a single virtual cpu. 26 27 vcpu ioctls should be issued from the same thread that was used to create 28 the vcpu, except for asynchronous vcpu ioctl that are marked as such in 29 the documentation. Otherwise, the first ioctl after switching threads 30 could see a performance impact. 31 32 - device ioctls: These query and set attributes that control the operation 33 of a single device. 34 35 device ioctls must be issued from the same process (address space) that 36 was used to create the VM. 37 382. File descriptors 39=================== 40 41The kvm API is centered around file descriptors. An initial 42open("/dev/kvm") obtains a handle to the kvm subsystem; this handle 43can be used to issue system ioctls. A KVM_CREATE_VM ioctl on this 44handle will create a VM file descriptor which can be used to issue VM 45ioctls. A KVM_CREATE_VCPU or KVM_CREATE_DEVICE ioctl on a VM fd will 46create a virtual cpu or device and return a file descriptor pointing to 47the new resource. Finally, ioctls on a vcpu or device fd can be used 48to control the vcpu or device. For vcpus, this includes the important 49task of actually running guest code. 50 51In general file descriptors can be migrated among processes by means 52of fork() and the SCM_RIGHTS facility of unix domain socket. These 53kinds of tricks are explicitly not supported by kvm. While they will 54not cause harm to the host, their actual behavior is not guaranteed by 55the API. See "General description" for details on the ioctl usage 56model that is supported by KVM. 57 58It is important to note that although VM ioctls may only be issued from 59the process that created the VM, a VM's lifecycle is associated with its 60file descriptor, not its creator (process). In other words, the VM and 61its resources, *including the associated address space*, are not freed 62until the last reference to the VM's file descriptor has been released. 63For example, if fork() is issued after ioctl(KVM_CREATE_VM), the VM will 64not be freed until both the parent (original) process and its child have 65put their references to the VM's file descriptor. 66 67Because a VM's resources are not freed until the last reference to its 68file descriptor is released, creating additional references to a VM 69via fork(), dup(), etc... without careful consideration is strongly 70discouraged and may have unwanted side effects, e.g. memory allocated 71by and on behalf of the VM's process may not be freed/unaccounted when 72the VM is shut down. 73 74 753. Extensions 76============= 77 78As of Linux 2.6.22, the KVM ABI has been stabilized: no backward 79incompatible change are allowed. However, there is an extension 80facility that allows backward-compatible extensions to the API to be 81queried and used. 82 83The extension mechanism is not based on the Linux version number. 84Instead, kvm defines extension identifiers and a facility to query 85whether a particular extension identifier is available. If it is, a 86set of ioctls is available for application use. 87 88 894. API description 90================== 91 92This section describes ioctls that can be used to control kvm guests. 93For each ioctl, the following information is provided along with a 94description: 95 96 Capability: 97 which KVM extension provides this ioctl. Can be 'basic', 98 which means that is will be provided by any kernel that supports 99 API version 12 (see section 4.1), a KVM_CAP_xyz constant, which 100 means availability needs to be checked with KVM_CHECK_EXTENSION 101 (see section 4.4), or 'none' which means that while not all kernels 102 support this ioctl, there's no capability bit to check its 103 availability: for kernels that don't support the ioctl, 104 the ioctl returns -ENOTTY. 105 106 Architectures: 107 which instruction set architectures provide this ioctl. 108 x86 includes both i386 and x86_64. 109 110 Type: 111 system, vm, or vcpu. 112 113 Parameters: 114 what parameters are accepted by the ioctl. 115 116 Returns: 117 the return value. General error numbers (EBADF, ENOMEM, EINVAL) 118 are not detailed, but errors with specific meanings are. 119 120 1214.1 KVM_GET_API_VERSION 122----------------------- 123 124:Capability: basic 125:Architectures: all 126:Type: system ioctl 127:Parameters: none 128:Returns: the constant KVM_API_VERSION (=12) 129 130This identifies the API version as the stable kvm API. It is not 131expected that this number will change. However, Linux 2.6.20 and 1322.6.21 report earlier versions; these are not documented and not 133supported. Applications should refuse to run if KVM_GET_API_VERSION 134returns a value other than 12. If this check passes, all ioctls 135described as 'basic' will be available. 136 137 1384.2 KVM_CREATE_VM 139----------------- 140 141:Capability: basic 142:Architectures: all 143:Type: system ioctl 144:Parameters: machine type identifier (KVM_VM_*) 145:Returns: a VM fd that can be used to control the new virtual machine. 146 147The new VM has no virtual cpus and no memory. 148You probably want to use 0 as machine type. 149 150In order to create user controlled virtual machines on S390, check 151KVM_CAP_S390_UCONTROL and use the flag KVM_VM_S390_UCONTROL as 152privileged user (CAP_SYS_ADMIN). 153 154To use hardware assisted virtualization on MIPS (VZ ASE) rather than 155the default trap & emulate implementation (which changes the virtual 156memory layout to fit in user mode), check KVM_CAP_MIPS_VZ and use the 157flag KVM_VM_MIPS_VZ. 158 159 160On arm64, the physical address size for a VM (IPA Size limit) is limited 161to 40bits by default. The limit can be configured if the host supports the 162extension KVM_CAP_ARM_VM_IPA_SIZE. When supported, use 163KVM_VM_TYPE_ARM_IPA_SIZE(IPA_Bits) to set the size in the machine type 164identifier, where IPA_Bits is the maximum width of any physical 165address used by the VM. The IPA_Bits is encoded in bits[7-0] of the 166machine type identifier. 167 168e.g, to configure a guest to use 48bit physical address size:: 169 170 vm_fd = ioctl(dev_fd, KVM_CREATE_VM, KVM_VM_TYPE_ARM_IPA_SIZE(48)); 171 172The requested size (IPA_Bits) must be: 173 174 == ========================================================= 175 0 Implies default size, 40bits (for backward compatibility) 176 N Implies N bits, where N is a positive integer such that, 177 32 <= N <= Host_IPA_Limit 178 == ========================================================= 179 180Host_IPA_Limit is the maximum possible value for IPA_Bits on the host and 181is dependent on the CPU capability and the kernel configuration. The limit can 182be retrieved using KVM_CAP_ARM_VM_IPA_SIZE of the KVM_CHECK_EXTENSION 183ioctl() at run-time. 184 185Creation of the VM will fail if the requested IPA size (whether it is 186implicit or explicit) is unsupported on the host. 187 188Please note that configuring the IPA size does not affect the capability 189exposed by the guest CPUs in ID_AA64MMFR0_EL1[PARange]. It only affects 190size of the address translated by the stage2 level (guest physical to 191host physical address translations). 192 193 1944.3 KVM_GET_MSR_INDEX_LIST, KVM_GET_MSR_FEATURE_INDEX_LIST 195---------------------------------------------------------- 196 197:Capability: basic, KVM_CAP_GET_MSR_FEATURES for KVM_GET_MSR_FEATURE_INDEX_LIST 198:Architectures: x86 199:Type: system ioctl 200:Parameters: struct kvm_msr_list (in/out) 201:Returns: 0 on success; -1 on error 202 203Errors: 204 205 ====== ============================================================ 206 EFAULT the msr index list cannot be read from or written to 207 E2BIG the msr index list is too big to fit in the array specified by 208 the user. 209 ====== ============================================================ 210 211:: 212 213 struct kvm_msr_list { 214 __u32 nmsrs; /* number of msrs in entries */ 215 __u32 indices[0]; 216 }; 217 218The user fills in the size of the indices array in nmsrs, and in return 219kvm adjusts nmsrs to reflect the actual number of msrs and fills in the 220indices array with their numbers. 221 222KVM_GET_MSR_INDEX_LIST returns the guest msrs that are supported. The list 223varies by kvm version and host processor, but does not change otherwise. 224 225Note: if kvm indicates supports MCE (KVM_CAP_MCE), then the MCE bank MSRs are 226not returned in the MSR list, as different vcpus can have a different number 227of banks, as set via the KVM_X86_SETUP_MCE ioctl. 228 229KVM_GET_MSR_FEATURE_INDEX_LIST returns the list of MSRs that can be passed 230to the KVM_GET_MSRS system ioctl. This lets userspace probe host capabilities 231and processor features that are exposed via MSRs (e.g., VMX capabilities). 232This list also varies by kvm version and host processor, but does not change 233otherwise. 234 235 2364.4 KVM_CHECK_EXTENSION 237----------------------- 238 239:Capability: basic, KVM_CAP_CHECK_EXTENSION_VM for vm ioctl 240:Architectures: all 241:Type: system ioctl, vm ioctl 242:Parameters: extension identifier (KVM_CAP_*) 243:Returns: 0 if unsupported; 1 (or some other positive integer) if supported 244 245The API allows the application to query about extensions to the core 246kvm API. Userspace passes an extension identifier (an integer) and 247receives an integer that describes the extension availability. 248Generally 0 means no and 1 means yes, but some extensions may report 249additional information in the integer return value. 250 251Based on their initialization different VMs may have different capabilities. 252It is thus encouraged to use the vm ioctl to query for capabilities (available 253with KVM_CAP_CHECK_EXTENSION_VM on the vm fd) 254 2554.5 KVM_GET_VCPU_MMAP_SIZE 256-------------------------- 257 258:Capability: basic 259:Architectures: all 260:Type: system ioctl 261:Parameters: none 262:Returns: size of vcpu mmap area, in bytes 263 264The KVM_RUN ioctl (cf.) communicates with userspace via a shared 265memory region. This ioctl returns the size of that region. See the 266KVM_RUN documentation for details. 267 268Besides the size of the KVM_RUN communication region, other areas of 269the VCPU file descriptor can be mmap-ed, including: 270 271- if KVM_CAP_COALESCED_MMIO is available, a page at 272 KVM_COALESCED_MMIO_PAGE_OFFSET * PAGE_SIZE; for historical reasons, 273 this page is included in the result of KVM_GET_VCPU_MMAP_SIZE. 274 KVM_CAP_COALESCED_MMIO is not documented yet. 275 276- if KVM_CAP_DIRTY_LOG_RING is available, a number of pages at 277 KVM_DIRTY_LOG_PAGE_OFFSET * PAGE_SIZE. For more information on 278 KVM_CAP_DIRTY_LOG_RING, see section 8.3. 279 280 2814.6 KVM_SET_MEMORY_REGION 282------------------------- 283 284:Capability: basic 285:Architectures: all 286:Type: vm ioctl 287:Parameters: struct kvm_memory_region (in) 288:Returns: 0 on success, -1 on error 289 290This ioctl is obsolete and has been removed. 291 292 2934.7 KVM_CREATE_VCPU 294------------------- 295 296:Capability: basic 297:Architectures: all 298:Type: vm ioctl 299:Parameters: vcpu id (apic id on x86) 300:Returns: vcpu fd on success, -1 on error 301 302This API adds a vcpu to a virtual machine. No more than max_vcpus may be added. 303The vcpu id is an integer in the range [0, max_vcpu_id). 304 305The recommended max_vcpus value can be retrieved using the KVM_CAP_NR_VCPUS of 306the KVM_CHECK_EXTENSION ioctl() at run-time. 307The maximum possible value for max_vcpus can be retrieved using the 308KVM_CAP_MAX_VCPUS of the KVM_CHECK_EXTENSION ioctl() at run-time. 309 310If the KVM_CAP_NR_VCPUS does not exist, you should assume that max_vcpus is 4 311cpus max. 312If the KVM_CAP_MAX_VCPUS does not exist, you should assume that max_vcpus is 313same as the value returned from KVM_CAP_NR_VCPUS. 314 315The maximum possible value for max_vcpu_id can be retrieved using the 316KVM_CAP_MAX_VCPU_ID of the KVM_CHECK_EXTENSION ioctl() at run-time. 317 318If the KVM_CAP_MAX_VCPU_ID does not exist, you should assume that max_vcpu_id 319is the same as the value returned from KVM_CAP_MAX_VCPUS. 320 321On powerpc using book3s_hv mode, the vcpus are mapped onto virtual 322threads in one or more virtual CPU cores. (This is because the 323hardware requires all the hardware threads in a CPU core to be in the 324same partition.) The KVM_CAP_PPC_SMT capability indicates the number 325of vcpus per virtual core (vcore). The vcore id is obtained by 326dividing the vcpu id by the number of vcpus per vcore. The vcpus in a 327given vcore will always be in the same physical core as each other 328(though that might be a different physical core from time to time). 329Userspace can control the threading (SMT) mode of the guest by its 330allocation of vcpu ids. For example, if userspace wants 331single-threaded guest vcpus, it should make all vcpu ids be a multiple 332of the number of vcpus per vcore. 333 334For virtual cpus that have been created with S390 user controlled virtual 335machines, the resulting vcpu fd can be memory mapped at page offset 336KVM_S390_SIE_PAGE_OFFSET in order to obtain a memory map of the virtual 337cpu's hardware control block. 338 339 3404.8 KVM_GET_DIRTY_LOG (vm ioctl) 341-------------------------------- 342 343:Capability: basic 344:Architectures: all 345:Type: vm ioctl 346:Parameters: struct kvm_dirty_log (in/out) 347:Returns: 0 on success, -1 on error 348 349:: 350 351 /* for KVM_GET_DIRTY_LOG */ 352 struct kvm_dirty_log { 353 __u32 slot; 354 __u32 padding; 355 union { 356 void __user *dirty_bitmap; /* one bit per page */ 357 __u64 padding; 358 }; 359 }; 360 361Given a memory slot, return a bitmap containing any pages dirtied 362since the last call to this ioctl. Bit 0 is the first page in the 363memory slot. Ensure the entire structure is cleared to avoid padding 364issues. 365 366If KVM_CAP_MULTI_ADDRESS_SPACE is available, bits 16-31 of slot field specifies 367the address space for which you want to return the dirty bitmap. See 368KVM_SET_USER_MEMORY_REGION for details on the usage of slot field. 369 370The bits in the dirty bitmap are cleared before the ioctl returns, unless 371KVM_CAP_MANUAL_DIRTY_LOG_PROTECT2 is enabled. For more information, 372see the description of the capability. 373 3744.9 KVM_SET_MEMORY_ALIAS 375------------------------ 376 377:Capability: basic 378:Architectures: x86 379:Type: vm ioctl 380:Parameters: struct kvm_memory_alias (in) 381:Returns: 0 (success), -1 (error) 382 383This ioctl is obsolete and has been removed. 384 385 3864.10 KVM_RUN 387------------ 388 389:Capability: basic 390:Architectures: all 391:Type: vcpu ioctl 392:Parameters: none 393:Returns: 0 on success, -1 on error 394 395Errors: 396 397 ======= ============================================================== 398 EINTR an unmasked signal is pending 399 ENOEXEC the vcpu hasn't been initialized or the guest tried to execute 400 instructions from device memory (arm64) 401 ENOSYS data abort outside memslots with no syndrome info and 402 KVM_CAP_ARM_NISV_TO_USER not enabled (arm64) 403 EPERM SVE feature set but not finalized (arm64) 404 ======= ============================================================== 405 406This ioctl is used to run a guest virtual cpu. While there are no 407explicit parameters, there is an implicit parameter block that can be 408obtained by mmap()ing the vcpu fd at offset 0, with the size given by 409KVM_GET_VCPU_MMAP_SIZE. The parameter block is formatted as a 'struct 410kvm_run' (see below). 411 412 4134.11 KVM_GET_REGS 414----------------- 415 416:Capability: basic 417:Architectures: all except ARM, arm64 418:Type: vcpu ioctl 419:Parameters: struct kvm_regs (out) 420:Returns: 0 on success, -1 on error 421 422Reads the general purpose registers from the vcpu. 423 424:: 425 426 /* x86 */ 427 struct kvm_regs { 428 /* out (KVM_GET_REGS) / in (KVM_SET_REGS) */ 429 __u64 rax, rbx, rcx, rdx; 430 __u64 rsi, rdi, rsp, rbp; 431 __u64 r8, r9, r10, r11; 432 __u64 r12, r13, r14, r15; 433 __u64 rip, rflags; 434 }; 435 436 /* mips */ 437 struct kvm_regs { 438 /* out (KVM_GET_REGS) / in (KVM_SET_REGS) */ 439 __u64 gpr[32]; 440 __u64 hi; 441 __u64 lo; 442 __u64 pc; 443 }; 444 445 4464.12 KVM_SET_REGS 447----------------- 448 449:Capability: basic 450:Architectures: all except ARM, arm64 451:Type: vcpu ioctl 452:Parameters: struct kvm_regs (in) 453:Returns: 0 on success, -1 on error 454 455Writes the general purpose registers into the vcpu. 456 457See KVM_GET_REGS for the data structure. 458 459 4604.13 KVM_GET_SREGS 461------------------ 462 463:Capability: basic 464:Architectures: x86, ppc 465:Type: vcpu ioctl 466:Parameters: struct kvm_sregs (out) 467:Returns: 0 on success, -1 on error 468 469Reads special registers from the vcpu. 470 471:: 472 473 /* x86 */ 474 struct kvm_sregs { 475 struct kvm_segment cs, ds, es, fs, gs, ss; 476 struct kvm_segment tr, ldt; 477 struct kvm_dtable gdt, idt; 478 __u64 cr0, cr2, cr3, cr4, cr8; 479 __u64 efer; 480 __u64 apic_base; 481 __u64 interrupt_bitmap[(KVM_NR_INTERRUPTS + 63) / 64]; 482 }; 483 484 /* ppc -- see arch/powerpc/include/uapi/asm/kvm.h */ 485 486interrupt_bitmap is a bitmap of pending external interrupts. At most 487one bit may be set. This interrupt has been acknowledged by the APIC 488but not yet injected into the cpu core. 489 490 4914.14 KVM_SET_SREGS 492------------------ 493 494:Capability: basic 495:Architectures: x86, ppc 496:Type: vcpu ioctl 497:Parameters: struct kvm_sregs (in) 498:Returns: 0 on success, -1 on error 499 500Writes special registers into the vcpu. See KVM_GET_SREGS for the 501data structures. 502 503 5044.15 KVM_TRANSLATE 505------------------ 506 507:Capability: basic 508:Architectures: x86 509:Type: vcpu ioctl 510:Parameters: struct kvm_translation (in/out) 511:Returns: 0 on success, -1 on error 512 513Translates a virtual address according to the vcpu's current address 514translation mode. 515 516:: 517 518 struct kvm_translation { 519 /* in */ 520 __u64 linear_address; 521 522 /* out */ 523 __u64 physical_address; 524 __u8 valid; 525 __u8 writeable; 526 __u8 usermode; 527 __u8 pad[5]; 528 }; 529 530 5314.16 KVM_INTERRUPT 532------------------ 533 534:Capability: basic 535:Architectures: x86, ppc, mips 536:Type: vcpu ioctl 537:Parameters: struct kvm_interrupt (in) 538:Returns: 0 on success, negative on failure. 539 540Queues a hardware interrupt vector to be injected. 541 542:: 543 544 /* for KVM_INTERRUPT */ 545 struct kvm_interrupt { 546 /* in */ 547 __u32 irq; 548 }; 549 550X86: 551^^^^ 552 553:Returns: 554 555 ========= =================================== 556 0 on success, 557 -EEXIST if an interrupt is already enqueued 558 -EINVAL the irq number is invalid 559 -ENXIO if the PIC is in the kernel 560 -EFAULT if the pointer is invalid 561 ========= =================================== 562 563Note 'irq' is an interrupt vector, not an interrupt pin or line. This 564ioctl is useful if the in-kernel PIC is not used. 565 566PPC: 567^^^^ 568 569Queues an external interrupt to be injected. This ioctl is overleaded 570with 3 different irq values: 571 572a) KVM_INTERRUPT_SET 573 574 This injects an edge type external interrupt into the guest once it's ready 575 to receive interrupts. When injected, the interrupt is done. 576 577b) KVM_INTERRUPT_UNSET 578 579 This unsets any pending interrupt. 580 581 Only available with KVM_CAP_PPC_UNSET_IRQ. 582 583c) KVM_INTERRUPT_SET_LEVEL 584 585 This injects a level type external interrupt into the guest context. The 586 interrupt stays pending until a specific ioctl with KVM_INTERRUPT_UNSET 587 is triggered. 588 589 Only available with KVM_CAP_PPC_IRQ_LEVEL. 590 591Note that any value for 'irq' other than the ones stated above is invalid 592and incurs unexpected behavior. 593 594This is an asynchronous vcpu ioctl and can be invoked from any thread. 595 596MIPS: 597^^^^^ 598 599Queues an external interrupt to be injected into the virtual CPU. A negative 600interrupt number dequeues the interrupt. 601 602This is an asynchronous vcpu ioctl and can be invoked from any thread. 603 604 6054.17 KVM_DEBUG_GUEST 606-------------------- 607 608:Capability: basic 609:Architectures: none 610:Type: vcpu ioctl 611:Parameters: none) 612:Returns: -1 on error 613 614Support for this has been removed. Use KVM_SET_GUEST_DEBUG instead. 615 616 6174.18 KVM_GET_MSRS 618----------------- 619 620:Capability: basic (vcpu), KVM_CAP_GET_MSR_FEATURES (system) 621:Architectures: x86 622:Type: system ioctl, vcpu ioctl 623:Parameters: struct kvm_msrs (in/out) 624:Returns: number of msrs successfully returned; 625 -1 on error 626 627When used as a system ioctl: 628Reads the values of MSR-based features that are available for the VM. This 629is similar to KVM_GET_SUPPORTED_CPUID, but it returns MSR indices and values. 630The list of msr-based features can be obtained using KVM_GET_MSR_FEATURE_INDEX_LIST 631in a system ioctl. 632 633When used as a vcpu ioctl: 634Reads model-specific registers from the vcpu. Supported msr indices can 635be obtained using KVM_GET_MSR_INDEX_LIST in a system ioctl. 636 637:: 638 639 struct kvm_msrs { 640 __u32 nmsrs; /* number of msrs in entries */ 641 __u32 pad; 642 643 struct kvm_msr_entry entries[0]; 644 }; 645 646 struct kvm_msr_entry { 647 __u32 index; 648 __u32 reserved; 649 __u64 data; 650 }; 651 652Application code should set the 'nmsrs' member (which indicates the 653size of the entries array) and the 'index' member of each array entry. 654kvm will fill in the 'data' member. 655 656 6574.19 KVM_SET_MSRS 658----------------- 659 660:Capability: basic 661:Architectures: x86 662:Type: vcpu ioctl 663:Parameters: struct kvm_msrs (in) 664:Returns: number of msrs successfully set (see below), -1 on error 665 666Writes model-specific registers to the vcpu. See KVM_GET_MSRS for the 667data structures. 668 669Application code should set the 'nmsrs' member (which indicates the 670size of the entries array), and the 'index' and 'data' members of each 671array entry. 672 673It tries to set the MSRs in array entries[] one by one. If setting an MSR 674fails, e.g., due to setting reserved bits, the MSR isn't supported/emulated 675by KVM, etc..., it stops processing the MSR list and returns the number of 676MSRs that have been set successfully. 677 678 6794.20 KVM_SET_CPUID 680------------------ 681 682:Capability: basic 683:Architectures: x86 684:Type: vcpu ioctl 685:Parameters: struct kvm_cpuid (in) 686:Returns: 0 on success, -1 on error 687 688Defines the vcpu responses to the cpuid instruction. Applications 689should use the KVM_SET_CPUID2 ioctl if available. 690 691Note, when this IOCTL fails, KVM gives no guarantees that previous valid CPUID 692configuration (if there is) is not corrupted. Userspace can get a copy of the 693resulting CPUID configuration through KVM_GET_CPUID2 in case. 694 695:: 696 697 struct kvm_cpuid_entry { 698 __u32 function; 699 __u32 eax; 700 __u32 ebx; 701 __u32 ecx; 702 __u32 edx; 703 __u32 padding; 704 }; 705 706 /* for KVM_SET_CPUID */ 707 struct kvm_cpuid { 708 __u32 nent; 709 __u32 padding; 710 struct kvm_cpuid_entry entries[0]; 711 }; 712 713 7144.21 KVM_SET_SIGNAL_MASK 715------------------------ 716 717:Capability: basic 718:Architectures: all 719:Type: vcpu ioctl 720:Parameters: struct kvm_signal_mask (in) 721:Returns: 0 on success, -1 on error 722 723Defines which signals are blocked during execution of KVM_RUN. This 724signal mask temporarily overrides the threads signal mask. Any 725unblocked signal received (except SIGKILL and SIGSTOP, which retain 726their traditional behaviour) will cause KVM_RUN to return with -EINTR. 727 728Note the signal will only be delivered if not blocked by the original 729signal mask. 730 731:: 732 733 /* for KVM_SET_SIGNAL_MASK */ 734 struct kvm_signal_mask { 735 __u32 len; 736 __u8 sigset[0]; 737 }; 738 739 7404.22 KVM_GET_FPU 741---------------- 742 743:Capability: basic 744:Architectures: x86 745:Type: vcpu ioctl 746:Parameters: struct kvm_fpu (out) 747:Returns: 0 on success, -1 on error 748 749Reads the floating point state from the vcpu. 750 751:: 752 753 /* for KVM_GET_FPU and KVM_SET_FPU */ 754 struct kvm_fpu { 755 __u8 fpr[8][16]; 756 __u16 fcw; 757 __u16 fsw; 758 __u8 ftwx; /* in fxsave format */ 759 __u8 pad1; 760 __u16 last_opcode; 761 __u64 last_ip; 762 __u64 last_dp; 763 __u8 xmm[16][16]; 764 __u32 mxcsr; 765 __u32 pad2; 766 }; 767 768 7694.23 KVM_SET_FPU 770---------------- 771 772:Capability: basic 773:Architectures: x86 774:Type: vcpu ioctl 775:Parameters: struct kvm_fpu (in) 776:Returns: 0 on success, -1 on error 777 778Writes the floating point state to the vcpu. 779 780:: 781 782 /* for KVM_GET_FPU and KVM_SET_FPU */ 783 struct kvm_fpu { 784 __u8 fpr[8][16]; 785 __u16 fcw; 786 __u16 fsw; 787 __u8 ftwx; /* in fxsave format */ 788 __u8 pad1; 789 __u16 last_opcode; 790 __u64 last_ip; 791 __u64 last_dp; 792 __u8 xmm[16][16]; 793 __u32 mxcsr; 794 __u32 pad2; 795 }; 796 797 7984.24 KVM_CREATE_IRQCHIP 799----------------------- 800 801:Capability: KVM_CAP_IRQCHIP, KVM_CAP_S390_IRQCHIP (s390) 802:Architectures: x86, ARM, arm64, s390 803:Type: vm ioctl 804:Parameters: none 805:Returns: 0 on success, -1 on error 806 807Creates an interrupt controller model in the kernel. 808On x86, creates a virtual ioapic, a virtual PIC (two PICs, nested), and sets up 809future vcpus to have a local APIC. IRQ routing for GSIs 0-15 is set to both 810PIC and IOAPIC; GSI 16-23 only go to the IOAPIC. 811On ARM/arm64, a GICv2 is created. Any other GIC versions require the usage of 812KVM_CREATE_DEVICE, which also supports creating a GICv2. Using 813KVM_CREATE_DEVICE is preferred over KVM_CREATE_IRQCHIP for GICv2. 814On s390, a dummy irq routing table is created. 815 816Note that on s390 the KVM_CAP_S390_IRQCHIP vm capability needs to be enabled 817before KVM_CREATE_IRQCHIP can be used. 818 819 8204.25 KVM_IRQ_LINE 821----------------- 822 823:Capability: KVM_CAP_IRQCHIP 824:Architectures: x86, arm, arm64 825:Type: vm ioctl 826:Parameters: struct kvm_irq_level 827:Returns: 0 on success, -1 on error 828 829Sets the level of a GSI input to the interrupt controller model in the kernel. 830On some architectures it is required that an interrupt controller model has 831been previously created with KVM_CREATE_IRQCHIP. Note that edge-triggered 832interrupts require the level to be set to 1 and then back to 0. 833 834On real hardware, interrupt pins can be active-low or active-high. This 835does not matter for the level field of struct kvm_irq_level: 1 always 836means active (asserted), 0 means inactive (deasserted). 837 838x86 allows the operating system to program the interrupt polarity 839(active-low/active-high) for level-triggered interrupts, and KVM used 840to consider the polarity. However, due to bitrot in the handling of 841active-low interrupts, the above convention is now valid on x86 too. 842This is signaled by KVM_CAP_X86_IOAPIC_POLARITY_IGNORED. Userspace 843should not present interrupts to the guest as active-low unless this 844capability is present (or unless it is not using the in-kernel irqchip, 845of course). 846 847 848ARM/arm64 can signal an interrupt either at the CPU level, or at the 849in-kernel irqchip (GIC), and for in-kernel irqchip can tell the GIC to 850use PPIs designated for specific cpus. The irq field is interpreted 851like this:: 852 853 bits: | 31 ... 28 | 27 ... 24 | 23 ... 16 | 15 ... 0 | 854 field: | vcpu2_index | irq_type | vcpu_index | irq_id | 855 856The irq_type field has the following values: 857 858- irq_type[0]: 859 out-of-kernel GIC: irq_id 0 is IRQ, irq_id 1 is FIQ 860- irq_type[1]: 861 in-kernel GIC: SPI, irq_id between 32 and 1019 (incl.) 862 (the vcpu_index field is ignored) 863- irq_type[2]: 864 in-kernel GIC: PPI, irq_id between 16 and 31 (incl.) 865 866(The irq_id field thus corresponds nicely to the IRQ ID in the ARM GIC specs) 867 868In both cases, level is used to assert/deassert the line. 869 870When KVM_CAP_ARM_IRQ_LINE_LAYOUT_2 is supported, the target vcpu is 871identified as (256 * vcpu2_index + vcpu_index). Otherwise, vcpu2_index 872must be zero. 873 874Note that on arm/arm64, the KVM_CAP_IRQCHIP capability only conditions 875injection of interrupts for the in-kernel irqchip. KVM_IRQ_LINE can always 876be used for a userspace interrupt controller. 877 878:: 879 880 struct kvm_irq_level { 881 union { 882 __u32 irq; /* GSI */ 883 __s32 status; /* not used for KVM_IRQ_LEVEL */ 884 }; 885 __u32 level; /* 0 or 1 */ 886 }; 887 888 8894.26 KVM_GET_IRQCHIP 890-------------------- 891 892:Capability: KVM_CAP_IRQCHIP 893:Architectures: x86 894:Type: vm ioctl 895:Parameters: struct kvm_irqchip (in/out) 896:Returns: 0 on success, -1 on error 897 898Reads the state of a kernel interrupt controller created with 899KVM_CREATE_IRQCHIP into a buffer provided by the caller. 900 901:: 902 903 struct kvm_irqchip { 904 __u32 chip_id; /* 0 = PIC1, 1 = PIC2, 2 = IOAPIC */ 905 __u32 pad; 906 union { 907 char dummy[512]; /* reserving space */ 908 struct kvm_pic_state pic; 909 struct kvm_ioapic_state ioapic; 910 } chip; 911 }; 912 913 9144.27 KVM_SET_IRQCHIP 915-------------------- 916 917:Capability: KVM_CAP_IRQCHIP 918:Architectures: x86 919:Type: vm ioctl 920:Parameters: struct kvm_irqchip (in) 921:Returns: 0 on success, -1 on error 922 923Sets the state of a kernel interrupt controller created with 924KVM_CREATE_IRQCHIP from a buffer provided by the caller. 925 926:: 927 928 struct kvm_irqchip { 929 __u32 chip_id; /* 0 = PIC1, 1 = PIC2, 2 = IOAPIC */ 930 __u32 pad; 931 union { 932 char dummy[512]; /* reserving space */ 933 struct kvm_pic_state pic; 934 struct kvm_ioapic_state ioapic; 935 } chip; 936 }; 937 938 9394.28 KVM_XEN_HVM_CONFIG 940----------------------- 941 942:Capability: KVM_CAP_XEN_HVM 943:Architectures: x86 944:Type: vm ioctl 945:Parameters: struct kvm_xen_hvm_config (in) 946:Returns: 0 on success, -1 on error 947 948Sets the MSR that the Xen HVM guest uses to initialize its hypercall 949page, and provides the starting address and size of the hypercall 950blobs in userspace. When the guest writes the MSR, kvm copies one 951page of a blob (32- or 64-bit, depending on the vcpu mode) to guest 952memory. 953 954:: 955 956 struct kvm_xen_hvm_config { 957 __u32 flags; 958 __u32 msr; 959 __u64 blob_addr_32; 960 __u64 blob_addr_64; 961 __u8 blob_size_32; 962 __u8 blob_size_64; 963 __u8 pad2[30]; 964 }; 965 966If the KVM_XEN_HVM_CONFIG_INTERCEPT_HCALL flag is returned from the 967KVM_CAP_XEN_HVM check, it may be set in the flags field of this ioctl. 968This requests KVM to generate the contents of the hypercall page 969automatically; hypercalls will be intercepted and passed to userspace 970through KVM_EXIT_XEN. In this case, all of the blob size and address 971fields must be zero. 972 973No other flags are currently valid in the struct kvm_xen_hvm_config. 974 9754.29 KVM_GET_CLOCK 976------------------ 977 978:Capability: KVM_CAP_ADJUST_CLOCK 979:Architectures: x86 980:Type: vm ioctl 981:Parameters: struct kvm_clock_data (out) 982:Returns: 0 on success, -1 on error 983 984Gets the current timestamp of kvmclock as seen by the current guest. In 985conjunction with KVM_SET_CLOCK, it is used to ensure monotonicity on scenarios 986such as migration. 987 988When KVM_CAP_ADJUST_CLOCK is passed to KVM_CHECK_EXTENSION, it returns the 989set of bits that KVM can return in struct kvm_clock_data's flag member. 990 991The only flag defined now is KVM_CLOCK_TSC_STABLE. If set, the returned 992value is the exact kvmclock value seen by all VCPUs at the instant 993when KVM_GET_CLOCK was called. If clear, the returned value is simply 994CLOCK_MONOTONIC plus a constant offset; the offset can be modified 995with KVM_SET_CLOCK. KVM will try to make all VCPUs follow this clock, 996but the exact value read by each VCPU could differ, because the host 997TSC is not stable. 