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 374Note that the Xen shared info page, if configured, shall always be assumed 375to be dirty. KVM will not explicitly mark it such. 376 3774.9 KVM_SET_MEMORY_ALIAS 378------------------------ 379 380:Capability: basic 381:Architectures: x86 382:Type: vm ioctl 383:Parameters: struct kvm_memory_alias (in) 384:Returns: 0 (success), -1 (error) 385 386This ioctl is obsolete and has been removed. 387 388 3894.10 KVM_RUN 390------------ 391 392:Capability: basic 393:Architectures: all 394:Type: vcpu ioctl 395:Parameters: none 396:Returns: 0 on success, -1 on error 397 398Errors: 399 400 ======= ============================================================== 401 EINTR an unmasked signal is pending 402 ENOEXEC the vcpu hasn't been initialized or the guest tried to execute 403 instructions from device memory (arm64) 404 ENOSYS data abort outside memslots with no syndrome info and 405 KVM_CAP_ARM_NISV_TO_USER not enabled (arm64) 406 EPERM SVE feature set but not finalized (arm64) 407 ======= ============================================================== 408 409This ioctl is used to run a guest virtual cpu. While there are no 410explicit parameters, there is an implicit parameter block that can be 411obtained by mmap()ing the vcpu fd at offset 0, with the size given by 412KVM_GET_VCPU_MMAP_SIZE. The parameter block is formatted as a 'struct 413kvm_run' (see below). 414 415 4164.11 KVM_GET_REGS 417----------------- 418 419:Capability: basic 420:Architectures: all except ARM, arm64 421:Type: vcpu ioctl 422:Parameters: struct kvm_regs (out) 423:Returns: 0 on success, -1 on error 424 425Reads the general purpose registers from the vcpu. 426 427:: 428 429 /* x86 */ 430 struct kvm_regs { 431 /* out (KVM_GET_REGS) / in (KVM_SET_REGS) */ 432 __u64 rax, rbx, rcx, rdx; 433 __u64 rsi, rdi, rsp, rbp; 434 __u64 r8, r9, r10, r11; 435 __u64 r12, r13, r14, r15; 436 __u64 rip, rflags; 437 }; 438 439 /* mips */ 440 struct kvm_regs { 441 /* out (KVM_GET_REGS) / in (KVM_SET_REGS) */ 442 __u64 gpr[32]; 443 __u64 hi; 444 __u64 lo; 445 __u64 pc; 446 }; 447 448 4494.12 KVM_SET_REGS 450----------------- 451 452:Capability: basic 453:Architectures: all except ARM, arm64 454:Type: vcpu ioctl 455:Parameters: struct kvm_regs (in) 456:Returns: 0 on success, -1 on error 457 458Writes the general purpose registers into the vcpu. 459 460See KVM_GET_REGS for the data structure. 461 462 4634.13 KVM_GET_SREGS 464------------------ 465 466:Capability: basic 467:Architectures: x86, ppc 468:Type: vcpu ioctl 469:Parameters: struct kvm_sregs (out) 470:Returns: 0 on success, -1 on error 471 472Reads special registers from the vcpu. 473 474:: 475 476 /* x86 */ 477 struct kvm_sregs { 478 struct kvm_segment cs, ds, es, fs, gs, ss; 479 struct kvm_segment tr, ldt; 480 struct kvm_dtable gdt, idt; 481 __u64 cr0, cr2, cr3, cr4, cr8; 482 __u64 efer; 483 __u64 apic_base; 484 __u64 interrupt_bitmap[(KVM_NR_INTERRUPTS + 63) / 64]; 485 }; 486 487 /* ppc -- see arch/powerpc/include/uapi/asm/kvm.h */ 488 489interrupt_bitmap is a bitmap of pending external interrupts. At most 490one bit may be set. This interrupt has been acknowledged by the APIC 491but not yet injected into the cpu core. 492 493 4944.14 KVM_SET_SREGS 495------------------ 496 497:Capability: basic 498:Architectures: x86, ppc 499:Type: vcpu ioctl 500:Parameters: struct kvm_sregs (in) 501:Returns: 0 on success, -1 on error 502 503Writes special registers into the vcpu. See KVM_GET_SREGS for the 504data structures. 505 506 5074.15 KVM_TRANSLATE 508------------------ 509 510:Capability: basic 511:Architectures: x86 512:Type: vcpu ioctl 513:Parameters: struct kvm_translation (in/out) 514:Returns: 0 on success, -1 on error 515 516Translates a virtual address according to the vcpu's current address 517translation mode. 518 519:: 520 521 struct kvm_translation { 522 /* in */ 523 __u64 linear_address; 524 525 /* out */ 526 __u64 physical_address; 527 __u8 valid; 528 __u8 writeable; 529 __u8 usermode; 530 __u8 pad[5]; 531 }; 532 533 5344.16 KVM_INTERRUPT 535------------------ 536 537:Capability: basic 538:Architectures: x86, ppc, mips, riscv 539:Type: vcpu ioctl 540:Parameters: struct kvm_interrupt (in) 541:Returns: 0 on success, negative on failure. 542 543Queues a hardware interrupt vector to be injected. 544 545:: 546 547 /* for KVM_INTERRUPT */ 548 struct kvm_interrupt { 549 /* in */ 550 __u32 irq; 551 }; 552 553X86: 554^^^^ 555 556:Returns: 557 558 ========= =================================== 559 0 on success, 560 -EEXIST if an interrupt is already enqueued 561 -EINVAL the irq number is invalid 562 -ENXIO if the PIC is in the kernel 563 -EFAULT if the pointer is invalid 564 ========= =================================== 565 566Note 'irq' is an interrupt vector, not an interrupt pin or line. This 567ioctl is useful if the in-kernel PIC is not used. 568 569PPC: 570^^^^ 571 572Queues an external interrupt to be injected. This ioctl is overleaded 573with 3 different irq values: 574 575a) KVM_INTERRUPT_SET 576 577 This injects an edge type external interrupt into the guest once it's ready 578 to receive interrupts. When injected, the interrupt is done. 579 580b) KVM_INTERRUPT_UNSET 581 582 This unsets any pending interrupt. 583 584 Only available with KVM_CAP_PPC_UNSET_IRQ. 585 586c) KVM_INTERRUPT_SET_LEVEL 587 588 This injects a level type external interrupt into the guest context. The 589 interrupt stays pending until a specific ioctl with KVM_INTERRUPT_UNSET 590 is triggered. 591 592 Only available with KVM_CAP_PPC_IRQ_LEVEL. 593 594Note that any value for 'irq' other than the ones stated above is invalid 595and incurs unexpected behavior. 596 597This is an asynchronous vcpu ioctl and can be invoked from any thread. 598 599MIPS: 600^^^^^ 601 602Queues an external interrupt to be injected into the virtual CPU. A negative 603interrupt number dequeues the interrupt. 604 605This is an asynchronous vcpu ioctl and can be invoked from any thread. 606 607RISC-V: 608^^^^^^^ 609 610Queues an external interrupt to be injected into the virutal CPU. This ioctl 611is overloaded with 2 different irq values: 612 613a) KVM_INTERRUPT_SET 614 615 This sets external interrupt for a virtual CPU and it will receive 616 once it is ready. 617 618b) KVM_INTERRUPT_UNSET 619 620 This clears pending external interrupt for a virtual CPU. 621 622This is an asynchronous vcpu ioctl and can be invoked from any thread. 623 624 6254.17 KVM_DEBUG_GUEST 626-------------------- 627 628:Capability: basic 629:Architectures: none 630:Type: vcpu ioctl 631:Parameters: none) 632:Returns: -1 on error 633 634Support for this has been removed. Use KVM_SET_GUEST_DEBUG instead. 635 636 6374.18 KVM_GET_MSRS 638----------------- 639 640:Capability: basic (vcpu), KVM_CAP_GET_MSR_FEATURES (system) 641:Architectures: x86 642:Type: system ioctl, vcpu ioctl 643:Parameters: struct kvm_msrs (in/out) 644:Returns: number of msrs successfully returned; 645 -1 on error 646 647When used as a system ioctl: 648Reads the values of MSR-based features that are available for the VM. This 649is similar to KVM_GET_SUPPORTED_CPUID, but it returns MSR indices and values. 650The list of msr-based features can be obtained using KVM_GET_MSR_FEATURE_INDEX_LIST 651in a system ioctl. 652 653When used as a vcpu ioctl: 654Reads model-specific registers from the vcpu. Supported msr indices can 655be obtained using KVM_GET_MSR_INDEX_LIST in a system ioctl. 656 657:: 658 659 struct kvm_msrs { 660 __u32 nmsrs; /* number of msrs in entries */ 661 __u32 pad; 662 663 struct kvm_msr_entry entries[0]; 664 }; 665 666 struct kvm_msr_entry { 667 __u32 index; 668 __u32 reserved; 669 __u64 data; 670 }; 671 672Application code should set the 'nmsrs' member (which indicates the 673size of the entries array) and the 'index' member of each array entry. 674kvm will fill in the 'data' member. 675 676 6774.19 KVM_SET_MSRS 678----------------- 679 680:Capability: basic 681:Architectures: x86 682:Type: vcpu ioctl 683:Parameters: struct kvm_msrs (in) 684:Returns: number of msrs successfully set (see below), -1 on error 685 686Writes model-specific registers to the vcpu. See KVM_GET_MSRS for the 687data structures. 688 689Application code should set the 'nmsrs' member (which indicates the 690size of the entries array), and the 'index' and 'data' members of each 691array entry. 692 693It tries to set the MSRs in array entries[] one by one. If setting an MSR 694fails, e.g., due to setting reserved bits, the MSR isn't supported/emulated 695by KVM, etc..., it stops processing the MSR list and returns the number of 696MSRs that have been set successfully. 697 698 6994.20 KVM_SET_CPUID 700------------------ 701 702:Capability: basic 703:Architectures: x86 704:Type: vcpu ioctl 705:Parameters: struct kvm_cpuid (in) 706:Returns: 0 on success, -1 on error 707 708Defines the vcpu responses to the cpuid instruction. Applications 709should use the KVM_SET_CPUID2 ioctl if available. 710 711Caveat emptor: 712 - If this IOCTL fails, KVM gives no guarantees that previous valid CPUID 713 configuration (if there is) is not corrupted. Userspace can get a copy 714 of the resulting CPUID configuration through KVM_GET_CPUID2 in case. 715 - Using KVM_SET_CPUID{,2} after KVM_RUN, i.e. changing the guest vCPU model 716 after running the guest, may cause guest instability. 717 - Using heterogeneous CPUID configurations, modulo APIC IDs, topology, etc... 718 may cause guest instability. 719 720:: 721 722 struct kvm_cpuid_entry { 723 __u32 function; 724 __u32 eax; 725 __u32 ebx; 726 __u32 ecx; 727 __u32 edx; 728 __u32 padding; 729 }; 730 731 /* for KVM_SET_CPUID */ 732 struct kvm_cpuid { 733 __u32 nent; 734 __u32 padding; 735 struct kvm_cpuid_entry entries[0]; 736 }; 737 738 7394.21 KVM_SET_SIGNAL_MASK 740------------------------ 741 742:Capability: basic 743:Architectures: all 744:Type: vcpu ioctl 745:Parameters: struct kvm_signal_mask (in) 746:Returns: 0 on success, -1 on error 747 748Defines which signals are blocked during execution of KVM_RUN. This 749signal mask temporarily overrides the threads signal mask. Any 750unblocked signal received (except SIGKILL and SIGSTOP, which retain 751their traditional behaviour) will cause KVM_RUN to return with -EINTR. 752 753Note the signal will only be delivered if not blocked by the original 754signal mask. 755 756:: 757 758 /* for KVM_SET_SIGNAL_MASK */ 759 struct kvm_signal_mask { 760 __u32 len; 761 __u8 sigset[0]; 762 }; 763 764 7654.22 KVM_GET_FPU 766---------------- 767 768:Capability: basic 769:Architectures: x86 770:Type: vcpu ioctl 771:Parameters: struct kvm_fpu (out) 772:Returns: 0 on success, -1 on error 773 774Reads the floating point state from the vcpu. 775 776:: 777 778 /* for KVM_GET_FPU and KVM_SET_FPU */ 779 struct kvm_fpu { 780 __u8 fpr[8][16]; 781 __u16 fcw; 782 __u16 fsw; 783 __u8 ftwx; /* in fxsave format */ 784 __u8 pad1; 785 __u16 last_opcode; 786 __u64 last_ip; 787 __u64 last_dp; 788 __u8 xmm[16][16]; 789 __u32 mxcsr; 790 __u32 pad2; 791 }; 792 793 7944.23 KVM_SET_FPU 795---------------- 796 797:Capability: basic 798:Architectures: x86 799:Type: vcpu ioctl 800:Parameters: struct kvm_fpu (in) 801:Returns: 0 on success, -1 on error 802 803Writes the floating point state to the vcpu. 804 805:: 806 807 /* for KVM_GET_FPU and KVM_SET_FPU */ 808 struct kvm_fpu { 809 __u8 fpr[8][16]; 810 __u16 fcw; 811 __u16 fsw; 812 __u8 ftwx; /* in fxsave format */ 813 __u8 pad1; 814 __u16 last_opcode; 815 __u64 last_ip; 816 __u64 last_dp; 817 __u8 xmm[16][16]; 818 __u32 mxcsr; 819 __u32 pad2; 820 }; 821 822 8234.24 KVM_CREATE_IRQCHIP 824----------------------- 825 826:Capability: KVM_CAP_IRQCHIP, KVM_CAP_S390_IRQCHIP (s390) 827:Architectures: x86, ARM, arm64, s390 828:Type: vm ioctl 829:Parameters: none 830:Returns: 0 on success, -1 on error 831 832Creates an interrupt controller model in the kernel. 833On x86, creates a virtual ioapic, a virtual PIC (two PICs, nested), and sets up 834future vcpus to have a local APIC. IRQ routing for GSIs 0-15 is set to both 835PIC and IOAPIC; GSI 16-23 only go to the IOAPIC. 836On ARM/arm64, a GICv2 is created. Any other GIC versions require the usage of 837KVM_CREATE_DEVICE, which also supports creating a GICv2. Using 838KVM_CREATE_DEVICE is preferred over KVM_CREATE_IRQCHIP for GICv2. 839On s390, a dummy irq routing table is created. 840 841Note that on s390 the KVM_CAP_S390_IRQCHIP vm capability needs to be enabled 842before KVM_CREATE_IRQCHIP can be used. 843 844 8454.25 KVM_IRQ_LINE 846----------------- 847 848:Capability: KVM_CAP_IRQCHIP 849:Architectures: x86, arm, arm64 850:Type: vm ioctl 851:Parameters: struct kvm_irq_level 852:Returns: 0 on success, -1 on error 853 854Sets the level of a GSI input to the interrupt controller model in the kernel. 855On some architectures it is required that an interrupt controller model has 856been previously created with KVM_CREATE_IRQCHIP. Note that edge-triggered 857interrupts require the level to be set to 1 and then back to 0. 858 859On real hardware, interrupt pins can be active-low or active-high. This 860does not matter for the level field of struct kvm_irq_level: 1 always 861means active (asserted), 0 means inactive (deasserted). 862 863x86 allows the operating system to program the interrupt polarity 864(active-low/active-high) for level-triggered interrupts, and KVM used 865to consider the polarity. However, due to bitrot in the handling of 866active-low interrupts, the above convention is now valid on x86 too. 867This is signaled by KVM_CAP_X86_IOAPIC_POLARITY_IGNORED. Userspace 868should not present interrupts to the guest as active-low unless this 869capability is present (or unless it is not using the in-kernel irqchip, 870of course). 871 872 873ARM/arm64 can signal an interrupt either at the CPU level, or at the 874in-kernel irqchip (GIC), and for in-kernel irqchip can tell the GIC to 875use PPIs designated for specific cpus. The irq field is interpreted 876like this:: 877 878 bits: | 31 ... 28 | 27 ... 24 | 23 ... 16 | 15 ... 0 | 879 field: | vcpu2_index | irq_type | vcpu_index | irq_id | 880 881The irq_type field has the following values: 882 883- irq_type[0]: 884 out-of-kernel GIC: irq_id 0 is IRQ, irq_id 1 is FIQ 885- irq_type[1]: 886 in-kernel GIC: SPI, irq_id between 32 and 1019 (incl.) 887 (the vcpu_index field is ignored) 888- irq_type[2]: 889 in-kernel GIC: PPI, irq_id between 16 and 31 (incl.) 890 891(The irq_id field thus corresponds nicely to the IRQ ID in the ARM GIC specs) 892 893In both cases, level is used to assert/deassert the line. 894 895When KVM_CAP_ARM_IRQ_LINE_LAYOUT_2 is supported, the target vcpu is 896identified as (256 * vcpu2_index + vcpu_index). Otherwise, vcpu2_index 897must be zero. 898 899Note that on arm/arm64, the KVM_CAP_IRQCHIP capability only conditions 900injection of interrupts for the in-kernel irqchip. KVM_IRQ_LINE can always 901be used for a userspace interrupt controller. 902 903:: 904 905 struct kvm_irq_level { 906 union { 907 __u32 irq; /* GSI */ 908 __s32 status; /* not used for KVM_IRQ_LEVEL */ 909 }; 910 __u32 level; /* 0 or 1 */ 911 }; 912 913 9144.26 KVM_GET_IRQCHIP 915-------------------- 916 917:Capability: KVM_CAP_IRQCHIP 918:Architectures: x86 919:Type: vm ioctl 920:Parameters: struct kvm_irqchip (in/out) 921:Returns: 0 on success, -1 on error 922 923Reads the state of a kernel interrupt controller created with 924KVM_CREATE_IRQCHIP into 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 9394.27 KVM_SET_IRQCHIP 940-------------------- 941 942:Capability: KVM_CAP_IRQCHIP 943:Architectures: x86 944:Type: vm ioctl 945:Parameters: struct kvm_irqchip (in) 946:Returns: 0 on success, -1 on error 947 948Sets the state of a kernel interrupt controller created with 949KVM_CREATE_IRQCHIP from a buffer provided by the caller. 950 951:: 952 953 struct kvm_irqchip { 954 __u32 chip_id; /* 0 = PIC1, 1 = PIC2, 2 = IOAPIC */ 955 __u32 pad; 956 union { 957 char dummy[512]; /* reserving space */ 958 struct kvm_pic_state pic; 959 struct kvm_ioapic_state ioapic; 960 } chip; 961 }; 962 963 9644.28 KVM_XEN_HVM_CONFIG 965----------------------- 966 967:Capability: KVM_CAP_XEN_HVM 968:Architectures: x86 969:Type: vm ioctl 970:Parameters: struct kvm_xen_hvm_config (in) 971:Returns: 0 on success, -1 on error 972 973Sets the MSR that the Xen HVM guest uses to initialize its hypercall 974page, and provides the starting address and size of the hypercall 975blobs in userspace. When the guest writes the MSR, kvm copies one 976page of a blob (32- or 64-bit, depending on the vcpu mode) to guest 977memory. 978 979:: 980 981 struct kvm_xen_hvm_config { 982 __u32 flags; 983 __u32 msr; 984 __u64 blob_addr_32; 985 __u64 blob_addr_64; 986 __u8 blob_size_32; 987 __u8 blob_size_64; 988 __u8 pad2[30]; 989 }; 990 991If the KVM_XEN_HVM_CONFIG_INTERCEPT_HCALL flag is returned from the 992KVM_CAP_XEN_HVM check, it may be set in the flags field of this ioctl. 993This requests KVM to generate the contents of the hypercall page 994automatically; hypercalls will be intercepted and passed to userspace 995through KVM_EXIT_XEN. In this case, all of the blob size and address 996fields must be zero. 997 998No other flags are currently valid in the struct kvm_xen_hvm_config. 999 10004.29 KVM_GET_CLOCK 1001------------------ 1002 1003:Capability: KVM_CAP_ADJUST_CLOCK 1004:Architectures: x86 1005:Type: vm ioctl 1006:Parameters: struct kvm_clock_data (out) 1007:Returns: 0 on success, -1 on error 1008 1009Gets the current timestamp of kvmclock as seen by the current guest. In 1010conjunction with KVM_SET_CLOCK, it is used to ensure monotonicity on scenarios 1011such as migration. 1012 1013When KVM_CAP_ADJUST_CLOCK is passed to KVM_CHECK_EXTENSION, it returns the 1014set of bits that KVM can return in struct kvm_clock_data's flag member. 1015 1016The following flags are defined: 1017 1018KVM_CLOCK_TSC_STABLE 1019 If set, the returned value is the exact kvmclock 1020 value seen by all VCPUs at the instant when KVM_GET_CLOCK was called. 1021 If clear, the returned value is simply CLOCK_MONOTONIC plus a constant 1022 offset; the offset can be modified with KVM_SET_CLOCK. KVM will try 1023 to make all VCPUs follow this clock, but the exact value read by each 1024 VCPU could differ, because the host TSC is not stable. 1025 1026KVM_CLOCK_REALTIME 1027 If set, the `realtime` field in the kvm_clock_data 1028 structure is populated with the value of the host's real time 1029 clocksource at the instant when KVM_GET_CLOCK was called. If clear, 1030 the `realtime` field does not contain a value. 1031 1032KVM_CLOCK_HOST_TSC 1033 If set, the `host_tsc` field in the kvm_clock_data 1034 structure is populated with the value of the host's timestamp counter (TSC) 1035 at the instant when KVM_GET_CLOCK was called. If clear, the `host_tsc` field 1036 does not contain a value. 1037 1038:: 1039 1040 struct kvm_clock_data { 1041 __u64 clock; /* kvmclock current value */ 1042 __u32 flags; 1043 __u32 pad0; 1044 __u64 realtime; 1045 __u64 host_tsc; 1046 __u32 pad[4]; 1047 }; 1048 1049 10504.30 KVM_SET_CLOCK 1051------------------ 1052 1053:Capability: KVM_CAP_ADJUST_CLOCK 1054:Architectures: x86 1055:Type: vm ioctl 1056:Parameters: struct kvm_clock_data (in) 1057:Returns: 0 on success, -1 on error 1058 1059Sets the current timestamp of kvmclock to the value specified in its parameter. 1060In conjunction with KVM_GET_CLOCK, it is used to ensure monotonicity on scenarios 1061such as migration. 1062 1063The following flags can be passed: 1064 1065KVM_CLOCK_REALTIME 1066 If set, KVM will compare the value of the `realtime` field 1067 with the value of the host's real time clocksource at the instant when 1068 KVM_SET_CLOCK was called. The difference in elapsed time is added to the final 1069 kvmclock value that will be provided to guests. 1070 1071Other flags returned by ``KVM_GET_CLOCK`` are accepted but ignored. 1072 1073:: 1074 1075 struct kvm_clock_data { 1076 __u64 clock; /* kvmclock current value */ 1077 __u32 flags; 1078 __u32 pad0; 1079 __u64 realtime; 1080 __u64 host_tsc; 1081 __u32 pad[4]; 1082 }; 1083 1084 10854.31 KVM_GET_VCPU_EVENTS 1086------------------------ 1087 1088:Capability: KVM_CAP_VCPU_EVENTS 1089:Extended by: KVM_CAP_INTR_SHADOW 1090:Architectures: x86, arm, arm64 1091:Type: vcpu ioctl 1092:Parameters: struct kvm_vcpu_event (out) 1093:Returns: 0 on success, -1 on error 1094 1095X86: 1096^^^^ 1097 1098Gets currently pending exceptions, interrupts, and NMIs as well as related 1099states of the vcpu. 1100 1101:: 1102 1103 struct kvm_vcpu_events { 1104 struct { 1105 __u8 injected; 1106 __u8 nr; 1107 __u8 has_error_code; 1108 __u8 pending; 1109 __u32 error_code; 1110 } exception; 1111 struct { 1112 __u8 injected; 1113 __u8 nr; 1114 __u8 soft; 1115 __u8 shadow; 1116 } interrupt; 1117 struct { 1118 __u8 injected; 1119 __u8 pending; 1120 __u8 masked; 1121 __u8 pad; 1122 } nmi; 1123 __u32 sipi_vector; 1124 __u32 flags; 1125 struct { 1126 __u8 smm; 1127 __u8 pending; 1128 __u8 smm_inside_nmi; 1129 __u8 latched_init; 1130 } smi; 1131 __u8 reserved[27]; 1132 __u8 exception_has_payload; 1133 __u64 exception_payload; 1134 }; 1135 1136The following bits are defined in the flags field: 1137 1138- KVM_VCPUEVENT_VALID_SHADOW may be set to signal that 1139 interrupt.shadow contains a valid state. 1140 1141- KVM_VCPUEVENT_VALID_SMM may be set to signal that smi contains a 1142 valid state. 1143 1144- KVM_VCPUEVENT_VALID_PAYLOAD may be set to signal that the 1145 exception_has_payload, exception_payload, and exception.pending 1146 fields contain a valid state. This bit will be set whenever 1147 KVM_CAP_EXCEPTION_PAYLOAD is enabled. 1148 1149ARM/ARM64: 1150^^^^^^^^^^ 1151 1152If the guest accesses a device that is being emulated by the host kernel in 1153such a way that a real device would generate a physical SError, KVM may make 1154a virtual SError pending for that VCPU. This system error interrupt remains 1155pending until the guest takes the exception by unmasking PSTATE.A. 1156 1157Running the VCPU may cause it to take a pending SError, or make an access that 1158causes an SError to become pending. The event's description is only valid while 1159the VPCU is not running. 1160 1161This API provides a way to read and write the pending 'event' state that is not 1162visible to the guest. To save, restore or migrate a VCPU the struct representing 1163the state can be read then written using this GET/SET API, along with the other 1164guest-visible registers. It is not possible to 'cancel' an SError that has been 1165made pending. 1166 1167A device being emulated in user-space may also wish to generate an SError. To do 1168this the events structure can be populated by user-space. The current state 1169should be read first, to ensure no existing SError is pending. If an existing 1170SError is pending, the architecture's 'Multiple SError interrupts' rules should 1171be followed. (2.5.3 of DDI0587.a "ARM Reliability, Availability, and 1172Serviceability (RAS) Specification"). 1173 1174SError exceptions always have an ESR value. Some CPUs have the ability to 1175specify what the virtual SError's ESR value should be. These systems will 1176advertise KVM_CAP_ARM_INJECT_SERROR_ESR. In this case exception.has_esr will 1177always have a non-zero value when read, and the agent making an SError pending 1178should specify the ISS field in the lower 24 bits of exception.serror_esr. If 1179the system supports KVM_CAP_ARM_INJECT_SERROR_ESR, but user-space sets the events 1180with exception.has_esr as zero, KVM will choose an ESR. 1181 1182Specifying exception.has_esr on a system that does not support it will return 1183-EINVAL. Setting anything other than the lower 24bits of exception.serror_esr 1184will return -EINVAL. 1185 1186It is not possible to read back a pending external abort (injected via 1187KVM_SET_VCPU_EVENTS or otherwise) because such an exception is always delivered 1188directly to the virtual CPU). 1189 1190:: 1191 1192 struct kvm_vcpu_events { 1193 struct { 1194 __u8 serror_pending; 1195 __u8 serror_has_esr; 1196 __u8 ext_dabt_pending; 1197 /* Align it to 8 bytes */ 1198 __u8 pad[5]; 1199 __u64 serror_esr; 1200 } exception; 1201 __u32 reserved[12]; 1202 }; 1203 12044.32 KVM_SET_VCPU_EVENTS 1205------------------------ 1206 1207:Capability: KVM_CAP_VCPU_EVENTS 1208:Extended by: KVM_CAP_INTR_SHADOW 1209:Architectures: x86, arm, arm64 1210:Type: vcpu ioctl 1211:Parameters: struct kvm_vcpu_event (in) 1212:Returns: 0 on success, -1 on error 1213 1214X86: 1215^^^^ 1216 1217Set pending exceptions, interrupts, and NMIs as well as related states of the 1218vcpu. 1219 1220See KVM_GET_VCPU_EVENTS for the data structure. 1221 1222Fields that may be modified asynchronously by running VCPUs can be excluded 1223from the update. These fields are nmi.pending, sipi_vector, smi.smm, 1224smi.pending. Keep the corresponding bits in the flags field cleared to 1225suppress overwriting the current in-kernel state. The bits are: 1226 1227=============================== ================================== 1228KVM_VCPUEVENT_VALID_NMI_PENDING transfer nmi.pending to the kernel 1229KVM_VCPUEVENT_VALID_SIPI_VECTOR transfer sipi_vector 1230KVM_VCPUEVENT_VALID_SMM transfer the smi sub-struct. 1231=============================== ================================== 1232 1233If KVM_CAP_INTR_SHADOW is available, KVM_VCPUEVENT_VALID_SHADOW can be set in 1234the flags field to signal that interrupt.shadow contains a valid state and 1235shall be written into the VCPU. 1236 1237KVM_VCPUEVENT_VALID_SMM can only be set if KVM_CAP_X86_SMM is available. 1238 1239If KVM_CAP_EXCEPTION_PAYLOAD is enabled, KVM_VCPUEVENT_VALID_PAYLOAD 1240can be set in the flags field to signal that the 1241exception_has_payload, exception_payload, and exception.pending fields 1242contain a valid state and shall be written into the VCPU. 1243 1244ARM/ARM64: 1245^^^^^^^^^^ 1246 1247User space may need to inject several types of events to the guest. 1248 1249Set the pending SError exception state for this VCPU. It is not possible to 1250'cancel' an Serror that has been made pending. 1251 1252If the guest performed an access to I/O memory which could not be handled by 1253userspace, for example because of missing instruction syndrome decode 1254information or because there is no device mapped at the accessed IPA, then 1255userspace can ask the kernel to inject an external abort using the address 1256from the exiting fault on the VCPU. It is a programming error to set 1257ext_dabt_pending after an exit which was not either KVM_EXIT_MMIO or 1258KVM_EXIT_ARM_NISV. This feature is only available if the system supports 1259KVM_CAP_ARM_INJECT_EXT_DABT. This is a helper which provides commonality in 1260how userspace reports accesses for the above cases to guests, across different 1261userspace implementations. Nevertheless, userspace can still emulate all Arm 1262exceptions by manipulating individual registers using the KVM_SET_ONE_REG API. 1263 1264See KVM_GET_VCPU_EVENTS for the data structure. 1265 1266 12674.33 KVM_GET_DEBUGREGS 1268---------------------- 1269 1270:Capability: KVM_CAP_DEBUGREGS 1271:Architectures: x86 1272:Type: vm ioctl 1273:Parameters: struct kvm_debugregs (out) 1274:Returns: 0 on success, -1 on error 1275 1276Reads debug registers from the vcpu. 1277 1278:: 1279 1280 struct kvm_debugregs { 1281 __u64 db[4]; 1282 __u64 dr6; 1283 __u64 dr7; 1284 __u64 flags; 1285 __u64 reserved[9]; 1286 }; 1287 1288 12894.34 KVM_SET_DEBUGREGS 1290---------------------- 1291 1292:Capability: KVM_CAP_DEBUGREGS 1293:Architectures: x86 1294:Type: vm ioctl 1295:Parameters: struct kvm_debugregs (in) 1296:Returns: 0 on success, -1 on error 1297 1298Writes debug registers into the vcpu. 1299 1300See KVM_GET_DEBUGREGS for the data structure. The flags field is unused 1301yet and must be cleared on entry. 