998 999:: 1000 1001 struct kvm_clock_data { 1002 __u64 clock; /* kvmclock current value */ 1003 __u32 flags; 1004 __u32 pad[9]; 1005 }; 1006 1007 10084.30 KVM_SET_CLOCK 1009------------------ 1010 1011:Capability: KVM_CAP_ADJUST_CLOCK 1012:Architectures: x86 1013:Type: vm ioctl 1014:Parameters: struct kvm_clock_data (in) 1015:Returns: 0 on success, -1 on error 1016 1017Sets the current timestamp of kvmclock to the value specified in its parameter. 1018In conjunction with KVM_GET_CLOCK, it is used to ensure monotonicity on scenarios 1019such as migration. 1020 1021:: 1022 1023 struct kvm_clock_data { 1024 __u64 clock; /* kvmclock current value */ 1025 __u32 flags; 1026 __u32 pad[9]; 1027 }; 1028 1029 10304.31 KVM_GET_VCPU_EVENTS 1031------------------------ 1032 1033:Capability: KVM_CAP_VCPU_EVENTS 1034:Extended by: KVM_CAP_INTR_SHADOW 1035:Architectures: x86, arm, arm64 1036:Type: vcpu ioctl 1037:Parameters: struct kvm_vcpu_event (out) 1038:Returns: 0 on success, -1 on error 1039 1040X86: 1041^^^^ 1042 1043Gets currently pending exceptions, interrupts, and NMIs as well as related 1044states of the vcpu. 1045 1046:: 1047 1048 struct kvm_vcpu_events { 1049 struct { 1050 __u8 injected; 1051 __u8 nr; 1052 __u8 has_error_code; 1053 __u8 pending; 1054 __u32 error_code; 1055 } exception; 1056 struct { 1057 __u8 injected; 1058 __u8 nr; 1059 __u8 soft; 1060 __u8 shadow; 1061 } interrupt; 1062 struct { 1063 __u8 injected; 1064 __u8 pending; 1065 __u8 masked; 1066 __u8 pad; 1067 } nmi; 1068 __u32 sipi_vector; 1069 __u32 flags; 1070 struct { 1071 __u8 smm; 1072 __u8 pending; 1073 __u8 smm_inside_nmi; 1074 __u8 latched_init; 1075 } smi; 1076 __u8 reserved[27]; 1077 __u8 exception_has_payload; 1078 __u64 exception_payload; 1079 }; 1080 1081The following bits are defined in the flags field: 1082 1083- KVM_VCPUEVENT_VALID_SHADOW may be set to signal that 1084 interrupt.shadow contains a valid state. 1085 1086- KVM_VCPUEVENT_VALID_SMM may be set to signal that smi contains a 1087 valid state. 1088 1089- KVM_VCPUEVENT_VALID_PAYLOAD may be set to signal that the 1090 exception_has_payload, exception_payload, and exception.pending 1091 fields contain a valid state. This bit will be set whenever 1092 KVM_CAP_EXCEPTION_PAYLOAD is enabled. 1093 1094ARM/ARM64: 1095^^^^^^^^^^ 1096 1097If the guest accesses a device that is being emulated by the host kernel in 1098such a way that a real device would generate a physical SError, KVM may make 1099a virtual SError pending for that VCPU. This system error interrupt remains 1100pending until the guest takes the exception by unmasking PSTATE.A. 1101 1102Running the VCPU may cause it to take a pending SError, or make an access that 1103causes an SError to become pending. The event's description is only valid while 1104the VPCU is not running. 1105 1106This API provides a way to read and write the pending 'event' state that is not 1107visible to the guest. To save, restore or migrate a VCPU the struct representing 1108the state can be read then written using this GET/SET API, along with the other 1109guest-visible registers. It is not possible to 'cancel' an SError that has been 1110made pending. 1111 1112A device being emulated in user-space may also wish to generate an SError. To do 1113this the events structure can be populated by user-space. The current state 1114should be read first, to ensure no existing SError is pending. If an existing 1115SError is pending, the architecture's 'Multiple SError interrupts' rules should 1116be followed. (2.5.3 of DDI0587.a "ARM Reliability, Availability, and 1117Serviceability (RAS) Specification"). 1118 1119SError exceptions always have an ESR value. Some CPUs have the ability to 1120specify what the virtual SError's ESR value should be. These systems will 1121advertise KVM_CAP_ARM_INJECT_SERROR_ESR. In this case exception.has_esr will 1122always have a non-zero value when read, and the agent making an SError pending 1123should specify the ISS field in the lower 24 bits of exception.serror_esr. If 1124the system supports KVM_CAP_ARM_INJECT_SERROR_ESR, but user-space sets the events 1125with exception.has_esr as zero, KVM will choose an ESR. 1126 1127Specifying exception.has_esr on a system that does not support it will return 1128-EINVAL. Setting anything other than the lower 24bits of exception.serror_esr 1129will return -EINVAL. 1130 1131It is not possible to read back a pending external abort (injected via 1132KVM_SET_VCPU_EVENTS or otherwise) because such an exception is always delivered 1133directly to the virtual CPU). 1134 1135:: 1136 1137 struct kvm_vcpu_events { 1138 struct { 1139 __u8 serror_pending; 1140 __u8 serror_has_esr; 1141 __u8 ext_dabt_pending; 1142 /* Align it to 8 bytes */ 1143 __u8 pad[5]; 1144 __u64 serror_esr; 1145 } exception; 1146 __u32 reserved[12]; 1147 }; 1148 11494.32 KVM_SET_VCPU_EVENTS 1150------------------------ 1151 1152:Capability: KVM_CAP_VCPU_EVENTS 1153:Extended by: KVM_CAP_INTR_SHADOW 1154:Architectures: x86, arm, arm64 1155:Type: vcpu ioctl 1156:Parameters: struct kvm_vcpu_event (in) 1157:Returns: 0 on success, -1 on error 1158 1159X86: 1160^^^^ 1161 1162Set pending exceptions, interrupts, and NMIs as well as related states of the 1163vcpu. 1164 1165See KVM_GET_VCPU_EVENTS for the data structure. 1166 1167Fields that may be modified asynchronously by running VCPUs can be excluded 1168from the update. These fields are nmi.pending, sipi_vector, smi.smm, 1169smi.pending. Keep the corresponding bits in the flags field cleared to 1170suppress overwriting the current in-kernel state. The bits are: 1171 1172=============================== ================================== 1173KVM_VCPUEVENT_VALID_NMI_PENDING transfer nmi.pending to the kernel 1174KVM_VCPUEVENT_VALID_SIPI_VECTOR transfer sipi_vector 1175KVM_VCPUEVENT_VALID_SMM transfer the smi sub-struct. 1176=============================== ================================== 1177 1178If KVM_CAP_INTR_SHADOW is available, KVM_VCPUEVENT_VALID_SHADOW can be set in 1179the flags field to signal that interrupt.shadow contains a valid state and 1180shall be written into the VCPU. 1181 1182KVM_VCPUEVENT_VALID_SMM can only be set if KVM_CAP_X86_SMM is available. 1183 1184If KVM_CAP_EXCEPTION_PAYLOAD is enabled, KVM_VCPUEVENT_VALID_PAYLOAD 1185can be set in the flags field to signal that the 1186exception_has_payload, exception_payload, and exception.pending fields 1187contain a valid state and shall be written into the VCPU. 1188 1189ARM/ARM64: 1190^^^^^^^^^^ 1191 1192User space may need to inject several types of events to the guest. 1193 1194Set the pending SError exception state for this VCPU. It is not possible to 1195'cancel' an Serror that has been made pending. 1196 1197If the guest performed an access to I/O memory which could not be handled by 1198userspace, for example because of missing instruction syndrome decode 1199information or because there is no device mapped at the accessed IPA, then 1200userspace can ask the kernel to inject an external abort using the address 1201from the exiting fault on the VCPU. It is a programming error to set 1202ext_dabt_pending after an exit which was not either KVM_EXIT_MMIO or 1203KVM_EXIT_ARM_NISV. This feature is only available if the system supports 1204KVM_CAP_ARM_INJECT_EXT_DABT. This is a helper which provides commonality in 1205how userspace reports accesses for the above cases to guests, across different 1206userspace implementations. Nevertheless, userspace can still emulate all Arm 1207exceptions by manipulating individual registers using the KVM_SET_ONE_REG API. 1208 1209See KVM_GET_VCPU_EVENTS for the data structure. 1210 1211 12124.33 KVM_GET_DEBUGREGS 1213---------------------- 1214 1215:Capability: KVM_CAP_DEBUGREGS 1216:Architectures: x86 1217:Type: vm ioctl 1218:Parameters: struct kvm_debugregs (out) 1219:Returns: 0 on success, -1 on error 1220 1221Reads debug registers from the vcpu. 1222 1223:: 1224 1225 struct kvm_debugregs { 1226 __u64 db[4]; 1227 __u64 dr6; 1228 __u64 dr7; 1229 __u64 flags; 1230 __u64 reserved[9]; 1231 }; 1232 1233 12344.34 KVM_SET_DEBUGREGS 1235---------------------- 1236 1237:Capability: KVM_CAP_DEBUGREGS 1238:Architectures: x86 1239:Type: vm ioctl 1240:Parameters: struct kvm_debugregs (in) 1241:Returns: 0 on success, -1 on error 1242 1243Writes debug registers into the vcpu. 1244 1245See KVM_GET_DEBUGREGS for the data structure. The flags field is unused 1246yet and must be cleared on entry. 1247 1248 12494.35 KVM_SET_USER_MEMORY_REGION 1250------------------------------- 1251 1252:Capability: KVM_CAP_USER_MEMORY 1253:Architectures: all 1254:Type: vm ioctl 1255:Parameters: struct kvm_userspace_memory_region (in) 1256:Returns: 0 on success, -1 on error 1257 1258:: 1259 1260 struct kvm_userspace_memory_region { 1261 __u32 slot; 1262 __u32 flags; 1263 __u64 guest_phys_addr; 1264 __u64 memory_size; /* bytes */ 1265 __u64 userspace_addr; /* start of the userspace allocated memory */ 1266 }; 1267 1268 /* for kvm_memory_region::flags */ 1269 #define KVM_MEM_LOG_DIRTY_PAGES (1UL << 0) 1270 #define KVM_MEM_READONLY (1UL << 1) 1271 1272This ioctl allows the user to create, modify or delete a guest physical 1273memory slot. Bits 0-15 of "slot" specify the slot id and this value 1274should be less than the maximum number of user memory slots supported per 1275VM. The maximum allowed slots can be queried using KVM_CAP_NR_MEMSLOTS. 1276Slots may not overlap in guest physical address space. 1277 1278If KVM_CAP_MULTI_ADDRESS_SPACE is available, bits 16-31 of "slot" 1279specifies the address space which is being modified. They must be 1280less than the value that KVM_CHECK_EXTENSION returns for the 1281KVM_CAP_MULTI_ADDRESS_SPACE capability. Slots in separate address spaces 1282are unrelated; the restriction on overlapping slots only applies within 1283each address space. 1284 1285Deleting a slot is done by passing zero for memory_size. When changing 1286an existing slot, it may be moved in the guest physical memory space, 1287or its flags may be modified, but it may not be resized. 1288 1289Memory for the region is taken starting at the address denoted by the 1290field userspace_addr, which must point at user addressable memory for 1291the entire memory slot size. Any object may back this memory, including 1292anonymous memory, ordinary files, and hugetlbfs. 1293 1294On architectures that support a form of address tagging, userspace_addr must 1295be an untagged address. 1296 1297It is recommended that the lower 21 bits of guest_phys_addr and userspace_addr 1298be identical. This allows large pages in the guest to be backed by large 1299pages in the host. 1300 1301The flags field supports two flags: KVM_MEM_LOG_DIRTY_PAGES and 1302KVM_MEM_READONLY. The former can be set to instruct KVM to keep track of 1303writes to memory within the slot. See KVM_GET_DIRTY_LOG ioctl to know how to 1304use it. The latter can be set, if KVM_CAP_READONLY_MEM capability allows it, 1305to make a new slot read-only. In this case, writes to this memory will be 1306posted to userspace as KVM_EXIT_MMIO exits. 1307 1308When the KVM_CAP_SYNC_MMU capability is available, changes in the backing of 1309the memory region are automatically reflected into the guest. For example, an 1310mmap() that affects the region will be made visible immediately. Another 1311example is madvise(MADV_DROP). 1312 1313It is recommended to use this API instead of the KVM_SET_MEMORY_REGION ioctl. 1314The KVM_SET_MEMORY_REGION does not allow fine grained control over memory 1315allocation and is deprecated. 1316 1317 13184.36 KVM_SET_TSS_ADDR 1319--------------------- 1320 1321:Capability: KVM_CAP_SET_TSS_ADDR 1322:Architectures: x86 1323:Type: vm ioctl 1324:Parameters: unsigned long tss_address (in) 1325:Returns: 0 on success, -1 on error 1326 1327This ioctl defines the physical address of a three-page region in the guest 1328physical address space. The region must be within the first 4GB of the 1329guest physical address space and must not conflict with any memory slot 1330or any mmio address. The guest may malfunction if it accesses this memory 1331region. 1332 1333This ioctl is required on Intel-based hosts. This is needed on Intel hardware 1334because of a quirk in the virtualization implementation (see the internals 1335documentation when it pops into existence). 1336 1337 13384.37 KVM_ENABLE_CAP 1339------------------- 1340 1341:Capability: KVM_CAP_ENABLE_CAP 1342:Architectures: mips, ppc, s390 1343:Type: vcpu ioctl 1344:Parameters: struct kvm_enable_cap (in) 1345:Returns: 0 on success; -1 on error 1346 1347:Capability: KVM_CAP_ENABLE_CAP_VM 1348:Architectures: all 1349:Type: vm ioctl 1350:Parameters: struct kvm_enable_cap (in) 1351:Returns: 0 on success; -1 on error 1352 1353.. note:: 1354 1355 Not all extensions are enabled by default. Using this ioctl the application 1356 can enable an extension, making it available to the guest. 1357 1358On systems that do not support this ioctl, it always fails. On systems that 1359do support it, it only works for extensions that are supported for enablement. 1360 1361To check if a capability can be enabled, the KVM_CHECK_EXTENSION ioctl should 1362be used. 1363 1364:: 1365 1366 struct kvm_enable_cap { 1367 /* in */ 1368 __u32 cap; 1369 1370The capability that is supposed to get enabled. 1371 1372:: 1373 1374 __u32 flags; 1375 1376A bitfield indicating future enhancements. Has to be 0 for now. 1377 1378:: 1379 1380 __u64 args[4]; 1381 1382Arguments for enabling a feature. If a feature needs initial values to 1383function properly, this is the place to put them. 1384 1385:: 1386 1387 __u8 pad[64]; 1388 }; 1389 1390The vcpu ioctl should be used for vcpu-specific capabilities, the vm ioctl 1391for vm-wide capabilities. 1392 13934.38 KVM_GET_MP_STATE 1394--------------------- 1395 1396:Capability: KVM_CAP_MP_STATE 1397:Architectures: x86, s390, arm, arm64 1398:Type: vcpu ioctl 1399:Parameters: struct kvm_mp_state (out) 1400:Returns: 0 on success; -1 on error 1401 1402:: 1403 1404 struct kvm_mp_state { 1405 __u32 mp_state; 1406 }; 1407 1408Returns the vcpu's current "multiprocessing state" (though also valid on 1409uniprocessor guests). 1410 1411Possible values are: 1412 1413 ========================== =============================================== 1414 KVM_MP_STATE_RUNNABLE the vcpu is currently running [x86,arm/arm64] 1415 KVM_MP_STATE_UNINITIALIZED the vcpu is an application processor (AP) 1416 which has not yet received an INIT signal [x86] 1417 KVM_MP_STATE_INIT_RECEIVED the vcpu has received an INIT signal, and is 1418 now ready for a SIPI [x86] 1419 KVM_MP_STATE_HALTED the vcpu has executed a HLT instruction and 1420 is waiting for an interrupt [x86] 1421 KVM_MP_STATE_SIPI_RECEIVED the vcpu has just received a SIPI (vector 1422 accessible via KVM_GET_VCPU_EVENTS) [x86] 1423 KVM_MP_STATE_STOPPED the vcpu is stopped [s390,arm/arm64] 1424 KVM_MP_STATE_CHECK_STOP the vcpu is in a special error state [s390] 1425 KVM_MP_STATE_OPERATING the vcpu is operating (running or halted) 1426 [s390] 1427 KVM_MP_STATE_LOAD the vcpu is in a special load/startup state 1428 [s390] 1429 ========================== =============================================== 1430 1431On x86, this ioctl is only useful after KVM_CREATE_IRQCHIP. Without an 1432in-kernel irqchip, the multiprocessing state must be maintained by userspace on 1433these architectures. 1434 1435For arm/arm64: 1436^^^^^^^^^^^^^^ 1437 1438The only states that are valid are KVM_MP_STATE_STOPPED and 1439KVM_MP_STATE_RUNNABLE which reflect if the vcpu is paused or not. 1440 14414.39 KVM_SET_MP_STATE 1442--------------------- 1443 1444:Capability: KVM_CAP_MP_STATE 1445:Architectures: x86, s390, arm, arm64 1446:Type: vcpu ioctl 1447:Parameters: struct kvm_mp_state (in) 1448:Returns: 0 on success; -1 on error 1449 1450Sets the vcpu's current "multiprocessing state"; see KVM_GET_MP_STATE for 1451arguments. 1452 1453On x86, this ioctl is only useful after KVM_CREATE_IRQCHIP. Without an 1454in-kernel irqchip, the multiprocessing state must be maintained by userspace on 1455these architectures. 1456 1457For arm/arm64: 1458^^^^^^^^^^^^^^ 1459 1460The only states that are valid are KVM_MP_STATE_STOPPED and 1461KVM_MP_STATE_RUNNABLE which reflect if the vcpu should be paused or not. 1462 14634.40 KVM_SET_IDENTITY_MAP_ADDR 1464------------------------------ 1465 1466:Capability: KVM_CAP_SET_IDENTITY_MAP_ADDR 1467:Architectures: x86 1468:Type: vm ioctl 1469:Parameters: unsigned long identity (in) 1470:Returns: 0 on success, -1 on error 1471 1472This ioctl defines the physical address of a one-page region in the guest 1473physical address space. The region must be within the first 4GB of the 1474guest physical address space and must not conflict with any memory slot 1475or any mmio address. The guest may malfunction if it accesses this memory 1476region. 1477 1478Setting the address to 0 will result in resetting the address to its default 1479(0xfffbc000). 1480 1481This ioctl is required on Intel-based hosts. This is needed on Intel hardware 1482because of a quirk in the virtualization implementation (see the internals 1483documentation when it pops into existence). 1484 1485Fails if any VCPU has already been created. 1486 14874.41 KVM_SET_BOOT_CPU_ID 1488------------------------ 1489 1490:Capability: KVM_CAP_SET_BOOT_CPU_ID 1491:Architectures: x86 1492:Type: vm ioctl 1493:Parameters: unsigned long vcpu_id 1494:Returns: 0 on success, -1 on error 1495 1496Define which vcpu is the Bootstrap Processor (BSP). Values are the same 1497as the vcpu id in KVM_CREATE_VCPU. If this ioctl is not called, the default 1498is vcpu 0. This ioctl has to be called before vcpu creation, 1499otherwise it will return EBUSY error. 1500 1501 15024.42 KVM_GET_XSAVE 1503------------------ 1504 1505:Capability: KVM_CAP_XSAVE 1506:Architectures: x86 1507:Type: vcpu ioctl 1508:Parameters: struct kvm_xsave (out) 1509:Returns: 0 on success, -1 on error 1510 1511 1512:: 1513 1514 struct kvm_xsave { 1515 __u32 region[1024]; 1516 }; 1517 1518This ioctl would copy current vcpu's xsave struct to the userspace. 1519 1520 15214.43 KVM_SET_XSAVE 1522------------------ 1523 1524:Capability: KVM_CAP_XSAVE 1525:Architectures: x86 1526:Type: vcpu ioctl 1527:Parameters: struct kvm_xsave (in) 1528:Returns: 0 on success, -1 on error 1529 1530:: 1531 1532 1533 struct kvm_xsave { 1534 __u32 region[1024]; 1535 }; 1536 1537This ioctl would copy userspace's xsave struct to the kernel. 1538 1539 15404.44 KVM_GET_XCRS 1541----------------- 1542 1543:Capability: KVM_CAP_XCRS 1544:Architectures: x86 1545:Type: vcpu ioctl 1546:Parameters: struct kvm_xcrs (out) 1547:Returns: 0 on success, -1 on error 1548 1549:: 1550 1551 struct kvm_xcr { 1552 __u32 xcr; 1553 __u32 reserved; 1554 __u64 value; 1555 }; 1556 1557 struct kvm_xcrs { 1558 __u32 nr_xcrs; 1559 __u32 flags; 1560 struct kvm_xcr xcrs[KVM_MAX_XCRS]; 1561 __u64 padding[16]; 1562 }; 1563 1564This ioctl would copy current vcpu's xcrs to the userspace. 1565 1566 15674.45 KVM_SET_XCRS 1568----------------- 1569 1570:Capability: KVM_CAP_XCRS 1571:Architectures: x86 1572:Type: vcpu ioctl 1573:Parameters: struct kvm_xcrs (in) 1574:Returns: 0 on success, -1 on error 1575 1576:: 1577 1578 struct kvm_xcr { 1579 __u32 xcr; 1580 __u32 reserved; 1581 __u64 value; 1582 }; 1583 1584 struct kvm_xcrs { 1585 __u32 nr_xcrs; 1586 __u32 flags; 1587 struct kvm_xcr xcrs[KVM_MAX_XCRS]; 1588 __u64 padding[16]; 1589 }; 1590 1591This ioctl would set vcpu's xcr to the value userspace specified. 1592 1593 15944.46 KVM_GET_SUPPORTED_CPUID 1595---------------------------- 1596 1597:Capability: KVM_CAP_EXT_CPUID 1598:Architectures: x86 1599:Type: system ioctl 1600:Parameters: struct kvm_cpuid2 (in/out) 1601:Returns: 0 on success, -1 on error 1602 1603:: 1604 1605 struct kvm_cpuid2 { 1606 __u32 nent; 1607 __u32 padding; 1608 struct kvm_cpuid_entry2 entries[0]; 1609 }; 1610 1611 #define KVM_CPUID_FLAG_SIGNIFCANT_INDEX BIT(0) 1612 #define KVM_CPUID_FLAG_STATEFUL_FUNC BIT(1) /* deprecated */ 1613 #define KVM_CPUID_FLAG_STATE_READ_NEXT BIT(2) /* deprecated */ 1614 1615 struct kvm_cpuid_entry2 { 1616 __u32 function; 1617 __u32 index; 1618 __u32 flags; 1619 __u32 eax; 1620 __u32 ebx; 1621 __u32 ecx; 1622 __u32 edx; 1623 __u32 padding[3]; 1624 }; 1625 1626This ioctl returns x86 cpuid features which are supported by both the 1627hardware and kvm in its default configuration. Userspace can use the 1628information returned by this ioctl to construct cpuid information (for 1629KVM_SET_CPUID2) that is consistent with hardware, kernel, and 1630userspace capabilities, and with user requirements (for example, the 1631user may wish to constrain cpuid to emulate older hardware, or for 1632feature consistency across a cluster). 1633 1634Note that certain capabilities, such as KVM_CAP_X86_DISABLE_EXITS, may 1635expose cpuid features (e.g. MONITOR) which are not supported by kvm in 1636its default configuration. If userspace enables such capabilities, it 1637is responsible for modifying the results of this ioctl appropriately. 1638 1639Userspace invokes KVM_GET_SUPPORTED_CPUID by passing a kvm_cpuid2 structure 1640with the 'nent' field indicating the number of entries in the variable-size 1641array 'entries'. If the number of entries is too low to describe the cpu 1642capabilities, an error (E2BIG) is returned. If the number is too high, 1643the 'nent' field is adjusted and an error (ENOMEM) is returned. If the 1644number is just right, the 'nent' field is adjusted to the number of valid 1645entries in the 'entries' array, which is then filled. 1646 1647The entries returned are the host cpuid as returned by the cpuid instruction, 1648with unknown or unsupported features masked out. Some features (for example, 1649x2apic), may not be present in the host cpu, but are exposed by kvm if it can 1650emulate them efficiently. The fields in each entry are defined as follows: 1651 1652 function: 1653 the eax value used to obtain the entry 1654 1655 index: 1656 the ecx value used to obtain the entry (for entries that are 1657 affected by ecx) 1658 1659 flags: 1660 an OR of zero or more of the following: 1661 1662 KVM_CPUID_FLAG_SIGNIFCANT_INDEX: 1663 if the index field is valid 1664 1665 eax, ebx, ecx, edx: 1666 the values returned by the cpuid instruction for 1667 this function/index combination 1668 1669The TSC deadline timer feature (CPUID leaf 1, ecx[24]) is always returned 1670as false, since the feature depends on KVM_CREATE_IRQCHIP for local APIC 1671support. Instead it is reported via:: 1672 1673 ioctl(KVM_CHECK_EXTENSION, KVM_CAP_TSC_DEADLINE_TIMER) 1674 1675if that returns true and you use KVM_CREATE_IRQCHIP, or if you emulate the 1676feature in userspace, then you can enable the feature for KVM_SET_CPUID2. 1677 1678 16794.47 KVM_PPC_GET_PVINFO 1680----------------------- 1681 1682:Capability: KVM_CAP_PPC_GET_PVINFO 1683:Architectures: ppc 1684:Type: vm ioctl 1685:Parameters: struct kvm_ppc_pvinfo (out) 1686:Returns: 0 on success, !0 on error 1687 1688:: 1689 1690 struct kvm_ppc_pvinfo { 1691 __u32 flags; 1692 __u32 hcall[4]; 1693 __u8 pad[108]; 1694 }; 1695 1696This ioctl fetches PV specific information that need to be passed to the guest 1697using the device tree or other means from vm context. 1698 1699The hcall array defines 4 instructions that make up a hypercall. 1700 1701If any additional field gets added to this structure later on, a bit for that 1702additional piece of information will be set in the flags bitmap. 1703 1704The flags bitmap is defined as:: 1705 1706 /* the host supports the ePAPR idle hcall 1707 #define KVM_PPC_PVINFO_FLAGS_EV_IDLE (1<<0) 1708 17094.52 KVM_SET_GSI_ROUTING 1710------------------------ 1711 1712:Capability: KVM_CAP_IRQ_ROUTING 1713:Architectures: x86 s390 arm arm64 1714:Type: vm ioctl 1715:Parameters: struct kvm_irq_routing (in) 1716:Returns: 0 on success, -1 on error 1717 1718Sets the GSI routing table entries, overwriting any previously set entries. 1719 1720On arm/arm64, GSI routing has the following limitation: 1721 1722- GSI routing does not apply to KVM_IRQ_LINE but only to KVM_IRQFD. 1723 1724:: 1725 1726 struct kvm_irq_routing { 1727 __u32 nr; 1728 __u32 flags; 1729 struct kvm_irq_routing_entry entries[0]; 1730 }; 1731 1732No flags are specified so far, the corresponding field must be set to zero. 1733 1734:: 1735 1736 struct kvm_irq_routing_entry { 1737 __u32 gsi; 1738 __u32 type; 1739 __u32 flags; 1740 __u32 pad; 1741 union { 1742 struct kvm_irq_routing_irqchip irqchip; 1743 struct kvm_irq_routing_msi msi; 1744 struct kvm_irq_routing_s390_adapter adapter; 1745 struct kvm_irq_routing_hv_sint hv_sint; 1746 __u32 pad[8]; 1747 } u; 1748 }; 1749 1750 /* gsi routing entry types */ 1751 #define KVM_IRQ_ROUTING_IRQCHIP 1 1752 #define KVM_IRQ_ROUTING_MSI 2 1753 #define KVM_IRQ_ROUTING_S390_ADAPTER 3 1754 #define KVM_IRQ_ROUTING_HV_SINT 4 1755 1756flags: 1757 1758- KVM_MSI_VALID_DEVID: used along with KVM_IRQ_ROUTING_MSI routing entry 1759 type, specifies that the devid field contains a valid value. The per-VM 1760 KVM_CAP_MSI_DEVID capability advertises the requirement to provide 1761 the device ID. If this capability is not available, userspace should 1762 never set the KVM_MSI_VALID_DEVID flag as the ioctl might fail. 1763- zero otherwise 1764 1765:: 1766 1767 struct kvm_irq_routing_irqchip { 1768 __u32 irqchip; 1769 __u32 pin; 1770 }; 1771 1772 struct kvm_irq_routing_msi { 1773 __u32 address_lo; 1774 __u32 address_hi; 1775 __u32 data; 1776 union { 1777 __u32 pad; 1778 __u32 devid; 1779 }; 1780 }; 1781 1782If KVM_MSI_VALID_DEVID is set, devid contains a unique device identifier 1783for the device that wrote the MSI message. For PCI, this is usually a 1784BFD identifier in the lower 16 bits. 1785 1786On x86, address_hi is ignored unless the KVM_X2APIC_API_USE_32BIT_IDS 1787feature of KVM_CAP_X2APIC_API capability is enabled. If it is enabled, 1788address_hi bits 31-8 provide bits 31-8 of the destination id. Bits 7-0 of 1789address_hi must be zero. 1790 1791:: 1792 1793 struct kvm_irq_routing_s390_adapter { 1794 __u64 ind_addr; 1795 __u64 summary_addr; 1796 __u64 ind_offset; 1797 __u32 summary_offset; 1798 __u32 adapter_id; 1799 }; 1800 1801 struct kvm_irq_routing_hv_sint { 1802 __u32 vcpu; 1803 __u32 sint; 1804 }; 1805 1806 18074.55 KVM_SET_TSC_KHZ 1808-------------------- 1809 1810:Capability: KVM_CAP_TSC_CONTROL 1811:Architectures: x86 1812:Type: vcpu ioctl 1813:Parameters: virtual tsc_khz 1814:Returns: 0 on success, -1 on error 1815 1816Specifies the tsc frequency for the virtual machine. The unit of the 1817frequency is KHz. 1818 1819 18204.56 KVM_GET_TSC_KHZ 1821-------------------- 1822 1823:Capability: KVM_CAP_GET_TSC_KHZ 1824:Architectures: x86 1825:Type: vcpu ioctl 1826:Parameters: none 1827:Returns: virtual tsc-khz on success, negative value on error 1828 1829Returns the tsc frequency of the guest. The unit of the return value is 1830KHz. If the host has unstable tsc this ioctl returns -EIO instead as an 1831error. 1832 1833 18344.57 KVM_GET_LAPIC 1835------------------ 1836 1837:Capability: KVM_CAP_IRQCHIP 1838:Architectures: x86 1839:Type: vcpu ioctl 1840:Parameters: struct kvm_lapic_state (out) 1841:Returns: 0 on success, -1 on error 1842 1843:: 1844 1845 #define KVM_APIC_REG_SIZE 0x400 1846 struct kvm_lapic_state { 1847 char regs[KVM_APIC_REG_SIZE]; 1848 }; 1849 1850Reads the Local APIC registers and copies them into the input argument. The 1851data format and layout are the same as documented in the architecture manual. 1852 1853If KVM_X2APIC_API_USE_32BIT_IDS feature of KVM_CAP_X2APIC_API is 1854enabled, then the format of APIC_ID register depends on the APIC mode 1855(reported by MSR_IA32_APICBASE) of its VCPU. x2APIC stores APIC ID in 1856the APIC_ID register (bytes 32-35). xAPIC only allows an 8-bit APIC ID 1857which is stored in bits 31-24 of the APIC register, or equivalently in 1858byte 35 of struct kvm_lapic_state's regs field. KVM_GET_LAPIC must then 1859be called after MSR_IA32_APICBASE has been set with KVM_SET_MSR. 1860 1861If KVM_X2APIC_API_USE_32BIT_IDS feature is disabled, struct kvm_lapic_state 1862always uses xAPIC format. 1863 1864 18654.58 KVM_SET_LAPIC 1866------------------ 1867 1868:Capability: KVM_CAP_IRQCHIP 1869:Architectures: x86 1870:Type: vcpu ioctl 1871:Parameters: struct kvm_lapic_state (in) 1872:Returns: 0 on success, -1 on error 1873 1874:: 1875 1876 #define KVM_APIC_REG_SIZE 0x400 1877 struct kvm_lapic_state { 1878 char regs[KVM_APIC_REG_SIZE]; 1879 }; 1880 1881Copies the input argument into the Local APIC registers. The data format 1882and layout are the same as documented in the architecture manual. 1883 1884The format of the APIC ID register (bytes 32-35 of struct kvm_lapic_state's 1885regs field) depends on the state of the KVM_CAP_X2APIC_API capability. 1886See the note in KVM_GET_LAPIC. 1887 1888 18894.59 KVM_IOEVENTFD 1890------------------ 1891 1892:Capability: KVM_CAP_IOEVENTFD 1893:Architectures: all 1894:Type: vm ioctl 1895:Parameters: struct kvm_ioeventfd (in) 1896:Returns: 0 on success, !0 on error 1897 1898This ioctl attaches or detaches an ioeventfd to a legal pio/mmio address 1899within the guest. A guest write in the registered address will signal the 1900provided event instead of triggering an exit. 1901 1902:: 1903 1904 struct kvm_ioeventfd { 1905 __u64 datamatch; 1906 __u64 addr; /* legal pio/mmio address */ 1907 __u32 len; /* 0, 1, 2, 4, or 8 bytes */ 1908 __s32 fd; 1909 __u32 flags; 1910 __u8 pad[36]; 1911 }; 1912 1913For the special case of virtio-ccw devices on s390, the ioevent is matched 1914to a subchannel/virtqueue tuple instead. 1915 1916The following flags are defined:: 1917 1918 #define KVM_IOEVENTFD_FLAG_DATAMATCH (1 << kvm_ioeventfd_flag_nr_datamatch) 1919 #define KVM_IOEVENTFD_FLAG_PIO (1 << kvm_ioeventfd_flag_nr_pio) 1920 #define KVM_IOEVENTFD_FLAG_DEASSIGN (1 << kvm_ioeventfd_flag_nr_deassign) 1921 #define KVM_IOEVENTFD_FLAG_VIRTIO_CCW_NOTIFY \ 1922 (1 << kvm_ioeventfd_flag_nr_virtio_ccw_notify) 1923 1924If datamatch flag is set, the event will be signaled only if the written value 1925to the registered address is equal to datamatch in struct kvm_ioeventfd. 1926 1927For virtio-ccw devices, addr contains the subchannel id and datamatch the 1928virtqueue index. 1929 1930With KVM_CAP_IOEVENTFD_ANY_LENGTH, a zero length ioeventfd is allowed, and 1931the kernel will ignore the length of guest write and may get a faster vmexit. 1932The speedup may only apply to specific architectures, but the ioeventfd will 1933work anyway. 1934 19354.60 KVM_DIRTY_TLB 1936------------------ 1937 1938:Capability: KVM_CAP_SW_TLB 1939:Architectures: ppc 1940:Type: vcpu ioctl 1941:Parameters: struct kvm_dirty_tlb (in) 1942:Returns: 0 on success, -1 on error 1943 1944:: 1945 1946 struct kvm_dirty_tlb { 1947 __u64 bitmap; 1948 __u32 num_dirty; 1949 }; 1950 1951This must be called whenever userspace has changed an entry in the shared 1952TLB, prior to calling KVM_RUN on the associated vcpu. 1953 1954The "bitmap" field is the userspace address of an array. This array 1955consists of a number of bits, equal to the total number of TLB entries as 1956determined by the last successful call to KVM_CONFIG_TLB, rounded up to the 1957nearest multiple of 64. 