1302 1303 13044.35 KVM_SET_USER_MEMORY_REGION 1305------------------------------- 1306 1307:Capability: KVM_CAP_USER_MEMORY 1308:Architectures: all 1309:Type: vm ioctl 1310:Parameters: struct kvm_userspace_memory_region (in) 1311:Returns: 0 on success, -1 on error 1312 1313:: 1314 1315 struct kvm_userspace_memory_region { 1316 __u32 slot; 1317 __u32 flags; 1318 __u64 guest_phys_addr; 1319 __u64 memory_size; /* bytes */ 1320 __u64 userspace_addr; /* start of the userspace allocated memory */ 1321 }; 1322 1323 /* for kvm_memory_region::flags */ 1324 #define KVM_MEM_LOG_DIRTY_PAGES (1UL << 0) 1325 #define KVM_MEM_READONLY (1UL << 1) 1326 1327This ioctl allows the user to create, modify or delete a guest physical 1328memory slot. Bits 0-15 of "slot" specify the slot id and this value 1329should be less than the maximum number of user memory slots supported per 1330VM. The maximum allowed slots can be queried using KVM_CAP_NR_MEMSLOTS. 1331Slots may not overlap in guest physical address space. 1332 1333If KVM_CAP_MULTI_ADDRESS_SPACE is available, bits 16-31 of "slot" 1334specifies the address space which is being modified. They must be 1335less than the value that KVM_CHECK_EXTENSION returns for the 1336KVM_CAP_MULTI_ADDRESS_SPACE capability. Slots in separate address spaces 1337are unrelated; the restriction on overlapping slots only applies within 1338each address space. 1339 1340Deleting a slot is done by passing zero for memory_size. When changing 1341an existing slot, it may be moved in the guest physical memory space, 1342or its flags may be modified, but it may not be resized. 1343 1344Memory for the region is taken starting at the address denoted by the 1345field userspace_addr, which must point at user addressable memory for 1346the entire memory slot size. Any object may back this memory, including 1347anonymous memory, ordinary files, and hugetlbfs. 1348 1349On architectures that support a form of address tagging, userspace_addr must 1350be an untagged address. 1351 1352It is recommended that the lower 21 bits of guest_phys_addr and userspace_addr 1353be identical. This allows large pages in the guest to be backed by large 1354pages in the host. 1355 1356The flags field supports two flags: KVM_MEM_LOG_DIRTY_PAGES and 1357KVM_MEM_READONLY. The former can be set to instruct KVM to keep track of 1358writes to memory within the slot. See KVM_GET_DIRTY_LOG ioctl to know how to 1359use it. The latter can be set, if KVM_CAP_READONLY_MEM capability allows it, 1360to make a new slot read-only. In this case, writes to this memory will be 1361posted to userspace as KVM_EXIT_MMIO exits. 1362 1363When the KVM_CAP_SYNC_MMU capability is available, changes in the backing of 1364the memory region are automatically reflected into the guest. For example, an 1365mmap() that affects the region will be made visible immediately. Another 1366example is madvise(MADV_DROP). 1367 1368It is recommended to use this API instead of the KVM_SET_MEMORY_REGION ioctl. 1369The KVM_SET_MEMORY_REGION does not allow fine grained control over memory 1370allocation and is deprecated. 1371 1372 13734.36 KVM_SET_TSS_ADDR 1374--------------------- 1375 1376:Capability: KVM_CAP_SET_TSS_ADDR 1377:Architectures: x86 1378:Type: vm ioctl 1379:Parameters: unsigned long tss_address (in) 1380:Returns: 0 on success, -1 on error 1381 1382This ioctl defines the physical address of a three-page region in the guest 1383physical address space. The region must be within the first 4GB of the 1384guest physical address space and must not conflict with any memory slot 1385or any mmio address. The guest may malfunction if it accesses this memory 1386region. 1387 1388This ioctl is required on Intel-based hosts. This is needed on Intel hardware 1389because of a quirk in the virtualization implementation (see the internals 1390documentation when it pops into existence). 1391 1392 13934.37 KVM_ENABLE_CAP 1394------------------- 1395 1396:Capability: KVM_CAP_ENABLE_CAP 1397:Architectures: mips, ppc, s390 1398:Type: vcpu ioctl 1399:Parameters: struct kvm_enable_cap (in) 1400:Returns: 0 on success; -1 on error 1401 1402:Capability: KVM_CAP_ENABLE_CAP_VM 1403:Architectures: all 1404:Type: vm ioctl 1405:Parameters: struct kvm_enable_cap (in) 1406:Returns: 0 on success; -1 on error 1407 1408.. note:: 1409 1410 Not all extensions are enabled by default. Using this ioctl the application 1411 can enable an extension, making it available to the guest. 1412 1413On systems that do not support this ioctl, it always fails. On systems that 1414do support it, it only works for extensions that are supported for enablement. 1415 1416To check if a capability can be enabled, the KVM_CHECK_EXTENSION ioctl should 1417be used. 1418 1419:: 1420 1421 struct kvm_enable_cap { 1422 /* in */ 1423 __u32 cap; 1424 1425The capability that is supposed to get enabled. 1426 1427:: 1428 1429 __u32 flags; 1430 1431A bitfield indicating future enhancements. Has to be 0 for now. 1432 1433:: 1434 1435 __u64 args[4]; 1436 1437Arguments for enabling a feature. If a feature needs initial values to 1438function properly, this is the place to put them. 1439 1440:: 1441 1442 __u8 pad[64]; 1443 }; 1444 1445The vcpu ioctl should be used for vcpu-specific capabilities, the vm ioctl 1446for vm-wide capabilities. 1447 14484.38 KVM_GET_MP_STATE 1449--------------------- 1450 1451:Capability: KVM_CAP_MP_STATE 1452:Architectures: x86, s390, arm, arm64, riscv 1453:Type: vcpu ioctl 1454:Parameters: struct kvm_mp_state (out) 1455:Returns: 0 on success; -1 on error 1456 1457:: 1458 1459 struct kvm_mp_state { 1460 __u32 mp_state; 1461 }; 1462 1463Returns the vcpu's current "multiprocessing state" (though also valid on 1464uniprocessor guests). 1465 1466Possible values are: 1467 1468 ========================== =============================================== 1469 KVM_MP_STATE_RUNNABLE the vcpu is currently running 1470 [x86,arm/arm64,riscv] 1471 KVM_MP_STATE_UNINITIALIZED the vcpu is an application processor (AP) 1472 which has not yet received an INIT signal [x86] 1473 KVM_MP_STATE_INIT_RECEIVED the vcpu has received an INIT signal, and is 1474 now ready for a SIPI [x86] 1475 KVM_MP_STATE_HALTED the vcpu has executed a HLT instruction and 1476 is waiting for an interrupt [x86] 1477 KVM_MP_STATE_SIPI_RECEIVED the vcpu has just received a SIPI (vector 1478 accessible via KVM_GET_VCPU_EVENTS) [x86] 1479 KVM_MP_STATE_STOPPED the vcpu is stopped [s390,arm/arm64,riscv] 1480 KVM_MP_STATE_CHECK_STOP the vcpu is in a special error state [s390] 1481 KVM_MP_STATE_OPERATING the vcpu is operating (running or halted) 1482 [s390] 1483 KVM_MP_STATE_LOAD the vcpu is in a special load/startup state 1484 [s390] 1485 ========================== =============================================== 1486 1487On x86, this ioctl is only useful after KVM_CREATE_IRQCHIP. Without an 1488in-kernel irqchip, the multiprocessing state must be maintained by userspace on 1489these architectures. 1490 1491For arm/arm64/riscv: 1492^^^^^^^^^^^^^^^^^^^^ 1493 1494The only states that are valid are KVM_MP_STATE_STOPPED and 1495KVM_MP_STATE_RUNNABLE which reflect if the vcpu is paused or not. 1496 14974.39 KVM_SET_MP_STATE 1498--------------------- 1499 1500:Capability: KVM_CAP_MP_STATE 1501:Architectures: x86, s390, arm, arm64, riscv 1502:Type: vcpu ioctl 1503:Parameters: struct kvm_mp_state (in) 1504:Returns: 0 on success; -1 on error 1505 1506Sets the vcpu's current "multiprocessing state"; see KVM_GET_MP_STATE for 1507arguments. 1508 1509On x86, this ioctl is only useful after KVM_CREATE_IRQCHIP. Without an 1510in-kernel irqchip, the multiprocessing state must be maintained by userspace on 1511these architectures. 1512 1513For arm/arm64/riscv: 1514^^^^^^^^^^^^^^^^^^^^ 1515 1516The only states that are valid are KVM_MP_STATE_STOPPED and 1517KVM_MP_STATE_RUNNABLE which reflect if the vcpu should be paused or not. 1518 15194.40 KVM_SET_IDENTITY_MAP_ADDR 1520------------------------------ 1521 1522:Capability: KVM_CAP_SET_IDENTITY_MAP_ADDR 1523:Architectures: x86 1524:Type: vm ioctl 1525:Parameters: unsigned long identity (in) 1526:Returns: 0 on success, -1 on error 1527 1528This ioctl defines the physical address of a one-page region in the guest 1529physical address space. The region must be within the first 4GB of the 1530guest physical address space and must not conflict with any memory slot 1531or any mmio address. The guest may malfunction if it accesses this memory 1532region. 1533 1534Setting the address to 0 will result in resetting the address to its default 1535(0xfffbc000). 1536 1537This ioctl is required on Intel-based hosts. This is needed on Intel hardware 1538because of a quirk in the virtualization implementation (see the internals 1539documentation when it pops into existence). 1540 1541Fails if any VCPU has already been created. 1542 15434.41 KVM_SET_BOOT_CPU_ID 1544------------------------ 1545 1546:Capability: KVM_CAP_SET_BOOT_CPU_ID 1547:Architectures: x86 1548:Type: vm ioctl 1549:Parameters: unsigned long vcpu_id 1550:Returns: 0 on success, -1 on error 1551 1552Define which vcpu is the Bootstrap Processor (BSP). Values are the same 1553as the vcpu id in KVM_CREATE_VCPU. If this ioctl is not called, the default 1554is vcpu 0. This ioctl has to be called before vcpu creation, 1555otherwise it will return EBUSY error. 1556 1557 15584.42 KVM_GET_XSAVE 1559------------------ 1560 1561:Capability: KVM_CAP_XSAVE 1562:Architectures: x86 1563:Type: vcpu ioctl 1564:Parameters: struct kvm_xsave (out) 1565:Returns: 0 on success, -1 on error 1566 1567 1568:: 1569 1570 struct kvm_xsave { 1571 __u32 region[1024]; 1572 __u32 extra[0]; 1573 }; 1574 1575This ioctl would copy current vcpu's xsave struct to the userspace. 1576 1577 15784.43 KVM_SET_XSAVE 1579------------------ 1580 1581:Capability: KVM_CAP_XSAVE and KVM_CAP_XSAVE2 1582:Architectures: x86 1583:Type: vcpu ioctl 1584:Parameters: struct kvm_xsave (in) 1585:Returns: 0 on success, -1 on error 1586 1587:: 1588 1589 1590 struct kvm_xsave { 1591 __u32 region[1024]; 1592 __u32 extra[0]; 1593 }; 1594 1595This ioctl would copy userspace's xsave struct to the kernel. It copies 1596as many bytes as are returned by KVM_CHECK_EXTENSION(KVM_CAP_XSAVE2), 1597when invoked on the vm file descriptor. The size value returned by 1598KVM_CHECK_EXTENSION(KVM_CAP_XSAVE2) will always be at least 4096. 1599Currently, it is only greater than 4096 if a dynamic feature has been 1600enabled with ``arch_prctl()``, but this may change in the future. 1601 1602The offsets of the state save areas in struct kvm_xsave follow the 1603contents of CPUID leaf 0xD on the host. 1604 1605 16064.44 KVM_GET_XCRS 1607----------------- 1608 1609:Capability: KVM_CAP_XCRS 1610:Architectures: x86 1611:Type: vcpu ioctl 1612:Parameters: struct kvm_xcrs (out) 1613:Returns: 0 on success, -1 on error 1614 1615:: 1616 1617 struct kvm_xcr { 1618 __u32 xcr; 1619 __u32 reserved; 1620 __u64 value; 1621 }; 1622 1623 struct kvm_xcrs { 1624 __u32 nr_xcrs; 1625 __u32 flags; 1626 struct kvm_xcr xcrs[KVM_MAX_XCRS]; 1627 __u64 padding[16]; 1628 }; 1629 1630This ioctl would copy current vcpu's xcrs to the userspace. 1631 1632 16334.45 KVM_SET_XCRS 1634----------------- 1635 1636:Capability: KVM_CAP_XCRS 1637:Architectures: x86 1638:Type: vcpu ioctl 1639:Parameters: struct kvm_xcrs (in) 1640:Returns: 0 on success, -1 on error 1641 1642:: 1643 1644 struct kvm_xcr { 1645 __u32 xcr; 1646 __u32 reserved; 1647 __u64 value; 1648 }; 1649 1650 struct kvm_xcrs { 1651 __u32 nr_xcrs; 1652 __u32 flags; 1653 struct kvm_xcr xcrs[KVM_MAX_XCRS]; 1654 __u64 padding[16]; 1655 }; 1656 1657This ioctl would set vcpu's xcr to the value userspace specified. 1658 1659 16604.46 KVM_GET_SUPPORTED_CPUID 1661---------------------------- 1662 1663:Capability: KVM_CAP_EXT_CPUID 1664:Architectures: x86 1665:Type: system ioctl 1666:Parameters: struct kvm_cpuid2 (in/out) 1667:Returns: 0 on success, -1 on error 1668 1669:: 1670 1671 struct kvm_cpuid2 { 1672 __u32 nent; 1673 __u32 padding; 1674 struct kvm_cpuid_entry2 entries[0]; 1675 }; 1676 1677 #define KVM_CPUID_FLAG_SIGNIFCANT_INDEX BIT(0) 1678 #define KVM_CPUID_FLAG_STATEFUL_FUNC BIT(1) /* deprecated */ 1679 #define KVM_CPUID_FLAG_STATE_READ_NEXT BIT(2) /* deprecated */ 1680 1681 struct kvm_cpuid_entry2 { 1682 __u32 function; 1683 __u32 index; 1684 __u32 flags; 1685 __u32 eax; 1686 __u32 ebx; 1687 __u32 ecx; 1688 __u32 edx; 1689 __u32 padding[3]; 1690 }; 1691 1692This ioctl returns x86 cpuid features which are supported by both the 1693hardware and kvm in its default configuration. Userspace can use the 1694information returned by this ioctl to construct cpuid information (for 1695KVM_SET_CPUID2) that is consistent with hardware, kernel, and 1696userspace capabilities, and with user requirements (for example, the 1697user may wish to constrain cpuid to emulate older hardware, or for 1698feature consistency across a cluster). 1699 1700Dynamically-enabled feature bits need to be requested with 1701``arch_prctl()`` before calling this ioctl. Feature bits that have not 1702been requested are excluded from the result. 1703 1704Note that certain capabilities, such as KVM_CAP_X86_DISABLE_EXITS, may 1705expose cpuid features (e.g. MONITOR) which are not supported by kvm in 1706its default configuration. If userspace enables such capabilities, it 1707is responsible for modifying the results of this ioctl appropriately. 1708 1709Userspace invokes KVM_GET_SUPPORTED_CPUID by passing a kvm_cpuid2 structure 1710with the 'nent' field indicating the number of entries in the variable-size 1711array 'entries'. If the number of entries is too low to describe the cpu 1712capabilities, an error (E2BIG) is returned. If the number is too high, 1713the 'nent' field is adjusted and an error (ENOMEM) is returned. If the 1714number is just right, the 'nent' field is adjusted to the number of valid 1715entries in the 'entries' array, which is then filled. 1716 1717The entries returned are the host cpuid as returned by the cpuid instruction, 1718with unknown or unsupported features masked out. Some features (for example, 1719x2apic), may not be present in the host cpu, but are exposed by kvm if it can 1720emulate them efficiently. The fields in each entry are defined as follows: 1721 1722 function: 1723 the eax value used to obtain the entry 1724 1725 index: 1726 the ecx value used to obtain the entry (for entries that are 1727 affected by ecx) 1728 1729 flags: 1730 an OR of zero or more of the following: 1731 1732 KVM_CPUID_FLAG_SIGNIFCANT_INDEX: 1733 if the index field is valid 1734 1735 eax, ebx, ecx, edx: 1736 the values returned by the cpuid instruction for 1737 this function/index combination 1738 1739The TSC deadline timer feature (CPUID leaf 1, ecx[24]) is always returned 1740as false, since the feature depends on KVM_CREATE_IRQCHIP for local APIC 1741support. Instead it is reported via:: 1742 1743 ioctl(KVM_CHECK_EXTENSION, KVM_CAP_TSC_DEADLINE_TIMER) 1744 1745if that returns true and you use KVM_CREATE_IRQCHIP, or if you emulate the 1746feature in userspace, then you can enable the feature for KVM_SET_CPUID2. 