1958 1959Each bit corresponds to one TLB entry, ordered the same as in the shared TLB 1960array. 1961 1962The array is little-endian: the bit 0 is the least significant bit of the 1963first byte, bit 8 is the least significant bit of the second byte, etc. 1964This avoids any complications with differing word sizes. 1965 1966The "num_dirty" field is a performance hint for KVM to determine whether it 1967should skip processing the bitmap and just invalidate everything. It must 1968be set to the number of set bits in the bitmap. 1969 1970 19714.62 KVM_CREATE_SPAPR_TCE 1972------------------------- 1973 1974:Capability: KVM_CAP_SPAPR_TCE 1975:Architectures: powerpc 1976:Type: vm ioctl 1977:Parameters: struct kvm_create_spapr_tce (in) 1978:Returns: file descriptor for manipulating the created TCE table 1979 1980This creates a virtual TCE (translation control entry) table, which 1981is an IOMMU for PAPR-style virtual I/O. It is used to translate 1982logical addresses used in virtual I/O into guest physical addresses, 1983and provides a scatter/gather capability for PAPR virtual I/O. 1984 1985:: 1986 1987 /* for KVM_CAP_SPAPR_TCE */ 1988 struct kvm_create_spapr_tce { 1989 __u64 liobn; 1990 __u32 window_size; 1991 }; 1992 1993The liobn field gives the logical IO bus number for which to create a 1994TCE table. The window_size field specifies the size of the DMA window 1995which this TCE table will translate - the table will contain one 64 1996bit TCE entry for every 4kiB of the DMA window. 1997 1998When the guest issues an H_PUT_TCE hcall on a liobn for which a TCE 1999table has been created using this ioctl(), the kernel will handle it 2000in real mode, updating the TCE table. H_PUT_TCE calls for other 2001liobns will cause a vm exit and must be handled by userspace. 2002 2003The return value is a file descriptor which can be passed to mmap(2) 2004to map the created TCE table into userspace. This lets userspace read 2005the entries written by kernel-handled H_PUT_TCE calls, and also lets 2006userspace update the TCE table directly which is useful in some 2007circumstances. 2008 2009 20104.63 KVM_ALLOCATE_RMA 2011--------------------- 2012 2013:Capability: KVM_CAP_PPC_RMA 2014:Architectures: powerpc 2015:Type: vm ioctl 2016:Parameters: struct kvm_allocate_rma (out) 2017:Returns: file descriptor for mapping the allocated RMA 2018 2019This allocates a Real Mode Area (RMA) from the pool allocated at boot 2020time by the kernel. An RMA is a physically-contiguous, aligned region 2021of memory used on older POWER processors to provide the memory which 2022will be accessed by real-mode (MMU off) accesses in a KVM guest. 2023POWER processors support a set of sizes for the RMA that usually 2024includes 64MB, 128MB, 256MB and some larger powers of two. 2025 2026:: 2027 2028 /* for KVM_ALLOCATE_RMA */ 2029 struct kvm_allocate_rma { 2030 __u64 rma_size; 2031 }; 2032 2033The return value is a file descriptor which can be passed to mmap(2) 2034to map the allocated RMA into userspace. The mapped area can then be 2035passed to the KVM_SET_USER_MEMORY_REGION ioctl to establish it as the 2036RMA for a virtual machine. The size of the RMA in bytes (which is 2037fixed at host kernel boot time) is returned in the rma_size field of 2038the argument structure. 2039 2040The KVM_CAP_PPC_RMA capability is 1 or 2 if the KVM_ALLOCATE_RMA ioctl 2041is supported; 2 if the processor requires all virtual machines to have 2042an RMA, or 1 if the processor can use an RMA but doesn't require it, 2043because it supports the Virtual RMA (VRMA) facility. 2044 2045 20464.64 KVM_NMI 2047------------ 2048 2049:Capability: KVM_CAP_USER_NMI 2050:Architectures: x86 2051:Type: vcpu ioctl 2052:Parameters: none 2053:Returns: 0 on success, -1 on error 2054 2055Queues an NMI on the thread's vcpu. Note this is well defined only 2056when KVM_CREATE_IRQCHIP has not been called, since this is an interface 2057between the virtual cpu core and virtual local APIC. After KVM_CREATE_IRQCHIP 2058has been called, this interface is completely emulated within the kernel. 2059 2060To use this to emulate the LINT1 input with KVM_CREATE_IRQCHIP, use the 2061following algorithm: 2062 2063 - pause the vcpu 2064 - read the local APIC's state (KVM_GET_LAPIC) 2065 - check whether changing LINT1 will queue an NMI (see the LVT entry for LINT1) 2066 - if so, issue KVM_NMI 2067 - resume the vcpu 2068 2069Some guests configure the LINT1 NMI input to cause a panic, aiding in 2070debugging. 2071 2072 20734.65 KVM_S390_UCAS_MAP 2074---------------------- 2075 2076:Capability: KVM_CAP_S390_UCONTROL 2077:Architectures: s390 2078:Type: vcpu ioctl 2079:Parameters: struct kvm_s390_ucas_mapping (in) 2080:Returns: 0 in case of success 2081 2082The parameter is defined like this:: 2083 2084 struct kvm_s390_ucas_mapping { 2085 __u64 user_addr; 2086 __u64 vcpu_addr; 2087 __u64 length; 2088 }; 2089 2090This ioctl maps the memory at "user_addr" with the length "length" to 2091the vcpu's address space starting at "vcpu_addr". All parameters need to 2092be aligned by 1 megabyte. 2093 2094 20954.66 KVM_S390_UCAS_UNMAP 2096------------------------ 2097 2098:Capability: KVM_CAP_S390_UCONTROL 2099:Architectures: s390 2100:Type: vcpu ioctl 2101:Parameters: struct kvm_s390_ucas_mapping (in) 2102:Returns: 0 in case of success 2103 2104The parameter is defined like this:: 2105 2106 struct kvm_s390_ucas_mapping { 2107 __u64 user_addr; 2108 __u64 vcpu_addr; 2109 __u64 length; 2110 }; 2111 2112This ioctl unmaps the memory in the vcpu's address space starting at 2113"vcpu_addr" with the length "length". The field "user_addr" is ignored. 2114All parameters need to be aligned by 1 megabyte. 2115 2116 21174.67 KVM_S390_VCPU_FAULT 2118------------------------ 2119 2120:Capability: KVM_CAP_S390_UCONTROL 2121:Architectures: s390 2122:Type: vcpu ioctl 2123:Parameters: vcpu absolute address (in) 2124:Returns: 0 in case of success 2125 2126This call creates a page table entry on the virtual cpu's address space 2127(for user controlled virtual machines) or the virtual machine's address 2128space (for regular virtual machines). This only works for minor faults, 2129thus it's recommended to access subject memory page via the user page 2130table upfront. This is useful to handle validity intercepts for user 2131controlled virtual machines to fault in the virtual cpu's lowcore pages 2132prior to calling the KVM_RUN ioctl. 2133 2134 21354.68 KVM_SET_ONE_REG 2136-------------------- 2137 2138:Capability: KVM_CAP_ONE_REG 2139:Architectures: all 2140:Type: vcpu ioctl 2141:Parameters: struct kvm_one_reg (in) 2142:Returns: 0 on success, negative value on failure 2143 2144Errors: 2145 2146 ====== ============================================================ 2147 ENOENT no such register 2148 EINVAL invalid register ID, or no such register or used with VMs in 2149 protected virtualization mode on s390 2150 EPERM (arm64) register access not allowed before vcpu finalization 2151 ====== ============================================================ 2152 2153(These error codes are indicative only: do not rely on a specific error 2154code being returned in a specific situation.) 2155 2156:: 2157 2158 struct kvm_one_reg { 2159 __u64 id; 2160 __u64 addr; 2161 }; 2162 2163Using this ioctl, a single vcpu register can be set to a specific value 2164defined by user space with the passed in struct kvm_one_reg, where id 2165refers to the register identifier as described below and addr is a pointer 2166to a variable with the respective size. There can be architecture agnostic 2167and architecture specific registers. Each have their own range of operation 2168and their own constants and width. To keep track of the implemented 2169registers, find a list below: 2170 2171 ======= =============================== ============ 2172 Arch Register Width (bits) 2173 ======= =============================== ============ 2174 PPC KVM_REG_PPC_HIOR 64 2175 PPC KVM_REG_PPC_IAC1 64 2176 PPC KVM_REG_PPC_IAC2 64 2177 PPC KVM_REG_PPC_IAC3 64 2178 PPC KVM_REG_PPC_IAC4 64 2179 PPC KVM_REG_PPC_DAC1 64 2180 PPC KVM_REG_PPC_DAC2 64 2181 PPC KVM_REG_PPC_DABR 64 2182 PPC KVM_REG_PPC_DSCR 64 2183 PPC KVM_REG_PPC_PURR 64 2184 PPC KVM_REG_PPC_SPURR 64 2185 PPC KVM_REG_PPC_DAR 64 2186 PPC KVM_REG_PPC_DSISR 32 2187 PPC KVM_REG_PPC_AMR 64 2188 PPC KVM_REG_PPC_UAMOR 64 2189 PPC KVM_REG_PPC_MMCR0 64 2190 PPC KVM_REG_PPC_MMCR1 64 2191 PPC KVM_REG_PPC_MMCRA 64 2192 PPC KVM_REG_PPC_MMCR2 64 2193 PPC KVM_REG_PPC_MMCRS 64 2194 PPC KVM_REG_PPC_MMCR3 64 2195 PPC KVM_REG_PPC_SIAR 64 2196 PPC KVM_REG_PPC_SDAR 64 2197 PPC KVM_REG_PPC_SIER 64 2198 PPC KVM_REG_PPC_SIER2 64 2199 PPC KVM_REG_PPC_SIER3 64 2200 PPC KVM_REG_PPC_PMC1 32 2201 PPC KVM_REG_PPC_PMC2 32 2202 PPC KVM_REG_PPC_PMC3 32 2203 PPC KVM_REG_PPC_PMC4 32 2204 PPC KVM_REG_PPC_PMC5 32 2205 PPC KVM_REG_PPC_PMC6 32 2206 PPC KVM_REG_PPC_PMC7 32 2207 PPC KVM_REG_PPC_PMC8 32 2208 PPC KVM_REG_PPC_FPR0 64 2209 ... 2210 PPC KVM_REG_PPC_FPR31 64 2211 PPC KVM_REG_PPC_VR0 128 2212 ... 2213 PPC KVM_REG_PPC_VR31 128 2214 PPC KVM_REG_PPC_VSR0 128 2215 ... 2216 PPC KVM_REG_PPC_VSR31 128 2217 PPC KVM_REG_PPC_FPSCR 64 2218 PPC KVM_REG_PPC_VSCR 32 2219 PPC KVM_REG_PPC_VPA_ADDR 64 2220 PPC KVM_REG_PPC_VPA_SLB 128 2221 PPC KVM_REG_PPC_VPA_DTL 128 2222 PPC KVM_REG_PPC_EPCR 32 2223 PPC KVM_REG_PPC_EPR 32 2224 PPC KVM_REG_PPC_TCR 32 2225 PPC KVM_REG_PPC_TSR 32 2226 PPC KVM_REG_PPC_OR_TSR 32 2227 PPC KVM_REG_PPC_CLEAR_TSR 32 2228 PPC KVM_REG_PPC_MAS0 32 2229 PPC KVM_REG_PPC_MAS1 32 2230 PPC KVM_REG_PPC_MAS2 64 2231 PPC KVM_REG_PPC_MAS7_3 64 2232 PPC KVM_REG_PPC_MAS4 32 2233 PPC KVM_REG_PPC_MAS6 32 2234 PPC KVM_REG_PPC_MMUCFG 32 2235 PPC KVM_REG_PPC_TLB0CFG 32 2236 PPC KVM_REG_PPC_TLB1CFG 32 2237 PPC KVM_REG_PPC_TLB2CFG 32 2238 PPC KVM_REG_PPC_TLB3CFG 32 2239 PPC KVM_REG_PPC_TLB0PS 32 2240 PPC KVM_REG_PPC_TLB1PS 32 2241 PPC KVM_REG_PPC_TLB2PS 32 2242 PPC KVM_REG_PPC_TLB3PS 32 2243 PPC KVM_REG_PPC_EPTCFG 32 2244 PPC KVM_REG_PPC_ICP_STATE 64 2245 PPC KVM_REG_PPC_VP_STATE 128 2246 PPC KVM_REG_PPC_TB_OFFSET 64 2247 PPC KVM_REG_PPC_SPMC1 32 2248 PPC KVM_REG_PPC_SPMC2 32 2249 PPC KVM_REG_PPC_IAMR 64 2250 PPC KVM_REG_PPC_TFHAR 64 2251 PPC KVM_REG_PPC_TFIAR 64 2252 PPC KVM_REG_PPC_TEXASR 64 2253 PPC KVM_REG_PPC_FSCR 64 2254 PPC KVM_REG_PPC_PSPB 32 2255 PPC KVM_REG_PPC_EBBHR 64 2256 PPC KVM_REG_PPC_EBBRR 64 2257 PPC KVM_REG_PPC_BESCR 64 2258 PPC KVM_REG_PPC_TAR 64 2259 PPC KVM_REG_PPC_DPDES 64 2260 PPC KVM_REG_PPC_DAWR 64 2261 PPC KVM_REG_PPC_DAWRX 64 2262 PPC KVM_REG_PPC_CIABR 64 2263 PPC KVM_REG_PPC_IC 64 2264 PPC KVM_REG_PPC_VTB 64 2265 PPC KVM_REG_PPC_CSIGR 64 2266 PPC KVM_REG_PPC_TACR 64 2267 PPC KVM_REG_PPC_TCSCR 64 2268 PPC KVM_REG_PPC_PID 64 2269 PPC KVM_REG_PPC_ACOP 64 2270 PPC KVM_REG_PPC_VRSAVE 32 2271 PPC KVM_REG_PPC_LPCR 32 2272 PPC KVM_REG_PPC_LPCR_64 64 2273 PPC KVM_REG_PPC_PPR 64 2274 PPC KVM_REG_PPC_ARCH_COMPAT 32 2275 PPC KVM_REG_PPC_DABRX 32 2276 PPC KVM_REG_PPC_WORT 64 2277 PPC KVM_REG_PPC_SPRG9 64 2278 PPC KVM_REG_PPC_DBSR 32 2279 PPC KVM_REG_PPC_TIDR 64 2280 PPC KVM_REG_PPC_PSSCR 64 2281 PPC KVM_REG_PPC_DEC_EXPIRY 64 2282 PPC KVM_REG_PPC_PTCR 64 2283 PPC KVM_REG_PPC_DAWR1 64 2284 PPC KVM_REG_PPC_DAWRX1 64 2285 PPC KVM_REG_PPC_TM_GPR0 64 2286 ... 2287 PPC KVM_REG_PPC_TM_GPR31 64 2288 PPC KVM_REG_PPC_TM_VSR0 128 2289 ... 2290 PPC KVM_REG_PPC_TM_VSR63 128 2291 PPC KVM_REG_PPC_TM_CR 64 2292 PPC KVM_REG_PPC_TM_LR 64 2293 PPC KVM_REG_PPC_TM_CTR 64 2294 PPC KVM_REG_PPC_TM_FPSCR 64 2295 PPC KVM_REG_PPC_TM_AMR 64 2296 PPC KVM_REG_PPC_TM_PPR 64 2297 PPC KVM_REG_PPC_TM_VRSAVE 64 2298 PPC KVM_REG_PPC_TM_VSCR 32 2299 PPC KVM_REG_PPC_TM_DSCR 64 2300 PPC KVM_REG_PPC_TM_TAR 64 2301 PPC KVM_REG_PPC_TM_XER 64 2302 2303 MIPS KVM_REG_MIPS_R0 64 2304 ... 2305 MIPS KVM_REG_MIPS_R31 64 2306 MIPS KVM_REG_MIPS_HI 64 2307 MIPS KVM_REG_MIPS_LO 64 2308 MIPS KVM_REG_MIPS_PC 64 2309 MIPS KVM_REG_MIPS_CP0_INDEX 32 2310 MIPS KVM_REG_MIPS_CP0_ENTRYLO0 64 2311 MIPS KVM_REG_MIPS_CP0_ENTRYLO1 64 2312 MIPS KVM_REG_MIPS_CP0_CONTEXT 64 2313 MIPS KVM_REG_MIPS_CP0_CONTEXTCONFIG 32 2314 MIPS KVM_REG_MIPS_CP0_USERLOCAL 64 2315 MIPS KVM_REG_MIPS_CP0_XCONTEXTCONFIG 64 2316 MIPS KVM_REG_MIPS_CP0_PAGEMASK 32 2317 MIPS KVM_REG_MIPS_CP0_PAGEGRAIN 32 2318 MIPS KVM_REG_MIPS_CP0_SEGCTL0 64 2319 MIPS KVM_REG_MIPS_CP0_SEGCTL1 64 2320 MIPS KVM_REG_MIPS_CP0_SEGCTL2 64 2321 MIPS KVM_REG_MIPS_CP0_PWBASE 64 2322 MIPS KVM_REG_MIPS_CP0_PWFIELD 64 2323 MIPS KVM_REG_MIPS_CP0_PWSIZE 64 2324 MIPS KVM_REG_MIPS_CP0_WIRED 32 2325 MIPS KVM_REG_MIPS_CP0_PWCTL 32 2326 MIPS KVM_REG_MIPS_CP0_HWRENA 32 2327 MIPS KVM_REG_MIPS_CP0_BADVADDR 64 2328 MIPS KVM_REG_MIPS_CP0_BADINSTR 32 2329 MIPS KVM_REG_MIPS_CP0_BADINSTRP 32 2330 MIPS KVM_REG_MIPS_CP0_COUNT 32 2331 MIPS KVM_REG_MIPS_CP0_ENTRYHI 64 2332 MIPS KVM_REG_MIPS_CP0_COMPARE 32 2333 MIPS KVM_REG_MIPS_CP0_STATUS 32 2334 MIPS KVM_REG_MIPS_CP0_INTCTL 32 2335 MIPS KVM_REG_MIPS_CP0_CAUSE 32 2336 MIPS KVM_REG_MIPS_CP0_EPC 64 2337 MIPS KVM_REG_MIPS_CP0_PRID 32 2338 MIPS KVM_REG_MIPS_CP0_EBASE 64 2339 MIPS KVM_REG_MIPS_CP0_CONFIG 32 2340 MIPS KVM_REG_MIPS_CP0_CONFIG1 32 2341 MIPS KVM_REG_MIPS_CP0_CONFIG2 32 2342 MIPS KVM_REG_MIPS_CP0_CONFIG3 32 2343 MIPS KVM_REG_MIPS_CP0_CONFIG4 32 2344 MIPS KVM_REG_MIPS_CP0_CONFIG5 32 2345 MIPS KVM_REG_MIPS_CP0_CONFIG7 32 2346 MIPS KVM_REG_MIPS_CP0_XCONTEXT 64 2347 MIPS KVM_REG_MIPS_CP0_ERROREPC 64 2348 MIPS KVM_REG_MIPS_CP0_KSCRATCH1 64 2349 MIPS KVM_REG_MIPS_CP0_KSCRATCH2 64 2350 MIPS KVM_REG_MIPS_CP0_KSCRATCH3 64 2351 MIPS KVM_REG_MIPS_CP0_KSCRATCH4 64 2352 MIPS KVM_REG_MIPS_CP0_KSCRATCH5 64 2353 MIPS KVM_REG_MIPS_CP0_KSCRATCH6 64 2354 MIPS KVM_REG_MIPS_CP0_MAAR(0..63) 64 2355 MIPS KVM_REG_MIPS_COUNT_CTL 64 2356 MIPS KVM_REG_MIPS_COUNT_RESUME 64 2357 MIPS KVM_REG_MIPS_COUNT_HZ 64 2358 MIPS KVM_REG_MIPS_FPR_32(0..31) 32 2359 MIPS KVM_REG_MIPS_FPR_64(0..31) 64 2360 MIPS KVM_REG_MIPS_VEC_128(0..31) 128 2361 MIPS KVM_REG_MIPS_FCR_IR 32 2362 MIPS KVM_REG_MIPS_FCR_CSR 32 2363 MIPS KVM_REG_MIPS_MSA_IR 32 2364 MIPS KVM_REG_MIPS_MSA_CSR 32 2365 ======= =============================== ============ 2366 2367ARM registers are mapped using the lower 32 bits. The upper 16 of that 2368is the register group type, or coprocessor number: 2369 2370ARM core registers have the following id bit patterns:: 2371 2372 0x4020 0000 0010 <index into the kvm_regs struct:16> 2373 2374ARM 32-bit CP15 registers have the following id bit patterns:: 2375 2376 0x4020 0000 000F <zero:1> <crn:4> <crm:4> <opc1:4> <opc2:3> 2377 2378ARM 64-bit CP15 registers have the following id bit patterns:: 2379 2380 0x4030 0000 000F <zero:1> <zero:4> <crm:4> <opc1:4> <zero:3> 2381 2382ARM CCSIDR registers are demultiplexed by CSSELR value:: 2383 2384 0x4020 0000 0011 00 <csselr:8> 2385 2386ARM 32-bit VFP control registers have the following id bit patterns:: 2387 2388 0x4020 0000 0012 1 <regno:12> 2389 2390ARM 64-bit FP registers have the following id bit patterns:: 2391 2392 0x4030 0000 0012 0 <regno:12> 2393 2394ARM firmware pseudo-registers have the following bit pattern:: 2395 2396 0x4030 0000 0014 <regno:16> 2397 2398 2399arm64 registers are mapped using the lower 32 bits. The upper 16 of 2400that is the register group type, or coprocessor number: 2401 2402arm64 core/FP-SIMD registers have the following id bit patterns. Note 2403that the size of the access is variable, as the kvm_regs structure 2404contains elements ranging from 32 to 128 bits. The index is a 32bit 2405value in the kvm_regs structure seen as a 32bit array:: 2406 2407 0x60x0 0000 0010 <index into the kvm_regs struct:16> 2408 2409Specifically: 2410 2411======================= ========= ===== ======================================= 2412 Encoding Register Bits kvm_regs member 2413======================= ========= ===== ======================================= 2414 0x6030 0000 0010 0000 X0 64 regs.regs[0] 2415 0x6030 0000 0010 0002 X1 64 regs.regs[1] 2416 ... 2417 0x6030 0000 0010 003c X30 64 regs.regs[30] 2418 0x6030 0000 0010 003e SP 64 regs.sp 2419 0x6030 0000 0010 0040 PC 64 regs.pc 2420 0x6030 0000 0010 0042 PSTATE 64 regs.pstate 2421 0x6030 0000 0010 0044 SP_EL1 64 sp_el1 2422 0x6030 0000 0010 0046 ELR_EL1 64 elr_el1 2423 0x6030 0000 0010 0048 SPSR_EL1 64 spsr[KVM_SPSR_EL1] (alias SPSR_SVC) 2424 0x6030 0000 0010 004a SPSR_ABT 64 spsr[KVM_SPSR_ABT] 2425 0x6030 0000 0010 004c SPSR_UND 64 spsr[KVM_SPSR_UND] 2426 0x6030 0000 0010 004e SPSR_IRQ 64 spsr[KVM_SPSR_IRQ] 2427 0x6060 0000 0010 0050 SPSR_FIQ 64 spsr[KVM_SPSR_FIQ] 2428 0x6040 0000 0010 0054 V0 128 fp_regs.vregs[0] [1]_ 2429 0x6040 0000 0010 0058 V1 128 fp_regs.vregs[1] [1]_ 2430 ... 2431 0x6040 0000 0010 00d0 V31 128 fp_regs.vregs[31] [1]_ 2432 0x6020 0000 0010 00d4 FPSR 32 fp_regs.fpsr 2433 0x6020 0000 0010 00d5 FPCR 32 fp_regs.fpcr 2434======================= ========= ===== ======================================= 2435 2436.. [1] These encodings are not accepted for SVE-enabled vcpus. See 2437 KVM_ARM_VCPU_INIT. 2438 2439 The equivalent register content can be accessed via bits [127:0] of 2440 the corresponding SVE Zn registers instead for vcpus that have SVE 2441 enabled (see below). 2442 2443arm64 CCSIDR registers are demultiplexed by CSSELR value:: 2444 2445 0x6020 0000 0011 00 <csselr:8> 2446 2447arm64 system registers have the following id bit patterns:: 2448 2449 0x6030 0000 0013 <op0:2> <op1:3> <crn:4> <crm:4> <op2:3> 2450 2451.. warning:: 2452 2453 Two system register IDs do not follow the specified pattern. These 2454 are KVM_REG_ARM_TIMER_CVAL and KVM_REG_ARM_TIMER_CNT, which map to 2455 system registers CNTV_CVAL_EL0 and CNTVCT_EL0 respectively. These 2456 two had their values accidentally swapped, which means TIMER_CVAL is 2457 derived from the register encoding for CNTVCT_EL0 and TIMER_CNT is 2458 derived from the register encoding for CNTV_CVAL_EL0. As this is 2459 API, it must remain this way. 2460 2461arm64 firmware pseudo-registers have the following bit pattern:: 2462 2463 0x6030 0000 0014 <regno:16> 2464 2465arm64 SVE registers have the following bit patterns:: 2466 2467 0x6080 0000 0015 00 <n:5> <slice:5> Zn bits[2048*slice + 2047 : 2048*slice] 2468 0x6050 0000 0015 04 <n:4> <slice:5> Pn bits[256*slice + 255 : 256*slice] 2469 0x6050 0000 0015 060 <slice:5> FFR bits[256*slice + 255 : 256*slice] 2470 0x6060 0000 0015 ffff KVM_REG_ARM64_SVE_VLS pseudo-register 2471 2472Access to register IDs where 2048 * slice >= 128 * max_vq will fail with 2473ENOENT. max_vq is the vcpu's maximum supported vector length in 128-bit 2474quadwords: see [2]_ below. 2475 2476These registers are only accessible on vcpus for which SVE is enabled. 2477See KVM_ARM_VCPU_INIT for details. 2478 2479In addition, except for KVM_REG_ARM64_SVE_VLS, these registers are not 2480accessible until the vcpu's SVE configuration has been finalized 2481using KVM_ARM_VCPU_FINALIZE(KVM_ARM_VCPU_SVE). See KVM_ARM_VCPU_INIT 2482and KVM_ARM_VCPU_FINALIZE for more information about this procedure. 2483 2484KVM_REG_ARM64_SVE_VLS is a pseudo-register that allows the set of vector 2485lengths supported by the vcpu to be discovered and configured by 2486userspace. When transferred to or from user memory via KVM_GET_ONE_REG 2487or KVM_SET_ONE_REG, the value of this register is of type 2488__u64[KVM_ARM64_SVE_VLS_WORDS], and encodes the set of vector lengths as 2489follows:: 2490 2491 __u64 vector_lengths[KVM_ARM64_SVE_VLS_WORDS]; 2492 2493 if (vq >= SVE_VQ_MIN && vq <= SVE_VQ_MAX && 2494 ((vector_lengths[(vq - KVM_ARM64_SVE_VQ_MIN) / 64] >> 2495 ((vq - KVM_ARM64_SVE_VQ_MIN) % 64)) & 1)) 2496 /* Vector length vq * 16 bytes supported */ 2497 else 2498 /* Vector length vq * 16 bytes not supported */ 2499 2500.. [2] The maximum value vq for which the above condition is true is 2501 max_vq. This is the maximum vector length available to the guest on 2502 this vcpu, and determines which register slices are visible through 2503 this ioctl interface. 2504 2505(See Documentation/arm64/sve.rst for an explanation of the "vq" 2506nomenclature.) 2507 2508KVM_REG_ARM64_SVE_VLS is only accessible after KVM_ARM_VCPU_INIT. 2509KVM_ARM_VCPU_INIT initialises it to the best set of vector lengths that 2510the host supports. 2511 2512Userspace may subsequently modify it if desired until the vcpu's SVE 2513configuration is finalized using KVM_ARM_VCPU_FINALIZE(KVM_ARM_VCPU_SVE). 2514 2515Apart from simply removing all vector lengths from the host set that 2516exceed some value, support for arbitrarily chosen sets of vector lengths 2517is hardware-dependent and may not be available. Attempting to configure 2518an invalid set of vector lengths via KVM_SET_ONE_REG will fail with 2519EINVAL. 2520 2521After the vcpu's SVE configuration is finalized, further attempts to 2522write this register will fail with EPERM. 2523 2524 2525MIPS registers are mapped using the lower 32 bits. The upper 16 of that is 2526the register group type: 2527 2528MIPS core registers (see above) have the following id bit patterns:: 2529 2530 0x7030 0000 0000 <reg:16> 2531 2532MIPS CP0 registers (see KVM_REG_MIPS_CP0_* above) have the following id bit 2533patterns depending on whether they're 32-bit or 64-bit registers:: 2534 2535 0x7020 0000 0001 00 <reg:5> <sel:3> (32-bit) 2536 0x7030 0000 0001 00 <reg:5> <sel:3> (64-bit) 2537 2538Note: KVM_REG_MIPS_CP0_ENTRYLO0 and KVM_REG_MIPS_CP0_ENTRYLO1 are the MIPS64 2539versions of the EntryLo registers regardless of the word size of the host 2540hardware, host kernel, guest, and whether XPA is present in the guest, i.e. 2541with the RI and XI bits (if they exist) in bits 63 and 62 respectively, and 2542the PFNX field starting at bit 30. 2543 2544MIPS MAARs (see KVM_REG_MIPS_CP0_MAAR(*) above) have the following id bit 2545patterns:: 2546 2547 0x7030 0000 0001 01 <reg:8> 2548 2549MIPS KVM control registers (see above) have the following id bit patterns:: 2550 2551 0x7030 0000 0002 <reg:16> 2552 2553MIPS FPU registers (see KVM_REG_MIPS_FPR_{32,64}() above) have the following 2554id bit patterns depending on the size of the register being accessed. They are 2555always accessed according to the current guest FPU mode (Status.FR and 2556Config5.FRE), i.e. as the guest would see them, and they become unpredictable 2557if the guest FPU mode is changed. MIPS SIMD Architecture (MSA) vector 2558registers (see KVM_REG_MIPS_VEC_128() above) have similar patterns as they 2559overlap the FPU registers:: 2560 2561 0x7020 0000 0003 00 <0:3> <reg:5> (32-bit FPU registers) 2562 0x7030 0000 0003 00 <0:3> <reg:5> (64-bit FPU registers) 2563 0x7040 0000 0003 00 <0:3> <reg:5> (128-bit MSA vector registers) 2564 2565MIPS FPU control registers (see KVM_REG_MIPS_FCR_{IR,CSR} above) have the 2566following id bit patterns:: 2567 2568 0x7020 0000 0003 01 <0:3> <reg:5> 2569 2570MIPS MSA control registers (see KVM_REG_MIPS_MSA_{IR,CSR} above) have the 2571following id bit patterns:: 2572 2573 0x7020 0000 0003 02 <0:3> <reg:5> 2574 2575 25764.69 KVM_GET_ONE_REG 2577-------------------- 2578 2579:Capability: KVM_CAP_ONE_REG 2580:Architectures: all 2581:Type: vcpu ioctl 2582:Parameters: struct kvm_one_reg (in and out) 2583:Returns: 0 on success, negative value on failure 2584 2585Errors include: 2586 2587 ======== ============================================================ 2588 ENOENT no such register 2589 EINVAL invalid register ID, or no such register or used with VMs in 2590 protected virtualization mode on s390 2591 EPERM (arm64) register access not allowed before vcpu finalization 2592 ======== ============================================================ 2593 2594(These error codes are indicative only: do not rely on a specific error 2595code being returned in a specific situation.) 2596 2597This ioctl allows to receive the value of a single register implemented 2598in a vcpu. The register to read is indicated by the "id" field of the 2599kvm_one_reg struct passed in. On success, the register value can be found 2600at the memory location pointed to by "addr". 2601 2602The list of registers accessible using this interface is identical to the 2603list in 4.68. 2604 2605 26064.70 KVM_KVMCLOCK_CTRL 2607---------------------- 2608 2609:Capability: KVM_CAP_KVMCLOCK_CTRL 2610:Architectures: Any that implement pvclocks (currently x86 only) 2611:Type: vcpu ioctl 2612:Parameters: None 2613:Returns: 0 on success, -1 on error 2614 2615This ioctl sets a flag accessible to the guest indicating that the specified 2616vCPU has been paused by the host userspace. 2617 2618The host will set a flag in the pvclock structure that is checked from the 2619soft lockup watchdog. The flag is part of the pvclock structure that is 2620shared between guest and host, specifically the second bit of the flags 2621field of the pvclock_vcpu_time_info structure. It will be set exclusively by 2622the host and read/cleared exclusively by the guest. The guest operation of 2623checking and clearing the flag must be an atomic operation so 2624load-link/store-conditional, or equivalent must be used. There are two cases 2625where the guest will clear the flag: when the soft lockup watchdog timer resets 2626itself or when a soft lockup is detected. This ioctl can be called any time 2627after pausing the vcpu, but before it is resumed. 2628 2629 26304.71 KVM_SIGNAL_MSI 2631------------------- 2632 2633:Capability: KVM_CAP_SIGNAL_MSI 2634:Architectures: x86 arm arm64 2635:Type: vm ioctl 2636:Parameters: struct kvm_msi (in) 2637:Returns: >0 on delivery, 0 if guest blocked the MSI, and -1 on error 2638 2639Directly inject a MSI message. Only valid with in-kernel irqchip that handles 2640MSI messages. 2641 2642:: 2643 2644 struct kvm_msi { 2645 __u32 address_lo; 2646 __u32 address_hi; 2647 __u32 data; 2648 __u32 flags; 2649 __u32 devid; 2650 __u8 pad[12]; 2651 }; 2652 2653flags: 2654 KVM_MSI_VALID_DEVID: devid contains a valid value. The per-VM 2655 KVM_CAP_MSI_DEVID capability advertises the requirement to provide 2656 the device ID. If this capability is not available, userspace 2657 should never set the KVM_MSI_VALID_DEVID flag as the ioctl might fail. 2658 2659If KVM_MSI_VALID_DEVID is set, devid contains a unique device identifier 2660for the device that wrote the MSI message. For PCI, this is usually a 2661BFD identifier in the lower 16 bits. 2662 2663On x86, address_hi is ignored unless the KVM_X2APIC_API_USE_32BIT_IDS 2664feature of KVM_CAP_X2APIC_API capability is enabled. If it is enabled, 2665address_hi bits 31-8 provide bits 31-8 of the destination id. Bits 7-0 of 2666address_hi must be zero. 2667 2668 26694.71 KVM_CREATE_PIT2 2670-------------------- 2671 2672:Capability: KVM_CAP_PIT2 2673:Architectures: x86 2674:Type: vm ioctl 2675:Parameters: struct kvm_pit_config (in) 2676:Returns: 0 on success, -1 on error 2677 2678Creates an in-kernel device model for the i8254 PIT. This call is only valid 2679after enabling in-kernel irqchip support via KVM_CREATE_IRQCHIP. The following 2680parameters have to be passed:: 2681 2682 struct kvm_pit_config { 2683 __u32 flags; 2684 __u32 pad[15]; 2685 }; 2686 2687Valid flags are:: 2688 2689 #define KVM_PIT_SPEAKER_DUMMY 1 /* emulate speaker port stub */ 2690 2691PIT timer interrupts may use a per-VM kernel thread for injection. If it 2692exists, this thread will have a name of the following pattern:: 2693 2694 kvm-pit/<owner-process-pid> 2695 2696When running a guest with elevated priorities, the scheduling parameters of 2697this thread may have to be adjusted accordingly. 2698 2699This IOCTL replaces the obsolete KVM_CREATE_PIT. 2700 2701 27024.72 KVM_GET_PIT2 2703----------------- 2704 2705:Capability: KVM_CAP_PIT_STATE2 2706:Architectures: x86 2707:Type: vm ioctl 2708:Parameters: struct kvm_pit_state2 (out) 2709:Returns: 0 on success, -1 on error 2710 2711Retrieves the state of the in-kernel PIT model. Only valid after 2712KVM_CREATE_PIT2. The state is returned in the following structure:: 2713 2714 struct kvm_pit_state2 { 2715 struct kvm_pit_channel_state channels[3]; 2716 __u32 flags; 2717 __u32 reserved[9]; 2718 }; 2719 2720Valid flags are:: 2721 2722 /* disable PIT in HPET legacy mode */ 2723 #define KVM_PIT_FLAGS_HPET_LEGACY 0x00000001 2724 2725This IOCTL replaces the obsolete KVM_GET_PIT. 2726 2727 27284.73 KVM_SET_PIT2 2729----------------- 2730 2731:Capability: KVM_CAP_PIT_STATE2 2732:Architectures: x86 2733:Type: vm ioctl 2734:Parameters: struct kvm_pit_state2 (in) 2735:Returns: 0 on success, -1 on error 2736 2737Sets the state of the in-kernel PIT model. Only valid after KVM_CREATE_PIT2. 2738See KVM_GET_PIT2 for details on struct kvm_pit_state2. 2739 2740This IOCTL replaces the obsolete KVM_SET_PIT. 2741 2742 27434.74 KVM_PPC_GET_SMMU_INFO 2744-------------------------- 2745 2746:Capability: KVM_CAP_PPC_GET_SMMU_INFO 2747:Architectures: powerpc 2748:Type: vm ioctl 2749:Parameters: None 2750:Returns: 0 on success, -1 on error 2751 2752This populates and returns a structure describing the features of 2753the "Server" class MMU emulation supported by KVM. 2754This can in turn be used by userspace to generate the appropriate 2755device-tree properties for the guest operating system. 2756 2757The structure contains some global information, followed by an 2758array of supported segment page sizes:: 2759 2760 struct kvm_ppc_smmu_info { 2761 __u64 flags; 2762 __u32 slb_size; 2763 __u32 pad; 2764 struct kvm_ppc_one_seg_page_size sps[KVM_PPC_PAGE_SIZES_MAX_SZ]; 2765 }; 2766 2767The supported flags are: 2768 2769 - KVM_PPC_PAGE_SIZES_REAL: 2770 When that flag is set, guest page sizes must "fit" the backing 2771 store page sizes. When not set, any page size in the list can 2772 be used regardless of how they are backed by userspace. 2773 2774 - KVM_PPC_1T_SEGMENTS 2775 The emulated MMU supports 1T segments in addition to the 2776 standard 256M ones. 2777 2778 - KVM_PPC_NO_HASH 2779 This flag indicates that HPT guests are not supported by KVM, 2780 thus all guests must use radix MMU mode. 2781 2782The "slb_size" field indicates how many SLB entries are supported 2783 2784The "sps" array contains 8 entries indicating the supported base 2785page sizes for a segment in increasing order. Each entry is defined 2786as follow:: 2787 2788 struct kvm_ppc_one_seg_page_size { 2789 __u32 page_shift; /* Base page shift of segment (or 0) */ 2790 __u32 slb_enc; /* SLB encoding for BookS */ 2791 struct kvm_ppc_one_page_size enc[KVM_PPC_PAGE_SIZES_MAX_SZ]; 2792 }; 2793 2794An entry with a "page_shift" of 0 is unused. Because the array is 2795organized in increasing order, a lookup can stop when encoutering 2796such an entry. 2797 2798The "slb_enc" field provides the encoding to use in the SLB for the 2799page size. The bits are in positions such as the value can directly 2800be OR'ed into the "vsid" argument of the slbmte instruction. 2801 2802The "enc" array is a list which for each of those segment base page 2803size provides the list of supported actual page sizes (which can be 2804only larger or equal to the base page size), along with the 2805corresponding encoding in the hash PTE. Similarly, the array is 28068 entries sorted by increasing sizes and an entry with a "0" shift 2807is an empty entry and a terminator:: 2808 2809 struct kvm_ppc_one_page_size { 2810 __u32 page_shift; /* Page shift (or 0) */ 2811 __u32 pte_enc; /* Encoding in the HPTE (>>12) */ 2812 }; 2813 2814The "pte_enc" field provides a value that can OR'ed into the hash 2815PTE's RPN field (ie, it needs to be shifted left by 12 to OR it 2816into the hash PTE second double word). 