1747 1748 17494.47 KVM_PPC_GET_PVINFO 1750----------------------- 1751 1752:Capability: KVM_CAP_PPC_GET_PVINFO 1753:Architectures: ppc 1754:Type: vm ioctl 1755:Parameters: struct kvm_ppc_pvinfo (out) 1756:Returns: 0 on success, !0 on error 1757 1758:: 1759 1760 struct kvm_ppc_pvinfo { 1761 __u32 flags; 1762 __u32 hcall[4]; 1763 __u8 pad[108]; 1764 }; 1765 1766This ioctl fetches PV specific information that need to be passed to the guest 1767using the device tree or other means from vm context. 1768 1769The hcall array defines 4 instructions that make up a hypercall. 1770 1771If any additional field gets added to this structure later on, a bit for that 1772additional piece of information will be set in the flags bitmap. 1773 1774The flags bitmap is defined as:: 1775 1776 /* the host supports the ePAPR idle hcall 1777 #define KVM_PPC_PVINFO_FLAGS_EV_IDLE (1<<0) 1778 17794.52 KVM_SET_GSI_ROUTING 1780------------------------ 1781 1782:Capability: KVM_CAP_IRQ_ROUTING 1783:Architectures: x86 s390 arm arm64 1784:Type: vm ioctl 1785:Parameters: struct kvm_irq_routing (in) 1786:Returns: 0 on success, -1 on error 1787 1788Sets the GSI routing table entries, overwriting any previously set entries. 1789 1790On arm/arm64, GSI routing has the following limitation: 1791 1792- GSI routing does not apply to KVM_IRQ_LINE but only to KVM_IRQFD. 1793 1794:: 1795 1796 struct kvm_irq_routing { 1797 __u32 nr; 1798 __u32 flags; 1799 struct kvm_irq_routing_entry entries[0]; 1800 }; 1801 1802No flags are specified so far, the corresponding field must be set to zero. 1803 1804:: 1805 1806 struct kvm_irq_routing_entry { 1807 __u32 gsi; 1808 __u32 type; 1809 __u32 flags; 1810 __u32 pad; 1811 union { 1812 struct kvm_irq_routing_irqchip irqchip; 1813 struct kvm_irq_routing_msi msi; 1814 struct kvm_irq_routing_s390_adapter adapter; 1815 struct kvm_irq_routing_hv_sint hv_sint; 1816 struct kvm_irq_routing_xen_evtchn xen_evtchn; 1817 __u32 pad[8]; 1818 } u; 1819 }; 1820 1821 /* gsi routing entry types */ 1822 #define KVM_IRQ_ROUTING_IRQCHIP 1 1823 #define KVM_IRQ_ROUTING_MSI 2 1824 #define KVM_IRQ_ROUTING_S390_ADAPTER 3 1825 #define KVM_IRQ_ROUTING_HV_SINT 4 1826 #define KVM_IRQ_ROUTING_XEN_EVTCHN 5 1827 1828flags: 1829 1830- KVM_MSI_VALID_DEVID: used along with KVM_IRQ_ROUTING_MSI routing entry 1831 type, specifies that the devid field contains a valid value. The per-VM 1832 KVM_CAP_MSI_DEVID capability advertises the requirement to provide 1833 the device ID. If this capability is not available, userspace should 1834 never set the KVM_MSI_VALID_DEVID flag as the ioctl might fail. 1835- zero otherwise 1836 1837:: 1838 1839 struct kvm_irq_routing_irqchip { 1840 __u32 irqchip; 1841 __u32 pin; 1842 }; 1843 1844 struct kvm_irq_routing_msi { 1845 __u32 address_lo; 1846 __u32 address_hi; 1847 __u32 data; 1848 union { 1849 __u32 pad; 1850 __u32 devid; 1851 }; 1852 }; 1853 1854If KVM_MSI_VALID_DEVID is set, devid contains a unique device identifier 1855for the device that wrote the MSI message. For PCI, this is usually a 1856BFD identifier in the lower 16 bits. 1857 1858On x86, address_hi is ignored unless the KVM_X2APIC_API_USE_32BIT_IDS 1859feature of KVM_CAP_X2APIC_API capability is enabled. If it is enabled, 1860address_hi bits 31-8 provide bits 31-8 of the destination id. Bits 7-0 of 1861address_hi must be zero. 1862 1863:: 1864 1865 struct kvm_irq_routing_s390_adapter { 1866 __u64 ind_addr; 1867 __u64 summary_addr; 1868 __u64 ind_offset; 1869 __u32 summary_offset; 1870 __u32 adapter_id; 1871 }; 1872 1873 struct kvm_irq_routing_hv_sint { 1874 __u32 vcpu; 1875 __u32 sint; 1876 }; 1877 1878 struct kvm_irq_routing_xen_evtchn { 1879 __u32 port; 1880 __u32 vcpu; 1881 __u32 priority; 1882 }; 1883 1884 1885When KVM_CAP_XEN_HVM includes the KVM_XEN_HVM_CONFIG_EVTCHN_2LEVEL bit 1886in its indication of supported features, routing to Xen event channels 1887is supported. Although the priority field is present, only the value 1888KVM_XEN_HVM_CONFIG_EVTCHN_2LEVEL is supported, which means delivery by 18892 level event channels. FIFO event channel support may be added in 1890the future. 1891 1892 18934.55 KVM_SET_TSC_KHZ 1894-------------------- 1895 1896:Capability: KVM_CAP_TSC_CONTROL 1897:Architectures: x86 1898:Type: vcpu ioctl 1899:Parameters: virtual tsc_khz 1900:Returns: 0 on success, -1 on error 1901 1902Specifies the tsc frequency for the virtual machine. The unit of the 1903frequency is KHz. 1904 1905 19064.56 KVM_GET_TSC_KHZ 1907-------------------- 1908 1909:Capability: KVM_CAP_GET_TSC_KHZ 1910:Architectures: x86 1911:Type: vcpu ioctl 1912:Parameters: none 1913:Returns: virtual tsc-khz on success, negative value on error 1914 1915Returns the tsc frequency of the guest. The unit of the return value is 1916KHz. If the host has unstable tsc this ioctl returns -EIO instead as an 1917error. 1918 1919 19204.57 KVM_GET_LAPIC 1921------------------ 1922 1923:Capability: KVM_CAP_IRQCHIP 1924:Architectures: x86 1925:Type: vcpu ioctl 1926:Parameters: struct kvm_lapic_state (out) 1927:Returns: 0 on success, -1 on error 1928 1929:: 1930 1931 #define KVM_APIC_REG_SIZE 0x400 1932 struct kvm_lapic_state { 1933 char regs[KVM_APIC_REG_SIZE]; 1934 }; 1935 1936Reads the Local APIC registers and copies them into the input argument. The 1937data format and layout are the same as documented in the architecture manual. 1938 1939If KVM_X2APIC_API_USE_32BIT_IDS feature of KVM_CAP_X2APIC_API is 1940enabled, then the format of APIC_ID register depends on the APIC mode 1941(reported by MSR_IA32_APICBASE) of its VCPU. x2APIC stores APIC ID in 1942the APIC_ID register (bytes 32-35). xAPIC only allows an 8-bit APIC ID 1943which is stored in bits 31-24 of the APIC register, or equivalently in 1944byte 35 of struct kvm_lapic_state's regs field. KVM_GET_LAPIC must then 1945be called after MSR_IA32_APICBASE has been set with KVM_SET_MSR. 1946 1947If KVM_X2APIC_API_USE_32BIT_IDS feature is disabled, struct kvm_lapic_state 1948always uses xAPIC format. 1949 1950 19514.58 KVM_SET_LAPIC 1952------------------ 1953 1954:Capability: KVM_CAP_IRQCHIP 1955:Architectures: x86 1956:Type: vcpu ioctl 1957:Parameters: struct kvm_lapic_state (in) 1958:Returns: 0 on success, -1 on error 1959 1960:: 1961 1962 #define KVM_APIC_REG_SIZE 0x400 1963 struct kvm_lapic_state { 1964 char regs[KVM_APIC_REG_SIZE]; 1965 }; 1966 1967Copies the input argument into the Local APIC registers. The data format 1968and layout are the same as documented in the architecture manual. 1969 1970The format of the APIC ID register (bytes 32-35 of struct kvm_lapic_state's 1971regs field) depends on the state of the KVM_CAP_X2APIC_API capability. 1972See the note in KVM_GET_LAPIC. 1973 1974 19754.59 KVM_IOEVENTFD 1976------------------ 1977 1978:Capability: KVM_CAP_IOEVENTFD 1979:Architectures: all 1980:Type: vm ioctl 1981:Parameters: struct kvm_ioeventfd (in) 1982:Returns: 0 on success, !0 on error 1983 1984This ioctl attaches or detaches an ioeventfd to a legal pio/mmio address 1985within the guest. A guest write in the registered address will signal the 1986provided event instead of triggering an exit. 1987 1988:: 1989 1990 struct kvm_ioeventfd { 1991 __u64 datamatch; 1992 __u64 addr; /* legal pio/mmio address */ 1993 __u32 len; /* 0, 1, 2, 4, or 8 bytes */ 1994 __s32 fd; 1995 __u32 flags; 1996 __u8 pad[36]; 1997 }; 1998 1999For the special case of virtio-ccw devices on s390, the ioevent is matched 2000to a subchannel/virtqueue tuple instead. 2001 2002The following flags are defined:: 2003 2004 #define KVM_IOEVENTFD_FLAG_DATAMATCH (1 << kvm_ioeventfd_flag_nr_datamatch) 2005 #define KVM_IOEVENTFD_FLAG_PIO (1 << kvm_ioeventfd_flag_nr_pio) 2006 #define KVM_IOEVENTFD_FLAG_DEASSIGN (1 << kvm_ioeventfd_flag_nr_deassign) 2007 #define KVM_IOEVENTFD_FLAG_VIRTIO_CCW_NOTIFY \ 2008 (1 << kvm_ioeventfd_flag_nr_virtio_ccw_notify) 2009 2010If datamatch flag is set, the event will be signaled only if the written value 2011to the registered address is equal to datamatch in struct kvm_ioeventfd. 2012 2013For virtio-ccw devices, addr contains the subchannel id and datamatch the 2014virtqueue index. 2015 2016With KVM_CAP_IOEVENTFD_ANY_LENGTH, a zero length ioeventfd is allowed, and 2017the kernel will ignore the length of guest write and may get a faster vmexit. 2018The speedup may only apply to specific architectures, but the ioeventfd will 2019work anyway. 2020 20214.60 KVM_DIRTY_TLB 2022------------------ 2023 2024:Capability: KVM_CAP_SW_TLB 2025:Architectures: ppc 2026:Type: vcpu ioctl 2027:Parameters: struct kvm_dirty_tlb (in) 2028:Returns: 0 on success, -1 on error 2029 2030:: 2031 2032 struct kvm_dirty_tlb { 2033 __u64 bitmap; 2034 __u32 num_dirty; 2035 }; 2036 2037This must be called whenever userspace has changed an entry in the shared 2038TLB, prior to calling KVM_RUN on the associated vcpu. 2039 2040The "bitmap" field is the userspace address of an array. This array 2041consists of a number of bits, equal to the total number of TLB entries as 2042determined by the last successful call to KVM_CONFIG_TLB, rounded up to the 2043nearest multiple of 64. 2044 2045Each bit corresponds to one TLB entry, ordered the same as in the shared TLB 2046array. 2047 2048The array is little-endian: the bit 0 is the least significant bit of the 2049first byte, bit 8 is the least significant bit of the second byte, etc. 2050This avoids any complications with differing word sizes. 2051 2052The "num_dirty" field is a performance hint for KVM to determine whether it 2053should skip processing the bitmap and just invalidate everything. It must 2054be set to the number of set bits in the bitmap. 2055 2056 20574.62 KVM_CREATE_SPAPR_TCE 2058------------------------- 2059 2060:Capability: KVM_CAP_SPAPR_TCE 2061:Architectures: powerpc 2062:Type: vm ioctl 2063:Parameters: struct kvm_create_spapr_tce (in) 2064:Returns: file descriptor for manipulating the created TCE table 2065 2066This creates a virtual TCE (translation control entry) table, which 2067is an IOMMU for PAPR-style virtual I/O. It is used to translate 2068logical addresses used in virtual I/O into guest physical addresses, 2069and provides a scatter/gather capability for PAPR virtual I/O. 2070 2071:: 2072 2073 /* for KVM_CAP_SPAPR_TCE */ 2074 struct kvm_create_spapr_tce { 2075 __u64 liobn; 2076 __u32 window_size; 2077 }; 2078 2079The liobn field gives the logical IO bus number for which to create a 2080TCE table. The window_size field specifies the size of the DMA window 2081which this TCE table will translate - the table will contain one 64 2082bit TCE entry for every 4kiB of the DMA window. 2083 2084When the guest issues an H_PUT_TCE hcall on a liobn for which a TCE 2085table has been created using this ioctl(), the kernel will handle it 2086in real mode, updating the TCE table. H_PUT_TCE calls for other 2087liobns will cause a vm exit and must be handled by userspace. 2088 2089The return value is a file descriptor which can be passed to mmap(2) 2090to map the created TCE table into userspace. This lets userspace read 2091the entries written by kernel-handled H_PUT_TCE calls, and also lets 2092userspace update the TCE table directly which is useful in some 2093circumstances. 2094 2095 20964.63 KVM_ALLOCATE_RMA 2097--------------------- 2098 2099:Capability: KVM_CAP_PPC_RMA 2100:Architectures: powerpc 2101:Type: vm ioctl 2102:Parameters: struct kvm_allocate_rma (out) 2103:Returns: file descriptor for mapping the allocated RMA 2104 2105This allocates a Real Mode Area (RMA) from the pool allocated at boot 2106time by the kernel. An RMA is a physically-contiguous, aligned region 2107of memory used on older POWER processors to provide the memory which 2108will be accessed by real-mode (MMU off) accesses in a KVM guest. 2109POWER processors support a set of sizes for the RMA that usually 2110includes 64MB, 128MB, 256MB and some larger powers of two. 2111 2112:: 2113 2114 /* for KVM_ALLOCATE_RMA */ 2115 struct kvm_allocate_rma { 2116 __u64 rma_size; 2117 }; 2118 2119The return value is a file descriptor which can be passed to mmap(2) 2120to map the allocated RMA into userspace. The mapped area can then be 2121passed to the KVM_SET_USER_MEMORY_REGION ioctl to establish it as the 2122RMA for a virtual machine. The size of the RMA in bytes (which is 2123fixed at host kernel boot time) is returned in the rma_size field of 2124the argument structure. 2125 2126The KVM_CAP_PPC_RMA capability is 1 or 2 if the KVM_ALLOCATE_RMA ioctl 2127is supported; 2 if the processor requires all virtual machines to have 2128an RMA, or 1 if the processor can use an RMA but doesn't require it, 2129because it supports the Virtual RMA (VRMA) facility. 2130 2131 21324.64 KVM_NMI 2133------------ 2134 2135:Capability: KVM_CAP_USER_NMI 2136:Architectures: x86 2137:Type: vcpu ioctl 2138:Parameters: none 2139:Returns: 0 on success, -1 on error 2140 2141Queues an NMI on the thread's vcpu. Note this is well defined only 2142when KVM_CREATE_IRQCHIP has not been called, since this is an interface 2143between the virtual cpu core and virtual local APIC. After KVM_CREATE_IRQCHIP 2144has been called, this interface is completely emulated within the kernel. 2145 2146To use this to emulate the LINT1 input with KVM_CREATE_IRQCHIP, use the 2147following algorithm: 2148 2149 - pause the vcpu 2150 - read the local APIC's state (KVM_GET_LAPIC) 2151 - check whether changing LINT1 will queue an NMI (see the LVT entry for LINT1) 2152 - if so, issue KVM_NMI 2153 - resume the vcpu 2154 2155Some guests configure the LINT1 NMI input to cause a panic, aiding in 2156debugging. 2157 2158 21594.65 KVM_S390_UCAS_MAP 2160---------------------- 2161 2162:Capability: KVM_CAP_S390_UCONTROL 2163:Architectures: s390 2164:Type: vcpu ioctl 2165:Parameters: struct kvm_s390_ucas_mapping (in) 2166:Returns: 0 in case of success 2167 2168The parameter is defined like this:: 2169 2170 struct kvm_s390_ucas_mapping { 2171 __u64 user_addr; 2172 __u64 vcpu_addr; 2173 __u64 length; 2174 }; 2175 2176This ioctl maps the memory at "user_addr" with the length "length" to 2177the vcpu's address space starting at "vcpu_addr". All parameters need to 2178be aligned by 1 megabyte. 2179 2180 21814.66 KVM_S390_UCAS_UNMAP 2182------------------------ 2183 2184:Capability: KVM_CAP_S390_UCONTROL 2185:Architectures: s390 2186:Type: vcpu ioctl 2187:Parameters: struct kvm_s390_ucas_mapping (in) 2188:Returns: 0 in case of success 2189 2190The parameter is defined like this:: 2191 2192 struct kvm_s390_ucas_mapping { 2193 __u64 user_addr; 2194 __u64 vcpu_addr; 2195 __u64 length; 2196 }; 2197 2198This ioctl unmaps the memory in the vcpu's address space starting at 2199"vcpu_addr" with the length "length". The field "user_addr" is ignored. 2200All parameters need to be aligned by 1 megabyte. 2201 2202 22034.67 KVM_S390_VCPU_FAULT 2204------------------------ 2205 2206:Capability: KVM_CAP_S390_UCONTROL 2207:Architectures: s390 2208:Type: vcpu ioctl 2209:Parameters: vcpu absolute address (in) 2210:Returns: 0 in case of success 2211 2212This call creates a page table entry on the virtual cpu's address space 2213(for user controlled virtual machines) or the virtual machine's address 2214space (for regular virtual machines). This only works for minor faults, 2215thus it's recommended to access subject memory page via the user page 2216table upfront. This is useful to handle validity intercepts for user 2217controlled virtual machines to fault in the virtual cpu's lowcore pages 2218prior to calling the KVM_RUN ioctl. 2219 2220 22214.68 KVM_SET_ONE_REG 2222-------------------- 2223 2224:Capability: KVM_CAP_ONE_REG 2225:Architectures: all 2226:Type: vcpu ioctl 2227:Parameters: struct kvm_one_reg (in) 2228:Returns: 0 on success, negative value on failure 2229 2230Errors: 2231 2232 ====== ============================================================ 2233 ENOENT no such register 2234 EINVAL invalid register ID, or no such register or used with VMs in 2235 protected virtualization mode on s390 2236 EPERM (arm64) register access not allowed before vcpu finalization 2237 ====== ============================================================ 2238 2239(These error codes are indicative only: do not rely on a specific error 2240code being returned in a specific situation.) 2241 2242:: 2243 2244 struct kvm_one_reg { 2245 __u64 id; 2246 __u64 addr; 2247 }; 2248 2249Using this ioctl, a single vcpu register can be set to a specific value 2250defined by user space with the passed in struct kvm_one_reg, where id 2251refers to the register identifier as described below and addr is a pointer 2252to a variable with the respective size. There can be architecture agnostic 2253and architecture specific registers. Each have their own range of operation 2254and their own constants and width. To keep track of the implemented 2255registers, find a list below: 2256 2257 ======= =============================== ============ 2258 Arch Register Width (bits) 2259 ======= =============================== ============ 2260 PPC KVM_REG_PPC_HIOR 64 2261 PPC KVM_REG_PPC_IAC1 64 2262 PPC KVM_REG_PPC_IAC2 64 2263 PPC KVM_REG_PPC_IAC3 64 2264 PPC KVM_REG_PPC_IAC4 64 2265 PPC KVM_REG_PPC_DAC1 64 2266 PPC KVM_REG_PPC_DAC2 64 2267 PPC KVM_REG_PPC_DABR 64 2268 PPC KVM_REG_PPC_DSCR 64 2269 PPC KVM_REG_PPC_PURR 64 2270 PPC KVM_REG_PPC_SPURR 64 2271 PPC KVM_REG_PPC_DAR 64 2272 PPC KVM_REG_PPC_DSISR 32 2273 PPC KVM_REG_PPC_AMR 64 2274 PPC KVM_REG_PPC_UAMOR 64 2275 PPC KVM_REG_PPC_MMCR0 64 2276 PPC KVM_REG_PPC_MMCR1 64 2277 PPC KVM_REG_PPC_MMCRA 64 2278 PPC KVM_REG_PPC_MMCR2 64 2279 PPC KVM_REG_PPC_MMCRS 64 2280 PPC KVM_REG_PPC_MMCR3 64 2281 PPC KVM_REG_PPC_SIAR 64 2282 PPC KVM_REG_PPC_SDAR 64 2283 PPC KVM_REG_PPC_SIER 64 2284 PPC KVM_REG_PPC_SIER2 64 2285 PPC KVM_REG_PPC_SIER3 64 2286 PPC KVM_REG_PPC_PMC1 32 2287 PPC KVM_REG_PPC_PMC2 32 2288 PPC KVM_REG_PPC_PMC3 32 2289 PPC KVM_REG_PPC_PMC4 32 2290 PPC KVM_REG_PPC_PMC5 32 2291 PPC KVM_REG_PPC_PMC6 32 2292 PPC KVM_REG_PPC_PMC7 32 2293 PPC KVM_REG_PPC_PMC8 32 2294 PPC KVM_REG_PPC_FPR0 64 2295 ... 2296 PPC KVM_REG_PPC_FPR31 64 2297 PPC KVM_REG_PPC_VR0 128 2298 ... 2299 PPC KVM_REG_PPC_VR31 128 2300 PPC KVM_REG_PPC_VSR0 128 2301 ... 2302 PPC KVM_REG_PPC_VSR31 128 2303 PPC KVM_REG_PPC_FPSCR 64 2304 PPC KVM_REG_PPC_VSCR 32 2305 PPC KVM_REG_PPC_VPA_ADDR 64 2306 PPC KVM_REG_PPC_VPA_SLB 128 2307 PPC KVM_REG_PPC_VPA_DTL 128 2308 PPC KVM_REG_PPC_EPCR 32 2309 PPC KVM_REG_PPC_EPR 32 2310 PPC KVM_REG_PPC_TCR 32 2311 PPC KVM_REG_PPC_TSR 32 2312 PPC KVM_REG_PPC_OR_TSR 32 2313 PPC KVM_REG_PPC_CLEAR_TSR 32 2314 PPC KVM_REG_PPC_MAS0 32 2315 PPC KVM_REG_PPC_MAS1 32 2316 PPC KVM_REG_PPC_MAS2 64 2317 PPC KVM_REG_PPC_MAS7_3 64 2318 PPC KVM_REG_PPC_MAS4 32 2319 PPC KVM_REG_PPC_MAS6 32 2320 PPC KVM_REG_PPC_MMUCFG 32 2321 PPC KVM_REG_PPC_TLB0CFG 32 2322 PPC KVM_REG_PPC_TLB1CFG 32 2323 PPC KVM_REG_PPC_TLB2CFG 32 2324 PPC KVM_REG_PPC_TLB3CFG 32 2325 PPC KVM_REG_PPC_TLB0PS 32 2326 PPC KVM_REG_PPC_TLB1PS 32 2327 PPC KVM_REG_PPC_TLB2PS 32 2328 PPC KVM_REG_PPC_TLB3PS 32 2329 PPC KVM_REG_PPC_EPTCFG 32 2330 PPC KVM_REG_PPC_ICP_STATE 64 2331 PPC KVM_REG_PPC_VP_STATE 128 2332 PPC KVM_REG_PPC_TB_OFFSET 64 2333 PPC KVM_REG_PPC_SPMC1 32 2334 PPC KVM_REG_PPC_SPMC2 32 2335 PPC KVM_REG_PPC_IAMR 64 2336 PPC KVM_REG_PPC_TFHAR 64 2337 PPC KVM_REG_PPC_TFIAR 64 2338 PPC KVM_REG_PPC_TEXASR 64 2339 PPC KVM_REG_PPC_FSCR 64 2340 PPC KVM_REG_PPC_PSPB 32 2341 PPC KVM_REG_PPC_EBBHR 64 2342 PPC KVM_REG_PPC_EBBRR 64 2343 PPC KVM_REG_PPC_BESCR 64 2344 PPC KVM_REG_PPC_TAR 64 2345 PPC KVM_REG_PPC_DPDES 64 2346 PPC KVM_REG_PPC_DAWR 64 2347 PPC KVM_REG_PPC_DAWRX 64 2348 PPC KVM_REG_PPC_CIABR 64 2349 PPC KVM_REG_PPC_IC 64 2350 PPC KVM_REG_PPC_VTB 64 2351 PPC KVM_REG_PPC_CSIGR 64 2352 PPC KVM_REG_PPC_TACR 64 2353 PPC KVM_REG_PPC_TCSCR 64 2354 PPC KVM_REG_PPC_PID 64 2355 PPC KVM_REG_PPC_ACOP 64 2356 PPC KVM_REG_PPC_VRSAVE 32 2357 PPC KVM_REG_PPC_LPCR 32 2358 PPC KVM_REG_PPC_LPCR_64 64 2359 PPC KVM_REG_PPC_PPR 64 2360 PPC KVM_REG_PPC_ARCH_COMPAT 32 2361 PPC KVM_REG_PPC_DABRX 32 2362 PPC KVM_REG_PPC_WORT 64 2363 PPC KVM_REG_PPC_SPRG9 64 2364 PPC KVM_REG_PPC_DBSR 32 2365 PPC KVM_REG_PPC_TIDR 64 2366 PPC KVM_REG_PPC_PSSCR 64 2367 PPC KVM_REG_PPC_DEC_EXPIRY 64 2368 PPC KVM_REG_PPC_PTCR 64 2369 PPC KVM_REG_PPC_DAWR1 64 2370 PPC KVM_REG_PPC_DAWRX1 64 2371 PPC KVM_REG_PPC_TM_GPR0 64 2372 ... 2373 PPC KVM_REG_PPC_TM_GPR31 64 2374 PPC KVM_REG_PPC_TM_VSR0 128 2375 ... 2376 PPC KVM_REG_PPC_TM_VSR63 128 2377 PPC KVM_REG_PPC_TM_CR 64 2378 PPC KVM_REG_PPC_TM_LR 64 2379 PPC KVM_REG_PPC_TM_CTR 64 2380 PPC KVM_REG_PPC_TM_FPSCR 64 2381 PPC KVM_REG_PPC_TM_AMR 64 2382 PPC KVM_REG_PPC_TM_PPR 64 2383 PPC KVM_REG_PPC_TM_VRSAVE 64 2384 PPC KVM_REG_PPC_TM_VSCR 32 2385 PPC KVM_REG_PPC_TM_DSCR 64 2386 PPC KVM_REG_PPC_TM_TAR 64 2387 PPC KVM_REG_PPC_TM_XER 64 2388 2389 MIPS KVM_REG_MIPS_R0 64 2390 ... 2391 MIPS KVM_REG_MIPS_R31 64 2392 MIPS KVM_REG_MIPS_HI 64 2393 MIPS KVM_REG_MIPS_LO 64 2394 MIPS KVM_REG_MIPS_PC 64 2395 MIPS KVM_REG_MIPS_CP0_INDEX 32 2396 MIPS KVM_REG_MIPS_CP0_ENTRYLO0 64 2397 MIPS KVM_REG_MIPS_CP0_ENTRYLO1 64 2398 MIPS KVM_REG_MIPS_CP0_CONTEXT 64 2399 MIPS KVM_REG_MIPS_CP0_CONTEXTCONFIG 32 2400 MIPS KVM_REG_MIPS_CP0_USERLOCAL 64 2401 MIPS KVM_REG_MIPS_CP0_XCONTEXTCONFIG 64 2402 MIPS KVM_REG_MIPS_CP0_PAGEMASK 32 2403 MIPS KVM_REG_MIPS_CP0_PAGEGRAIN 32 2404 MIPS KVM_REG_MIPS_CP0_SEGCTL0 64 2405 MIPS KVM_REG_MIPS_CP0_SEGCTL1 64 2406 MIPS KVM_REG_MIPS_CP0_SEGCTL2 64 2407 MIPS KVM_REG_MIPS_CP0_PWBASE 64 2408 MIPS KVM_REG_MIPS_CP0_PWFIELD 64 2409 MIPS KVM_REG_MIPS_CP0_PWSIZE 64 2410 MIPS KVM_REG_MIPS_CP0_WIRED 32 2411 MIPS KVM_REG_MIPS_CP0_PWCTL 32 2412 MIPS KVM_REG_MIPS_CP0_HWRENA 32 2413 MIPS KVM_REG_MIPS_CP0_BADVADDR 64 2414 MIPS KVM_REG_MIPS_CP0_BADINSTR 32 2415 MIPS KVM_REG_MIPS_CP0_BADINSTRP 32 2416 MIPS KVM_REG_MIPS_CP0_COUNT 32 2417 MIPS KVM_REG_MIPS_CP0_ENTRYHI 64 2418 MIPS KVM_REG_MIPS_CP0_COMPARE 32 2419 MIPS KVM_REG_MIPS_CP0_STATUS 32 2420 MIPS KVM_REG_MIPS_CP0_INTCTL 32 2421 MIPS KVM_REG_MIPS_CP0_CAUSE 32 2422 MIPS KVM_REG_MIPS_CP0_EPC 64 2423 MIPS KVM_REG_MIPS_CP0_PRID 32 2424 MIPS KVM_REG_MIPS_CP0_EBASE 64 2425 MIPS KVM_REG_MIPS_CP0_CONFIG 32 2426 MIPS KVM_REG_MIPS_CP0_CONFIG1 32 2427 MIPS KVM_REG_MIPS_CP0_CONFIG2 32 2428 MIPS KVM_REG_MIPS_CP0_CONFIG3 32 2429 MIPS KVM_REG_MIPS_CP0_CONFIG4 32 2430 MIPS KVM_REG_MIPS_CP0_CONFIG5 32 2431 MIPS KVM_REG_MIPS_CP0_CONFIG7 32 2432 MIPS KVM_REG_MIPS_CP0_XCONTEXT 64 2433 MIPS KVM_REG_MIPS_CP0_ERROREPC 64 2434 MIPS KVM_REG_MIPS_CP0_KSCRATCH1 64 2435 MIPS KVM_REG_MIPS_CP0_KSCRATCH2 64 2436 MIPS KVM_REG_MIPS_CP0_KSCRATCH3 64 2437 MIPS KVM_REG_MIPS_CP0_KSCRATCH4 64 2438 MIPS KVM_REG_MIPS_CP0_KSCRATCH5 64 2439 MIPS KVM_REG_MIPS_CP0_KSCRATCH6 64 2440 MIPS KVM_REG_MIPS_CP0_MAAR(0..63) 64 2441 MIPS KVM_REG_MIPS_COUNT_CTL 64 2442 MIPS KVM_REG_MIPS_COUNT_RESUME 64 2443 MIPS KVM_REG_MIPS_COUNT_HZ 64 2444 MIPS KVM_REG_MIPS_FPR_32(0..31) 32 2445 MIPS KVM_REG_MIPS_FPR_64(0..31) 64 2446 MIPS KVM_REG_MIPS_VEC_128(0..31) 128 2447 MIPS KVM_REG_MIPS_FCR_IR 32 2448 MIPS KVM_REG_MIPS_FCR_CSR 32 2449 MIPS KVM_REG_MIPS_MSA_IR 32 2450 MIPS KVM_REG_MIPS_MSA_CSR 32 2451 ======= =============================== ============ 2452 2453ARM registers are mapped using the lower 32 bits. The upper 16 of that 2454is the register group type, or coprocessor number: 2455 2456ARM core registers have the following id bit patterns:: 2457 2458 0x4020 0000 0010 <index into the kvm_regs struct:16> 2459 2460ARM 32-bit CP15 registers have the following id bit patterns:: 2461 2462 0x4020 0000 000F <zero:1> <crn:4> <crm:4> <opc1:4> <opc2:3> 2463 2464ARM 64-bit CP15 registers have the following id bit patterns:: 2465 2466 0x4030 0000 000F <zero:1> <zero:4> <crm:4> <opc1:4> <zero:3> 2467 2468ARM CCSIDR registers are demultiplexed by CSSELR value:: 2469 2470 0x4020 0000 0011 00 <csselr:8> 2471 2472ARM 32-bit VFP control registers have the following id bit patterns:: 2473 2474 0x4020 0000 0012 1 <regno:12> 2475 2476ARM 64-bit FP registers have the following id bit patterns:: 2477 2478 0x4030 0000 0012 0 <regno:12> 2479 2480ARM firmware pseudo-registers have the following bit pattern:: 2481 2482 0x4030 0000 0014 <regno:16> 2483 2484 2485arm64 registers are mapped using the lower 32 bits. The upper 16 of 2486that is the register group type, or coprocessor number: 2487 2488arm64 core/FP-SIMD registers have the following id bit patterns. Note 2489that the size of the access is variable, as the kvm_regs structure 2490contains elements ranging from 32 to 128 bits. The index is a 32bit 2491value in the kvm_regs structure seen as a 32bit array:: 2492 2493 0x60x0 0000 0010 <index into the kvm_regs struct:16> 2494 2495Specifically: 2496 2497======================= ========= ===== ======================================= 2498 Encoding Register Bits kvm_regs member 2499======================= ========= ===== ======================================= 2500 0x6030 0000 0010 0000 X0 64 regs.regs[0] 2501 0x6030 0000 0010 0002 X1 64 regs.regs[1] 2502 ... 2503 0x6030 0000 0010 003c X30 64 regs.regs[30] 2504 0x6030 0000 0010 003e SP 64 regs.sp 2505 0x6030 0000 0010 0040 PC 64 regs.pc 2506 0x6030 0000 0010 0042 PSTATE 64 regs.pstate 2507 0x6030 0000 0010 0044 SP_EL1 64 sp_el1 2508 0x6030 0000 0010 0046 ELR_EL1 64 elr_el1 2509 0x6030 0000 0010 0048 SPSR_EL1 64 spsr[KVM_SPSR_EL1] (alias SPSR_SVC) 2510 0x6030 0000 0010 004a SPSR_ABT 64 spsr[KVM_SPSR_ABT] 2511 0x6030 0000 0010 004c SPSR_UND 64 spsr[KVM_SPSR_UND] 2512 0x6030 0000 0010 004e SPSR_IRQ 64 spsr[KVM_SPSR_IRQ] 2513 0x6060 0000 0010 0050 SPSR_FIQ 64 spsr[KVM_SPSR_FIQ] 2514 0x6040 0000 0010 0054 V0 128 fp_regs.vregs[0] [1]_ 2515 0x6040 0000 0010 0058 V1 128 fp_regs.vregs[1] [1]_ 2516 ... 2517 0x6040 0000 0010 00d0 V31 128 fp_regs.vregs[31] [1]_ 2518 0x6020 0000 0010 00d4 FPSR 32 fp_regs.fpsr 2519 0x6020 0000 0010 00d5 FPCR 32 fp_regs.fpcr 2520======================= ========= ===== ======================================= 2521 2522.. [1] These encodings are not accepted for SVE-enabled vcpus. See 2523 KVM_ARM_VCPU_INIT. 2524 2525 The equivalent register content can be accessed via bits [127:0] of 2526 the corresponding SVE Zn registers instead for vcpus that have SVE 2527 enabled (see below). 2528 2529arm64 CCSIDR registers are demultiplexed by CSSELR value:: 2530 2531 0x6020 0000 0011 00 <csselr:8> 2532 2533arm64 system registers have the following id bit patterns:: 2534 2535 0x6030 0000 0013 <op0:2> <op1:3> <crn:4> <crm:4> <op2:3> 2536 2537.. warning:: 2538 2539 Two system register IDs do not follow the specified pattern. These 2540 are KVM_REG_ARM_TIMER_CVAL and KVM_REG_ARM_TIMER_CNT, which map to 2541 system registers CNTV_CVAL_EL0 and CNTVCT_EL0 respectively. These 2542 two had their values accidentally swapped, which means TIMER_CVAL is 2543 derived from the register encoding for CNTVCT_EL0 and TIMER_CNT is 2544 derived from the register encoding for CNTV_CVAL_EL0. As this is 2545 API, it must remain this way. 2546 2547arm64 firmware pseudo-registers have the following bit pattern:: 2548 2549 0x6030 0000 0014 <regno:16> 2550 2551arm64 SVE registers have the following bit patterns:: 2552 2553 0x6080 0000 0015 00 <n:5> <slice:5> Zn bits[2048*slice + 2047 : 2048*slice] 2554 0x6050 0000 0015 04 <n:4> <slice:5> Pn bits[256*slice + 255 : 256*slice] 2555 0x6050 0000 0015 060 <slice:5> FFR bits[256*slice + 255 : 256*slice] 2556 0x6060 0000 0015 ffff KVM_REG_ARM64_SVE_VLS pseudo-register 2557 2558Access to register IDs where 2048 * slice >= 128 * max_vq will fail with 2559ENOENT. max_vq is the vcpu's maximum supported vector length in 128-bit 2560quadwords: see [2]_ below. 2561 2562These registers are only accessible on vcpus for which SVE is enabled. 2563See KVM_ARM_VCPU_INIT for details. 