2817 28184.75 KVM_IRQFD 2819-------------- 2820 2821:Capability: KVM_CAP_IRQFD 2822:Architectures: x86 s390 arm arm64 2823:Type: vm ioctl 2824:Parameters: struct kvm_irqfd (in) 2825:Returns: 0 on success, -1 on error 2826 2827Allows setting an eventfd to directly trigger a guest interrupt. 2828kvm_irqfd.fd specifies the file descriptor to use as the eventfd and 2829kvm_irqfd.gsi specifies the irqchip pin toggled by this event. When 2830an event is triggered on the eventfd, an interrupt is injected into 2831the guest using the specified gsi pin. The irqfd is removed using 2832the KVM_IRQFD_FLAG_DEASSIGN flag, specifying both kvm_irqfd.fd 2833and kvm_irqfd.gsi. 2834 2835With KVM_CAP_IRQFD_RESAMPLE, KVM_IRQFD supports a de-assert and notify 2836mechanism allowing emulation of level-triggered, irqfd-based 2837interrupts. When KVM_IRQFD_FLAG_RESAMPLE is set the user must pass an 2838additional eventfd in the kvm_irqfd.resamplefd field. When operating 2839in resample mode, posting of an interrupt through kvm_irq.fd asserts 2840the specified gsi in the irqchip. When the irqchip is resampled, such 2841as from an EOI, the gsi is de-asserted and the user is notified via 2842kvm_irqfd.resamplefd. It is the user's responsibility to re-queue 2843the interrupt if the device making use of it still requires service. 2844Note that closing the resamplefd is not sufficient to disable the 2845irqfd. The KVM_IRQFD_FLAG_RESAMPLE is only necessary on assignment 2846and need not be specified with KVM_IRQFD_FLAG_DEASSIGN. 2847 2848On arm/arm64, gsi routing being supported, the following can happen: 2849 2850- in case no routing entry is associated to this gsi, injection fails 2851- in case the gsi is associated to an irqchip routing entry, 2852 irqchip.pin + 32 corresponds to the injected SPI ID. 2853- in case the gsi is associated to an MSI routing entry, the MSI 2854 message and device ID are translated into an LPI (support restricted 2855 to GICv3 ITS in-kernel emulation). 2856 28574.76 KVM_PPC_ALLOCATE_HTAB 2858-------------------------- 2859 2860:Capability: KVM_CAP_PPC_ALLOC_HTAB 2861:Architectures: powerpc 2862:Type: vm ioctl 2863:Parameters: Pointer to u32 containing hash table order (in/out) 2864:Returns: 0 on success, -1 on error 2865 2866This requests the host kernel to allocate an MMU hash table for a 2867guest using the PAPR paravirtualization interface. This only does 2868anything if the kernel is configured to use the Book 3S HV style of 2869virtualization. Otherwise the capability doesn't exist and the ioctl 2870returns an ENOTTY error. The rest of this description assumes Book 3S 2871HV. 2872 2873There must be no vcpus running when this ioctl is called; if there 2874are, it will do nothing and return an EBUSY error. 2875 2876The parameter is a pointer to a 32-bit unsigned integer variable 2877containing the order (log base 2) of the desired size of the hash 2878table, which must be between 18 and 46. On successful return from the 2879ioctl, the value will not be changed by the kernel. 2880 2881If no hash table has been allocated when any vcpu is asked to run 2882(with the KVM_RUN ioctl), the host kernel will allocate a 2883default-sized hash table (16 MB). 2884 2885If this ioctl is called when a hash table has already been allocated, 2886with a different order from the existing hash table, the existing hash 2887table will be freed and a new one allocated. If this is ioctl is 2888called when a hash table has already been allocated of the same order 2889as specified, the kernel will clear out the existing hash table (zero 2890all HPTEs). In either case, if the guest is using the virtualized 2891real-mode area (VRMA) facility, the kernel will re-create the VMRA 2892HPTEs on the next KVM_RUN of any vcpu. 2893 28944.77 KVM_S390_INTERRUPT 2895----------------------- 2896 2897:Capability: basic 2898:Architectures: s390 2899:Type: vm ioctl, vcpu ioctl 2900:Parameters: struct kvm_s390_interrupt (in) 2901:Returns: 0 on success, -1 on error 2902 2903Allows to inject an interrupt to the guest. Interrupts can be floating 2904(vm ioctl) or per cpu (vcpu ioctl), depending on the interrupt type. 2905 2906Interrupt parameters are passed via kvm_s390_interrupt:: 2907 2908 struct kvm_s390_interrupt { 2909 __u32 type; 2910 __u32 parm; 2911 __u64 parm64; 2912 }; 2913 2914type can be one of the following: 2915 2916KVM_S390_SIGP_STOP (vcpu) 2917 - sigp stop; optional flags in parm 2918KVM_S390_PROGRAM_INT (vcpu) 2919 - program check; code in parm 2920KVM_S390_SIGP_SET_PREFIX (vcpu) 2921 - sigp set prefix; prefix address in parm 2922KVM_S390_RESTART (vcpu) 2923 - restart 2924KVM_S390_INT_CLOCK_COMP (vcpu) 2925 - clock comparator interrupt 2926KVM_S390_INT_CPU_TIMER (vcpu) 2927 - CPU timer interrupt 2928KVM_S390_INT_VIRTIO (vm) 2929 - virtio external interrupt; external interrupt 2930 parameters in parm and parm64 2931KVM_S390_INT_SERVICE (vm) 2932 - sclp external interrupt; sclp parameter in parm 2933KVM_S390_INT_EMERGENCY (vcpu) 2934 - sigp emergency; source cpu in parm 2935KVM_S390_INT_EXTERNAL_CALL (vcpu) 2936 - sigp external call; source cpu in parm 2937KVM_S390_INT_IO(ai,cssid,ssid,schid) (vm) 2938 - compound value to indicate an 2939 I/O interrupt (ai - adapter interrupt; cssid,ssid,schid - subchannel); 2940 I/O interruption parameters in parm (subchannel) and parm64 (intparm, 2941 interruption subclass) 2942KVM_S390_MCHK (vm, vcpu) 2943 - machine check interrupt; cr 14 bits in parm, machine check interrupt 2944 code in parm64 (note that machine checks needing further payload are not 2945 supported by this ioctl) 2946 2947This is an asynchronous vcpu ioctl and can be invoked from any thread. 2948 29494.78 KVM_PPC_GET_HTAB_FD 2950------------------------ 2951 2952:Capability: KVM_CAP_PPC_HTAB_FD 2953:Architectures: powerpc 2954:Type: vm ioctl 2955:Parameters: Pointer to struct kvm_get_htab_fd (in) 2956:Returns: file descriptor number (>= 0) on success, -1 on error 2957 2958This returns a file descriptor that can be used either to read out the 2959entries in the guest's hashed page table (HPT), or to write entries to 2960initialize the HPT. The returned fd can only be written to if the 2961KVM_GET_HTAB_WRITE bit is set in the flags field of the argument, and 2962can only be read if that bit is clear. The argument struct looks like 2963this:: 2964 2965 /* For KVM_PPC_GET_HTAB_FD */ 2966 struct kvm_get_htab_fd { 2967 __u64 flags; 2968 __u64 start_index; 2969 __u64 reserved[2]; 2970 }; 2971 2972 /* Values for kvm_get_htab_fd.flags */ 2973 #define KVM_GET_HTAB_BOLTED_ONLY ((__u64)0x1) 2974 #define KVM_GET_HTAB_WRITE ((__u64)0x2) 2975 2976The 'start_index' field gives the index in the HPT of the entry at 2977which to start reading. It is ignored when writing. 2978 2979Reads on the fd will initially supply information about all 2980"interesting" HPT entries. Interesting entries are those with the 2981bolted bit set, if the KVM_GET_HTAB_BOLTED_ONLY bit is set, otherwise 2982all entries. When the end of the HPT is reached, the read() will 2983return. If read() is called again on the fd, it will start again from 2984the beginning of the HPT, but will only return HPT entries that have 2985changed since they were last read. 2986 2987Data read or written is structured as a header (8 bytes) followed by a 2988series of valid HPT entries (16 bytes) each. The header indicates how 2989many valid HPT entries there are and how many invalid entries follow 2990the valid entries. The invalid entries are not represented explicitly 2991in the stream. The header format is:: 2992 2993 struct kvm_get_htab_header { 2994 __u32 index; 2995 __u16 n_valid; 2996 __u16 n_invalid; 2997 }; 2998 2999Writes to the fd create HPT entries starting at the index given in the 3000header; first 'n_valid' valid entries with contents from the data 3001written, then 'n_invalid' invalid entries, invalidating any previously 3002valid entries found. 3003 30044.79 KVM_CREATE_DEVICE 3005---------------------- 3006 3007:Capability: KVM_CAP_DEVICE_CTRL 3008:Type: vm ioctl 3009:Parameters: struct kvm_create_device (in/out) 3010:Returns: 0 on success, -1 on error 3011 3012Errors: 3013 3014 ====== ======================================================= 3015 ENODEV The device type is unknown or unsupported 3016 EEXIST Device already created, and this type of device may not 3017 be instantiated multiple times 3018 ====== ======================================================= 3019 3020 Other error conditions may be defined by individual device types or 3021 have their standard meanings. 3022 3023Creates an emulated device in the kernel. The file descriptor returned 3024in fd can be used with KVM_SET/GET/HAS_DEVICE_ATTR. 3025 3026If the KVM_CREATE_DEVICE_TEST flag is set, only test whether the 3027device type is supported (not necessarily whether it can be created 3028in the current vm). 3029 3030Individual devices should not define flags. Attributes should be used 3031for specifying any behavior that is not implied by the device type 3032number. 3033 3034:: 3035 3036 struct kvm_create_device { 3037 __u32 type; /* in: KVM_DEV_TYPE_xxx */ 3038 __u32 fd; /* out: device handle */ 3039 __u32 flags; /* in: KVM_CREATE_DEVICE_xxx */ 3040 }; 3041 30424.80 KVM_SET_DEVICE_ATTR/KVM_GET_DEVICE_ATTR 3043-------------------------------------------- 3044 3045:Capability: KVM_CAP_DEVICE_CTRL, KVM_CAP_VM_ATTRIBUTES for vm device, 3046 KVM_CAP_VCPU_ATTRIBUTES for vcpu device 3047:Type: device ioctl, vm ioctl, vcpu ioctl 3048:Parameters: struct kvm_device_attr 3049:Returns: 0 on success, -1 on error 3050 3051Errors: 3052 3053 ===== ============================================================= 3054 ENXIO The group or attribute is unknown/unsupported for this device 3055 or hardware support is missing. 3056 EPERM The attribute cannot (currently) be accessed this way 3057 (e.g. read-only attribute, or attribute that only makes 3058 sense when the device is in a different state) 3059 ===== ============================================================= 3060 3061 Other error conditions may be defined by individual device types. 3062 3063Gets/sets a specified piece of device configuration and/or state. The 3064semantics are device-specific. See individual device documentation in 3065the "devices" directory. As with ONE_REG, the size of the data 3066transferred is defined by the particular attribute. 3067 3068:: 3069 3070 struct kvm_device_attr { 3071 __u32 flags; /* no flags currently defined */ 3072 __u32 group; /* device-defined */ 3073 __u64 attr; /* group-defined */ 3074 __u64 addr; /* userspace address of attr data */ 3075 }; 3076 30774.81 KVM_HAS_DEVICE_ATTR 3078------------------------ 3079 3080:Capability: KVM_CAP_DEVICE_CTRL, KVM_CAP_VM_ATTRIBUTES for vm device, 3081 KVM_CAP_VCPU_ATTRIBUTES for vcpu device 3082:Type: device ioctl, vm ioctl, vcpu ioctl 3083:Parameters: struct kvm_device_attr 3084:Returns: 0 on success, -1 on error 3085 3086Errors: 3087 3088 ===== ============================================================= 3089 ENXIO The group or attribute is unknown/unsupported for this device 3090 or hardware support is missing. 3091 ===== ============================================================= 3092 3093Tests whether a device supports a particular attribute. A successful 3094return indicates the attribute is implemented. It does not necessarily 3095indicate that the attribute can be read or written in the device's 3096current state. "addr" is ignored. 3097 30984.82 KVM_ARM_VCPU_INIT 3099---------------------- 3100 3101:Capability: basic 3102:Architectures: arm, arm64 3103:Type: vcpu ioctl 3104:Parameters: struct kvm_vcpu_init (in) 3105:Returns: 0 on success; -1 on error 3106 3107Errors: 3108 3109 ====== ================================================================= 3110 EINVAL the target is unknown, or the combination of features is invalid. 3111 ENOENT a features bit specified is unknown. 3112 ====== ================================================================= 3113 3114This tells KVM what type of CPU to present to the guest, and what 3115optional features it should have. This will cause a reset of the cpu 3116registers to their initial values. If this is not called, KVM_RUN will 3117return ENOEXEC for that vcpu. 3118 3119The initial values are defined as: 3120 - Processor state: 3121 * AArch64: EL1h, D, A, I and F bits set. All other bits 3122 are cleared. 3123 * AArch32: SVC, A, I and F bits set. All other bits are 3124 cleared. 3125 - General Purpose registers, including PC and SP: set to 0 3126 - FPSIMD/NEON registers: set to 0 3127 - SVE registers: set to 0 3128 - System registers: Reset to their architecturally defined 3129 values as for a warm reset to EL1 (resp. SVC) 3130 3131Note that because some registers reflect machine topology, all vcpus 3132should be created before this ioctl is invoked. 3133 3134Userspace can call this function multiple times for a given vcpu, including 3135after the vcpu has been run. This will reset the vcpu to its initial 3136state. All calls to this function after the initial call must use the same 3137target and same set of feature flags, otherwise EINVAL will be returned. 3138 3139Possible features: 3140 3141 - KVM_ARM_VCPU_POWER_OFF: Starts the CPU in a power-off state. 3142 Depends on KVM_CAP_ARM_PSCI. If not set, the CPU will be powered on 3143 and execute guest code when KVM_RUN is called. 3144 - KVM_ARM_VCPU_EL1_32BIT: Starts the CPU in a 32bit mode. 3145 Depends on KVM_CAP_ARM_EL1_32BIT (arm64 only). 3146 - KVM_ARM_VCPU_PSCI_0_2: Emulate PSCI v0.2 (or a future revision 3147 backward compatible with v0.2) for the CPU. 3148 Depends on KVM_CAP_ARM_PSCI_0_2. 3149 - KVM_ARM_VCPU_PMU_V3: Emulate PMUv3 for the CPU. 3150 Depends on KVM_CAP_ARM_PMU_V3. 3151 3152 - KVM_ARM_VCPU_PTRAUTH_ADDRESS: Enables Address Pointer authentication 3153 for arm64 only. 3154 Depends on KVM_CAP_ARM_PTRAUTH_ADDRESS. 3155 If KVM_CAP_ARM_PTRAUTH_ADDRESS and KVM_CAP_ARM_PTRAUTH_GENERIC are 3156 both present, then both KVM_ARM_VCPU_PTRAUTH_ADDRESS and 3157 KVM_ARM_VCPU_PTRAUTH_GENERIC must be requested or neither must be 3158 requested. 3159 3160 - KVM_ARM_VCPU_PTRAUTH_GENERIC: Enables Generic Pointer authentication 3161 for arm64 only. 3162 Depends on KVM_CAP_ARM_PTRAUTH_GENERIC. 3163 If KVM_CAP_ARM_PTRAUTH_ADDRESS and KVM_CAP_ARM_PTRAUTH_GENERIC are 3164 both present, then both KVM_ARM_VCPU_PTRAUTH_ADDRESS and 3165 KVM_ARM_VCPU_PTRAUTH_GENERIC must be requested or neither must be 3166 requested. 3167 3168 - KVM_ARM_VCPU_SVE: Enables SVE for the CPU (arm64 only). 3169 Depends on KVM_CAP_ARM_SVE. 3170 Requires KVM_ARM_VCPU_FINALIZE(KVM_ARM_VCPU_SVE): 3171 3172 * After KVM_ARM_VCPU_INIT: 3173 3174 - KVM_REG_ARM64_SVE_VLS may be read using KVM_GET_ONE_REG: the 3175 initial value of this pseudo-register indicates the best set of 3176 vector lengths possible for a vcpu on this host. 3177 3178 * Before KVM_ARM_VCPU_FINALIZE(KVM_ARM_VCPU_SVE): 3179 3180 - KVM_RUN and KVM_GET_REG_LIST are not available; 3181 3182 - KVM_GET_ONE_REG and KVM_SET_ONE_REG cannot be used to access 3183 the scalable archietctural SVE registers 3184 KVM_REG_ARM64_SVE_ZREG(), KVM_REG_ARM64_SVE_PREG() or 3185 KVM_REG_ARM64_SVE_FFR; 3186 3187 - KVM_REG_ARM64_SVE_VLS may optionally be written using 3188 KVM_SET_ONE_REG, to modify the set of vector lengths available 3189 for the vcpu. 3190 3191 * After KVM_ARM_VCPU_FINALIZE(KVM_ARM_VCPU_SVE): 3192 3193 - the KVM_REG_ARM64_SVE_VLS pseudo-register is immutable, and can 3194 no longer be written using KVM_SET_ONE_REG. 3195 31964.83 KVM_ARM_PREFERRED_TARGET 3197----------------------------- 3198 3199:Capability: basic 3200:Architectures: arm, arm64 3201:Type: vm ioctl 3202:Parameters: struct kvm_vcpu_init (out) 3203:Returns: 0 on success; -1 on error 3204 3205Errors: 3206 3207 ====== ========================================== 3208 ENODEV no preferred target available for the host 3209 ====== ========================================== 3210 3211This queries KVM for preferred CPU target type which can be emulated 3212by KVM on underlying host. 3213 3214The ioctl returns struct kvm_vcpu_init instance containing information 3215about preferred CPU target type and recommended features for it. The 3216kvm_vcpu_init->features bitmap returned will have feature bits set if 3217the preferred target recommends setting these features, but this is 3218not mandatory. 3219 3220The information returned by this ioctl can be used to prepare an instance 3221of struct kvm_vcpu_init for KVM_ARM_VCPU_INIT ioctl which will result in 3222VCPU matching underlying host. 3223 3224 32254.84 KVM_GET_REG_LIST 3226--------------------- 3227 3228:Capability: basic 3229:Architectures: arm, arm64, mips 3230:Type: vcpu ioctl 3231:Parameters: struct kvm_reg_list (in/out) 3232:Returns: 0 on success; -1 on error 3233 3234Errors: 3235 3236 ===== ============================================================== 3237 E2BIG the reg index list is too big to fit in the array specified by 3238 the user (the number required will be written into n). 3239 ===== ============================================================== 3240 3241:: 3242 3243 struct kvm_reg_list { 3244 __u64 n; /* number of registers in reg[] */ 3245 __u64 reg[0]; 3246 }; 3247 3248This ioctl returns the guest registers that are supported for the 3249KVM_GET_ONE_REG/KVM_SET_ONE_REG calls. 3250 3251 32524.85 KVM_ARM_SET_DEVICE_ADDR (deprecated) 3253----------------------------------------- 3254 3255:Capability: KVM_CAP_ARM_SET_DEVICE_ADDR 3256:Architectures: arm, arm64 3257:Type: vm ioctl 3258:Parameters: struct kvm_arm_device_address (in) 3259:Returns: 0 on success, -1 on error 3260 3261Errors: 3262 3263 ====== ============================================ 3264 ENODEV The device id is unknown 3265 ENXIO Device not supported on current system 3266 EEXIST Address already set 3267 E2BIG Address outside guest physical address space 3268 EBUSY Address overlaps with other device range 3269 ====== ============================================ 3270 3271:: 3272 3273 struct kvm_arm_device_addr { 3274 __u64 id; 3275 __u64 addr; 3276 }; 3277 3278Specify a device address in the guest's physical address space where guests 3279can access emulated or directly exposed devices, which the host kernel needs 3280to know about. The id field is an architecture specific identifier for a 3281specific device. 3282 3283ARM/arm64 divides the id field into two parts, a device id and an 3284address type id specific to the individual device:: 3285 3286 bits: | 63 ... 32 | 31 ... 16 | 15 ... 0 | 3287 field: | 0x00000000 | device id | addr type id | 3288 3289ARM/arm64 currently only require this when using the in-kernel GIC 3290support for the hardware VGIC features, using KVM_ARM_DEVICE_VGIC_V2 3291as the device id. When setting the base address for the guest's 3292mapping of the VGIC virtual CPU and distributor interface, the ioctl 3293must be called after calling KVM_CREATE_IRQCHIP, but before calling 3294KVM_RUN on any of the VCPUs. Calling this ioctl twice for any of the 3295base addresses will return -EEXIST. 3296 3297Note, this IOCTL is deprecated and the more flexible SET/GET_DEVICE_ATTR API 3298should be used instead. 3299 3300 33014.86 KVM_PPC_RTAS_DEFINE_TOKEN 3302------------------------------ 3303 3304:Capability: KVM_CAP_PPC_RTAS 3305:Architectures: ppc 3306:Type: vm ioctl 3307:Parameters: struct kvm_rtas_token_args 3308:Returns: 0 on success, -1 on error 3309 3310Defines a token value for a RTAS (Run Time Abstraction Services) 3311service in order to allow it to be handled in the kernel. The 3312argument struct gives the name of the service, which must be the name 3313of a service that has a kernel-side implementation. If the token 3314value is non-zero, it will be associated with that service, and 3315subsequent RTAS calls by the guest specifying that token will be 3316handled by the kernel. If the token value is 0, then any token 3317associated with the service will be forgotten, and subsequent RTAS 3318calls by the guest for that service will be passed to userspace to be 3319handled. 3320 33214.87 KVM_SET_GUEST_DEBUG 3322------------------------ 3323 3324:Capability: KVM_CAP_SET_GUEST_DEBUG 3325:Architectures: x86, s390, ppc, arm64 3326:Type: vcpu ioctl 3327:Parameters: struct kvm_guest_debug (in) 3328:Returns: 0 on success; -1 on error 3329 3330:: 3331 3332 struct kvm_guest_debug { 3333 __u32 control; 3334 __u32 pad; 3335 struct kvm_guest_debug_arch arch; 3336 }; 3337 3338Set up the processor specific debug registers and configure vcpu for 3339handling guest debug events. There are two parts to the structure, the 3340first a control bitfield indicates the type of debug events to handle 3341when running. Common control bits are: 3342 3343 - KVM_GUESTDBG_ENABLE: guest debugging is enabled 3344 - KVM_GUESTDBG_SINGLESTEP: the next run should single-step 3345 3346The top 16 bits of the control field are architecture specific control 3347flags which can include the following: 3348 3349 - KVM_GUESTDBG_USE_SW_BP: using software breakpoints [x86, arm64] 3350 - KVM_GUESTDBG_USE_HW_BP: using hardware breakpoints [x86, s390] 3351 - KVM_GUESTDBG_USE_HW: using hardware debug events [arm64] 3352 - KVM_GUESTDBG_INJECT_DB: inject DB type exception [x86] 3353 - KVM_GUESTDBG_INJECT_BP: inject BP type exception [x86] 3354 - KVM_GUESTDBG_EXIT_PENDING: trigger an immediate guest exit [s390] 3355 3356For example KVM_GUESTDBG_USE_SW_BP indicates that software breakpoints 3357are enabled in memory so we need to ensure breakpoint exceptions are 3358correctly trapped and the KVM run loop exits at the breakpoint and not 3359running off into the normal guest vector. For KVM_GUESTDBG_USE_HW_BP 3360we need to ensure the guest vCPUs architecture specific registers are 3361updated to the correct (supplied) values. 3362 3363The second part of the structure is architecture specific and 3364typically contains a set of debug registers. 3365 3366For arm64 the number of debug registers is implementation defined and 3367can be determined by querying the KVM_CAP_GUEST_DEBUG_HW_BPS and 3368KVM_CAP_GUEST_DEBUG_HW_WPS capabilities which return a positive number 3369indicating the number of supported registers. 3370 3371For ppc, the KVM_CAP_PPC_GUEST_DEBUG_SSTEP capability indicates whether 3372the single-step debug event (KVM_GUESTDBG_SINGLESTEP) is supported. 3373 3374Also when supported, KVM_CAP_SET_GUEST_DEBUG2 capability indicates the 3375supported KVM_GUESTDBG_* bits in the control field. 3376 3377When debug events exit the main run loop with the reason 3378KVM_EXIT_DEBUG with the kvm_debug_exit_arch part of the kvm_run 3379structure containing architecture specific debug information. 3380 33814.88 KVM_GET_EMULATED_CPUID 3382--------------------------- 3383 3384:Capability: KVM_CAP_EXT_EMUL_CPUID 3385:Architectures: x86 3386:Type: system ioctl 3387:Parameters: struct kvm_cpuid2 (in/out) 3388:Returns: 0 on success, -1 on error 3389 3390:: 3391 3392 struct kvm_cpuid2 { 3393 __u32 nent; 3394 __u32 flags; 3395 struct kvm_cpuid_entry2 entries[0]; 3396 }; 3397 3398The member 'flags' is used for passing flags from userspace. 3399 3400:: 3401 3402 #define KVM_CPUID_FLAG_SIGNIFCANT_INDEX BIT(0) 3403 #define KVM_CPUID_FLAG_STATEFUL_FUNC BIT(1) /* deprecated */ 3404 #define KVM_CPUID_FLAG_STATE_READ_NEXT BIT(2) /* deprecated */ 3405 3406 struct kvm_cpuid_entry2 { 3407 __u32 function; 3408 __u32 index; 3409 __u32 flags; 3410 __u32 eax; 3411 __u32 ebx; 3412 __u32 ecx; 3413 __u32 edx; 3414 __u32 padding[3]; 3415 }; 3416 3417This ioctl returns x86 cpuid features which are emulated by 3418kvm.Userspace can use the information returned by this ioctl to query 3419which features are emulated by kvm instead of being present natively. 3420 3421Userspace invokes KVM_GET_EMULATED_CPUID by passing a kvm_cpuid2 3422structure with the 'nent' field indicating the number of entries in 3423the variable-size array 'entries'. If the number of entries is too low 3424to describe the cpu capabilities, an error (E2BIG) is returned. If the 3425number is too high, the 'nent' field is adjusted and an error (ENOMEM) 3426is returned. If the number is just right, the 'nent' field is adjusted 3427to the number of valid entries in the 'entries' array, which is then 3428filled. 3429 3430The entries returned are the set CPUID bits of the respective features 3431which kvm emulates, as returned by the CPUID instruction, with unknown 3432or unsupported feature bits cleared. 3433 3434Features like x2apic, for example, may not be present in the host cpu 3435but are exposed by kvm in KVM_GET_SUPPORTED_CPUID because they can be 3436emulated efficiently and thus not included here. 3437 3438The fields in each entry are defined as follows: 3439 3440 function: 3441 the eax value used to obtain the entry 3442 index: 3443 the ecx value used to obtain the entry (for entries that are 3444 affected by ecx) 3445 flags: 3446 an OR of zero or more of the following: 3447 3448 KVM_CPUID_FLAG_SIGNIFCANT_INDEX: 3449 if the index field is valid 3450 3451 eax, ebx, ecx, edx: 3452 3453 the values returned by the cpuid instruction for 3454 this function/index combination 3455 34564.89 KVM_S390_MEM_OP 3457-------------------- 3458 3459:Capability: KVM_CAP_S390_MEM_OP 3460:Architectures: s390 3461:Type: vcpu ioctl 3462:Parameters: struct kvm_s390_mem_op (in) 3463:Returns: = 0 on success, 3464 < 0 on generic error (e.g. -EFAULT or -ENOMEM), 3465 > 0 if an exception occurred while walking the page tables 3466 3467Read or write data from/to the logical (virtual) memory of a VCPU. 3468 3469Parameters are specified via the following structure:: 3470 3471 struct kvm_s390_mem_op { 3472 __u64 gaddr; /* the guest address */ 3473 __u64 flags; /* flags */ 3474 __u32 size; /* amount of bytes */ 3475 __u32 op; /* type of operation */ 3476 __u64 buf; /* buffer in userspace */ 3477 __u8 ar; /* the access register number */ 3478 __u8 reserved[31]; /* should be set to 0 */ 3479 }; 3480 3481The type of operation is specified in the "op" field. It is either 3482KVM_S390_MEMOP_LOGICAL_READ for reading from logical memory space or 3483KVM_S390_MEMOP_LOGICAL_WRITE for writing to logical memory space. The 3484KVM_S390_MEMOP_F_CHECK_ONLY flag can be set in the "flags" field to check 3485whether the corresponding memory access would create an access exception 3486(without touching the data in the memory at the destination). In case an 3487access exception occurred while walking the MMU tables of the guest, the 3488ioctl returns a positive error number to indicate the type of exception. 3489This exception is also raised directly at the corresponding VCPU if the 3490flag KVM_S390_MEMOP_F_INJECT_EXCEPTION is set in the "flags" field. 3491 3492The start address of the memory region has to be specified in the "gaddr" 3493field, and the length of the region in the "size" field (which must not 3494be 0). The maximum value for "size" can be obtained by checking the 3495KVM_CAP_S390_MEM_OP capability. "buf" is the buffer supplied by the 3496userspace application where the read data should be written to for 3497KVM_S390_MEMOP_LOGICAL_READ, or where the data that should be written is 3498stored for a KVM_S390_MEMOP_LOGICAL_WRITE. When KVM_S390_MEMOP_F_CHECK_ONLY 3499is specified, "buf" is unused and can be NULL. "ar" designates the access 3500register number to be used; the valid range is 0..15. 3501 3502The "reserved" field is meant for future extensions. It is not used by 3503KVM with the currently defined set of flags. 3504 35054.90 KVM_S390_GET_SKEYS 3506----------------------- 3507 3508:Capability: KVM_CAP_S390_SKEYS 3509:Architectures: s390 3510:Type: vm ioctl 3511:Parameters: struct kvm_s390_skeys 3512:Returns: 0 on success, KVM_S390_GET_KEYS_NONE if guest is not using storage 3513 keys, negative value on error 3514 3515This ioctl is used to get guest storage key values on the s390 3516architecture. The ioctl takes parameters via the kvm_s390_skeys struct:: 3517 3518 struct kvm_s390_skeys { 3519 __u64 start_gfn; 3520 __u64 count; 3521 __u64 skeydata_addr; 3522 __u32 flags; 3523 __u32 reserved[9]; 3524 }; 3525 3526The start_gfn field is the number of the first guest frame whose storage keys 3527you want to get. 3528 3529The count field is the number of consecutive frames (starting from start_gfn) 3530whose storage keys to get. The count field must be at least 1 and the maximum 3531allowed value is defined as KVM_S390_SKEYS_ALLOC_MAX. Values outside this range 3532will cause the ioctl to return -EINVAL. 3533 3534The skeydata_addr field is the address to a buffer large enough to hold count 3535bytes. This buffer will be filled with storage key data by the ioctl. 3536 35374.91 KVM_S390_SET_SKEYS 3538----------------------- 3539 3540:Capability: KVM_CAP_S390_SKEYS 3541:Architectures: s390 3542:Type: vm ioctl 3543:Parameters: struct kvm_s390_skeys 3544:Returns: 0 on success, negative value on error 3545 3546This ioctl is used to set guest storage key values on the s390 3547architecture. The ioctl takes parameters via the kvm_s390_skeys struct. 3548See section on KVM_S390_GET_SKEYS for struct definition. 3549 3550The start_gfn field is the number of the first guest frame whose storage keys 3551you want to set. 3552 3553The count field is the number of consecutive frames (starting from start_gfn) 3554whose storage keys to get. The count field must be at least 1 and the maximum 3555allowed value is defined as KVM_S390_SKEYS_ALLOC_MAX. Values outside this range 3556will cause the ioctl to return -EINVAL. 3557 3558The skeydata_addr field is the address to a buffer containing count bytes of 3559storage keys. Each byte in the buffer will be set as the storage key for a 3560single frame starting at start_gfn for count frames. 3561 3562Note: If any architecturally invalid key value is found in the given data then 3563the ioctl will return -EINVAL. 3564 35654.92 KVM_S390_IRQ 3566----------------- 3567 3568:Capability: KVM_CAP_S390_INJECT_IRQ 3569:Architectures: s390 3570:Type: vcpu ioctl 3571:Parameters: struct kvm_s390_irq (in) 3572:Returns: 0 on success, -1 on error 3573 3574Errors: 3575 3576 3577 ====== ================================================================= 3578 EINVAL interrupt type is invalid 3579 type is KVM_S390_SIGP_STOP and flag parameter is invalid value, 3580 type is KVM_S390_INT_EXTERNAL_CALL and code is bigger 3581 than the maximum of VCPUs 3582 EBUSY type is KVM_S390_SIGP_SET_PREFIX and vcpu is not stopped, 3583 type is KVM_S390_SIGP_STOP and a stop irq is already pending, 3584 type is KVM_S390_INT_EXTERNAL_CALL and an external call interrupt 3585 is already pending 3586 ====== ================================================================= 3587 3588Allows to inject an interrupt to the guest. 3589 3590Using struct kvm_s390_irq as a parameter allows 3591to inject additional payload which is not 3592possible via KVM_S390_INTERRUPT. 3593 3594Interrupt parameters are passed via kvm_s390_irq:: 3595 3596 struct kvm_s390_irq { 3597 __u64 type; 3598 union { 3599 struct kvm_s390_io_info io; 3600 struct kvm_s390_ext_info ext; 3601 struct kvm_s390_pgm_info pgm; 3602 struct kvm_s390_emerg_info emerg; 3603 struct kvm_s390_extcall_info extcall; 3604 struct kvm_s390_prefix_info prefix; 3605 struct kvm_s390_stop_info stop; 3606 struct kvm_s390_mchk_info mchk; 3607 char reserved[64]; 3608 } u; 3609 }; 3610 3611type can be one of the following: 3612 3613- KVM_S390_SIGP_STOP - sigp stop; parameter in .stop 3614- KVM_S390_PROGRAM_INT - program check; parameters in .pgm 3615- KVM_S390_SIGP_SET_PREFIX - sigp set prefix; parameters in .prefix 3616- KVM_S390_RESTART - restart; no parameters 3617- KVM_S390_INT_CLOCK_COMP - clock comparator interrupt; no parameters 3618- KVM_S390_INT_CPU_TIMER - CPU timer interrupt; no parameters 3619- KVM_S390_INT_EMERGENCY - sigp emergency; parameters in .emerg 3620- KVM_S390_INT_EXTERNAL_CALL - sigp external call; parameters in .extcall 3621- KVM_S390_MCHK - machine check interrupt; parameters in .mchk 3622 3623This is an asynchronous vcpu ioctl and can be invoked from any thread. 3624 36254.94 KVM_S390_GET_IRQ_STATE 3626--------------------------- 3627 3628:Capability: KVM_CAP_S390_IRQ_STATE 3629:Architectures: s390 3630:Type: vcpu ioctl 3631:Parameters: struct kvm_s390_irq_state (out) 3632:Returns: >= number of bytes copied into buffer, 3633 -EINVAL if buffer size is 0, 3634 -ENOBUFS if buffer size is too small to fit all pending interrupts, 3635 -EFAULT if the buffer address was invalid 3636 3637This ioctl allows userspace to retrieve the complete state of all currently 3638pending interrupts in a single buffer. Use cases include migration 3639and introspection. The parameter structure contains the address of a 3640userspace buffer and its length:: 3641 3642 struct kvm_s390_irq_state { 3643 __u64 buf; 3644 __u32 flags; /* will stay unused for compatibility reasons */ 3645 __u32 len; 3646 __u32 reserved[4]; /* will stay unused for compatibility reasons */ 3647 }; 3648 3649Userspace passes in the above struct and for each pending interrupt a 3650struct kvm_s390_irq is copied to the provided buffer. 3651 3652The structure contains a flags and a reserved field for future extensions. As 3653the kernel never checked for flags == 0 and QEMU never pre-zeroed flags and 3654reserved, these fields can not be used in the future without breaking 3655compatibility. 3656 3657If -ENOBUFS is returned the buffer provided was too small and userspace 3658may retry with a bigger buffer. 3659 36604.95 KVM_S390_SET_IRQ_STATE 3661--------------------------- 3662 3663:Capability: KVM_CAP_S390_IRQ_STATE 3664:Architectures: s390 3665:Type: vcpu ioctl 3666:Parameters: struct kvm_s390_irq_state (in) 3667:Returns: 0 on success, 3668 -EFAULT if the buffer address was invalid, 3669 -EINVAL for an invalid buffer length (see below), 3670 -EBUSY if there were already interrupts pending, 3671 errors occurring when actually injecting the 3672 interrupt. See KVM_S390_IRQ. 3673 3674This ioctl allows userspace to set the complete state of all cpu-local 3675interrupts currently pending for the vcpu. It is intended for restoring 3676interrupt state after a migration. The input parameter is a userspace buffer 3677containing a struct kvm_s390_irq_state:: 3678 3679 struct kvm_s390_irq_state { 3680 __u64 buf; 3681 __u32 flags; /* will stay unused for compatibility reasons */ 3682 __u32 len; 3683 __u32 reserved[4]; /* will stay unused for compatibility reasons */ 3684 }; 3685 3686The restrictions for flags and reserved apply as well. 3687(see KVM_S390_GET_IRQ_STATE) 3688 3689The userspace memory referenced by buf contains a struct kvm_s390_irq 3690for each interrupt to be injected into the guest. 3691If one of the interrupts could not be injected for some reason the 3692ioctl aborts. 3693 3694len must be a multiple of sizeof(struct kvm_s390_irq). It must be > 0 3695and it must not exceed (max_vcpus + 32) * sizeof(struct kvm_s390_irq), 3696which is the maximum number of possibly pending cpu-local interrupts. 3697 36984.96 KVM_SMI 3699------------ 3700 3701:Capability: KVM_CAP_X86_SMM 3702:Architectures: x86 3703:Type: vcpu ioctl 3704:Parameters: none 3705:Returns: 0 on success, -1 on error 3706 3707Queues an SMI on the thread's vcpu. 3708 37094.97 KVM_X86_SET_MSR_FILTER 3710---------------------------- 3711 3712:Capability: KVM_X86_SET_MSR_FILTER 3713:Architectures: x86 3714:Type: vm ioctl 3715:Parameters: struct kvm_msr_filter 3716:Returns: 0 on success, < 0 on error 3717 3718:: 3719 3720 struct kvm_msr_filter_range { 3721 #define KVM_MSR_FILTER_READ (1 << 0) 3722 #define KVM_MSR_FILTER_WRITE (1 << 1) 3723 __u32 flags; 3724 __u32 nmsrs; /* number of msrs in bitmap */ 3725 __u32 base; /* MSR index the bitmap starts at */ 3726 __u8 *bitmap; /* a 1 bit allows the operations in flags, 0 denies */ 3727 }; 3728 3729 #define KVM_MSR_FILTER_MAX_RANGES 16 3730 struct kvm_msr_filter { 3731 #define KVM_MSR_FILTER_DEFAULT_ALLOW (0 << 0) 3732 #define KVM_MSR_FILTER_DEFAULT_DENY (1 << 0) 3733 __u32 flags; 3734 struct kvm_msr_filter_range ranges[KVM_MSR_FILTER_MAX_RANGES]; 3735 }; 3736 3737flags values for ``struct kvm_msr_filter_range``: 3738 3739``KVM_MSR_FILTER_READ`` 3740 3741 Filter read accesses to MSRs using the given bitmap. A 0 in the bitmap 3742 indicates that a read should immediately fail, while a 1 indicates that 3743 a read for a particular MSR should be handled regardless of the default 3744 filter action. 3745 3746``KVM_MSR_FILTER_WRITE`` 3747 3748 Filter write accesses to MSRs using the given bitmap. A 0 in the bitmap 3749 indicates that a write should immediately fail, while a 1 indicates that 3750 a write for a particular MSR should be handled regardless of the default 3751 filter action. 3752 3753``KVM_MSR_FILTER_READ | KVM_MSR_FILTER_WRITE`` 3754 3755 Filter both read and write accesses to MSRs using the given bitmap. A 0 3756 in the bitmap indicates that both reads and writes should immediately fail, 3757 while a 1 indicates that reads and writes for a particular MSR are not 3758 filtered by this range. 3759 3760flags values for ``struct kvm_msr_filter``: 3761 3762``KVM_MSR_FILTER_DEFAULT_ALLOW`` 3763 3764 If no filter range matches an MSR index that is getting accessed, KVM will 3765 fall back to allowing access to the MSR. 3766 3767``KVM_MSR_FILTER_DEFAULT_DENY`` 3768 3769 If no filter range matches an MSR index that is getting accessed, KVM will 3770 fall back to rejecting access to the MSR. In this mode, all MSRs that should 3771 be processed by KVM need to explicitly be marked as allowed in the bitmaps. 3772 3773This ioctl allows user space to define up to 16 bitmaps of MSR ranges to 3774specify whether a certain MSR access should be explicitly filtered for or not. 3775 3776If this ioctl has never been invoked, MSR accesses are not guarded and the 3777default KVM in-kernel emulation behavior is fully preserved. 3778 3779Calling this ioctl with an empty set of ranges (all nmsrs == 0) disables MSR 3780filtering. In that mode, ``KVM_MSR_FILTER_DEFAULT_DENY`` is invalid and causes 3781an error. 3782 3783As soon as the filtering is in place, every MSR access is processed through 3784the filtering except for accesses to the x2APIC MSRs (from 0x800 to 0x8ff); 3785x2APIC MSRs are always allowed, independent of the ``default_allow`` setting, 3786and their behavior depends on the ``X2APIC_ENABLE`` bit of the APIC base 3787register. 3788 3789If a bit is within one of the defined ranges, read and write accesses are 3790guarded by the bitmap's value for the MSR index if the kind of access 3791is included in the ``struct kvm_msr_filter_range`` flags. If no range 3792cover this particular access, the behavior is determined by the flags 3793field in the kvm_msr_filter struct: ``KVM_MSR_FILTER_DEFAULT_ALLOW`` 3794and ``KVM_MSR_FILTER_DEFAULT_DENY``. 3795 3796Each bitmap range specifies a range of MSRs to potentially allow access on. 3797The range goes from MSR index [base .. base+nmsrs]. The flags field 3798indicates whether reads, writes or both reads and writes are filtered 3799by setting a 1 bit in the bitmap for the corresponding MSR index. 3800 3801If an MSR access is not permitted through the filtering, it generates a 3802#GP inside the guest. When combined with KVM_CAP_X86_USER_SPACE_MSR, that 3803allows user space to deflect and potentially handle various MSR accesses 3804into user space. 3805 3806If a vCPU is in running state while this ioctl is invoked, the vCPU may 3807experience inconsistent filtering behavior on MSR accesses. 3808 38094.98 KVM_CREATE_SPAPR_TCE_64 3810---------------------------- 3811 3812:Capability: KVM_CAP_SPAPR_TCE_64 3813:Architectures: powerpc 3814:Type: vm ioctl 3815:Parameters: struct kvm_create_spapr_tce_64 (in) 3816:Returns: file descriptor for manipulating the created TCE table 3817 3818This is an extension for KVM_CAP_SPAPR_TCE which only supports 32bit 3819windows, described in 4.62 KVM_CREATE_SPAPR_TCE 3820 3821This capability uses extended struct in ioctl interface:: 3822 3823 /* for KVM_CAP_SPAPR_TCE_64 */ 3824 struct kvm_create_spapr_tce_64 { 3825 __u64 liobn; 3826 __u32 page_shift; 3827 __u32 flags; 3828 __u64 offset; /* in pages */ 3829 __u64 size; /* in pages */ 3830 }; 3831 3832The aim of extension is to support an additional bigger DMA window with 3833a variable page size. 3834KVM_CREATE_SPAPR_TCE_64 receives a 64bit window size, an IOMMU page shift and 3835a bus offset of the corresponding DMA window, @size and @offset are numbers 3836of IOMMU pages. 3837 3838@flags are not used at the moment. 3839 3840The rest of functionality is identical to KVM_CREATE_SPAPR_TCE. 3841 38424.99 KVM_REINJECT_CONTROL 3843------------------------- 3844 3845:Capability: KVM_CAP_REINJECT_CONTROL 3846:Architectures: x86 3847:Type: vm ioctl 3848:Parameters: struct kvm_reinject_control (in) 3849:Returns: 0 on success, 3850 -EFAULT if struct kvm_reinject_control cannot be read, 3851 -ENXIO if KVM_CREATE_PIT or KVM_CREATE_PIT2 didn't succeed earlier. 3852 3853i8254 (PIT) has two modes, reinject and !reinject. The default is reinject, 3854where KVM queues elapsed i8254 ticks and monitors completion of interrupt from 3855vector(s) that i8254 injects. Reinject mode dequeues a tick and injects its 3856interrupt whenever there isn't a pending interrupt from i8254. 3857!reinject mode injects an interrupt as soon as a tick arrives. 3858 3859:: 3860 3861 struct kvm_reinject_control { 3862 __u8 pit_reinject; 3863 __u8 reserved[31]; 3864 }; 3865 3866pit_reinject = 0 (!reinject mode) is recommended, unless running an old 3867operating system that uses the PIT for timing (e.g. Linux 2.4.x). 3868 38694.100 KVM_PPC_CONFIGURE_V3_MMU 3870------------------------------ 3871 3872:Capability: KVM_CAP_PPC_RADIX_MMU or KVM_CAP_PPC_HASH_MMU_V3 3873:Architectures: ppc 3874:Type: vm ioctl 3875:Parameters: struct kvm_ppc_mmuv3_cfg (in) 3876:Returns: 0 on success, 3877 -EFAULT if struct kvm_ppc_mmuv3_cfg cannot be read, 3878 -EINVAL if the configuration is invalid 3879 3880This ioctl controls whether the guest will use radix or HPT (hashed 3881page table) translation, and sets the pointer to the process table for 3882the guest. 3883 3884:: 3885 3886 struct kvm_ppc_mmuv3_cfg { 3887 __u64 flags; 3888 __u64 process_table; 3889 }; 3890 3891There are two bits that can be set in flags; KVM_PPC_MMUV3_RADIX and 3892KVM_PPC_MMUV3_GTSE. KVM_PPC_MMUV3_RADIX, if set, configures the guest 3893to use radix tree translation, and if clear, to use HPT translation. 3894KVM_PPC_MMUV3_GTSE, if set and if KVM permits it, configures the guest 3895to be able to use the global TLB and SLB invalidation instructions; 3896if clear, the guest may not use these instructions. 3897 3898The process_table field specifies the address and size of the guest 3899process table, which is in the guest's space. This field is formatted 3900as the second doubleword of the partition table entry, as defined in 3901the Power ISA V3.00, Book III section 5.7.6.1. 3902 39034.101 KVM_PPC_GET_RMMU_INFO 3904--------------------------- 3905 3906:Capability: KVM_CAP_PPC_RADIX_MMU 3907:Architectures: ppc 3908:Type: vm ioctl 3909:Parameters: struct kvm_ppc_rmmu_info (out) 3910:Returns: 0 on success, 3911 -EFAULT if struct kvm_ppc_rmmu_info cannot be written, 3912 -EINVAL if no useful information can be returned 3913 3914This ioctl returns a structure containing two things: (a) a list 3915containing supported radix tree geometries, and (b) a list that maps 3916page sizes to put in the "AP" (actual page size) field for the tlbie 3917(TLB invalidate entry) instruction. 3918 3919:: 3920 3921 struct kvm_ppc_rmmu_info { 3922 struct kvm_ppc_radix_geom { 3923 __u8 page_shift; 3924 __u8 level_bits[4]; 3925 __u8 pad[3]; 3926 } geometries[8]; 3927 __u32 ap_encodings[8]; 3928 }; 3929 3930The geometries[] field gives up to 8 supported geometries for the 3931radix page table, in terms of the log base 2 of the smallest page 3932size, and the number of bits indexed at each level of the tree, from 3933the PTE level up to the PGD level in that order. Any unused entries 3934will have 0 in the page_shift field. 3935 3936The ap_encodings gives the supported page sizes and their AP field 3937encodings, encoded with the AP value in the top 3 bits and the log 3938base 2 of the page size in the bottom 6 bits. 3939 39404.102 KVM_PPC_RESIZE_HPT_PREPARE 3941-------------------------------- 3942 3943:Capability: KVM_CAP_SPAPR_RESIZE_HPT 3944:Architectures: powerpc 3945:Type: vm ioctl 3946:Parameters: struct kvm_ppc_resize_hpt (in) 3947:Returns: 0 on successful completion, 3948 >0 if a new HPT is being prepared, the value is an estimated 3949 number of milliseconds until preparation is complete, 3950 -EFAULT if struct kvm_reinject_control cannot be read, 3951 -EINVAL if the supplied shift or flags are invalid, 3952 -ENOMEM if unable to allocate the new HPT, 3953 3954Used to implement the PAPR extension for runtime resizing of a guest's 3955Hashed Page Table (HPT). Specifically this starts, stops or monitors 3956the preparation of a new potential HPT for the guest, essentially 3957implementing the H_RESIZE_HPT_PREPARE hypercall. 3958 3959:: 3960 3961 struct kvm_ppc_resize_hpt { 3962 __u64 flags; 3963 __u32 shift; 3964 __u32 pad; 3965 }; 3966 3967If called with shift > 0 when there is no pending HPT for the guest, 3968this begins preparation of a new pending HPT of size 2^(shift) bytes. 3969It then returns a positive integer with the estimated number of 3970milliseconds until preparation is complete. 3971 3972If called when there is a pending HPT whose size does not match that 3973requested in the parameters, discards the existing pending HPT and 3974creates a new one as above. 3975 3976If called when there is a pending HPT of the size requested, will: 3977 3978 * If preparation of the pending HPT is already complete, return 0 3979 * If preparation of the pending HPT has failed, return an error 3980 code, then discard the pending HPT. 3981 * If preparation of the pending HPT is still in progress, return an 3982 estimated number of milliseconds until preparation is complete. 3983 3984If called with shift == 0, discards any currently pending HPT and 3985returns 0 (i.e. cancels any in-progress preparation). 3986 3987flags is reserved for future expansion, currently setting any bits in 3988flags will result in an -EINVAL. 3989 3990Normally this will be called repeatedly with the same parameters until 3991it returns <= 0. The first call will initiate preparation, subsequent 3992ones will monitor preparation until it completes or fails. 3993 39944.103 KVM_PPC_RESIZE_HPT_COMMIT 3995------------------------------- 3996 3997:Capability: KVM_CAP_SPAPR_RESIZE_HPT 3998:Architectures: powerpc 3999:Type: vm ioctl 4000:Parameters: struct kvm_ppc_resize_hpt (in) 4001:Returns: 0 on successful completion, 4002 -EFAULT if struct kvm_reinject_control cannot be read, 4003 -EINVAL if the supplied shift or flags are invalid, 4004 -ENXIO is there is no pending HPT, or the pending HPT doesn't 4005 have the requested size, 4006 -EBUSY if the pending HPT is not fully prepared, 4007 -ENOSPC if there was a hash collision when moving existing 4008 HPT entries to the new HPT, 4009 -EIO on other error conditions 4010 4011Used to implement the PAPR extension for runtime resizing of a guest's 4012Hashed Page Table (HPT). Specifically this requests that the guest be 4013transferred to working with the new HPT, essentially implementing the 4014H_RESIZE_HPT_COMMIT hypercall. 4015 4016:: 4017 4018 struct kvm_ppc_resize_hpt { 4019 __u64 flags; 4020 __u32 shift; 4021 __u32 pad; 4022 }; 4023 4024This should only be called after KVM_PPC_RESIZE_HPT_PREPARE has 4025returned 0 with the same parameters. In other cases 4026KVM_PPC_RESIZE_HPT_COMMIT will return an error (usually -ENXIO or 4027-EBUSY, though others may be possible if the preparation was started, 4028but failed). 4029 4030This will have undefined effects on the guest if it has not already 4031placed itself in a quiescent state where no vcpu will make MMU enabled 4032memory accesses. 4033 4034On succsful completion, the pending HPT will become the guest's active 4035HPT and the previous HPT will be discarded. 4036 4037On failure, the guest will still be operating on its previous HPT. 4038 40394.104 KVM_X86_GET_MCE_CAP_SUPPORTED 4040----------------------------------- 4041 4042:Capability: KVM_CAP_MCE 4043:Architectures: x86 4044:Type: system ioctl 4045:Parameters: u64 mce_cap (out) 4046:Returns: 0 on success, -1 on error 4047 4048Returns supported MCE capabilities. The u64 mce_cap parameter 4049has the same format as the MSR_IA32_MCG_CAP register. Supported 4050capabilities will have the corresponding bits set. 4051 40524.105 KVM_X86_SETUP_MCE 4053----------------------- 4054 4055:Capability: KVM_CAP_MCE 4056:Architectures: x86 4057:Type: vcpu ioctl 4058:Parameters: u64 mcg_cap (in) 4059:Returns: 0 on success, 4060 -EFAULT if u64 mcg_cap cannot be read, 4061 -EINVAL if the requested number of banks is invalid, 4062 -EINVAL if requested MCE capability is not supported. 4063 4064Initializes MCE support for use. The u64 mcg_cap parameter 4065has the same format as the MSR_IA32_MCG_CAP register and 4066specifies which capabilities should be enabled. The maximum 4067supported number of error-reporting banks can be retrieved when 4068checking for KVM_CAP_MCE. The supported capabilities can be 4069retrieved with KVM_X86_GET_MCE_CAP_SUPPORTED. 4070 40714.106 KVM_X86_SET_MCE 4072--------------------- 4073 4074:Capability: KVM_CAP_MCE 4075:Architectures: x86 4076:Type: vcpu ioctl 4077:Parameters: struct kvm_x86_mce (in) 4078:Returns: 0 on success, 4079 -EFAULT if struct kvm_x86_mce cannot be read, 4080 -EINVAL if the bank number is invalid, 4081 -EINVAL if VAL bit is not set in status field. 4082 4083Inject a machine check error (MCE) into the guest. The input 4084parameter is:: 4085 4086 struct kvm_x86_mce { 4087 __u64 status; 4088 __u64 addr; 4089 __u64 misc; 4090 __u64 mcg_status; 4091 __u8 bank; 4092 __u8 pad1[7]; 4093 __u64 pad2[3]; 4094 }; 4095 4096If the MCE being reported is an uncorrected error, KVM will 4097inject it as an MCE exception into the guest. If the guest 4098MCG_STATUS register reports that an MCE is in progress, KVM 4099causes an KVM_EXIT_SHUTDOWN vmexit. 4100 4101Otherwise, if the MCE is a corrected error, KVM will just 4102store it in the corresponding bank (provided this bank is 4103not holding a previously reported uncorrected error). 4104 41054.107 KVM_S390_GET_CMMA_BITS 4106---------------------------- 4107 4108:Capability: KVM_CAP_S390_CMMA_MIGRATION 4109:Architectures: s390 4110:Type: vm ioctl 4111:Parameters: struct kvm_s390_cmma_log (in, out) 4112:Returns: 0 on success, a negative value on error 4113 4114This ioctl is used to get the values of the CMMA bits on the s390 4115architecture. It is meant to be used in two scenarios: 4116 4117- During live migration to save the CMMA values. Live migration needs 4118 to be enabled via the KVM_REQ_START_MIGRATION VM property. 4119- To non-destructively peek at the CMMA values, with the flag 4120 KVM_S390_CMMA_PEEK set. 4121 4122The ioctl takes parameters via the kvm_s390_cmma_log struct. The desired 4123values are written to a buffer whose location is indicated via the "values" 4124member in the kvm_s390_cmma_log struct. The values in the input struct are 4125also updated as needed. 4126 4127Each CMMA value takes up one byte. 4128 4129:: 4130 4131 struct kvm_s390_cmma_log { 4132 __u64 start_gfn; 4133 __u32 count; 4134 __u32 flags; 4135 union { 4136 __u64 remaining; 4137 __u64 mask; 4138 }; 4139 __u64 values; 4140 }; 4141 4142start_gfn is the number of the first guest frame whose CMMA values are 4143to be retrieved, 4144 4145count is the length of the buffer in bytes, 4146 4147values points to the buffer where the result will be written to. 4148 4149If count is greater than KVM_S390_SKEYS_MAX, then it is considered to be 4150KVM_S390_SKEYS_MAX. KVM_S390_SKEYS_MAX is re-used for consistency with 4151other ioctls. 4152 4153The result is written in the buffer pointed to by the field values, and 4154the values of the input parameter are updated as follows. 4155 4156Depending on the flags, different actions are performed. The only 4157supported flag so far is KVM_S390_CMMA_PEEK. 4158 4159The default behaviour if KVM_S390_CMMA_PEEK is not set is: 4160start_gfn will indicate the first page frame whose CMMA bits were dirty. 4161It is not necessarily the same as the one passed as input, as clean pages 4162are skipped. 4163 4164count will indicate the number of bytes actually written in the buffer. 4165It can (and very often will) be smaller than the input value, since the 4166buffer is only filled until 16 bytes of clean values are found (which 4167are then not copied in the buffer). Since a CMMA migration block needs 4168the base address and the length, for a total of 16 bytes, we will send 4169back some clean data if there is some dirty data afterwards, as long as 4170the size of the clean data does not exceed the size of the header. This 4171allows to minimize the amount of data to be saved or transferred over 4172the network at the expense of more roundtrips to userspace. The next 4173invocation of the ioctl will skip over all the clean values, saving 4174potentially more than just the 16 bytes we found. 4175 4176If KVM_S390_CMMA_PEEK is set: 4177the existing storage attributes are read even when not in migration 4178mode, and no other action is performed; 4179 4180the output start_gfn will be equal to the input start_gfn, 4181 4182the output count will be equal to the input count, except if the end of 4183memory has been reached. 4184 4185In both cases: 4186the field "remaining" will indicate the total number of dirty CMMA values 4187still remaining, or 0 if KVM_S390_CMMA_PEEK is set and migration mode is 4188not enabled. 4189 4190mask is unused. 4191 4192values points to the userspace buffer where the result will be stored. 4193 4194This ioctl can fail with -ENOMEM if not enough memory can be allocated to 4195complete the task, with -ENXIO if CMMA is not enabled, with -EINVAL if 4196KVM_S390_CMMA_PEEK is not set but migration mode was not enabled, with 4197-EFAULT if the userspace address is invalid or if no page table is 4198present for the addresses (e.g. when using hugepages). 4199 42004.108 KVM_S390_SET_CMMA_BITS 4201---------------------------- 4202 4203:Capability: KVM_CAP_S390_CMMA_MIGRATION 4204:Architectures: s390 4205:Type: vm ioctl 4206:Parameters: struct kvm_s390_cmma_log (in) 4207:Returns: 0 on success, a negative value on error 4208 4209This ioctl is used to set the values of the CMMA bits on the s390 4210architecture. It is meant to be used during live migration to restore 4211the CMMA values, but there are no restrictions on its use. 4212The ioctl takes parameters via the kvm_s390_cmma_values struct. 4213Each CMMA value takes up one byte. 4214 4215:: 4216 4217 struct kvm_s390_cmma_log { 4218 __u64 start_gfn; 4219 __u32 count; 4220 __u32 flags; 4221 union { 4222 __u64 remaining; 4223 __u64 mask; 4224 }; 4225 __u64 values; 4226 }; 4227 4228start_gfn indicates the starting guest frame number, 4229 4230count indicates how many values are to be considered in the buffer, 4231 4232flags is not used and must be 0. 4233 4234mask indicates which PGSTE bits are to be considered. 4235 4236remaining is not used. 4237 4238values points to the buffer in userspace where to store the values. 4239 4240This ioctl can fail with -ENOMEM if not enough memory can be allocated to 4241complete the task, with -ENXIO if CMMA is not enabled, with -EINVAL if 4242the count field is too large (e.g. more than KVM_S390_CMMA_SIZE_MAX) or 4243if the flags field was not 0, with -EFAULT if the userspace address is 4244invalid, if invalid pages are written to (e.g. after the end of memory) 4245or if no page table is present for the addresses (e.g. when using 4246hugepages). 4247 42484.109 KVM_PPC_GET_CPU_CHAR 4249-------------------------- 4250 4251:Capability: KVM_CAP_PPC_GET_CPU_CHAR 4252:Architectures: powerpc 4253:Type: vm ioctl 4254:Parameters: struct kvm_ppc_cpu_char (out) 4255:Returns: 0 on successful completion, 4256 -EFAULT if struct kvm_ppc_cpu_char cannot be written 4257 4258This ioctl gives userspace information about certain characteristics 4259of the CPU relating to speculative execution of instructions and 4260possible information leakage resulting from speculative execution (see 4261CVE-2017-5715, CVE-2017-5753 and CVE-2017-5754). The information is 4262returned in struct kvm_ppc_cpu_char, which looks like this:: 4263 4264 struct kvm_ppc_cpu_char { 4265 __u64 character; /* characteristics of the CPU */ 4266 __u64 behaviour; /* recommended software behaviour */ 4267 __u64 character_mask; /* valid bits in character */ 4268 __u64 behaviour_mask; /* valid bits in behaviour */ 4269 }; 4270 4271For extensibility, the character_mask and behaviour_mask fields 4272indicate which bits of character and behaviour have been filled in by 4273the kernel. If the set of defined bits is extended in future then 4274userspace will be able to tell whether it is running on a kernel that 4275knows about the new bits. 4276 4277The character field describes attributes of the CPU which can help 4278with preventing inadvertent information disclosure - specifically, 4279whether there is an instruction to flash-invalidate the L1 data cache 4280(ori 30,30,0 or mtspr SPRN_TRIG2,rN), whether the L1 data cache is set 4281to a mode where entries can only be used by the thread that created 4282them, whether the bcctr[l] instruction prevents speculation, and 4283whether a speculation barrier instruction (ori 31,31,0) is provided. 4284 4285The behaviour field describes actions that software should take to 4286prevent inadvertent information disclosure, and thus describes which 4287vulnerabilities the hardware is subject to; specifically whether the 4288L1 data cache should be flushed when returning to user mode from the 4289kernel, and whether a speculation barrier should be placed between an 4290array bounds check and the array access. 4291 4292These fields use the same bit definitions as the new 4293H_GET_CPU_CHARACTERISTICS hypercall. 4294 42954.110 KVM_MEMORY_ENCRYPT_OP 4296--------------------------- 4297 4298:Capability: basic 4299:Architectures: x86 4300:Type: vm 4301:Parameters: an opaque platform specific structure (in/out) 4302:Returns: 0 on success; -1 on error 4303 4304If the platform supports creating encrypted VMs then this ioctl can be used 4305for issuing platform-specific memory encryption commands to manage those 4306encrypted VMs. 4307 4308Currently, this ioctl is used for issuing Secure Encrypted Virtualization 4309(SEV) commands on AMD Processors. The SEV commands are defined in 4310Documentation/virt/kvm/amd-memory-encryption.rst. 4311 43124.111 KVM_MEMORY_ENCRYPT_REG_REGION 4313----------------------------------- 4314 4315:Capability: basic 4316:Architectures: x86 4317:Type: system 4318:Parameters: struct kvm_enc_region (in) 4319:Returns: 0 on success; -1 on error 4320 4321This ioctl can be used to register a guest memory region which may 4322contain encrypted data (e.g. guest RAM, SMRAM etc). 4323 4324It is used in the SEV-enabled guest. When encryption is enabled, a guest 4325memory region may contain encrypted data. The SEV memory encryption 4326engine uses a tweak such that two identical plaintext pages, each at 4327different locations will have differing ciphertexts. So swapping or 4328moving ciphertext of those pages will not result in plaintext being 4329swapped. So relocating (or migrating) physical backing pages for the SEV 4330guest will require some additional steps. 4331 4332Note: The current SEV key management spec does not provide commands to 4333swap or migrate (move) ciphertext pages. Hence, for now we pin the guest 4334memory region registered with the ioctl. 4335 43364.112 KVM_MEMORY_ENCRYPT_UNREG_REGION 4337------------------------------------- 4338 4339:Capability: basic 4340:Architectures: x86 4341:Type: system 4342:Parameters: struct kvm_enc_region (in) 4343:Returns: 0 on success; -1 on error 4344 4345This ioctl can be used to unregister the guest memory region registered 4346with KVM_MEMORY_ENCRYPT_REG_REGION ioctl above. 4347 43484.113 KVM_HYPERV_EVENTFD 4349------------------------ 4350 4351:Capability: KVM_CAP_HYPERV_EVENTFD 4352:Architectures: x86 4353:Type: vm ioctl 4354:Parameters: struct kvm_hyperv_eventfd (in) 4355 4356This ioctl (un)registers an eventfd to receive notifications from the guest on 4357the specified Hyper-V connection id through the SIGNAL_EVENT hypercall, without 4358causing a user exit. SIGNAL_EVENT hypercall with non-zero event flag number 4359(bits 24-31) still triggers a KVM_EXIT_HYPERV_HCALL user exit. 4360 4361:: 4362 4363 struct kvm_hyperv_eventfd { 4364 __u32 conn_id; 4365 __s32 fd; 4366 __u32 flags; 4367 __u32 padding[3]; 4368 }; 4369 4370The conn_id field should fit within 24 bits:: 4371 4372 #define KVM_HYPERV_CONN_ID_MASK 0x00ffffff 4373 4374The acceptable values for the flags field are:: 4375 4376 #define KVM_HYPERV_EVENTFD_DEASSIGN (1 << 0) 4377 4378:Returns: 0 on success, 4379 -EINVAL if conn_id or flags is outside the allowed range, 4380 -ENOENT on deassign if the conn_id isn't registered, 4381 -EEXIST on assign if the conn_id is already registered 4382 43834.114 KVM_GET_NESTED_STATE 4384-------------------------- 4385 4386:Capability: KVM_CAP_NESTED_STATE 4387:Architectures: x86 4388:Type: vcpu ioctl 4389:Parameters: struct kvm_nested_state (in/out) 4390:Returns: 0 on success, -1 on error 4391 4392Errors: 4393 4394 ===== ============================================================= 4395 E2BIG the total state size exceeds the value of 'size' specified by 4396 the user; the size required will be written into size. 4397 ===== ============================================================= 4398 4399:: 4400 4401 struct kvm_nested_state { 4402 __u16 flags; 4403 __u16 format; 4404 __u32 size; 4405 4406 union { 4407 struct kvm_vmx_nested_state_hdr vmx; 4408 struct kvm_svm_nested_state_hdr svm; 4409 4410 /* Pad the header to 128 bytes. */ 4411 __u8 pad[120]; 4412 } hdr; 4413 4414 union { 4415 struct kvm_vmx_nested_state_data vmx[0]; 4416 struct kvm_svm_nested_state_data svm[0]; 4417 } data; 4418 }; 4419 4420 #define KVM_STATE_NESTED_GUEST_MODE 0x00000001 4421 #define KVM_STATE_NESTED_RUN_PENDING 0x00000002 4422 #define KVM_STATE_NESTED_EVMCS 0x00000004 4423 4424 #define KVM_STATE_NESTED_FORMAT_VMX 0 4425 #define KVM_STATE_NESTED_FORMAT_SVM 1 4426 4427 #define KVM_STATE_NESTED_VMX_VMCS_SIZE 0x1000 4428 4429 #define KVM_STATE_NESTED_VMX_SMM_GUEST_MODE 0x00000001 4430 #define KVM_STATE_NESTED_VMX_SMM_VMXON 0x00000002 4431 4432 #define KVM_STATE_VMX_PREEMPTION_TIMER_DEADLINE 0x00000001 4433 4434 struct kvm_vmx_nested_state_hdr { 4435 __u64 vmxon_pa; 4436 __u64 vmcs12_pa; 4437 4438 struct { 4439 __u16 flags; 4440 } smm; 4441 4442 __u32 flags; 4443 __u64 preemption_timer_deadline; 4444 }; 4445 4446 struct kvm_vmx_nested_state_data { 4447 __u8 vmcs12[KVM_STATE_NESTED_VMX_VMCS_SIZE]; 4448 __u8 shadow_vmcs12[KVM_STATE_NESTED_VMX_VMCS_SIZE]; 4449 }; 4450 4451This ioctl copies the vcpu's nested virtualization state from the kernel to 4452userspace. 