2564 2565In addition, except for KVM_REG_ARM64_SVE_VLS, these registers are not 2566accessible until the vcpu's SVE configuration has been finalized 2567using KVM_ARM_VCPU_FINALIZE(KVM_ARM_VCPU_SVE). See KVM_ARM_VCPU_INIT 2568and KVM_ARM_VCPU_FINALIZE for more information about this procedure. 2569 2570KVM_REG_ARM64_SVE_VLS is a pseudo-register that allows the set of vector 2571lengths supported by the vcpu to be discovered and configured by 2572userspace. When transferred to or from user memory via KVM_GET_ONE_REG 2573or KVM_SET_ONE_REG, the value of this register is of type 2574__u64[KVM_ARM64_SVE_VLS_WORDS], and encodes the set of vector lengths as 2575follows:: 2576 2577 __u64 vector_lengths[KVM_ARM64_SVE_VLS_WORDS]; 2578 2579 if (vq >= SVE_VQ_MIN && vq <= SVE_VQ_MAX && 2580 ((vector_lengths[(vq - KVM_ARM64_SVE_VQ_MIN) / 64] >> 2581 ((vq - KVM_ARM64_SVE_VQ_MIN) % 64)) & 1)) 2582 /* Vector length vq * 16 bytes supported */ 2583 else 2584 /* Vector length vq * 16 bytes not supported */ 2585 2586.. [2] The maximum value vq for which the above condition is true is 2587 max_vq. This is the maximum vector length available to the guest on 2588 this vcpu, and determines which register slices are visible through 2589 this ioctl interface. 2590 2591(See Documentation/arm64/sve.rst for an explanation of the "vq" 2592nomenclature.) 2593 2594KVM_REG_ARM64_SVE_VLS is only accessible after KVM_ARM_VCPU_INIT. 2595KVM_ARM_VCPU_INIT initialises it to the best set of vector lengths that 2596the host supports. 2597 2598Userspace may subsequently modify it if desired until the vcpu's SVE 2599configuration is finalized using KVM_ARM_VCPU_FINALIZE(KVM_ARM_VCPU_SVE). 2600 2601Apart from simply removing all vector lengths from the host set that 2602exceed some value, support for arbitrarily chosen sets of vector lengths 2603is hardware-dependent and may not be available. Attempting to configure 2604an invalid set of vector lengths via KVM_SET_ONE_REG will fail with 2605EINVAL. 2606 2607After the vcpu's SVE configuration is finalized, further attempts to 2608write this register will fail with EPERM. 2609 2610 2611MIPS registers are mapped using the lower 32 bits. The upper 16 of that is 2612the register group type: 2613 2614MIPS core registers (see above) have the following id bit patterns:: 2615 2616 0x7030 0000 0000 <reg:16> 2617 2618MIPS CP0 registers (see KVM_REG_MIPS_CP0_* above) have the following id bit 2619patterns depending on whether they're 32-bit or 64-bit registers:: 2620 2621 0x7020 0000 0001 00 <reg:5> <sel:3> (32-bit) 2622 0x7030 0000 0001 00 <reg:5> <sel:3> (64-bit) 2623 2624Note: KVM_REG_MIPS_CP0_ENTRYLO0 and KVM_REG_MIPS_CP0_ENTRYLO1 are the MIPS64 2625versions of the EntryLo registers regardless of the word size of the host 2626hardware, host kernel, guest, and whether XPA is present in the guest, i.e. 2627with the RI and XI bits (if they exist) in bits 63 and 62 respectively, and 2628the PFNX field starting at bit 30. 2629 2630MIPS MAARs (see KVM_REG_MIPS_CP0_MAAR(*) above) have the following id bit 2631patterns:: 2632 2633 0x7030 0000 0001 01 <reg:8> 2634 2635MIPS KVM control registers (see above) have the following id bit patterns:: 2636 2637 0x7030 0000 0002 <reg:16> 2638 2639MIPS FPU registers (see KVM_REG_MIPS_FPR_{32,64}() above) have the following 2640id bit patterns depending on the size of the register being accessed. They are 2641always accessed according to the current guest FPU mode (Status.FR and 2642Config5.FRE), i.e. as the guest would see them, and they become unpredictable 2643if the guest FPU mode is changed. MIPS SIMD Architecture (MSA) vector 2644registers (see KVM_REG_MIPS_VEC_128() above) have similar patterns as they 2645overlap the FPU registers:: 2646 2647 0x7020 0000 0003 00 <0:3> <reg:5> (32-bit FPU registers) 2648 0x7030 0000 0003 00 <0:3> <reg:5> (64-bit FPU registers) 2649 0x7040 0000 0003 00 <0:3> <reg:5> (128-bit MSA vector registers) 2650 2651MIPS FPU control registers (see KVM_REG_MIPS_FCR_{IR,CSR} above) have the 2652following id bit patterns:: 2653 2654 0x7020 0000 0003 01 <0:3> <reg:5> 2655 2656MIPS MSA control registers (see KVM_REG_MIPS_MSA_{IR,CSR} above) have the 2657following id bit patterns:: 2658 2659 0x7020 0000 0003 02 <0:3> <reg:5> 2660 2661RISC-V registers are mapped using the lower 32 bits. The upper 8 bits of 2662that is the register group type. 2663 2664RISC-V config registers are meant for configuring a Guest VCPU and it has 2665the following id bit patterns:: 2666 2667 0x8020 0000 01 <index into the kvm_riscv_config struct:24> (32bit Host) 2668 0x8030 0000 01 <index into the kvm_riscv_config struct:24> (64bit Host) 2669 2670Following are the RISC-V config registers: 2671 2672======================= ========= ============================================= 2673 Encoding Register Description 2674======================= ========= ============================================= 2675 0x80x0 0000 0100 0000 isa ISA feature bitmap of Guest VCPU 2676======================= ========= ============================================= 2677 2678The isa config register can be read anytime but can only be written before 2679a Guest VCPU runs. It will have ISA feature bits matching underlying host 2680set by default. 2681 2682RISC-V core registers represent the general excution state of a Guest VCPU 2683and it has the following id bit patterns:: 2684 2685 0x8020 0000 02 <index into the kvm_riscv_core struct:24> (32bit Host) 2686 0x8030 0000 02 <index into the kvm_riscv_core struct:24> (64bit Host) 2687 2688Following are the RISC-V core registers: 2689 2690======================= ========= ============================================= 2691 Encoding Register Description 2692======================= ========= ============================================= 2693 0x80x0 0000 0200 0000 regs.pc Program counter 2694 0x80x0 0000 0200 0001 regs.ra Return address 2695 0x80x0 0000 0200 0002 regs.sp Stack pointer 2696 0x80x0 0000 0200 0003 regs.gp Global pointer 2697 0x80x0 0000 0200 0004 regs.tp Task pointer 2698 0x80x0 0000 0200 0005 regs.t0 Caller saved register 0 2699 0x80x0 0000 0200 0006 regs.t1 Caller saved register 1 2700 0x80x0 0000 0200 0007 regs.t2 Caller saved register 2 2701 0x80x0 0000 0200 0008 regs.s0 Callee saved register 0 2702 0x80x0 0000 0200 0009 regs.s1 Callee saved register 1 2703 0x80x0 0000 0200 000a regs.a0 Function argument (or return value) 0 2704 0x80x0 0000 0200 000b regs.a1 Function argument (or return value) 1 2705 0x80x0 0000 0200 000c regs.a2 Function argument 2 2706 0x80x0 0000 0200 000d regs.a3 Function argument 3 2707 0x80x0 0000 0200 000e regs.a4 Function argument 4 2708 0x80x0 0000 0200 000f regs.a5 Function argument 5 2709 0x80x0 0000 0200 0010 regs.a6 Function argument 6 2710 0x80x0 0000 0200 0011 regs.a7 Function argument 7 2711 0x80x0 0000 0200 0012 regs.s2 Callee saved register 2 2712 0x80x0 0000 0200 0013 regs.s3 Callee saved register 3 2713 0x80x0 0000 0200 0014 regs.s4 Callee saved register 4 2714 0x80x0 0000 0200 0015 regs.s5 Callee saved register 5 2715 0x80x0 0000 0200 0016 regs.s6 Callee saved register 6 2716 0x80x0 0000 0200 0017 regs.s7 Callee saved register 7 2717 0x80x0 0000 0200 0018 regs.s8 Callee saved register 8 2718 0x80x0 0000 0200 0019 regs.s9 Callee saved register 9 2719 0x80x0 0000 0200 001a regs.s10 Callee saved register 10 2720 0x80x0 0000 0200 001b regs.s11 Callee saved register 11 2721 0x80x0 0000 0200 001c regs.t3 Caller saved register 3 2722 0x80x0 0000 0200 001d regs.t4 Caller saved register 4 2723 0x80x0 0000 0200 001e regs.t5 Caller saved register 5 2724 0x80x0 0000 0200 001f regs.t6 Caller saved register 6 2725 0x80x0 0000 0200 0020 mode Privilege mode (1 = S-mode or 0 = U-mode) 2726======================= ========= ============================================= 2727 2728RISC-V csr registers represent the supervisor mode control/status registers 2729of a Guest VCPU and it has the following id bit patterns:: 2730 2731 0x8020 0000 03 <index into the kvm_riscv_csr struct:24> (32bit Host) 2732 0x8030 0000 03 <index into the kvm_riscv_csr struct:24> (64bit Host) 2733 2734Following are the RISC-V csr registers: 2735 2736======================= ========= ============================================= 2737 Encoding Register Description 2738======================= ========= ============================================= 2739 0x80x0 0000 0300 0000 sstatus Supervisor status 2740 0x80x0 0000 0300 0001 sie Supervisor interrupt enable 2741 0x80x0 0000 0300 0002 stvec Supervisor trap vector base 2742 0x80x0 0000 0300 0003 sscratch Supervisor scratch register 2743 0x80x0 0000 0300 0004 sepc Supervisor exception program counter 2744 0x80x0 0000 0300 0005 scause Supervisor trap cause 2745 0x80x0 0000 0300 0006 stval Supervisor bad address or instruction 2746 0x80x0 0000 0300 0007 sip Supervisor interrupt pending 2747 0x80x0 0000 0300 0008 satp Supervisor address translation and protection 2748======================= ========= ============================================= 2749 2750RISC-V timer registers represent the timer state of a Guest VCPU and it has 2751the following id bit patterns:: 2752 2753 0x8030 0000 04 <index into the kvm_riscv_timer struct:24> 2754 2755Following are the RISC-V timer registers: 2756 2757======================= ========= ============================================= 2758 Encoding Register Description 2759======================= ========= ============================================= 2760 0x8030 0000 0400 0000 frequency Time base frequency (read-only) 2761 0x8030 0000 0400 0001 time Time value visible to Guest 2762 0x8030 0000 0400 0002 compare Time compare programmed by Guest 2763 0x8030 0000 0400 0003 state Time compare state (1 = ON or 0 = OFF) 2764======================= ========= ============================================= 2765 2766RISC-V F-extension registers represent the single precision floating point 2767state of a Guest VCPU and it has the following id bit patterns:: 2768 2769 0x8020 0000 05 <index into the __riscv_f_ext_state struct:24> 2770 2771Following are the RISC-V F-extension registers: 2772 2773======================= ========= ============================================= 2774 Encoding Register Description 2775======================= ========= ============================================= 2776 0x8020 0000 0500 0000 f[0] Floating point register 0 2777 ... 2778 0x8020 0000 0500 001f f[31] Floating point register 31 2779 0x8020 0000 0500 0020 fcsr Floating point control and status register 2780======================= ========= ============================================= 2781 2782RISC-V D-extension registers represent the double precision floating point 2783state of a Guest VCPU and it has the following id bit patterns:: 2784 2785 0x8020 0000 06 <index into the __riscv_d_ext_state struct:24> (fcsr) 2786 0x8030 0000 06 <index into the __riscv_d_ext_state struct:24> (non-fcsr) 2787 2788Following are the RISC-V D-extension registers: 2789 2790======================= ========= ============================================= 2791 Encoding Register Description 2792======================= ========= ============================================= 2793 0x8030 0000 0600 0000 f[0] Floating point register 0 2794 ... 2795 0x8030 0000 0600 001f f[31] Floating point register 31 2796 0x8020 0000 0600 0020 fcsr Floating point control and status register 2797======================= ========= ============================================= 2798 2799 28004.69 KVM_GET_ONE_REG 2801-------------------- 2802 2803:Capability: KVM_CAP_ONE_REG 2804:Architectures: all 2805:Type: vcpu ioctl 2806:Parameters: struct kvm_one_reg (in and out) 2807:Returns: 0 on success, negative value on failure 2808 2809Errors include: 2810 2811 ======== ============================================================ 2812 ENOENT no such register 2813 EINVAL invalid register ID, or no such register or used with VMs in 2814 protected virtualization mode on s390 2815 EPERM (arm64) register access not allowed before vcpu finalization 2816 ======== ============================================================ 2817 2818(These error codes are indicative only: do not rely on a specific error 2819code being returned in a specific situation.) 2820 2821This ioctl allows to receive the value of a single register implemented 2822in a vcpu. The register to read is indicated by the "id" field of the 2823kvm_one_reg struct passed in. On success, the register value can be found 2824at the memory location pointed to by "addr". 2825 2826The list of registers accessible using this interface is identical to the 2827list in 4.68. 2828 2829 28304.70 KVM_KVMCLOCK_CTRL 2831---------------------- 2832 2833:Capability: KVM_CAP_KVMCLOCK_CTRL 2834:Architectures: Any that implement pvclocks (currently x86 only) 2835:Type: vcpu ioctl 2836:Parameters: None 2837:Returns: 0 on success, -1 on error 2838 2839This ioctl sets a flag accessible to the guest indicating that the specified 2840vCPU has been paused by the host userspace. 2841 2842The host will set a flag in the pvclock structure that is checked from the 2843soft lockup watchdog. The flag is part of the pvclock structure that is 2844shared between guest and host, specifically the second bit of the flags 2845field of the pvclock_vcpu_time_info structure. It will be set exclusively by 2846the host and read/cleared exclusively by the guest. The guest operation of 2847checking and clearing the flag must be an atomic operation so 2848load-link/store-conditional, or equivalent must be used. There are two cases 2849where the guest will clear the flag: when the soft lockup watchdog timer resets 2850itself or when a soft lockup is detected. This ioctl can be called any time 2851after pausing the vcpu, but before it is resumed. 2852 2853 28544.71 KVM_SIGNAL_MSI 2855------------------- 2856 2857:Capability: KVM_CAP_SIGNAL_MSI 2858:Architectures: x86 arm arm64 2859:Type: vm ioctl 2860:Parameters: struct kvm_msi (in) 2861:Returns: >0 on delivery, 0 if guest blocked the MSI, and -1 on error 2862 2863Directly inject a MSI message. Only valid with in-kernel irqchip that handles 2864MSI messages. 