4453 4454The maximum size of the state can be retrieved by passing KVM_CAP_NESTED_STATE 4455to the KVM_CHECK_EXTENSION ioctl(). 4456 44574.115 KVM_SET_NESTED_STATE 4458-------------------------- 4459 4460:Capability: KVM_CAP_NESTED_STATE 4461:Architectures: x86 4462:Type: vcpu ioctl 4463:Parameters: struct kvm_nested_state (in) 4464:Returns: 0 on success, -1 on error 4465 4466This copies the vcpu's kvm_nested_state struct from userspace to the kernel. 4467For the definition of struct kvm_nested_state, see KVM_GET_NESTED_STATE. 4468 44694.116 KVM_(UN)REGISTER_COALESCED_MMIO 4470------------------------------------- 4471 4472:Capability: KVM_CAP_COALESCED_MMIO (for coalesced mmio) 4473 KVM_CAP_COALESCED_PIO (for coalesced pio) 4474:Architectures: all 4475:Type: vm ioctl 4476:Parameters: struct kvm_coalesced_mmio_zone 4477:Returns: 0 on success, < 0 on error 4478 4479Coalesced I/O is a performance optimization that defers hardware 4480register write emulation so that userspace exits are avoided. It is 4481typically used to reduce the overhead of emulating frequently accessed 4482hardware registers. 4483 4484When a hardware register is configured for coalesced I/O, write accesses 4485do not exit to userspace and their value is recorded in a ring buffer 4486that is shared between kernel and userspace. 4487 4488Coalesced I/O is used if one or more write accesses to a hardware 4489register can be deferred until a read or a write to another hardware 4490register on the same device. This last access will cause a vmexit and 4491userspace will process accesses from the ring buffer before emulating 4492it. That will avoid exiting to userspace on repeated writes. 4493 4494Coalesced pio is based on coalesced mmio. There is little difference 4495between coalesced mmio and pio except that coalesced pio records accesses 4496to I/O ports. 4497 44984.117 KVM_CLEAR_DIRTY_LOG (vm ioctl) 4499------------------------------------ 4500 4501:Capability: KVM_CAP_MANUAL_DIRTY_LOG_PROTECT2 4502:Architectures: x86, arm, arm64, mips 4503:Type: vm ioctl 4504:Parameters: struct kvm_clear_dirty_log (in) 4505:Returns: 0 on success, -1 on error 4506 4507:: 4508 4509 /* for KVM_CLEAR_DIRTY_LOG */ 4510 struct kvm_clear_dirty_log { 4511 __u32 slot; 4512 __u32 num_pages; 4513 __u64 first_page; 4514 union { 4515 void __user *dirty_bitmap; /* one bit per page */ 4516 __u64 padding; 4517 }; 4518 }; 4519 4520The ioctl clears the dirty status of pages in a memory slot, according to 4521the bitmap that is passed in struct kvm_clear_dirty_log's dirty_bitmap 4522field. Bit 0 of the bitmap corresponds to page "first_page" in the 4523memory slot, and num_pages is the size in bits of the input bitmap. 4524first_page must be a multiple of 64; num_pages must also be a multiple of 452564 unless first_page + num_pages is the size of the memory slot. For each 4526bit that is set in the input bitmap, the corresponding page is marked "clean" 4527in KVM's dirty bitmap, and dirty tracking is re-enabled for that page 4528(for example via write-protection, or by clearing the dirty bit in 4529a page table entry). 4530 4531If KVM_CAP_MULTI_ADDRESS_SPACE is available, bits 16-31 of slot field specifies 4532the address space for which you want to clear the dirty status. See 4533KVM_SET_USER_MEMORY_REGION for details on the usage of slot field. 4534 4535This ioctl is mostly useful when KVM_CAP_MANUAL_DIRTY_LOG_PROTECT2 4536is enabled; for more information, see the description of the capability. 4537However, it can always be used as long as KVM_CHECK_EXTENSION confirms 4538that KVM_CAP_MANUAL_DIRTY_LOG_PROTECT2 is present. 4539 45404.118 KVM_GET_SUPPORTED_HV_CPUID 4541-------------------------------- 4542 4543:Capability: KVM_CAP_HYPERV_CPUID (vcpu), KVM_CAP_SYS_HYPERV_CPUID (system) 4544:Architectures: x86 4545:Type: system ioctl, vcpu ioctl 4546:Parameters: struct kvm_cpuid2 (in/out) 4547:Returns: 0 on success, -1 on error 4548 4549:: 4550 4551 struct kvm_cpuid2 { 4552 __u32 nent; 4553 __u32 padding; 4554 struct kvm_cpuid_entry2 entries[0]; 4555 }; 4556 4557 struct kvm_cpuid_entry2 { 4558 __u32 function; 4559 __u32 index; 4560 __u32 flags; 4561 __u32 eax; 4562 __u32 ebx; 4563 __u32 ecx; 4564 __u32 edx; 4565 __u32 padding[3]; 4566 }; 4567 4568This ioctl returns x86 cpuid features leaves related to Hyper-V emulation in 4569KVM. Userspace can use the information returned by this ioctl to construct 4570cpuid information presented to guests consuming Hyper-V enlightenments (e.g. 4571Windows or Hyper-V guests). 4572 4573CPUID feature leaves returned by this ioctl are defined by Hyper-V Top Level 4574Functional Specification (TLFS). These leaves can't be obtained with 4575KVM_GET_SUPPORTED_CPUID ioctl because some of them intersect with KVM feature 4576leaves (0x40000000, 0x40000001). 4577 4578Currently, the following list of CPUID leaves are returned: 4579 4580 - HYPERV_CPUID_VENDOR_AND_MAX_FUNCTIONS 4581 - HYPERV_CPUID_INTERFACE 4582 - HYPERV_CPUID_VERSION 4583 - HYPERV_CPUID_FEATURES 4584 - HYPERV_CPUID_ENLIGHTMENT_INFO 4585 - HYPERV_CPUID_IMPLEMENT_LIMITS 4586 - HYPERV_CPUID_NESTED_FEATURES 4587 - HYPERV_CPUID_SYNDBG_VENDOR_AND_MAX_FUNCTIONS 4588 - HYPERV_CPUID_SYNDBG_INTERFACE 4589 - HYPERV_CPUID_SYNDBG_PLATFORM_CAPABILITIES 4590 4591Userspace invokes KVM_GET_SUPPORTED_HV_CPUID by passing a kvm_cpuid2 structure 4592with the 'nent' field indicating the number of entries in the variable-size 4593array 'entries'. If the number of entries is too low to describe all Hyper-V 4594feature leaves, an error (E2BIG) is returned. If the number is more or equal 4595to the number of Hyper-V feature leaves, the 'nent' field is adjusted to the 4596number of valid entries in the 'entries' array, which is then filled. 4597 4598'index' and 'flags' fields in 'struct kvm_cpuid_entry2' are currently reserved, 4599userspace should not expect to get any particular value there. 4600 4601Note, vcpu version of KVM_GET_SUPPORTED_HV_CPUID is currently deprecated. Unlike 4602system ioctl which exposes all supported feature bits unconditionally, vcpu 4603version has the following quirks: 4604 4605- HYPERV_CPUID_NESTED_FEATURES leaf and HV_X64_ENLIGHTENED_VMCS_RECOMMENDED 4606 feature bit are only exposed when Enlightened VMCS was previously enabled 4607 on the corresponding vCPU (KVM_CAP_HYPERV_ENLIGHTENED_VMCS). 4608- HV_STIMER_DIRECT_MODE_AVAILABLE bit is only exposed with in-kernel LAPIC. 4609 (presumes KVM_CREATE_IRQCHIP has already been called). 4610 46114.119 KVM_ARM_VCPU_FINALIZE 4612--------------------------- 4613 4614:Architectures: arm, arm64 4615:Type: vcpu ioctl 4616:Parameters: int feature (in) 4617:Returns: 0 on success, -1 on error 4618 4619Errors: 4620 4621 ====== ============================================================== 4622 EPERM feature not enabled, needs configuration, or already finalized 4623 EINVAL feature unknown or not present 4624 ====== ============================================================== 4625 4626Recognised values for feature: 4627 4628 ===== =========================================== 4629 arm64 KVM_ARM_VCPU_SVE (requires KVM_CAP_ARM_SVE) 4630 ===== =========================================== 4631 4632Finalizes the configuration of the specified vcpu feature. 4633 4634The vcpu must already have been initialised, enabling the affected feature, by 4635means of a successful KVM_ARM_VCPU_INIT call with the appropriate flag set in 4636features[]. 4637 4638For affected vcpu features, this is a mandatory step that must be performed 4639before the vcpu is fully usable. 4640 4641Between KVM_ARM_VCPU_INIT and KVM_ARM_VCPU_FINALIZE, the feature may be 4642configured by use of ioctls such as KVM_SET_ONE_REG. The exact configuration 4643that should be performaned and how to do it are feature-dependent. 4644 4645Other calls that depend on a particular feature being finalized, such as 4646KVM_RUN, KVM_GET_REG_LIST, KVM_GET_ONE_REG and KVM_SET_ONE_REG, will fail with 4647-EPERM unless the feature has already been finalized by means of a 4648KVM_ARM_VCPU_FINALIZE call. 4649 4650See KVM_ARM_VCPU_INIT for details of vcpu features that require finalization 4651using this ioctl. 4652 46534.120 KVM_SET_PMU_EVENT_FILTER 4654------------------------------ 4655 4656:Capability: KVM_CAP_PMU_EVENT_FILTER 4657:Architectures: x86 4658:Type: vm ioctl 4659:Parameters: struct kvm_pmu_event_filter (in) 4660:Returns: 0 on success, -1 on error 4661 4662:: 4663 4664 struct kvm_pmu_event_filter { 4665 __u32 action; 4666 __u32 nevents; 4667 __u32 fixed_counter_bitmap; 4668 __u32 flags; 4669 __u32 pad[4]; 4670 __u64 events[0]; 4671 }; 4672 4673This ioctl restricts the set of PMU events that the guest can program. 4674The argument holds a list of events which will be allowed or denied. 4675The eventsel+umask of each event the guest attempts to program is compared 4676against the events field to determine whether the guest should have access. 4677The events field only controls general purpose counters; fixed purpose 4678counters are controlled by the fixed_counter_bitmap. 4679 4680No flags are defined yet, the field must be zero. 4681 4682Valid values for 'action':: 4683 4684 #define KVM_PMU_EVENT_ALLOW 0 4685 #define KVM_PMU_EVENT_DENY 1 4686 46874.121 KVM_PPC_SVM_OFF 4688--------------------- 4689 4690:Capability: basic 4691:Architectures: powerpc 4692:Type: vm ioctl 4693:Parameters: none 4694:Returns: 0 on successful completion, 4695 4696Errors: 4697 4698 ====== ================================================================ 4699 EINVAL if ultravisor failed to terminate the secure guest 4700 ENOMEM if hypervisor failed to allocate new radix page tables for guest 4701 ====== ================================================================ 4702 4703This ioctl is used to turn off the secure mode of the guest or transition 4704the guest from secure mode to normal mode. This is invoked when the guest 4705is reset. This has no effect if called for a normal guest. 4706 4707This ioctl issues an ultravisor call to terminate the secure guest, 4708unpins the VPA pages and releases all the device pages that are used to 4709track the secure pages by hypervisor. 4710 47114.122 KVM_S390_NORMAL_RESET 4712--------------------------- 4713 4714:Capability: KVM_CAP_S390_VCPU_RESETS 4715:Architectures: s390 4716:Type: vcpu ioctl 4717:Parameters: none 4718:Returns: 0 4719 4720This ioctl resets VCPU registers and control structures according to 4721the cpu reset definition in the POP (Principles Of Operation). 4722 47234.123 KVM_S390_INITIAL_RESET 4724---------------------------- 4725 4726:Capability: none 4727:Architectures: s390 4728:Type: vcpu ioctl 4729:Parameters: none 4730:Returns: 0 4731 4732This ioctl resets VCPU registers and control structures according to 4733the initial cpu reset definition in the POP. However, the cpu is not 4734put into ESA mode. This reset is a superset of the normal reset. 4735 47364.124 KVM_S390_CLEAR_RESET 4737-------------------------- 4738 4739:Capability: KVM_CAP_S390_VCPU_RESETS 4740:Architectures: s390 4741:Type: vcpu ioctl 4742:Parameters: none 4743:Returns: 0 4744 4745This ioctl resets VCPU registers and control structures according to 4746the clear cpu reset definition in the POP. However, the cpu is not put 4747into ESA mode. This reset is a superset of the initial reset. 4748 4749 47504.125 KVM_S390_PV_COMMAND 4751------------------------- 4752 4753:Capability: KVM_CAP_S390_PROTECTED 4754:Architectures: s390 4755:Type: vm ioctl 4756:Parameters: struct kvm_pv_cmd 4757:Returns: 0 on success, < 0 on error 4758 4759:: 4760 4761 struct kvm_pv_cmd { 4762 __u32 cmd; /* Command to be executed */ 4763 __u16 rc; /* Ultravisor return code */ 4764 __u16 rrc; /* Ultravisor return reason code */ 4765 __u64 data; /* Data or address */ 4766 __u32 flags; /* flags for future extensions. Must be 0 for now */ 4767 __u32 reserved[3]; 4768 }; 4769 4770cmd values: 4771 4772KVM_PV_ENABLE 4773 Allocate memory and register the VM with the Ultravisor, thereby 4774 donating memory to the Ultravisor that will become inaccessible to 4775 KVM. All existing CPUs are converted to protected ones. After this 4776 command has succeeded, any CPU added via hotplug will become 4777 protected during its creation as well. 4778 4779 Errors: 4780 4781 ===== ============================= 4782 EINTR an unmasked signal is pending 4783 ===== ============================= 4784 4785KVM_PV_DISABLE 4786 4787 Deregister the VM from the Ultravisor and reclaim the memory that 4788 had been donated to the Ultravisor, making it usable by the kernel 4789 again. All registered VCPUs are converted back to non-protected 4790 ones. 4791 4792KVM_PV_VM_SET_SEC_PARMS 4793 Pass the image header from VM memory to the Ultravisor in 4794 preparation of image unpacking and verification. 4795 4796KVM_PV_VM_UNPACK 4797 Unpack (protect and decrypt) a page of the encrypted boot image. 4798 4799KVM_PV_VM_VERIFY 4800 Verify the integrity of the unpacked image. Only if this succeeds, 4801 KVM is allowed to start protected VCPUs. 4802 48034.126 KVM_X86_SET_MSR_FILTER 4804---------------------------- 4805 4806:Capability: KVM_X86_SET_MSR_FILTER 4807:Architectures: x86 4808:Type: vm ioctl 4809:Parameters: struct kvm_msr_filter 4810:Returns: 0 on success, < 0 on error 4811 4812:: 4813 4814 struct kvm_msr_filter_range { 4815 #define KVM_MSR_FILTER_READ (1 << 0) 4816 #define KVM_MSR_FILTER_WRITE (1 << 1) 4817 __u32 flags; 4818 __u32 nmsrs; /* number of msrs in bitmap */ 4819 __u32 base; /* MSR index the bitmap starts at */ 4820 __u8 *bitmap; /* a 1 bit allows the operations in flags, 0 denies */ 4821 }; 4822 4823 #define KVM_MSR_FILTER_MAX_RANGES 16 4824 struct kvm_msr_filter { 4825 #define KVM_MSR_FILTER_DEFAULT_ALLOW (0 << 0) 4826 #define KVM_MSR_FILTER_DEFAULT_DENY (1 << 0) 4827 __u32 flags; 4828 struct kvm_msr_filter_range ranges[KVM_MSR_FILTER_MAX_RANGES]; 4829 }; 4830 4831flags values for ``struct kvm_msr_filter_range``: 4832 4833``KVM_MSR_FILTER_READ`` 4834 4835 Filter read accesses to MSRs using the given bitmap. A 0 in the bitmap 4836 indicates that a read should immediately fail, while a 1 indicates that 4837 a read for a particular MSR should be handled regardless of the default 4838 filter action. 4839 4840``KVM_MSR_FILTER_WRITE`` 4841 4842 Filter write accesses to MSRs using the given bitmap. A 0 in the bitmap 4843 indicates that a write should immediately fail, while a 1 indicates that 4844 a write for a particular MSR should be handled regardless of the default 4845 filter action. 4846 4847``KVM_MSR_FILTER_READ | KVM_MSR_FILTER_WRITE`` 4848 4849 Filter both read and write accesses to MSRs using the given bitmap. A 0 4850 in the bitmap indicates that both reads and writes should immediately fail, 4851 while a 1 indicates that reads and writes for a particular MSR are not 4852 filtered by this range. 4853 4854flags values for ``struct kvm_msr_filter``: 4855 4856``KVM_MSR_FILTER_DEFAULT_ALLOW`` 4857 4858 If no filter range matches an MSR index that is getting accessed, KVM will 4859 fall back to allowing access to the MSR. 4860 4861``KVM_MSR_FILTER_DEFAULT_DENY`` 4862 4863 If no filter range matches an MSR index that is getting accessed, KVM will 4864 fall back to rejecting access to the MSR. In this mode, all MSRs that should 4865 be processed by KVM need to explicitly be marked as allowed in the bitmaps. 4866 4867This ioctl allows user space to define up to 16 bitmaps of MSR ranges to 4868specify whether a certain MSR access should be explicitly filtered for or not. 4869 4870If this ioctl has never been invoked, MSR accesses are not guarded and the 4871default KVM in-kernel emulation behavior is fully preserved. 4872 4873Calling this ioctl with an empty set of ranges (all nmsrs == 0) disables MSR 4874filtering. In that mode, ``KVM_MSR_FILTER_DEFAULT_DENY`` is invalid and causes 4875an error. 4876 4877As soon as the filtering is in place, every MSR access is processed through 4878the filtering except for accesses to the x2APIC MSRs (from 0x800 to 0x8ff); 4879x2APIC MSRs are always allowed, independent of the ``default_allow`` setting, 4880and their behavior depends on the ``X2APIC_ENABLE`` bit of the APIC base 4881register. 4882 4883If a bit is within one of the defined ranges, read and write accesses are 4884guarded by the bitmap's value for the MSR index if the kind of access 4885is included in the ``struct kvm_msr_filter_range`` flags. If no range 4886cover this particular access, the behavior is determined by the flags 4887field in the kvm_msr_filter struct: ``KVM_MSR_FILTER_DEFAULT_ALLOW`` 4888and ``KVM_MSR_FILTER_DEFAULT_DENY``. 4889 4890Each bitmap range specifies a range of MSRs to potentially allow access on. 4891The range goes from MSR index [base .. base+nmsrs]. The flags field 4892indicates whether reads, writes or both reads and writes are filtered 4893by setting a 1 bit in the bitmap for the corresponding MSR index. 4894 4895If an MSR access is not permitted through the filtering, it generates a 4896#GP inside the guest. When combined with KVM_CAP_X86_USER_SPACE_MSR, that 4897allows user space to deflect and potentially handle various MSR accesses 4898into user space. 4899 4900Note, invoking this ioctl with a vCPU is running is inherently racy. However, 4901KVM does guarantee that vCPUs will see either the previous filter or the new 4902filter, e.g. MSRs with identical settings in both the old and new filter will 4903have deterministic behavior. 4904 49054.127 KVM_XEN_HVM_SET_ATTR 4906-------------------------- 4907 4908:Capability: KVM_CAP_XEN_HVM / KVM_XEN_HVM_CONFIG_SHARED_INFO 4909:Architectures: x86 4910:Type: vm ioctl 4911:Parameters: struct kvm_xen_hvm_attr 4912:Returns: 0 on success, < 0 on error 4913 4914:: 4915 4916 struct kvm_xen_hvm_attr { 4917 __u16 type; 4918 __u16 pad[3]; 4919 union { 4920 __u8 long_mode; 4921 __u8 vector; 4922 struct { 4923 __u64 gfn; 4924 } shared_info; 4925 __u64 pad[4]; 4926 } u; 4927 }; 4928 4929type values: 4930 4931KVM_XEN_ATTR_TYPE_LONG_MODE 4932 Sets the ABI mode of the VM to 32-bit or 64-bit (long mode). This 4933 determines the layout of the shared info pages exposed to the VM. 4934 4935KVM_XEN_ATTR_TYPE_SHARED_INFO 4936 Sets the guest physical frame number at which the Xen "shared info" 4937 page resides. Note that although Xen places vcpu_info for the first 4938 32 vCPUs in the shared_info page, KVM does not automatically do so 4939 and instead requires that KVM_XEN_VCPU_ATTR_TYPE_VCPU_INFO be used 4940 explicitly even when the vcpu_info for a given vCPU resides at the 4941 "default" location in the shared_info page. This is because KVM is 4942 not aware of the Xen CPU id which is used as the index into the 4943 vcpu_info[] array, so cannot know the correct default location. 4944 4945KVM_XEN_ATTR_TYPE_UPCALL_VECTOR 4946 Sets the exception vector used to deliver Xen event channel upcalls. 4947 49484.127 KVM_XEN_HVM_GET_ATTR 4949-------------------------- 4950 4951:Capability: KVM_CAP_XEN_HVM / KVM_XEN_HVM_CONFIG_SHARED_INFO 4952:Architectures: x86 4953:Type: vm ioctl 4954:Parameters: struct kvm_xen_hvm_attr 4955:Returns: 0 on success, < 0 on error 4956 4957Allows Xen VM attributes to be read. For the structure and types, 4958see KVM_XEN_HVM_SET_ATTR above. 4959 49604.128 KVM_XEN_VCPU_SET_ATTR 4961--------------------------- 4962 4963:Capability: KVM_CAP_XEN_HVM / KVM_XEN_HVM_CONFIG_SHARED_INFO 4964:Architectures: x86 4965:Type: vcpu ioctl 4966:Parameters: struct kvm_xen_vcpu_attr 4967:Returns: 0 on success, < 0 on error 4968 4969:: 4970 4971 struct kvm_xen_vcpu_attr { 4972 __u16 type; 4973 __u16 pad[3]; 4974 union { 4975 __u64 gpa; 4976 __u64 pad[4]; 4977 struct { 4978 __u64 state; 4979 __u64 state_entry_time; 4980 __u64 time_running; 4981 __u64 time_runnable; 4982 __u64 time_blocked; 4983 __u64 time_offline; 4984 } runstate; 4985 } u; 4986 }; 4987 4988type values: 4989 4990KVM_XEN_VCPU_ATTR_TYPE_VCPU_INFO 4991 Sets the guest physical address of the vcpu_info for a given vCPU. 4992 4993KVM_XEN_VCPU_ATTR_TYPE_VCPU_TIME_INFO 4994 Sets the guest physical address of an additional pvclock structure 4995 for a given vCPU. This is typically used for guest vsyscall support. 4996 4997KVM_XEN_VCPU_ATTR_TYPE_RUNSTATE_ADDR 4998 Sets the guest physical address of the vcpu_runstate_info for a given 4999 vCPU. This is how a Xen guest tracks CPU state such as steal time. 5000 5001KVM_XEN_VCPU_ATTR_TYPE_RUNSTATE_CURRENT 5002 Sets the runstate (RUNSTATE_running/_runnable/_blocked/_offline) of 5003 the given vCPU from the .u.runstate.state member of the structure. 5004 KVM automatically accounts running and runnable time but blocked 5005 and offline states are only entered explicitly. 5006 5007KVM_XEN_VCPU_ATTR_TYPE_RUNSTATE_DATA 5008 Sets all fields of the vCPU runstate data from the .u.runstate member 5009 of the structure, including the current runstate. The state_entry_time 5010 must equal the sum of the other four times. 5011 5012KVM_XEN_VCPU_ATTR_TYPE_RUNSTATE_ADJUST 5013 This *adds* the contents of the .u.runstate members of the structure 5014 to the corresponding members of the given vCPU's runstate data, thus 5015 permitting atomic adjustments to the runstate times. The adjustment 5016 to the state_entry_time must equal the sum of the adjustments to the 5017 other four times. The state field must be set to -1, or to a valid 5018 runstate value (RUNSTATE_running, RUNSTATE_runnable, RUNSTATE_blocked 5019 or RUNSTATE_offline) to set the current accounted state as of the 5020 adjusted state_entry_time. 5021 50224.129 KVM_XEN_VCPU_GET_ATTR 5023--------------------------- 5024 5025:Capability: KVM_CAP_XEN_HVM / KVM_XEN_HVM_CONFIG_SHARED_INFO 5026:Architectures: x86 5027:Type: vcpu ioctl 5028:Parameters: struct kvm_xen_vcpu_attr 5029:Returns: 0 on success, < 0 on error 5030 5031Allows Xen vCPU attributes to be read. For the structure and types, 5032see KVM_XEN_VCPU_SET_ATTR above. 5033 5034The KVM_XEN_VCPU_ATTR_TYPE_RUNSTATE_ADJUST type may not be used 5035with the KVM_XEN_VCPU_GET_ATTR ioctl. 5036 50375. The kvm_run structure 5038======================== 5039 5040Application code obtains a pointer to the kvm_run structure by 5041mmap()ing a vcpu fd. From that point, application code can control 5042execution by changing fields in kvm_run prior to calling the KVM_RUN 5043ioctl, and obtain information about the reason KVM_RUN returned by 5044looking up structure members. 5045 5046:: 5047 5048 struct kvm_run { 5049 /* in */ 5050 __u8 request_interrupt_window; 5051 5052Request that KVM_RUN return when it becomes possible to inject external 5053interrupts into the guest. Useful in conjunction with KVM_INTERRUPT. 5054 5055:: 5056 5057 __u8 immediate_exit; 5058 5059This field is polled once when KVM_RUN starts; if non-zero, KVM_RUN 5060exits immediately, returning -EINTR. In the common scenario where a 5061signal is used to "kick" a VCPU out of KVM_RUN, this field can be used 5062to avoid usage of KVM_SET_SIGNAL_MASK, which has worse scalability. 5063Rather than blocking the signal outside KVM_RUN, userspace can set up 5064a signal handler that sets run->immediate_exit to a non-zero value. 5065 5066This field is ignored if KVM_CAP_IMMEDIATE_EXIT is not available. 5067 5068:: 5069 5070 __u8 padding1[6]; 5071 5072 /* out */ 5073 __u32 exit_reason; 5074 5075When KVM_RUN has returned successfully (return value 0), this informs 5076application code why KVM_RUN has returned. Allowable values for this 5077field are detailed below. 5078 5079:: 5080 5081 __u8 ready_for_interrupt_injection; 5082 5083If request_interrupt_window has been specified, this field indicates 5084an interrupt can be injected now with KVM_INTERRUPT. 5085 5086:: 5087 5088 __u8 if_flag; 5089 5090The value of the current interrupt flag. Only valid if in-kernel 5091local APIC is not used. 5092 5093:: 5094 5095 __u16 flags; 5096 5097More architecture-specific flags detailing state of the VCPU that may 5098affect the device's behavior. Current defined flags:: 5099 5100 /* x86, set if the VCPU is in system management mode */ 5101 #define KVM_RUN_X86_SMM (1 << 0) 5102 /* x86, set if bus lock detected in VM */ 5103 #define KVM_RUN_BUS_LOCK (1 << 1) 5104 5105:: 5106 5107 /* in (pre_kvm_run), out (post_kvm_run) */ 5108 __u64 cr8; 5109 5110The value of the cr8 register. Only valid if in-kernel local APIC is 5111not used. Both input and output. 5112 5113:: 5114 5115 __u64 apic_base; 5116 5117The value of the APIC BASE msr. Only valid if in-kernel local 5118APIC is not used. Both input and output. 5119 5120:: 5121 5122 union { 5123 /* KVM_EXIT_UNKNOWN */ 5124 struct { 5125 __u64 hardware_exit_reason; 5126 } hw; 5127 5128If exit_reason is KVM_EXIT_UNKNOWN, the vcpu has exited due to unknown 5129reasons. Further architecture-specific information is available in 5130hardware_exit_reason. 5131 5132:: 5133 5134 /* KVM_EXIT_FAIL_ENTRY */ 5135 struct { 5136 __u64 hardware_entry_failure_reason; 5137 __u32 cpu; /* if KVM_LAST_CPU */ 5138 } fail_entry; 5139 5140If exit_reason is KVM_EXIT_FAIL_ENTRY, the vcpu could not be run due 5141to unknown reasons. Further architecture-specific information is 5142available in hardware_entry_failure_reason. 5143 5144:: 5145 5146 /* KVM_EXIT_EXCEPTION */ 5147 struct { 5148 __u32 exception; 5149 __u32 error_code; 5150 } ex; 5151 5152Unused. 5153 5154:: 5155 5156 /* KVM_EXIT_IO */ 5157 struct { 5158 #define KVM_EXIT_IO_IN 0 5159 #define KVM_EXIT_IO_OUT 1 5160 __u8 direction; 5161 __u8 size; /* bytes */ 5162 __u16 port; 5163 __u32 count; 5164 __u64 data_offset; /* relative to kvm_run start */ 5165 } io; 5166 5167If exit_reason is KVM_EXIT_IO, then the vcpu has 5168executed a port I/O instruction which could not be satisfied by kvm. 5169data_offset describes where the data is located (KVM_EXIT_IO_OUT) or 5170where kvm expects application code to place the data for the next 5171KVM_RUN invocation (KVM_EXIT_IO_IN). Data format is a packed array. 5172 5173:: 5174 5175 /* KVM_EXIT_DEBUG */ 5176 struct { 5177 struct kvm_debug_exit_arch arch; 5178 } debug; 5179 5180If the exit_reason is KVM_EXIT_DEBUG, then a vcpu is processing a debug event 5181for which architecture specific information is returned. 5182 5183:: 5184 5185 /* KVM_EXIT_MMIO */ 5186 struct { 5187 __u64 phys_addr; 5188 __u8 data[8]; 5189 __u32 len; 5190 __u8 is_write; 5191 } mmio; 5192 5193If exit_reason is KVM_EXIT_MMIO, then the vcpu has 5194executed a memory-mapped I/O instruction which could not be satisfied 5195by kvm. The 'data' member contains the written data if 'is_write' is 5196true, and should be filled by application code otherwise. 5197 5198The 'data' member contains, in its first 'len' bytes, the value as it would 5199appear if the VCPU performed a load or store of the appropriate width directly 5200to the byte array. 5201 5202.. note:: 5203 5204 For KVM_EXIT_IO, KVM_EXIT_MMIO, KVM_EXIT_OSI, KVM_EXIT_PAPR, KVM_EXIT_XEN, 5205 KVM_EXIT_EPR, KVM_EXIT_X86_RDMSR and KVM_EXIT_X86_WRMSR the corresponding 5206 operations are complete (and guest state is consistent) only after userspace 5207 has re-entered the kernel with KVM_RUN. The kernel side will first finish 5208 incomplete operations and then check for pending signals. 5209 5210 The pending state of the operation is not preserved in state which is 5211 visible to userspace, thus userspace should ensure that the operation is 5212 completed before performing a live migration. Userspace can re-enter the 5213 guest with an unmasked signal pending or with the immediate_exit field set 5214 to complete pending operations without allowing any further instructions 5215 to be executed. 5216 5217:: 5218 5219 /* KVM_EXIT_HYPERCALL */ 5220 struct { 5221 __u64 nr; 5222 __u64 args[6]; 5223 __u64 ret; 5224 __u32 longmode; 5225 __u32 pad; 5226 } hypercall; 5227 5228Unused. This was once used for 'hypercall to userspace'. To implement 5229such functionality, use KVM_EXIT_IO (x86) or KVM_EXIT_MMIO (all except s390). 5230 5231.. note:: KVM_EXIT_IO is significantly faster than KVM_EXIT_MMIO. 5232 5233:: 5234 5235 /* KVM_EXIT_TPR_ACCESS */ 5236 struct { 5237 __u64 rip; 5238 __u32 is_write; 5239 __u32 pad; 5240 } tpr_access; 5241 5242To be documented (KVM_TPR_ACCESS_REPORTING). 5243 5244:: 5245 5246 /* KVM_EXIT_S390_SIEIC */ 5247 struct { 5248 __u8 icptcode; 5249 __u64 mask; /* psw upper half */ 5250 __u64 addr; /* psw lower half */ 5251 __u16 ipa; 5252 __u32 ipb; 5253 } s390_sieic; 5254 5255s390 specific. 5256 5257:: 5258 5259 /* KVM_EXIT_S390_RESET */ 5260 #define KVM_S390_RESET_POR 1 5261 #define KVM_S390_RESET_CLEAR 2 5262 #define KVM_S390_RESET_SUBSYSTEM 4 5263 #define KVM_S390_RESET_CPU_INIT 8 5264 #define KVM_S390_RESET_IPL 16 5265 __u64 s390_reset_flags; 5266 5267s390 specific. 5268 5269:: 5270 5271 /* KVM_EXIT_S390_UCONTROL */ 5272 struct { 5273 __u64 trans_exc_code; 5274 __u32 pgm_code; 5275 } s390_ucontrol; 5276 5277s390 specific. A page fault has occurred for a user controlled virtual 5278machine (KVM_VM_S390_UNCONTROL) on it's host page table that cannot be 5279resolved by the kernel. 5280The program code and the translation exception code that were placed 5281in the cpu's lowcore are presented here as defined by the z Architecture 5282Principles of Operation Book in the Chapter for Dynamic Address Translation 5283(DAT) 5284 5285:: 5286 5287 /* KVM_EXIT_DCR */ 5288 struct { 5289 __u32 dcrn; 5290 __u32 data; 5291 __u8 is_write; 5292 } dcr; 5293 5294Deprecated - was used for 440 KVM. 5295 5296:: 5297 5298 /* KVM_EXIT_OSI */ 5299 struct { 5300 __u64 gprs[32]; 5301 } osi; 5302 5303MOL uses a special hypercall interface it calls 'OSI'. To enable it, we catch 5304hypercalls and exit with this exit struct that contains all the guest gprs. 5305 5306If exit_reason is KVM_EXIT_OSI, then the vcpu has triggered such a hypercall. 5307Userspace can now handle the hypercall and when it's done modify the gprs as 5308necessary. Upon guest entry all guest GPRs will then be replaced by the values 5309in this struct. 5310 5311:: 5312 5313 /* KVM_EXIT_PAPR_HCALL */ 5314 struct { 5315 __u64 nr; 5316 __u64 ret; 5317 __u64 args[9]; 5318 } papr_hcall; 5319 5320This is used on 64-bit PowerPC when emulating a pSeries partition, 5321e.g. with the 'pseries' machine type in qemu. It occurs when the 5322guest does a hypercall using the 'sc 1' instruction. The 'nr' field 5323contains the hypercall number (from the guest R3), and 'args' contains 5324the arguments (from the guest R4 - R12). Userspace should put the 5325return code in 'ret' and any extra returned values in args[]. 5326The possible hypercalls are defined in the Power Architecture Platform 5327Requirements (PAPR) document available from www.power.org (free 5328developer registration required to access it). 5329 5330:: 5331 5332 /* KVM_EXIT_S390_TSCH */ 5333 struct { 5334 __u16 subchannel_id; 5335 __u16 subchannel_nr; 5336 __u32 io_int_parm; 5337 __u32 io_int_word; 5338 __u32 ipb; 5339 __u8 dequeued; 5340 } s390_tsch; 5341 5342s390 specific. This exit occurs when KVM_CAP_S390_CSS_SUPPORT has been enabled 5343and TEST SUBCHANNEL was intercepted. If dequeued is set, a pending I/O 5344interrupt for the target subchannel has been dequeued and subchannel_id, 5345subchannel_nr, io_int_parm and io_int_word contain the parameters for that 5346interrupt. ipb is needed for instruction parameter decoding. 