2865 2866:: 2867 2868 struct kvm_msi { 2869 __u32 address_lo; 2870 __u32 address_hi; 2871 __u32 data; 2872 __u32 flags; 2873 __u32 devid; 2874 __u8 pad[12]; 2875 }; 2876 2877flags: 2878 KVM_MSI_VALID_DEVID: devid contains a valid value. The per-VM 2879 KVM_CAP_MSI_DEVID capability advertises the requirement to provide 2880 the device ID. If this capability is not available, userspace 2881 should never set the KVM_MSI_VALID_DEVID flag as the ioctl might fail. 2882 2883If KVM_MSI_VALID_DEVID is set, devid contains a unique device identifier 2884for the device that wrote the MSI message. For PCI, this is usually a 2885BFD identifier in the lower 16 bits. 2886 2887On x86, address_hi is ignored unless the KVM_X2APIC_API_USE_32BIT_IDS 2888feature of KVM_CAP_X2APIC_API capability is enabled. If it is enabled, 2889address_hi bits 31-8 provide bits 31-8 of the destination id. Bits 7-0 of 2890address_hi must be zero. 2891 2892 28934.71 KVM_CREATE_PIT2 2894-------------------- 2895 2896:Capability: KVM_CAP_PIT2 2897:Architectures: x86 2898:Type: vm ioctl 2899:Parameters: struct kvm_pit_config (in) 2900:Returns: 0 on success, -1 on error 2901 2902Creates an in-kernel device model for the i8254 PIT. This call is only valid 2903after enabling in-kernel irqchip support via KVM_CREATE_IRQCHIP. The following 2904parameters have to be passed:: 2905 2906 struct kvm_pit_config { 2907 __u32 flags; 2908 __u32 pad[15]; 2909 }; 2910 2911Valid flags are:: 2912 2913 #define KVM_PIT_SPEAKER_DUMMY 1 /* emulate speaker port stub */ 2914 2915PIT timer interrupts may use a per-VM kernel thread for injection. If it 2916exists, this thread will have a name of the following pattern:: 2917 2918 kvm-pit/<owner-process-pid> 2919 2920When running a guest with elevated priorities, the scheduling parameters of 2921this thread may have to be adjusted accordingly. 2922 2923This IOCTL replaces the obsolete KVM_CREATE_PIT. 2924 2925 29264.72 KVM_GET_PIT2 2927----------------- 2928 2929:Capability: KVM_CAP_PIT_STATE2 2930:Architectures: x86 2931:Type: vm ioctl 2932:Parameters: struct kvm_pit_state2 (out) 2933:Returns: 0 on success, -1 on error 2934 2935Retrieves the state of the in-kernel PIT model. Only valid after 2936KVM_CREATE_PIT2. The state is returned in the following structure:: 2937 2938 struct kvm_pit_state2 { 2939 struct kvm_pit_channel_state channels[3]; 2940 __u32 flags; 2941 __u32 reserved[9]; 2942 }; 2943 2944Valid flags are:: 2945 2946 /* disable PIT in HPET legacy mode */ 2947 #define KVM_PIT_FLAGS_HPET_LEGACY 0x00000001 2948 2949This IOCTL replaces the obsolete KVM_GET_PIT. 2950 2951 29524.73 KVM_SET_PIT2 2953----------------- 2954 2955:Capability: KVM_CAP_PIT_STATE2 2956:Architectures: x86 2957:Type: vm ioctl 2958:Parameters: struct kvm_pit_state2 (in) 2959:Returns: 0 on success, -1 on error 2960 2961Sets the state of the in-kernel PIT model. Only valid after KVM_CREATE_PIT2. 2962See KVM_GET_PIT2 for details on struct kvm_pit_state2. 2963 2964This IOCTL replaces the obsolete KVM_SET_PIT. 2965 2966 29674.74 KVM_PPC_GET_SMMU_INFO 2968-------------------------- 2969 2970:Capability: KVM_CAP_PPC_GET_SMMU_INFO 2971:Architectures: powerpc 2972:Type: vm ioctl 2973:Parameters: None 2974:Returns: 0 on success, -1 on error 2975 2976This populates and returns a structure describing the features of 2977the "Server" class MMU emulation supported by KVM. 2978This can in turn be used by userspace to generate the appropriate 2979device-tree properties for the guest operating system. 2980 2981The structure contains some global information, followed by an 2982array of supported segment page sizes:: 2983 2984 struct kvm_ppc_smmu_info { 2985 __u64 flags; 2986 __u32 slb_size; 2987 __u32 pad; 2988 struct kvm_ppc_one_seg_page_size sps[KVM_PPC_PAGE_SIZES_MAX_SZ]; 2989 }; 2990 2991The supported flags are: 2992 2993 - KVM_PPC_PAGE_SIZES_REAL: 2994 When that flag is set, guest page sizes must "fit" the backing 2995 store page sizes. When not set, any page size in the list can 2996 be used regardless of how they are backed by userspace. 2997 2998 - KVM_PPC_1T_SEGMENTS 2999 The emulated MMU supports 1T segments in addition to the 3000 standard 256M ones. 3001 3002 - KVM_PPC_NO_HASH 3003 This flag indicates that HPT guests are not supported by KVM, 3004 thus all guests must use radix MMU mode. 3005 3006The "slb_size" field indicates how many SLB entries are supported 3007 3008The "sps" array contains 8 entries indicating the supported base 3009page sizes for a segment in increasing order. Each entry is defined 3010as follow:: 3011 3012 struct kvm_ppc_one_seg_page_size { 3013 __u32 page_shift; /* Base page shift of segment (or 0) */ 3014 __u32 slb_enc; /* SLB encoding for BookS */ 3015 struct kvm_ppc_one_page_size enc[KVM_PPC_PAGE_SIZES_MAX_SZ]; 3016 }; 3017 3018An entry with a "page_shift" of 0 is unused. Because the array is 3019organized in increasing order, a lookup can stop when encoutering 3020such an entry. 3021 3022The "slb_enc" field provides the encoding to use in the SLB for the 3023page size. The bits are in positions such as the value can directly 3024be OR'ed into the "vsid" argument of the slbmte instruction. 3025 3026The "enc" array is a list which for each of those segment base page 3027size provides the list of supported actual page sizes (which can be 3028only larger or equal to the base page size), along with the 3029corresponding encoding in the hash PTE. Similarly, the array is 30308 entries sorted by increasing sizes and an entry with a "0" shift 3031is an empty entry and a terminator:: 3032 3033 struct kvm_ppc_one_page_size { 3034 __u32 page_shift; /* Page shift (or 0) */ 3035 __u32 pte_enc; /* Encoding in the HPTE (>>12) */ 3036 }; 3037 3038The "pte_enc" field provides a value that can OR'ed into the hash 3039PTE's RPN field (ie, it needs to be shifted left by 12 to OR it 3040into the hash PTE second double word). 3041 30424.75 KVM_IRQFD 3043-------------- 3044 3045:Capability: KVM_CAP_IRQFD 3046:Architectures: x86 s390 arm arm64 3047:Type: vm ioctl 3048:Parameters: struct kvm_irqfd (in) 3049:Returns: 0 on success, -1 on error 3050 3051Allows setting an eventfd to directly trigger a guest interrupt. 3052kvm_irqfd.fd specifies the file descriptor to use as the eventfd and 3053kvm_irqfd.gsi specifies the irqchip pin toggled by this event. When 3054an event is triggered on the eventfd, an interrupt is injected into 3055the guest using the specified gsi pin. The irqfd is removed using 3056the KVM_IRQFD_FLAG_DEASSIGN flag, specifying both kvm_irqfd.fd 3057and kvm_irqfd.gsi. 3058 3059With KVM_CAP_IRQFD_RESAMPLE, KVM_IRQFD supports a de-assert and notify 3060mechanism allowing emulation of level-triggered, irqfd-based 3061interrupts. When KVM_IRQFD_FLAG_RESAMPLE is set the user must pass an 3062additional eventfd in the kvm_irqfd.resamplefd field. When operating 3063in resample mode, posting of an interrupt through kvm_irq.fd asserts 3064the specified gsi in the irqchip. When the irqchip is resampled, such 3065as from an EOI, the gsi is de-asserted and the user is notified via 3066kvm_irqfd.resamplefd. It is the user's responsibility to re-queue 3067the interrupt if the device making use of it still requires service. 3068Note that closing the resamplefd is not sufficient to disable the 3069irqfd. The KVM_IRQFD_FLAG_RESAMPLE is only necessary on assignment 3070and need not be specified with KVM_IRQFD_FLAG_DEASSIGN. 3071 3072On arm/arm64, gsi routing being supported, the following can happen: 3073 3074- in case no routing entry is associated to this gsi, injection fails 3075- in case the gsi is associated to an irqchip routing entry, 3076 irqchip.pin + 32 corresponds to the injected SPI ID. 3077- in case the gsi is associated to an MSI routing entry, the MSI 3078 message and device ID are translated into an LPI (support restricted 3079 to GICv3 ITS in-kernel emulation). 3080 30814.76 KVM_PPC_ALLOCATE_HTAB 3082-------------------------- 3083 3084:Capability: KVM_CAP_PPC_ALLOC_HTAB 3085:Architectures: powerpc 3086:Type: vm ioctl 3087:Parameters: Pointer to u32 containing hash table order (in/out) 3088:Returns: 0 on success, -1 on error 3089 3090This requests the host kernel to allocate an MMU hash table for a 3091guest using the PAPR paravirtualization interface. This only does 3092anything if the kernel is configured to use the Book 3S HV style of 3093virtualization. Otherwise the capability doesn't exist and the ioctl 3094returns an ENOTTY error. The rest of this description assumes Book 3S 3095HV. 3096 3097There must be no vcpus running when this ioctl is called; if there 3098are, it will do nothing and return an EBUSY error. 3099 3100The parameter is a pointer to a 32-bit unsigned integer variable 3101containing the order (log base 2) of the desired size of the hash 3102table, which must be between 18 and 46. On successful return from the 3103ioctl, the value will not be changed by the kernel. 3104 3105If no hash table has been allocated when any vcpu is asked to run 3106(with the KVM_RUN ioctl), the host kernel will allocate a 3107default-sized hash table (16 MB). 3108 3109If this ioctl is called when a hash table has already been allocated, 3110with a different order from the existing hash table, the existing hash 3111table will be freed and a new one allocated. If this is ioctl is 3112called when a hash table has already been allocated of the same order 3113as specified, the kernel will clear out the existing hash table (zero 3114all HPTEs). In either case, if the guest is using the virtualized 3115real-mode area (VRMA) facility, the kernel will re-create the VMRA 3116HPTEs on the next KVM_RUN of any vcpu. 3117 31184.77 KVM_S390_INTERRUPT 3119----------------------- 3120 3121:Capability: basic 3122:Architectures: s390 3123:Type: vm ioctl, vcpu ioctl 3124:Parameters: struct kvm_s390_interrupt (in) 3125:Returns: 0 on success, -1 on error 3126 3127Allows to inject an interrupt to the guest. Interrupts can be floating 3128(vm ioctl) or per cpu (vcpu ioctl), depending on the interrupt type. 3129 3130Interrupt parameters are passed via kvm_s390_interrupt:: 3131 3132 struct kvm_s390_interrupt { 3133 __u32 type; 3134 __u32 parm; 3135 __u64 parm64; 3136 }; 3137 3138type can be one of the following: 3139 3140KVM_S390_SIGP_STOP (vcpu) 3141 - sigp stop; optional flags in parm 3142KVM_S390_PROGRAM_INT (vcpu) 3143 - program check; code in parm 3144KVM_S390_SIGP_SET_PREFIX (vcpu) 3145 - sigp set prefix; prefix address in parm 3146KVM_S390_RESTART (vcpu) 3147 - restart 3148KVM_S390_INT_CLOCK_COMP (vcpu) 3149 - clock comparator interrupt 3150KVM_S390_INT_CPU_TIMER (vcpu) 3151 - CPU timer interrupt 3152KVM_S390_INT_VIRTIO (vm) 3153 - virtio external interrupt; external interrupt 3154 parameters in parm and parm64 3155KVM_S390_INT_SERVICE (vm) 3156 - sclp external interrupt; sclp parameter in parm 3157KVM_S390_INT_EMERGENCY (vcpu) 3158 - sigp emergency; source cpu in parm 3159KVM_S390_INT_EXTERNAL_CALL (vcpu) 3160 - sigp external call; source cpu in parm 3161KVM_S390_INT_IO(ai,cssid,ssid,schid) (vm) 3162 - compound value to indicate an 3163 I/O interrupt (ai - adapter interrupt; cssid,ssid,schid - subchannel); 3164 I/O interruption parameters in parm (subchannel) and parm64 (intparm, 3165 interruption subclass) 3166KVM_S390_MCHK (vm, vcpu) 3167 - machine check interrupt; cr 14 bits in parm, machine check interrupt 3168 code in parm64 (note that machine checks needing further payload are not 3169 supported by this ioctl) 3170 3171This is an asynchronous vcpu ioctl and can be invoked from any thread. 3172 31734.78 KVM_PPC_GET_HTAB_FD 3174------------------------ 3175 3176:Capability: KVM_CAP_PPC_HTAB_FD 3177:Architectures: powerpc 3178:Type: vm ioctl 3179:Parameters: Pointer to struct kvm_get_htab_fd (in) 3180:Returns: file descriptor number (>= 0) on success, -1 on error 3181 3182This returns a file descriptor that can be used either to read out the 3183entries in the guest's hashed page table (HPT), or to write entries to 3184initialize the HPT. The returned fd can only be written to if the 3185KVM_GET_HTAB_WRITE bit is set in the flags field of the argument, and 3186can only be read if that bit is clear. The argument struct looks like 3187this:: 3188 3189 /* For KVM_PPC_GET_HTAB_FD */ 3190 struct kvm_get_htab_fd { 3191 __u64 flags; 3192 __u64 start_index; 3193 __u64 reserved[2]; 3194 }; 3195 3196 /* Values for kvm_get_htab_fd.flags */ 3197 #define KVM_GET_HTAB_BOLTED_ONLY ((__u64)0x1) 3198 #define KVM_GET_HTAB_WRITE ((__u64)0x2) 3199 3200The 'start_index' field gives the index in the HPT of the entry at 3201which to start reading. It is ignored when writing. 3202 3203Reads on the fd will initially supply information about all 3204"interesting" HPT entries. Interesting entries are those with the 3205bolted bit set, if the KVM_GET_HTAB_BOLTED_ONLY bit is set, otherwise 3206all entries. When the end of the HPT is reached, the read() will 3207return. If read() is called again on the fd, it will start again from 3208the beginning of the HPT, but will only return HPT entries that have 3209changed since they were last read. 3210 3211Data read or written is structured as a header (8 bytes) followed by a 3212series of valid HPT entries (16 bytes) each. The header indicates how 3213many valid HPT entries there are and how many invalid entries follow 3214the valid entries. The invalid entries are not represented explicitly 3215in the stream. The header format is:: 3216 3217 struct kvm_get_htab_header { 3218 __u32 index; 3219 __u16 n_valid; 3220 __u16 n_invalid; 3221 }; 3222 3223Writes to the fd create HPT entries starting at the index given in the 3224header; first 'n_valid' valid entries with contents from the data 3225written, then 'n_invalid' invalid entries, invalidating any previously 3226valid entries found. 3227 32284.79 KVM_CREATE_DEVICE 3229---------------------- 3230 3231:Capability: KVM_CAP_DEVICE_CTRL 3232:Type: vm ioctl 3233:Parameters: struct kvm_create_device (in/out) 3234:Returns: 0 on success, -1 on error 3235 3236Errors: 3237 3238 ====== ======================================================= 3239 ENODEV The device type is unknown or unsupported 3240 EEXIST Device already created, and this type of device may not 3241 be instantiated multiple times 3242 ====== ======================================================= 3243 3244 Other error conditions may be defined by individual device types or 3245 have their standard meanings. 