5347 5348:: 5349 5350 /* KVM_EXIT_EPR */ 5351 struct { 5352 __u32 epr; 5353 } epr; 5354 5355On FSL BookE PowerPC chips, the interrupt controller has a fast patch 5356interrupt acknowledge path to the core. When the core successfully 5357delivers an interrupt, it automatically populates the EPR register with 5358the interrupt vector number and acknowledges the interrupt inside 5359the interrupt controller. 5360 5361In case the interrupt controller lives in user space, we need to do 5362the interrupt acknowledge cycle through it to fetch the next to be 5363delivered interrupt vector using this exit. 5364 5365It gets triggered whenever both KVM_CAP_PPC_EPR are enabled and an 5366external interrupt has just been delivered into the guest. User space 5367should put the acknowledged interrupt vector into the 'epr' field. 5368 5369:: 5370 5371 /* KVM_EXIT_SYSTEM_EVENT */ 5372 struct { 5373 #define KVM_SYSTEM_EVENT_SHUTDOWN 1 5374 #define KVM_SYSTEM_EVENT_RESET 2 5375 #define KVM_SYSTEM_EVENT_CRASH 3 5376 __u32 type; 5377 __u64 flags; 5378 } system_event; 5379 5380If exit_reason is KVM_EXIT_SYSTEM_EVENT then the vcpu has triggered 5381a system-level event using some architecture specific mechanism (hypercall 5382or some special instruction). In case of ARM/ARM64, this is triggered using 5383HVC instruction based PSCI call from the vcpu. The 'type' field describes 5384the system-level event type. The 'flags' field describes architecture 5385specific flags for the system-level event. 5386 5387Valid values for 'type' are: 5388 5389 - KVM_SYSTEM_EVENT_SHUTDOWN -- the guest has requested a shutdown of the 5390 VM. Userspace is not obliged to honour this, and if it does honour 5391 this does not need to destroy the VM synchronously (ie it may call 5392 KVM_RUN again before shutdown finally occurs). 5393 - KVM_SYSTEM_EVENT_RESET -- the guest has requested a reset of the VM. 5394 As with SHUTDOWN, userspace can choose to ignore the request, or 5395 to schedule the reset to occur in the future and may call KVM_RUN again. 5396 - KVM_SYSTEM_EVENT_CRASH -- the guest crash occurred and the guest 5397 has requested a crash condition maintenance. Userspace can choose 5398 to ignore the request, or to gather VM memory core dump and/or 5399 reset/shutdown of the VM. 5400 5401:: 5402 5403 /* KVM_EXIT_IOAPIC_EOI */ 5404 struct { 5405 __u8 vector; 5406 } eoi; 5407 5408Indicates that the VCPU's in-kernel local APIC received an EOI for a 5409level-triggered IOAPIC interrupt. This exit only triggers when the 5410IOAPIC is implemented in userspace (i.e. KVM_CAP_SPLIT_IRQCHIP is enabled); 5411the userspace IOAPIC should process the EOI and retrigger the interrupt if 5412it is still asserted. Vector is the LAPIC interrupt vector for which the 5413EOI was received. 5414 5415:: 5416 5417 struct kvm_hyperv_exit { 5418 #define KVM_EXIT_HYPERV_SYNIC 1 5419 #define KVM_EXIT_HYPERV_HCALL 2 5420 #define KVM_EXIT_HYPERV_SYNDBG 3 5421 __u32 type; 5422 __u32 pad1; 5423 union { 5424 struct { 5425 __u32 msr; 5426 __u32 pad2; 5427 __u64 control; 5428 __u64 evt_page; 5429 __u64 msg_page; 5430 } synic; 5431 struct { 5432 __u64 input; 5433 __u64 result; 5434 __u64 params[2]; 5435 } hcall; 5436 struct { 5437 __u32 msr; 5438 __u32 pad2; 5439 __u64 control; 5440 __u64 status; 5441 __u64 send_page; 5442 __u64 recv_page; 5443 __u64 pending_page; 5444 } syndbg; 5445 } u; 5446 }; 5447 /* KVM_EXIT_HYPERV */ 5448 struct kvm_hyperv_exit hyperv; 5449 5450Indicates that the VCPU exits into userspace to process some tasks 5451related to Hyper-V emulation. 5452 5453Valid values for 'type' are: 5454 5455 - KVM_EXIT_HYPERV_SYNIC -- synchronously notify user-space about 5456 5457Hyper-V SynIC state change. Notification is used to remap SynIC 5458event/message pages and to enable/disable SynIC messages/events processing 5459in userspace. 5460 5461 - KVM_EXIT_HYPERV_SYNDBG -- synchronously notify user-space about 5462 5463Hyper-V Synthetic debugger state change. Notification is used to either update 5464the pending_page location or to send a control command (send the buffer located 5465in send_page or recv a buffer to recv_page). 5466 5467:: 5468 5469 /* KVM_EXIT_ARM_NISV */ 5470 struct { 5471 __u64 esr_iss; 5472 __u64 fault_ipa; 5473 } arm_nisv; 5474 5475Used on arm and arm64 systems. If a guest accesses memory not in a memslot, 5476KVM will typically return to userspace and ask it to do MMIO emulation on its 5477behalf. However, for certain classes of instructions, no instruction decode 5478(direction, length of memory access) is provided, and fetching and decoding 5479the instruction from the VM is overly complicated to live in the kernel. 5480 5481Historically, when this situation occurred, KVM would print a warning and kill 5482the VM. KVM assumed that if the guest accessed non-memslot memory, it was 5483trying to do I/O, which just couldn't be emulated, and the warning message was 5484phrased accordingly. However, what happened more often was that a guest bug 5485caused access outside the guest memory areas which should lead to a more 5486meaningful warning message and an external abort in the guest, if the access 5487did not fall within an I/O window. 5488 5489Userspace implementations can query for KVM_CAP_ARM_NISV_TO_USER, and enable 5490this capability at VM creation. Once this is done, these types of errors will 5491instead return to userspace with KVM_EXIT_ARM_NISV, with the valid bits from 5492the HSR (arm) and ESR_EL2 (arm64) in the esr_iss field, and the faulting IPA 5493in the fault_ipa field. Userspace can either fix up the access if it's 5494actually an I/O access by decoding the instruction from guest memory (if it's 5495very brave) and continue executing the guest, or it can decide to suspend, 5496dump, or restart the guest. 5497 5498Note that KVM does not skip the faulting instruction as it does for 5499KVM_EXIT_MMIO, but userspace has to emulate any change to the processing state 5500if it decides to decode and emulate the instruction. 5501 5502:: 5503 5504 /* KVM_EXIT_X86_RDMSR / KVM_EXIT_X86_WRMSR */ 5505 struct { 5506 __u8 error; /* user -> kernel */ 5507 __u8 pad[7]; 5508 __u32 reason; /* kernel -> user */ 5509 __u32 index; /* kernel -> user */ 5510 __u64 data; /* kernel <-> user */ 5511 } msr; 5512 5513Used on x86 systems. When the VM capability KVM_CAP_X86_USER_SPACE_MSR is 5514enabled, MSR accesses to registers that would invoke a #GP by KVM kernel code 5515will instead trigger a KVM_EXIT_X86_RDMSR exit for reads and KVM_EXIT_X86_WRMSR 5516exit for writes. 5517 5518The "reason" field specifies why the MSR trap occurred. User space will only 5519receive MSR exit traps when a particular reason was requested during through 5520ENABLE_CAP. Currently valid exit reasons are: 5521 5522 KVM_MSR_EXIT_REASON_UNKNOWN - access to MSR that is unknown to KVM 5523 KVM_MSR_EXIT_REASON_INVAL - access to invalid MSRs or reserved bits 5524 KVM_MSR_EXIT_REASON_FILTER - access blocked by KVM_X86_SET_MSR_FILTER 5525 5526For KVM_EXIT_X86_RDMSR, the "index" field tells user space which MSR the guest 5527wants to read. To respond to this request with a successful read, user space 5528writes the respective data into the "data" field and must continue guest 5529execution to ensure the read data is transferred into guest register state. 5530 5531If the RDMSR request was unsuccessful, user space indicates that with a "1" in 5532the "error" field. This will inject a #GP into the guest when the VCPU is 5533executed again. 5534 5535For KVM_EXIT_X86_WRMSR, the "index" field tells user space which MSR the guest 5536wants to write. Once finished processing the event, user space must continue 5537vCPU execution. If the MSR write was unsuccessful, user space also sets the 5538"error" field to "1". 5539 5540:: 5541 5542 5543 struct kvm_xen_exit { 5544 #define KVM_EXIT_XEN_HCALL 1 5545 __u32 type; 5546 union { 5547 struct { 5548 __u32 longmode; 5549 __u32 cpl; 5550 __u64 input; 5551 __u64 result; 5552 __u64 params[6]; 5553 } hcall; 5554 } u; 5555 }; 5556 /* KVM_EXIT_XEN */ 5557 struct kvm_hyperv_exit xen; 5558 5559Indicates that the VCPU exits into userspace to process some tasks 5560related to Xen emulation. 5561 5562Valid values for 'type' are: 5563 5564 - KVM_EXIT_XEN_HCALL -- synchronously notify user-space about Xen hypercall. 5565 Userspace is expected to place the hypercall result into the appropriate 5566 field before invoking KVM_RUN again. 5567 5568:: 5569 5570 /* Fix the size of the union. */ 5571 char padding[256]; 5572 }; 5573 5574 /* 5575 * shared registers between kvm and userspace. 5576 * kvm_valid_regs specifies the register classes set by the host 5577 * kvm_dirty_regs specified the register classes dirtied by userspace 5578 * struct kvm_sync_regs is architecture specific, as well as the 5579 * bits for kvm_valid_regs and kvm_dirty_regs 5580 */ 5581 __u64 kvm_valid_regs; 5582 __u64 kvm_dirty_regs; 5583 union { 5584 struct kvm_sync_regs regs; 5585 char padding[SYNC_REGS_SIZE_BYTES]; 5586 } s; 5587 5588If KVM_CAP_SYNC_REGS is defined, these fields allow userspace to access 5589certain guest registers without having to call SET/GET_*REGS. Thus we can 5590avoid some system call overhead if userspace has to handle the exit. 5591Userspace can query the validity of the structure by checking 5592kvm_valid_regs for specific bits. These bits are architecture specific 5593and usually define the validity of a groups of registers. (e.g. one bit 5594for general purpose registers) 5595 5596Please note that the kernel is allowed to use the kvm_run structure as the 5597primary storage for certain register types. Therefore, the kernel may use the 5598values in kvm_run even if the corresponding bit in kvm_dirty_regs is not set. 5599 5600:: 5601 5602 }; 5603 5604 5605 56066. Capabilities that can be enabled on vCPUs 5607============================================ 5608 5609There are certain capabilities that change the behavior of the virtual CPU or 5610the virtual machine when enabled. To enable them, please see section 4.37. 5611Below you can find a list of capabilities and what their effect on the vCPU or 5612the virtual machine is when enabling them. 5613 5614The following information is provided along with the description: 5615 5616 Architectures: 5617 which instruction set architectures provide this ioctl. 5618 x86 includes both i386 and x86_64. 5619 5620 Target: 5621 whether this is a per-vcpu or per-vm capability. 5622 5623 Parameters: 5624 what parameters are accepted by the capability. 5625 5626 Returns: 5627 the return value. General error numbers (EBADF, ENOMEM, EINVAL) 5628 are not detailed, but errors with specific meanings are. 5629 5630 56316.1 KVM_CAP_PPC_OSI 5632------------------- 5633 5634:Architectures: ppc 5635:Target: vcpu 5636:Parameters: none 5637:Returns: 0 on success; -1 on error 5638 5639This capability enables interception of OSI hypercalls that otherwise would 5640be treated as normal system calls to be injected into the guest. OSI hypercalls 5641were invented by Mac-on-Linux to have a standardized communication mechanism 5642between the guest and the host. 5643 5644When this capability is enabled, KVM_EXIT_OSI can occur. 5645 5646 56476.2 KVM_CAP_PPC_PAPR 5648-------------------- 5649 5650:Architectures: ppc 5651:Target: vcpu 5652:Parameters: none 5653:Returns: 0 on success; -1 on error 5654 5655This capability enables interception of PAPR hypercalls. PAPR hypercalls are 5656done using the hypercall instruction "sc 1". 5657 5658It also sets the guest privilege level to "supervisor" mode. Usually the guest 5659runs in "hypervisor" privilege mode with a few missing features. 5660 5661In addition to the above, it changes the semantics of SDR1. In this mode, the 5662HTAB address part of SDR1 contains an HVA instead of a GPA, as PAPR keeps the 5663HTAB invisible to the guest. 5664 5665When this capability is enabled, KVM_EXIT_PAPR_HCALL can occur. 5666 5667 56686.3 KVM_CAP_SW_TLB 5669------------------ 5670 5671:Architectures: ppc 5672:Target: vcpu 5673:Parameters: args[0] is the address of a struct kvm_config_tlb 5674:Returns: 0 on success; -1 on error 5675 5676:: 5677 5678 struct kvm_config_tlb { 5679 __u64 params; 5680 __u64 array; 5681 __u32 mmu_type; 5682 __u32 array_len; 5683 }; 5684 5685Configures the virtual CPU's TLB array, establishing a shared memory area 5686between userspace and KVM. The "params" and "array" fields are userspace 5687addresses of mmu-type-specific data structures. The "array_len" field is an 5688safety mechanism, and should be set to the size in bytes of the memory that 5689userspace has reserved for the array. It must be at least the size dictated 5690by "mmu_type" and "params". 5691 5692While KVM_RUN is active, the shared region is under control of KVM. Its 5693contents are undefined, and any modification by userspace results in 5694boundedly undefined behavior. 5695 5696On return from KVM_RUN, the shared region will reflect the current state of 5697the guest's TLB. If userspace makes any changes, it must call KVM_DIRTY_TLB 5698to tell KVM which entries have been changed, prior to calling KVM_RUN again 5699on this vcpu. 5700 5701For mmu types KVM_MMU_FSL_BOOKE_NOHV and KVM_MMU_FSL_BOOKE_HV: 5702 5703 - The "params" field is of type "struct kvm_book3e_206_tlb_params". 5704 - The "array" field points to an array of type "struct 5705 kvm_book3e_206_tlb_entry". 5706 - The array consists of all entries in the first TLB, followed by all 5707 entries in the second TLB. 5708 - Within a TLB, entries are ordered first by increasing set number. Within a 5709 set, entries are ordered by way (increasing ESEL). 5710 - The hash for determining set number in TLB0 is: (MAS2 >> 12) & (num_sets - 1) 5711 where "num_sets" is the tlb_sizes[] value divided by the tlb_ways[] value. 5712 - The tsize field of mas1 shall be set to 4K on TLB0, even though the 5713 hardware ignores this value for TLB0. 5714 57156.4 KVM_CAP_S390_CSS_SUPPORT 5716---------------------------- 5717 5718:Architectures: s390 5719:Target: vcpu 5720:Parameters: none 5721:Returns: 0 on success; -1 on error 5722 5723This capability enables support for handling of channel I/O instructions. 5724 5725TEST PENDING INTERRUPTION and the interrupt portion of TEST SUBCHANNEL are 5726handled in-kernel, while the other I/O instructions are passed to userspace. 5727 5728When this capability is enabled, KVM_EXIT_S390_TSCH will occur on TEST 5729SUBCHANNEL intercepts. 5730 5731Note that even though this capability is enabled per-vcpu, the complete 5732virtual machine is affected. 5733 57346.5 KVM_CAP_PPC_EPR 5735------------------- 5736 5737:Architectures: ppc 5738:Target: vcpu 5739:Parameters: args[0] defines whether the proxy facility is active 5740:Returns: 0 on success; -1 on error 5741 5742This capability enables or disables the delivery of interrupts through the 5743external proxy facility. 5744 5745When enabled (args[0] != 0), every time the guest gets an external interrupt 5746delivered, it automatically exits into user space with a KVM_EXIT_EPR exit 5747to receive the topmost interrupt vector. 5748 5749When disabled (args[0] == 0), behavior is as if this facility is unsupported. 5750 5751When this capability is enabled, KVM_EXIT_EPR can occur. 5752 57536.6 KVM_CAP_IRQ_MPIC 5754-------------------- 5755 5756:Architectures: ppc 5757:Parameters: args[0] is the MPIC device fd; 5758 args[1] is the MPIC CPU number for this vcpu 5759 5760This capability connects the vcpu to an in-kernel MPIC device. 5761 57626.7 KVM_CAP_IRQ_XICS 5763-------------------- 5764 5765:Architectures: ppc 5766:Target: vcpu 5767:Parameters: args[0] is the XICS device fd; 5768 args[1] is the XICS CPU number (server ID) for this vcpu 5769 5770This capability connects the vcpu to an in-kernel XICS device. 5771 57726.8 KVM_CAP_S390_IRQCHIP 5773------------------------ 5774 5775:Architectures: s390 5776:Target: vm 5777:Parameters: none 5778 5779This capability enables the in-kernel irqchip for s390. Please refer to 5780"4.24 KVM_CREATE_IRQCHIP" for details. 5781 57826.9 KVM_CAP_MIPS_FPU 5783-------------------- 5784 5785:Architectures: mips 5786:Target: vcpu 5787:Parameters: args[0] is reserved for future use (should be 0). 5788 5789This capability allows the use of the host Floating Point Unit by the guest. It 5790allows the Config1.FP bit to be set to enable the FPU in the guest. Once this is 5791done the ``KVM_REG_MIPS_FPR_*`` and ``KVM_REG_MIPS_FCR_*`` registers can be 5792accessed (depending on the current guest FPU register mode), and the Status.FR, 5793Config5.FRE bits are accessible via the KVM API and also from the guest, 5794depending on them being supported by the FPU. 5795 57966.10 KVM_CAP_MIPS_MSA 5797--------------------- 5798 5799:Architectures: mips 5800:Target: vcpu 5801:Parameters: args[0] is reserved for future use (should be 0). 5802 5803This capability allows the use of the MIPS SIMD Architecture (MSA) by the guest. 5804It allows the Config3.MSAP bit to be set to enable the use of MSA by the guest. 5805Once this is done the ``KVM_REG_MIPS_VEC_*`` and ``KVM_REG_MIPS_MSA_*`` 5806registers can be accessed, and the Config5.MSAEn bit is accessible via the 5807KVM API and also from the guest. 5808 58096.74 KVM_CAP_SYNC_REGS 5810---------------------- 5811 5812:Architectures: s390, x86 5813:Target: s390: always enabled, x86: vcpu 5814:Parameters: none 5815:Returns: x86: KVM_CHECK_EXTENSION returns a bit-array indicating which register 5816 sets are supported 5817 (bitfields defined in arch/x86/include/uapi/asm/kvm.h). 5818 5819As described above in the kvm_sync_regs struct info in section 5 (kvm_run): 5820KVM_CAP_SYNC_REGS "allow[s] userspace to access certain guest registers 5821without having to call SET/GET_*REGS". This reduces overhead by eliminating 5822repeated ioctl calls for setting and/or getting register values. This is 5823particularly important when userspace is making synchronous guest state 5824modifications, e.g. when emulating and/or intercepting instructions in 5825userspace. 5826 5827For s390 specifics, please refer to the source code. 5828 5829For x86: 5830 5831- the register sets to be copied out to kvm_run are selectable 5832 by userspace (rather that all sets being copied out for every exit). 5833- vcpu_events are available in addition to regs and sregs. 5834 5835For x86, the 'kvm_valid_regs' field of struct kvm_run is overloaded to 5836function as an input bit-array field set by userspace to indicate the 5837specific register sets to be copied out on the next exit. 5838 5839To indicate when userspace has modified values that should be copied into 5840the vCPU, the all architecture bitarray field, 'kvm_dirty_regs' must be set. 5841This is done using the same bitflags as for the 'kvm_valid_regs' field. 5842If the dirty bit is not set, then the register set values will not be copied 5843into the vCPU even if they've been modified. 5844 5845Unused bitfields in the bitarrays must be set to zero. 5846 5847:: 5848 5849 struct kvm_sync_regs { 5850 struct kvm_regs regs; 5851 struct kvm_sregs sregs; 5852 struct kvm_vcpu_events events; 5853 }; 5854 58556.75 KVM_CAP_PPC_IRQ_XIVE 5856------------------------- 5857 5858:Architectures: ppc 5859:Target: vcpu 5860:Parameters: args[0] is the XIVE device fd; 5861 args[1] is the XIVE CPU number (server ID) for this vcpu 5862 5863This capability connects the vcpu to an in-kernel XIVE device. 5864 58657. Capabilities that can be enabled on VMs 5866========================================== 5867 5868There are certain capabilities that change the behavior of the virtual 5869machine when enabled. To enable them, please see section 4.37. Below 5870you can find a list of capabilities and what their effect on the VM 5871is when enabling them. 5872 5873The following information is provided along with the description: 5874 5875 Architectures: 5876 which instruction set architectures provide this ioctl. 5877 x86 includes both i386 and x86_64. 5878 5879 Parameters: 5880 what parameters are accepted by the capability. 5881 5882 Returns: 5883 the return value. General error numbers (EBADF, ENOMEM, EINVAL) 5884 are not detailed, but errors with specific meanings are. 5885 5886 58877.1 KVM_CAP_PPC_ENABLE_HCALL 5888---------------------------- 5889 5890:Architectures: ppc 5891:Parameters: args[0] is the sPAPR hcall number; 5892 args[1] is 0 to disable, 1 to enable in-kernel handling 5893 5894This capability controls whether individual sPAPR hypercalls (hcalls) 5895get handled by the kernel or not. Enabling or disabling in-kernel 5896handling of an hcall is effective across the VM. On creation, an 5897initial set of hcalls are enabled for in-kernel handling, which 5898consists of those hcalls for which in-kernel handlers were implemented 5899before this capability was implemented. If disabled, the kernel will 5900not to attempt to handle the hcall, but will always exit to userspace 5901to handle it. Note that it may not make sense to enable some and 5902disable others of a group of related hcalls, but KVM does not prevent 5903userspace from doing that. 5904 5905If the hcall number specified is not one that has an in-kernel 5906implementation, the KVM_ENABLE_CAP ioctl will fail with an EINVAL 5907error. 5908 59097.2 KVM_CAP_S390_USER_SIGP 5910-------------------------- 5911 5912:Architectures: s390 5913:Parameters: none 5914 5915This capability controls which SIGP orders will be handled completely in user 5916space. With this capability enabled, all fast orders will be handled completely 5917in the kernel: 5918 5919- SENSE 5920- SENSE RUNNING 5921- EXTERNAL CALL 5922- EMERGENCY SIGNAL 5923- CONDITIONAL EMERGENCY SIGNAL 5924 5925All other orders will be handled completely in user space. 5926 5927Only privileged operation exceptions will be checked for in the kernel (or even 5928in the hardware prior to interception). If this capability is not enabled, the 5929old way of handling SIGP orders is used (partially in kernel and user space). 5930 59317.3 KVM_CAP_S390_VECTOR_REGISTERS 5932--------------------------------- 5933 5934:Architectures: s390 5935:Parameters: none 5936:Returns: 0 on success, negative value on error 5937 5938Allows use of the vector registers introduced with z13 processor, and 5939provides for the synchronization between host and user space. Will 5940return -EINVAL if the machine does not support vectors. 5941 59427.4 KVM_CAP_S390_USER_STSI 5943-------------------------- 5944 5945:Architectures: s390 5946:Parameters: none 5947 5948This capability allows post-handlers for the STSI instruction. After 5949initial handling in the kernel, KVM exits to user space with 5950KVM_EXIT_S390_STSI to allow user space to insert further data. 5951 5952Before exiting to userspace, kvm handlers should fill in s390_stsi field of 5953vcpu->run:: 5954 5955 struct { 5956 __u64 addr; 5957 __u8 ar; 5958 __u8 reserved; 5959 __u8 fc; 5960 __u8 sel1; 5961 __u16 sel2; 5962 } s390_stsi; 5963 5964 @addr - guest address of STSI SYSIB 5965 @fc - function code 5966 @sel1 - selector 1 5967 @sel2 - selector 2 5968 @ar - access register number 5969 5970KVM handlers should exit to userspace with rc = -EREMOTE. 5971 59727.5 KVM_CAP_SPLIT_IRQCHIP 5973------------------------- 5974 5975:Architectures: x86 5976:Parameters: args[0] - number of routes reserved for userspace IOAPICs 5977:Returns: 0 on success, -1 on error 5978 5979Create a local apic for each processor in the kernel. This can be used 5980instead of KVM_CREATE_IRQCHIP if the userspace VMM wishes to emulate the 5981IOAPIC and PIC (and also the PIT, even though this has to be enabled 5982separately). 5983 5984This capability also enables in kernel routing of interrupt requests; 5985when KVM_CAP_SPLIT_IRQCHIP only routes of KVM_IRQ_ROUTING_MSI type are 5986used in the IRQ routing table. The first args[0] MSI routes are reserved 5987for the IOAPIC pins. Whenever the LAPIC receives an EOI for these routes, 5988a KVM_EXIT_IOAPIC_EOI vmexit will be reported to userspace. 5989 5990Fails if VCPU has already been created, or if the irqchip is already in the 5991kernel (i.e. KVM_CREATE_IRQCHIP has already been called). 5992 59937.6 KVM_CAP_S390_RI 5994------------------- 5995 5996:Architectures: s390 5997:Parameters: none 5998 5999Allows use of runtime-instrumentation introduced with zEC12 processor. 6000Will return -EINVAL if the machine does not support runtime-instrumentation. 6001Will return -EBUSY if a VCPU has already been created. 6002 60037.7 KVM_CAP_X2APIC_API 6004---------------------- 6005 6006:Architectures: x86 6007:Parameters: args[0] - features that should be enabled 6008:Returns: 0 on success, -EINVAL when args[0] contains invalid features 6009 6010Valid feature flags in args[0] are:: 6011 6012 #define KVM_X2APIC_API_USE_32BIT_IDS (1ULL << 0) 6013 #define KVM_X2APIC_API_DISABLE_BROADCAST_QUIRK (1ULL << 1) 6014 6015Enabling KVM_X2APIC_API_USE_32BIT_IDS changes the behavior of 6016KVM_SET_GSI_ROUTING, KVM_SIGNAL_MSI, KVM_SET_LAPIC, and KVM_GET_LAPIC, 6017allowing the use of 32-bit APIC IDs. See KVM_CAP_X2APIC_API in their 6018respective sections. 6019 6020KVM_X2APIC_API_DISABLE_BROADCAST_QUIRK must be enabled for x2APIC to work 6021in logical mode or with more than 255 VCPUs. Otherwise, KVM treats 0xff 6022as a broadcast even in x2APIC mode in order to support physical x2APIC 6023without interrupt remapping. This is undesirable in logical mode, 6024where 0xff represents CPUs 0-7 in cluster 0. 6025 60267.8 KVM_CAP_S390_USER_INSTR0 6027---------------------------- 6028 6029:Architectures: s390 6030:Parameters: none 6031 6032With this capability enabled, all illegal instructions 0x0000 (2 bytes) will 6033be intercepted and forwarded to user space. User space can use this 6034mechanism e.g. to realize 2-byte software breakpoints. The kernel will 6035not inject an operating exception for these instructions, user space has 6036to take care of that. 6037 6038This capability can be enabled dynamically even if VCPUs were already 6039created and are running. 6040 60417.9 KVM_CAP_S390_GS 6042------------------- 6043 6044:Architectures: s390 6045:Parameters: none 6046:Returns: 0 on success; -EINVAL if the machine does not support 6047 guarded storage; -EBUSY if a VCPU has already been created. 6048 6049Allows use of guarded storage for the KVM guest. 6050 60517.10 KVM_CAP_S390_AIS 6052--------------------- 6053 6054:Architectures: s390 6055:Parameters: none 6056 6057Allow use of adapter-interruption suppression. 6058:Returns: 0 on success; -EBUSY if a VCPU has already been created. 6059 60607.11 KVM_CAP_PPC_SMT 6061-------------------- 6062 6063:Architectures: ppc 6064:Parameters: vsmt_mode, flags 6065 6066Enabling this capability on a VM provides userspace with a way to set 6067the desired virtual SMT mode (i.e. the number of virtual CPUs per 6068virtual core). The virtual SMT mode, vsmt_mode, must be a power of 2 6069between 1 and 8. On POWER8, vsmt_mode must also be no greater than 6070the number of threads per subcore for the host. Currently flags must 6071be 0. A successful call to enable this capability will result in 6072vsmt_mode being returned when the KVM_CAP_PPC_SMT capability is 6073subsequently queried for the VM. This capability is only supported by 6074HV KVM, and can only be set before any VCPUs have been created. 6075The KVM_CAP_PPC_SMT_POSSIBLE capability indicates which virtual SMT 6076modes are available. 6077 60787.12 KVM_CAP_PPC_FWNMI 6079---------------------- 6080 6081:Architectures: ppc 6082:Parameters: none 6083 6084With this capability a machine check exception in the guest address 6085space will cause KVM to exit the guest with NMI exit reason. This 6086enables QEMU to build error log and branch to guest kernel registered 6087machine check handling routine. Without this capability KVM will 6088branch to guests' 0x200 interrupt vector. 6089 60907.13 KVM_CAP_X86_DISABLE_EXITS 6091------------------------------ 6092 6093:Architectures: x86 6094:Parameters: args[0] defines which exits are disabled 6095:Returns: 0 on success, -EINVAL when args[0] contains invalid exits 6096 6097Valid bits in args[0] are:: 6098 6099 #define KVM_X86_DISABLE_EXITS_MWAIT (1 << 0) 6100 #define KVM_X86_DISABLE_EXITS_HLT (1 << 1) 6101 #define KVM_X86_DISABLE_EXITS_PAUSE (1 << 2) 6102 #define KVM_X86_DISABLE_EXITS_CSTATE (1 << 3) 6103 6104Enabling this capability on a VM provides userspace with a way to no 6105longer intercept some instructions for improved latency in some 6106workloads, and is suggested when vCPUs are associated to dedicated 6107physical CPUs. More bits can be added in the future; userspace can 6108just pass the KVM_CHECK_EXTENSION result to KVM_ENABLE_CAP to disable 6109all such vmexits. 6110 6111Do not enable KVM_FEATURE_PV_UNHALT if you disable HLT exits. 6112 61137.14 KVM_CAP_S390_HPAGE_1M 6114-------------------------- 6115 6116:Architectures: s390 6117:Parameters: none 6118:Returns: 0 on success, -EINVAL if hpage module parameter was not set 6119 or cmma is enabled, or the VM has the KVM_VM_S390_UCONTROL 6120 flag set 6121 6122With this capability the KVM support for memory backing with 1m pages 6123through hugetlbfs can be enabled for a VM. After the capability is 6124enabled, cmma can't be enabled anymore and pfmfi and the storage key 6125interpretation are disabled. If cmma has already been enabled or the 6126hpage module parameter is not set to 1, -EINVAL is returned. 6127 6128While it is generally possible to create a huge page backed VM without 6129this capability, the VM will not be able to run. 6130 61317.15 KVM_CAP_MSR_PLATFORM_INFO 6132------------------------------ 6133 6134:Architectures: x86 6135:Parameters: args[0] whether feature should be enabled or not 6136 6137With this capability, a guest may read the MSR_PLATFORM_INFO MSR. Otherwise, 6138a #GP would be raised when the guest tries to access. Currently, this 6139capability does not enable write permissions of this MSR for the guest. 6140 61417.16 KVM_CAP_PPC_NESTED_HV 6142-------------------------- 6143 6144:Architectures: ppc 6145:Parameters: none 6146:Returns: 0 on success, -EINVAL when the implementation doesn't support 6147 nested-HV virtualization. 6148 6149HV-KVM on POWER9 and later systems allows for "nested-HV" 6150virtualization, which provides a way for a guest VM to run guests that 6151can run using the CPU's supervisor mode (privileged non-hypervisor 6152state). Enabling this capability on a VM depends on the CPU having 6153the necessary functionality and on the facility being enabled with a 6154kvm-hv module parameter. 6155 61567.17 KVM_CAP_EXCEPTION_PAYLOAD 6157------------------------------ 6158 6159:Architectures: x86 6160:Parameters: args[0] whether feature should be enabled or not 6161 6162With this capability enabled, CR2 will not be modified prior to the 6163emulated VM-exit when L1 intercepts a #PF exception that occurs in 6164L2. Similarly, for kvm-intel only, DR6 will not be modified prior to 6165the emulated VM-exit when L1 intercepts a #DB exception that occurs in 6166L2. As a result, when KVM_GET_VCPU_EVENTS reports a pending #PF (or 6167#DB) exception for L2, exception.has_payload will be set and the 6168faulting address (or the new DR6 bits*) will be reported in the 6169exception_payload field. Similarly, when userspace injects a #PF (or 6170#DB) into L2 using KVM_SET_VCPU_EVENTS, it is expected to set 6171exception.has_payload and to put the faulting address - or the new DR6 6172bits\ [#]_ - in the exception_payload field. 6173 6174This capability also enables exception.pending in struct 6175kvm_vcpu_events, which allows userspace to distinguish between pending 6176and injected exceptions. 6177 6178 6179.. [#] For the new DR6 bits, note that bit 16 is set iff the #DB exception 6180 will clear DR6.RTM. 6181 61827.18 KVM_CAP_MANUAL_DIRTY_LOG_PROTECT2 6183 6184:Architectures: x86, arm, arm64, mips 6185:Parameters: args[0] whether feature should be enabled or not 6186 6187Valid flags are:: 6188 6189 #define KVM_DIRTY_LOG_MANUAL_PROTECT_ENABLE (1 << 0) 6190 #define KVM_DIRTY_LOG_INITIALLY_SET (1 << 1) 6191 6192With KVM_DIRTY_LOG_MANUAL_PROTECT_ENABLE is set, KVM_GET_DIRTY_LOG will not 6193automatically clear and write-protect all pages that are returned as dirty. 6194Rather, userspace will have to do this operation separately using 6195KVM_CLEAR_DIRTY_LOG. 6196 6197At the cost of a slightly more complicated operation, this provides better 6198scalability and responsiveness for two reasons. First, 6199KVM_CLEAR_DIRTY_LOG ioctl can operate on a 64-page granularity rather 6200than requiring to sync a full memslot; this ensures that KVM does not 6201take spinlocks for an extended period of time. Second, in some cases a 6202large amount of time can pass between a call to KVM_GET_DIRTY_LOG and 6203userspace actually using the data in the page. Pages can be modified 6204during this time, which is inefficient for both the guest and userspace: 6205the guest will incur a higher penalty due to write protection faults, 6206while userspace can see false reports of dirty pages. Manual reprotection 6207helps reducing this time, improving guest performance and reducing the 6208number of dirty log false positives. 6209 6210With KVM_DIRTY_LOG_INITIALLY_SET set, all the bits of the dirty bitmap 6211will be initialized to 1 when created. This also improves performance because 6212dirty logging can be enabled gradually in small chunks on the first call 6213to KVM_CLEAR_DIRTY_LOG. KVM_DIRTY_LOG_INITIALLY_SET depends on 6214KVM_DIRTY_LOG_MANUAL_PROTECT_ENABLE (it is also only available on 6215x86 and arm64 for now). 6216 6217KVM_CAP_MANUAL_DIRTY_LOG_PROTECT2 was previously available under the name 6218KVM_CAP_MANUAL_DIRTY_LOG_PROTECT, but the implementation had bugs that make 6219it hard or impossible to use it correctly. The availability of 6220KVM_CAP_MANUAL_DIRTY_LOG_PROTECT2 signals that those bugs are fixed. 6221Userspace should not try to use KVM_CAP_MANUAL_DIRTY_LOG_PROTECT. 6222 62237.19 KVM_CAP_PPC_SECURE_GUEST 6224------------------------------ 6225 6226:Architectures: ppc 6227 6228This capability indicates that KVM is running on a host that has 6229ultravisor firmware and thus can support a secure guest. On such a 6230system, a guest can ask the ultravisor to make it a secure guest, 6231one whose memory is inaccessible to the host except for pages which 6232are explicitly requested to be shared with the host. The ultravisor 6233notifies KVM when a guest requests to become a secure guest, and KVM 6234has the opportunity to veto the transition. 6235 6236If present, this capability can be enabled for a VM, meaning that KVM 6237will allow the transition to secure guest mode. Otherwise KVM will 6238veto the transition. 6239 62407.20 KVM_CAP_HALT_POLL 6241---------------------- 6242 6243:Architectures: all 6244:Target: VM 6245:Parameters: args[0] is the maximum poll time in nanoseconds 6246:Returns: 0 on success; -1 on error 6247 6248This capability overrides the kvm module parameter halt_poll_ns for the 6249target VM. 6250 6251VCPU polling allows a VCPU to poll for wakeup events instead of immediately 6252scheduling during guest halts. The maximum time a VCPU can spend polling is 6253controlled by the kvm module parameter halt_poll_ns. This capability allows 6254the maximum halt time to specified on a per-VM basis, effectively overriding 6255the module parameter for the target VM. 6256 62577.21 KVM_CAP_X86_USER_SPACE_MSR 6258------------------------------- 6259 6260:Architectures: x86 6261:Target: VM 6262:Parameters: args[0] contains the mask of KVM_MSR_EXIT_REASON_* events to report 6263:Returns: 0 on success; -1 on error 6264 6265This capability enables trapping of #GP invoking RDMSR and WRMSR instructions 6266into user space. 6267 6268When a guest requests to read or write an MSR, KVM may not implement all MSRs 6269that are relevant to a respective system. It also does not differentiate by 6270CPU type. 6271 6272To allow more fine grained control over MSR handling, user space may enable 6273this capability. With it enabled, MSR accesses that match the mask specified in 6274args[0] and trigger a #GP event inside the guest by KVM will instead trigger 6275KVM_EXIT_X86_RDMSR and KVM_EXIT_X86_WRMSR exit notifications which user space 6276can then handle to implement model specific MSR handling and/or user notifications 6277to inform a user that an MSR was not handled. 6278 62797.22 KVM_CAP_X86_BUS_LOCK_EXIT 6280------------------------------- 6281 6282:Architectures: x86 6283:Target: VM 6284:Parameters: args[0] defines the policy used when bus locks detected in guest 6285:Returns: 0 on success, -EINVAL when args[0] contains invalid bits 6286 6287Valid bits in args[0] are:: 6288 6289 #define KVM_BUS_LOCK_DETECTION_OFF (1 << 0) 6290 #define KVM_BUS_LOCK_DETECTION_EXIT (1 << 1) 6291 6292Enabling this capability on a VM provides userspace with a way to select 6293a policy to handle the bus locks detected in guest. Userspace can obtain 6294the supported modes from the result of KVM_CHECK_EXTENSION and define it 6295through the KVM_ENABLE_CAP. 6296 6297KVM_BUS_LOCK_DETECTION_OFF and KVM_BUS_LOCK_DETECTION_EXIT are supported 6298currently and mutually exclusive with each other. More bits can be added in 6299the future. 6300 6301With KVM_BUS_LOCK_DETECTION_OFF set, bus locks in guest will not cause vm exits 6302so that no additional actions are needed. This is the default mode. 6303 6304With KVM_BUS_LOCK_DETECTION_EXIT set, vm exits happen when bus lock detected 6305in VM. KVM just exits to userspace when handling them. Userspace can enforce 6306its own throttling or other policy based mitigations. 6307 6308This capability is aimed to address the thread that VM can exploit bus locks to 6309degree the performance of the whole system. Once the userspace enable this 6310capability and select the KVM_BUS_LOCK_DETECTION_EXIT mode, KVM will set the 6311KVM_RUN_BUS_LOCK flag in vcpu-run->flags field and exit to userspace. Concerning 6312the bus lock vm exit can be preempted by a higher priority VM exit, the exit 6313notifications to userspace can be KVM_EXIT_BUS_LOCK or other reasons. 6314KVM_RUN_BUS_LOCK flag is used to distinguish between them. 6315 63167.23 KVM_CAP_PPC_DAWR1 6317---------------------- 6318 6319:Architectures: ppc 6320:Parameters: none 6321:Returns: 0 on success, -EINVAL when CPU doesn't support 2nd DAWR 6322 6323This capability can be used to check / enable 2nd DAWR feature provided 6324by POWER10 processor. 6325 63267.24 KVM_CAP_VM_COPY_ENC_CONTEXT_FROM 6327------------------------------------- 6328 6329Architectures: x86 SEV enabled 6330Type: vm 6331Parameters: args[0] is the fd of the source vm 6332Returns: 0 on success; ENOTTY on error 6333 6334This capability enables userspace to copy encryption context from the vm 6335indicated by the fd to the vm this is called on. 6336 6337This is intended to support in-guest workloads scheduled by the host. This 6338allows the in-guest workload to maintain its own NPTs and keeps the two vms 6339from accidentally clobbering each other with interrupts and the like (separate 6340APIC/MSRs/etc). 6341 63427.25 KVM_CAP_SGX_ATTRIBUTE 6343-------------------------- 6344 6345:Architectures: x86 6346:Target: VM 6347:Parameters: args[0] is a file handle of a SGX attribute file in securityfs 6348:Returns: 0 on success, -EINVAL if the file handle is invalid or if a requested 6349 attribute is not supported by KVM. 6350 6351KVM_CAP_SGX_ATTRIBUTE enables a userspace VMM to grant a VM access to one or 6352more priveleged enclave attributes. args[0] must hold a file handle to a valid 6353SGX attribute file corresponding to an attribute that is supported/restricted 6354by KVM (currently only PROVISIONKEY). 6355 6356The SGX subsystem restricts access to a subset of enclave attributes to provide 6357additional security for an uncompromised kernel, e.g. use of the PROVISIONKEY 6358is restricted to deter malware from using the PROVISIONKEY to obtain a stable 6359system fingerprint. To prevent userspace from circumventing such restrictions 6360by running an enclave in a VM, KVM prevents access to privileged attributes by 6361default. 6362 6363See Documentation/x86/sgx/2.Kernel-internals.rst for more details. 6364 63658. Other capabilities. 6366====================== 6367 6368This section lists capabilities that give information about other 6369features of the KVM implementation. 6370 63718.1 KVM_CAP_PPC_HWRNG 6372--------------------- 6373 6374:Architectures: ppc 6375 6376This capability, if KVM_CHECK_EXTENSION indicates that it is 6377available, means that the kernel has an implementation of the 6378H_RANDOM hypercall backed by a hardware random-number generator. 6379If present, the kernel H_RANDOM handler can be enabled for guest use 6380with the KVM_CAP_PPC_ENABLE_HCALL capability. 6381 63828.2 KVM_CAP_HYPERV_SYNIC 6383------------------------ 6384 6385:Architectures: x86 6386 6387This capability, if KVM_CHECK_EXTENSION indicates that it is 6388available, means that the kernel has an implementation of the 6389Hyper-V Synthetic interrupt controller(SynIC). Hyper-V SynIC is 6390used to support Windows Hyper-V based guest paravirt drivers(VMBus). 6391 6392In order to use SynIC, it has to be activated by setting this 6393capability via KVM_ENABLE_CAP ioctl on the vcpu fd. Note that this 6394will disable the use of APIC hardware virtualization even if supported 6395by the CPU, as it's incompatible with SynIC auto-EOI behavior. 6396 63978.3 KVM_CAP_PPC_RADIX_MMU 6398------------------------- 6399 6400:Architectures: ppc 6401 6402This capability, if KVM_CHECK_EXTENSION indicates that it is 6403available, means that the kernel can support guests using the 6404radix MMU defined in Power ISA V3.00 (as implemented in the POWER9 6405processor). 6406 64078.4 KVM_CAP_PPC_HASH_MMU_V3 6408--------------------------- 6409 6410:Architectures: ppc 6411 6412This capability, if KVM_CHECK_EXTENSION indicates that it is 6413available, means that the kernel can support guests using the 6414hashed page table MMU defined in Power ISA V3.00 (as implemented in 6415the POWER9 processor), including in-memory segment tables. 6416 64178.5 KVM_CAP_MIPS_VZ 6418------------------- 6419 6420:Architectures: mips 6421 6422This capability, if KVM_CHECK_EXTENSION on the main kvm handle indicates that 6423it is available, means that full hardware assisted virtualization capabilities 6424of the hardware are available for use through KVM. An appropriate 6425KVM_VM_MIPS_* type must be passed to KVM_CREATE_VM to create a VM which 6426utilises it. 6427 6428If KVM_CHECK_EXTENSION on a kvm VM handle indicates that this capability is 6429available, it means that the VM is using full hardware assisted virtualization 6430capabilities of the hardware. This is useful to check after creating a VM with 6431KVM_VM_MIPS_DEFAULT. 6432 6433The value returned by KVM_CHECK_EXTENSION should be compared against known 6434values (see below). All other values are reserved. This is to allow for the 6435possibility of other hardware assisted virtualization implementations which 6436may be incompatible with the MIPS VZ ASE. 6437 6438== ========================================================================== 6439 0 The trap & emulate implementation is in use to run guest code in user 6440 mode. Guest virtual memory segments are rearranged to fit the guest in the 6441 user mode address space. 6442 6443 1 The MIPS VZ ASE is in use, providing full hardware assisted 6444 virtualization, including standard guest virtual memory segments. 6445== ========================================================================== 6446 64478.6 KVM_CAP_MIPS_TE 6448------------------- 6449 6450:Architectures: mips 6451 6452This capability, if KVM_CHECK_EXTENSION on the main kvm handle indicates that 6453it is available, means that the trap & emulate implementation is available to 6454run guest code in user mode, even if KVM_CAP_MIPS_VZ indicates that hardware 6455assisted virtualisation is also available. KVM_VM_MIPS_TE (0) must be passed 6456to KVM_CREATE_VM to create a VM which utilises it. 6457 6458If KVM_CHECK_EXTENSION on a kvm VM handle indicates that this capability is 6459available, it means that the VM is using trap & emulate. 6460 64618.7 KVM_CAP_MIPS_64BIT 6462---------------------- 6463 6464:Architectures: mips 6465 6466This capability indicates the supported architecture type of the guest, i.e. the 6467supported register and address width. 6468 6469The values returned when this capability is checked by KVM_CHECK_EXTENSION on a 6470kvm VM handle correspond roughly to the CP0_Config.AT register field, and should 6471be checked specifically against known values (see below). All other values are 6472reserved. 6473 6474== ======================================================================== 6475 0 MIPS32 or microMIPS32. 6476 Both registers and addresses are 32-bits wide. 6477 It will only be possible to run 32-bit guest code. 6478 6479 1 MIPS64 or microMIPS64 with access only to 32-bit compatibility segments. 6480 Registers are 64-bits wide, but addresses are 32-bits wide. 6481 64-bit guest code may run but cannot access MIPS64 memory segments. 6482 It will also be possible to run 32-bit guest code. 6483 6484 2 MIPS64 or microMIPS64 with access to all address segments. 6485 Both registers and addresses are 64-bits wide. 6486 It will be possible to run 64-bit or 32-bit guest code. 6487== ======================================================================== 6488 64898.9 KVM_CAP_ARM_USER_IRQ 6490------------------------ 6491 6492:Architectures: arm, arm64 6493 6494This capability, if KVM_CHECK_EXTENSION indicates that it is available, means 6495that if userspace creates a VM without an in-kernel interrupt controller, it 6496will be notified of changes to the output level of in-kernel emulated devices, 6497which can generate virtual interrupts, presented to the VM. 6498For such VMs, on every return to userspace, the kernel 6499updates the vcpu's run->s.regs.device_irq_level field to represent the actual 6500output level of the device. 6501 6502Whenever kvm detects a change in the device output level, kvm guarantees at 6503least one return to userspace before running the VM. This exit could either 6504be a KVM_EXIT_INTR or any other exit event, like KVM_EXIT_MMIO. This way, 6505userspace can always sample the device output level and re-compute the state of 6506the userspace interrupt controller. Userspace should always check the state 6507of run->s.regs.device_irq_level on every kvm exit. 6508The value in run->s.regs.device_irq_level can represent both level and edge 6509triggered interrupt signals, depending on the device. Edge triggered interrupt 6510signals will exit to userspace with the bit in run->s.regs.device_irq_level 6511set exactly once per edge signal. 6512 6513The field run->s.regs.device_irq_level is available independent of 6514run->kvm_valid_regs or run->kvm_dirty_regs bits. 6515 6516If KVM_CAP_ARM_USER_IRQ is supported, the KVM_CHECK_EXTENSION ioctl returns a 6517number larger than 0 indicating the version of this capability is implemented 6518and thereby which bits in run->s.regs.device_irq_level can signal values. 6519 6520Currently the following bits are defined for the device_irq_level bitmap:: 6521 6522 KVM_CAP_ARM_USER_IRQ >= 1: 6523 6524 KVM_ARM_DEV_EL1_VTIMER - EL1 virtual timer 6525 KVM_ARM_DEV_EL1_PTIMER - EL1 physical timer 6526 KVM_ARM_DEV_PMU - ARM PMU overflow interrupt signal 6527 6528Future versions of kvm may implement additional events. These will get 6529indicated by returning a higher number from KVM_CHECK_EXTENSION and will be 6530listed above. 6531 65328.10 KVM_CAP_PPC_SMT_POSSIBLE 6533----------------------------- 6534 6535:Architectures: ppc 6536 6537Querying this capability returns a bitmap indicating the possible 6538virtual SMT modes that can be set using KVM_CAP_PPC_SMT. If bit N 6539(counting from the right) is set, then a virtual SMT mode of 2^N is 6540available. 6541 65428.11 KVM_CAP_HYPERV_SYNIC2 6543-------------------------- 6544 6545:Architectures: x86 6546 6547This capability enables a newer version of Hyper-V Synthetic interrupt 6548controller (SynIC). The only difference with KVM_CAP_HYPERV_SYNIC is that KVM 6549doesn't clear SynIC message and event flags pages when they are enabled by 6550writing to the respective MSRs. 6551 65528.12 KVM_CAP_HYPERV_VP_INDEX 6553---------------------------- 6554 6555:Architectures: x86 6556 6557This capability indicates that userspace can load HV_X64_MSR_VP_INDEX msr. Its 6558value is used to denote the target vcpu for a SynIC interrupt. For 6559compatibilty, KVM initializes this msr to KVM's internal vcpu index. When this 6560capability is absent, userspace can still query this msr's value. 6561 65628.13 KVM_CAP_S390_AIS_MIGRATION 6563------------------------------- 6564 6565:Architectures: s390 6566:Parameters: none 6567 6568This capability indicates if the flic device will be able to get/set the 6569AIS states for migration via the KVM_DEV_FLIC_AISM_ALL attribute and allows 6570to discover this without having to create a flic device. 6571 65728.14 KVM_CAP_S390_PSW 6573--------------------- 6574 6575:Architectures: s390 6576 6577This capability indicates that the PSW is exposed via the kvm_run structure. 6578 65798.15 KVM_CAP_S390_GMAP 6580---------------------- 6581 6582:Architectures: s390 6583 6584This capability indicates that the user space memory used as guest mapping can 6585be anywhere in the user memory address space, as long as the memory slots are 6586aligned and sized to a segment (1MB) boundary. 6587 65888.16 KVM_CAP_S390_COW 6589--------------------- 6590 6591:Architectures: s390 6592 6593This capability indicates that the user space memory used as guest mapping can 6594use copy-on-write semantics as well as dirty pages tracking via read-only page 6595tables. 6596 65978.17 KVM_CAP_S390_BPB 6598--------------------- 6599 6600:Architectures: s390 6601 6602This capability indicates that kvm will implement the interfaces to handle 6603reset, migration and nested KVM for branch prediction blocking. The stfle 6604facility 82 should not be provided to the guest without this capability. 6605 66068.18 KVM_CAP_HYPERV_TLBFLUSH 6607---------------------------- 6608 6609:Architectures: x86 6610 6611This capability indicates that KVM supports paravirtualized Hyper-V TLB Flush 6612hypercalls: 6613HvFlushVirtualAddressSpace, HvFlushVirtualAddressSpaceEx, 6614HvFlushVirtualAddressList, HvFlushVirtualAddressListEx. 6615 66168.19 KVM_CAP_ARM_INJECT_SERROR_ESR 6617---------------------------------- 6618 6619:Architectures: arm, arm64 6620 6621This capability indicates that userspace can specify (via the 6622KVM_SET_VCPU_EVENTS ioctl) the syndrome value reported to the guest when it 6623takes a virtual SError interrupt exception. 6624If KVM advertises this capability, userspace can only specify the ISS field for 6625the ESR syndrome. Other parts of the ESR, such as the EC are generated by the 6626CPU when the exception is taken. If this virtual SError is taken to EL1 using 6627AArch64, this value will be reported in the ISS field of ESR_ELx. 6628 6629See KVM_CAP_VCPU_EVENTS for more details. 6630 66318.20 KVM_CAP_HYPERV_SEND_IPI 6632---------------------------- 6633 6634:Architectures: x86 6635 6636This capability indicates that KVM supports paravirtualized Hyper-V IPI send 6637hypercalls: 6638HvCallSendSyntheticClusterIpi, HvCallSendSyntheticClusterIpiEx. 6639 66408.21 KVM_CAP_HYPERV_DIRECT_TLBFLUSH 6641----------------------------------- 6642 6643:Architectures: x86 6644 6645This capability indicates that KVM running on top of Hyper-V hypervisor 6646enables Direct TLB flush for its guests meaning that TLB flush 6647hypercalls are handled by Level 0 hypervisor (Hyper-V) bypassing KVM. 6648Due to the different ABI for hypercall parameters between Hyper-V and 6649KVM, enabling this capability effectively disables all hypercall 6650handling by KVM (as some KVM hypercall may be mistakenly treated as TLB 6651flush hypercalls by Hyper-V) so userspace should disable KVM identification 6652in CPUID and only exposes Hyper-V identification. In this case, guest 6653thinks it's running on Hyper-V and only use Hyper-V hypercalls. 6654 66558.22 KVM_CAP_S390_VCPU_RESETS 6656----------------------------- 6657 6658:Architectures: s390 6659 6660This capability indicates that the KVM_S390_NORMAL_RESET and 6661KVM_S390_CLEAR_RESET ioctls are available. 6662 66638.23 KVM_CAP_S390_PROTECTED 6664--------------------------- 6665 6666:Architectures: s390 6667 6668This capability indicates that the Ultravisor has been initialized and 6669KVM can therefore start protected VMs. 6670This capability governs the KVM_S390_PV_COMMAND ioctl and the 6671KVM_MP_STATE_LOAD MP_STATE. KVM_SET_MP_STATE can fail for protected 6672guests when the state change is invalid. 6673 66748.24 KVM_CAP_STEAL_TIME 6675----------------------- 6676 6677:Architectures: arm64, x86 6678 6679This capability indicates that KVM supports steal time accounting. 6680When steal time accounting is supported it may be enabled with 6681architecture-specific interfaces. This capability and the architecture- 6682specific interfaces must be consistent, i.e. if one says the feature 6683is supported, than the other should as well and vice versa. For arm64 6684see Documentation/virt/kvm/devices/vcpu.rst "KVM_ARM_VCPU_PVTIME_CTRL". 6685For x86 see Documentation/virt/kvm/msr.rst "MSR_KVM_STEAL_TIME". 6686 66878.25 KVM_CAP_S390_DIAG318 6688------------------------- 6689 6690:Architectures: s390 6691 6692This capability enables a guest to set information about its control program 6693(i.e. guest kernel type and version). The information is helpful during 6694system/firmware service events, providing additional data about the guest 6695environments running on the machine. 6696 6697The information is associated with the DIAGNOSE 0x318 instruction, which sets 6698an 8-byte value consisting of a one-byte Control Program Name Code (CPNC) and 6699a 7-byte Control Program Version Code (CPVC). The CPNC determines what 6700environment the control program is running in (e.g. Linux, z/VM...), and the 6701CPVC is used for information specific to OS (e.g. Linux version, Linux 6702distribution...) 6703 6704If this capability is available, then the CPNC and CPVC can be synchronized 6705between KVM and userspace via the sync regs mechanism (KVM_SYNC_DIAG318). 6706 67078.26 KVM_CAP_X86_USER_SPACE_MSR 6708------------------------------- 6709 6710:Architectures: x86 6711 6712This capability indicates that KVM supports deflection of MSR reads and 6713writes to user space. It can be enabled on a VM level. If enabled, MSR 6714accesses that would usually trigger a #GP by KVM into the guest will 6715instead get bounced to user space through the KVM_EXIT_X86_RDMSR and 6716KVM_EXIT_X86_WRMSR exit notifications. 6717 67188.27 KVM_X86_SET_MSR_FILTER 6719--------------------------- 6720 6721:Architectures: x86 6722 6723This capability indicates that KVM supports that accesses to user defined MSRs 6724may be rejected. With this capability exposed, KVM exports new VM ioctl 6725KVM_X86_SET_MSR_FILTER which user space can call to specify bitmaps of MSR 6726ranges that KVM should reject access to. 6727 6728In combination with KVM_CAP_X86_USER_SPACE_MSR, this allows user space to 6729trap and emulate MSRs that are outside of the scope of KVM as well as 6730limit the attack surface on KVM's MSR emulation code. 6731 67328.28 KVM_CAP_ENFORCE_PV_CPUID 6733----------------------------- 6734 6735Architectures: x86 6736 6737When enabled, KVM will disable paravirtual features provided to the 6738guest according to the bits in the KVM_CPUID_FEATURES CPUID leaf 6739(0x40000001). Otherwise, a guest may use the paravirtual features 6740regardless of what has actually been exposed through the CPUID leaf. 6741 67428.29 KVM_CAP_DIRTY_LOG_RING 6743--------------------------- 6744 6745:Architectures: x86 6746:Parameters: args[0] - size of the dirty log ring 6747 6748KVM is capable of tracking dirty memory using ring buffers that are 6749mmaped into userspace; there is one dirty ring per vcpu. 6750 6751The dirty ring is available to userspace as an array of 6752``struct kvm_dirty_gfn``. Each dirty entry it's defined as:: 6753 6754 struct kvm_dirty_gfn { 6755 __u32 flags; 6756 __u32 slot; /* as_id | slot_id */ 6757 __u64 offset; 6758 }; 6759 6760The following values are defined for the flags field to define the 6761current state of the entry:: 6762 6763 #define KVM_DIRTY_GFN_F_DIRTY BIT(0) 6764 #define KVM_DIRTY_GFN_F_RESET BIT(1) 6765 #define KVM_DIRTY_GFN_F_MASK 0x3 6766 6767Userspace should call KVM_ENABLE_CAP ioctl right after KVM_CREATE_VM 6768ioctl to enable this capability for the new guest and set the size of 6769the rings. Enabling the capability is only allowed before creating any 6770vCPU, and the size of the ring must be a power of two. The larger the 6771ring buffer, the less likely the ring is full and the VM is forced to 6772exit to userspace. The optimal size depends on the workload, but it is 6773recommended that it be at least 64 KiB (4096 entries). 6774 6775Just like for dirty page bitmaps, the buffer tracks writes to 6776all user memory regions for which the KVM_MEM_LOG_DIRTY_PAGES flag was 6777set in KVM_SET_USER_MEMORY_REGION. Once a memory region is registered 6778with the flag set, userspace can start harvesting dirty pages from the 6779ring buffer. 6780 6781An entry in the ring buffer can be unused (flag bits ``00``), 6782dirty (flag bits ``01``) or harvested (flag bits ``1X``). The 6783state machine for the entry is as follows:: 6784 6785 dirtied harvested reset 6786 00 -----------> 01 -------------> 1X -------+ 6787 ^ | 6788 | | 6789 +------------------------------------------+ 6790 6791To harvest the dirty pages, userspace accesses the mmaped ring buffer 6792to read the dirty GFNs. If the flags has the DIRTY bit set (at this stage 6793the RESET bit must be cleared), then it means this GFN is a dirty GFN. 6794The userspace should harvest this GFN and mark the flags from state 6795``01b`` to ``1Xb`` (bit 0 will be ignored by KVM, but bit 1 must be set 6796to show that this GFN is harvested and waiting for a reset), and move 6797on to the next GFN. The userspace should continue to do this until the 6798flags of a GFN have the DIRTY bit cleared, meaning that it has harvested 6799all the dirty GFNs that were available. 6800 6801It's not necessary for userspace to harvest the all dirty GFNs at once. 6802However it must collect the dirty GFNs in sequence, i.e., the userspace 6803program cannot skip one dirty GFN to collect the one next to it. 6804 6805After processing one or more entries in the ring buffer, userspace 6806calls the VM ioctl KVM_RESET_DIRTY_RINGS to notify the kernel about 6807it, so that the kernel will reprotect those collected GFNs. 6808Therefore, the ioctl must be called *before* reading the content of 6809the dirty pages. 6810 6811The dirty ring can get full. When it happens, the KVM_RUN of the 6812vcpu will return with exit reason KVM_EXIT_DIRTY_LOG_FULL. 6813 6814The dirty ring interface has a major difference comparing to the 6815KVM_GET_DIRTY_LOG interface in that, when reading the dirty ring from 6816userspace, it's still possible that the kernel has not yet flushed the 6817processor's dirty page buffers into the kernel buffer (with dirty bitmaps, the 6818flushing is done by the KVM_GET_DIRTY_LOG ioctl). To achieve that, one 6819needs to kick the vcpu out of KVM_RUN using a signal. The resulting 6820vmexit ensures that all dirty GFNs are flushed to the dirty rings. 6821 6822NOTE: the capability KVM_CAP_DIRTY_LOG_RING and the corresponding 6823ioctl KVM_RESET_DIRTY_RINGS are mutual exclusive to the existing ioctls 6824KVM_GET_DIRTY_LOG and KVM_CLEAR_DIRTY_LOG. After enabling 6825KVM_CAP_DIRTY_LOG_RING with an acceptable dirty ring size, the virtual 6826machine will switch to ring-buffer dirty page tracking and further 6827KVM_GET_DIRTY_LOG or KVM_CLEAR_DIRTY_LOG ioctls will fail. 6828 68298.30 KVM_CAP_XEN_HVM 6830-------------------- 6831 6832:Architectures: x86 6833 6834This capability indicates the features that Xen supports for hosting Xen 6835PVHVM guests. Valid flags are:: 6836 6837 #define KVM_XEN_HVM_CONFIG_HYPERCALL_MSR (1 << 0) 6838 #define KVM_XEN_HVM_CONFIG_INTERCEPT_HCALL (1 << 1) 6839 #define KVM_XEN_HVM_CONFIG_SHARED_INFO (1 << 2) 6840 #define KVM_XEN_HVM_CONFIG_RUNSTATE (1 << 2) 6841 6842The KVM_XEN_HVM_CONFIG_HYPERCALL_MSR flag indicates that the KVM_XEN_HVM_CONFIG 6843ioctl is available, for the guest to set its hypercall page. 6844 6845If KVM_XEN_HVM_CONFIG_INTERCEPT_HCALL is also set, the same flag may also be 6846provided in the flags to KVM_XEN_HVM_CONFIG, without providing hypercall page 6847contents, to request that KVM generate hypercall page content automatically 6848and also enable interception of guest hypercalls with KVM_EXIT_XEN. 6849 6850The KVM_XEN_HVM_CONFIG_SHARED_INFO flag indicates the availability of the 6851KVM_XEN_HVM_SET_ATTR, KVM_XEN_HVM_GET_ATTR, KVM_XEN_VCPU_SET_ATTR and 6852KVM_XEN_VCPU_GET_ATTR ioctls, as well as the delivery of exception vectors 6853for event channel upcalls when the evtchn_upcall_pending field of a vcpu's 6854vcpu_info is set. 6855 6856The KVM_XEN_HVM_CONFIG_RUNSTATE flag indicates that the runstate-related 6857features KVM_XEN_VCPU_ATTR_TYPE_RUNSTATE_ADDR/_CURRENT/_DATA/_ADJUST are 6858supported by the KVM_XEN_VCPU_SET_ATTR/KVM_XEN_VCPU_GET_ATTR ioctls. 6859 68608.31 KVM_CAP_PPC_MULTITCE 6861------------------------- 6862 6863:Capability: KVM_CAP_PPC_MULTITCE 6864:Architectures: ppc 6865:Type: vm 6866 6867This capability means the kernel is capable of handling hypercalls 6868H_PUT_TCE_INDIRECT and H_STUFF_TCE without passing those into the user 6869space. This significantly accelerates DMA operations for PPC KVM guests. 6870User space should expect that its handlers for these hypercalls 6871are not going to be called if user space previously registered LIOBN 6872in KVM (via KVM_CREATE_SPAPR_TCE or similar calls). 6873 6874In order to enable H_PUT_TCE_INDIRECT and H_STUFF_TCE use in the guest, 6875user space might have to advertise it for the guest. For example, 6876IBM pSeries (sPAPR) guest starts using them if "hcall-multi-tce" is 6877present in the "ibm,hypertas-functions" device-tree property. 6878 6879The hypercalls mentioned above may or may not be processed successfully 6880in the kernel based fast path. If they can not be handled by the kernel, 6881they will get passed on to user space. So user space still has to have 6882an implementation for these despite the in kernel acceleration. 6883 6884This capability is always enabled. 6885 68868.32 KVM_CAP_PTP_KVM 6887-------------------- 6888 6889:Architectures: arm64 6890 6891This capability indicates that the KVM virtual PTP service is 6892supported in the host. A VMM can check whether the service is 6893available to the guest on migration. 6894