3246 3247Creates an emulated device in the kernel. The file descriptor returned 3248in fd can be used with KVM_SET/GET/HAS_DEVICE_ATTR. 3249 3250If the KVM_CREATE_DEVICE_TEST flag is set, only test whether the 3251device type is supported (not necessarily whether it can be created 3252in the current vm). 3253 3254Individual devices should not define flags. Attributes should be used 3255for specifying any behavior that is not implied by the device type 3256number. 3257 3258:: 3259 3260 struct kvm_create_device { 3261 __u32 type; /* in: KVM_DEV_TYPE_xxx */ 3262 __u32 fd; /* out: device handle */ 3263 __u32 flags; /* in: KVM_CREATE_DEVICE_xxx */ 3264 }; 3265 32664.80 KVM_SET_DEVICE_ATTR/KVM_GET_DEVICE_ATTR 3267-------------------------------------------- 3268 3269:Capability: KVM_CAP_DEVICE_CTRL, KVM_CAP_VM_ATTRIBUTES for vm device, 3270 KVM_CAP_VCPU_ATTRIBUTES for vcpu device 3271:Type: device ioctl, vm ioctl, vcpu ioctl 3272:Parameters: struct kvm_device_attr 3273:Returns: 0 on success, -1 on error 3274 3275Errors: 3276 3277 ===== ============================================================= 3278 ENXIO The group or attribute is unknown/unsupported for this device 3279 or hardware support is missing. 3280 EPERM The attribute cannot (currently) be accessed this way 3281 (e.g. read-only attribute, or attribute that only makes 3282 sense when the device is in a different state) 3283 ===== ============================================================= 3284 3285 Other error conditions may be defined by individual device types. 3286 3287Gets/sets a specified piece of device configuration and/or state. The 3288semantics are device-specific. See individual device documentation in 3289the "devices" directory. As with ONE_REG, the size of the data 3290transferred is defined by the particular attribute. 3291 3292:: 3293 3294 struct kvm_device_attr { 3295 __u32 flags; /* no flags currently defined */ 3296 __u32 group; /* device-defined */ 3297 __u64 attr; /* group-defined */ 3298 __u64 addr; /* userspace address of attr data */ 3299 }; 3300 33014.81 KVM_HAS_DEVICE_ATTR 3302------------------------ 3303 3304:Capability: KVM_CAP_DEVICE_CTRL, KVM_CAP_VM_ATTRIBUTES for vm device, 3305 KVM_CAP_VCPU_ATTRIBUTES for vcpu device 3306:Type: device ioctl, vm ioctl, vcpu ioctl 3307:Parameters: struct kvm_device_attr 3308:Returns: 0 on success, -1 on error 3309 3310Errors: 3311 3312 ===== ============================================================= 3313 ENXIO The group or attribute is unknown/unsupported for this device 3314 or hardware support is missing. 3315 ===== ============================================================= 3316 3317Tests whether a device supports a particular attribute. A successful 3318return indicates the attribute is implemented. It does not necessarily 3319indicate that the attribute can be read or written in the device's 3320current state. "addr" is ignored. 3321 33224.82 KVM_ARM_VCPU_INIT 3323---------------------- 3324 3325:Capability: basic 3326:Architectures: arm, arm64 3327:Type: vcpu ioctl 3328:Parameters: struct kvm_vcpu_init (in) 3329:Returns: 0 on success; -1 on error 3330 3331Errors: 3332 3333 ====== ================================================================= 3334 EINVAL the target is unknown, or the combination of features is invalid. 3335 ENOENT a features bit specified is unknown. 3336 ====== ================================================================= 3337 3338This tells KVM what type of CPU to present to the guest, and what 3339optional features it should have. This will cause a reset of the cpu 3340registers to their initial values. If this is not called, KVM_RUN will 3341return ENOEXEC for that vcpu. 3342 3343The initial values are defined as: 3344 - Processor state: 3345 * AArch64: EL1h, D, A, I and F bits set. All other bits 3346 are cleared. 3347 * AArch32: SVC, A, I and F bits set. All other bits are 3348 cleared. 3349 - General Purpose registers, including PC and SP: set to 0 3350 - FPSIMD/NEON registers: set to 0 3351 - SVE registers: set to 0 3352 - System registers: Reset to their architecturally defined 3353 values as for a warm reset to EL1 (resp. SVC) 3354 3355Note that because some registers reflect machine topology, all vcpus 3356should be created before this ioctl is invoked. 3357 3358Userspace can call this function multiple times for a given vcpu, including 3359after the vcpu has been run. This will reset the vcpu to its initial 3360state. All calls to this function after the initial call must use the same 3361target and same set of feature flags, otherwise EINVAL will be returned. 3362 3363Possible features: 3364 3365 - KVM_ARM_VCPU_POWER_OFF: Starts the CPU in a power-off state. 3366 Depends on KVM_CAP_ARM_PSCI. If not set, the CPU will be powered on 3367 and execute guest code when KVM_RUN is called. 3368 - KVM_ARM_VCPU_EL1_32BIT: Starts the CPU in a 32bit mode. 3369 Depends on KVM_CAP_ARM_EL1_32BIT (arm64 only). 3370 - KVM_ARM_VCPU_PSCI_0_2: Emulate PSCI v0.2 (or a future revision 3371 backward compatible with v0.2) for the CPU. 3372 Depends on KVM_CAP_ARM_PSCI_0_2. 3373 - KVM_ARM_VCPU_PMU_V3: Emulate PMUv3 for the CPU. 3374 Depends on KVM_CAP_ARM_PMU_V3. 3375 3376 - KVM_ARM_VCPU_PTRAUTH_ADDRESS: Enables Address Pointer authentication 3377 for arm64 only. 3378 Depends on KVM_CAP_ARM_PTRAUTH_ADDRESS. 3379 If KVM_CAP_ARM_PTRAUTH_ADDRESS and KVM_CAP_ARM_PTRAUTH_GENERIC are 3380 both present, then both KVM_ARM_VCPU_PTRAUTH_ADDRESS and 3381 KVM_ARM_VCPU_PTRAUTH_GENERIC must be requested or neither must be 3382 requested. 3383 3384 - KVM_ARM_VCPU_PTRAUTH_GENERIC: Enables Generic Pointer authentication 3385 for arm64 only. 3386 Depends on KVM_CAP_ARM_PTRAUTH_GENERIC. 3387 If KVM_CAP_ARM_PTRAUTH_ADDRESS and KVM_CAP_ARM_PTRAUTH_GENERIC are 3388 both present, then both KVM_ARM_VCPU_PTRAUTH_ADDRESS and 3389 KVM_ARM_VCPU_PTRAUTH_GENERIC must be requested or neither must be 3390 requested. 3391 3392 - KVM_ARM_VCPU_SVE: Enables SVE for the CPU (arm64 only). 3393 Depends on KVM_CAP_ARM_SVE. 3394 Requires KVM_ARM_VCPU_FINALIZE(KVM_ARM_VCPU_SVE): 3395 3396 * After KVM_ARM_VCPU_INIT: 3397 3398 - KVM_REG_ARM64_SVE_VLS may be read using KVM_GET_ONE_REG: the 3399 initial value of this pseudo-register indicates the best set of 3400 vector lengths possible for a vcpu on this host. 3401 3402 * Before KVM_ARM_VCPU_FINALIZE(KVM_ARM_VCPU_SVE): 3403 3404 - KVM_RUN and KVM_GET_REG_LIST are not available; 3405 3406 - KVM_GET_ONE_REG and KVM_SET_ONE_REG cannot be used to access 3407 the scalable archietctural SVE registers 3408 KVM_REG_ARM64_SVE_ZREG(), KVM_REG_ARM64_SVE_PREG() or 3409 KVM_REG_ARM64_SVE_FFR; 3410 3411 - KVM_REG_ARM64_SVE_VLS may optionally be written using 3412 KVM_SET_ONE_REG, to modify the set of vector lengths available 3413 for the vcpu. 3414 3415 * After KVM_ARM_VCPU_FINALIZE(KVM_ARM_VCPU_SVE): 3416 3417 - the KVM_REG_ARM64_SVE_VLS pseudo-register is immutable, and can 3418 no longer be written using KVM_SET_ONE_REG. 3419 34204.83 KVM_ARM_PREFERRED_TARGET 3421----------------------------- 3422 3423:Capability: basic 3424:Architectures: arm, arm64 3425:Type: vm ioctl 3426:Parameters: struct kvm_vcpu_init (out) 3427:Returns: 0 on success; -1 on error 3428 3429Errors: 3430 3431 ====== ========================================== 3432 ENODEV no preferred target available for the host 3433 ====== ========================================== 3434 3435This queries KVM for preferred CPU target type which can be emulated 3436by KVM on underlying host. 3437 3438The ioctl returns struct kvm_vcpu_init instance containing information 3439about preferred CPU target type and recommended features for it. The 3440kvm_vcpu_init->features bitmap returned will have feature bits set if 3441the preferred target recommends setting these features, but this is 3442not mandatory. 3443 3444The information returned by this ioctl can be used to prepare an instance 3445of struct kvm_vcpu_init for KVM_ARM_VCPU_INIT ioctl which will result in 3446VCPU matching underlying host. 3447 3448 34494.84 KVM_GET_REG_LIST 3450--------------------- 3451 3452:Capability: basic 3453:Architectures: arm, arm64, mips 3454:Type: vcpu ioctl 3455:Parameters: struct kvm_reg_list (in/out) 3456:Returns: 0 on success; -1 on error 3457 3458Errors: 3459 3460 ===== ============================================================== 3461 E2BIG the reg index list is too big to fit in the array specified by 3462 the user (the number required will be written into n). 3463 ===== ============================================================== 3464 3465:: 3466 3467 struct kvm_reg_list { 3468 __u64 n; /* number of registers in reg[] */ 3469 __u64 reg[0]; 3470 }; 3471 3472This ioctl returns the guest registers that are supported for the 3473KVM_GET_ONE_REG/KVM_SET_ONE_REG calls. 3474 3475 34764.85 KVM_ARM_SET_DEVICE_ADDR (deprecated) 3477----------------------------------------- 3478 3479:Capability: KVM_CAP_ARM_SET_DEVICE_ADDR 3480:Architectures: arm, arm64 3481:Type: vm ioctl 3482:Parameters: struct kvm_arm_device_address (in) 3483:Returns: 0 on success, -1 on error 3484 3485Errors: 3486 3487 ====== ============================================ 3488 ENODEV The device id is unknown 3489 ENXIO Device not supported on current system 3490 EEXIST Address already set 3491 E2BIG Address outside guest physical address space 3492 EBUSY Address overlaps with other device range 3493 ====== ============================================ 3494 3495:: 3496 3497 struct kvm_arm_device_addr { 3498 __u64 id; 3499 __u64 addr; 3500 }; 3501 3502Specify a device address in the guest's physical address space where guests 3503can access emulated or directly exposed devices, which the host kernel needs 3504to know about. The id field is an architecture specific identifier for a 3505specific device. 3506 3507ARM/arm64 divides the id field into two parts, a device id and an 3508address type id specific to the individual device:: 3509 3510 bits: | 63 ... 32 | 31 ... 16 | 15 ... 0 | 3511 field: | 0x00000000 | device id | addr type id | 3512 3513ARM/arm64 currently only require this when using the in-kernel GIC 3514support for the hardware VGIC features, using KVM_ARM_DEVICE_VGIC_V2 3515as the device id. When setting the base address for the guest's 3516mapping of the VGIC virtual CPU and distributor interface, the ioctl 3517must be called after calling KVM_CREATE_IRQCHIP, but before calling 3518KVM_RUN on any of the VCPUs. Calling this ioctl twice for any of the 3519base addresses will return -EEXIST. 3520 3521Note, this IOCTL is deprecated and the more flexible SET/GET_DEVICE_ATTR API 3522should be used instead. 3523 3524 35254.86 KVM_PPC_RTAS_DEFINE_TOKEN 3526------------------------------ 3527 3528:Capability: KVM_CAP_PPC_RTAS 3529:Architectures: ppc 3530:Type: vm ioctl 3531:Parameters: struct kvm_rtas_token_args 3532:Returns: 0 on success, -1 on error 3533 3534Defines a token value for a RTAS (Run Time Abstraction Services) 3535service in order to allow it to be handled in the kernel. The 3536argument struct gives the name of the service, which must be the name 3537of a service that has a kernel-side implementation. If the token 3538value is non-zero, it will be associated with that service, and 3539subsequent RTAS calls by the guest specifying that token will be 3540handled by the kernel. If the token value is 0, then any token 3541associated with the service will be forgotten, and subsequent RTAS 3542calls by the guest for that service will be passed to userspace to be 3543handled. 3544 35454.87 KVM_SET_GUEST_DEBUG 3546------------------------ 3547 3548:Capability: KVM_CAP_SET_GUEST_DEBUG 3549:Architectures: x86, s390, ppc, arm64 3550:Type: vcpu ioctl 3551:Parameters: struct kvm_guest_debug (in) 3552:Returns: 0 on success; -1 on error 3553 3554:: 3555 3556 struct kvm_guest_debug { 3557 __u32 control; 3558 __u32 pad; 3559 struct kvm_guest_debug_arch arch; 3560 }; 3561 3562Set up the processor specific debug registers and configure vcpu for 3563handling guest debug events. There are two parts to the structure, the 3564first a control bitfield indicates the type of debug events to handle 3565when running. Common control bits are: 3566 3567 - KVM_GUESTDBG_ENABLE: guest debugging is enabled 3568 - KVM_GUESTDBG_SINGLESTEP: the next run should single-step 3569 3570The top 16 bits of the control field are architecture specific control 3571flags which can include the following: 3572 3573 - KVM_GUESTDBG_USE_SW_BP: using software breakpoints [x86, arm64] 3574 - KVM_GUESTDBG_USE_HW_BP: using hardware breakpoints [x86, s390] 3575 - KVM_GUESTDBG_USE_HW: using hardware debug events [arm64] 3576 - KVM_GUESTDBG_INJECT_DB: inject DB type exception [x86] 3577 - KVM_GUESTDBG_INJECT_BP: inject BP type exception [x86] 3578 - KVM_GUESTDBG_EXIT_PENDING: trigger an immediate guest exit [s390] 3579 - KVM_GUESTDBG_BLOCKIRQ: avoid injecting interrupts/NMI/SMI [x86] 3580 3581For example KVM_GUESTDBG_USE_SW_BP indicates that software breakpoints 3582are enabled in memory so we need to ensure breakpoint exceptions are 3583correctly trapped and the KVM run loop exits at the breakpoint and not 3584running off into the normal guest vector. For KVM_GUESTDBG_USE_HW_BP 3585we need to ensure the guest vCPUs architecture specific registers are 3586updated to the correct (supplied) values. 3587 3588The second part of the structure is architecture specific and 3589typically contains a set of debug registers. 3590 3591For arm64 the number of debug registers is implementation defined and 3592can be determined by querying the KVM_CAP_GUEST_DEBUG_HW_BPS and 3593KVM_CAP_GUEST_DEBUG_HW_WPS capabilities which return a positive number 3594indicating the number of supported registers. 3595 3596For ppc, the KVM_CAP_PPC_GUEST_DEBUG_SSTEP capability indicates whether 3597the single-step debug event (KVM_GUESTDBG_SINGLESTEP) is supported. 3598 3599Also when supported, KVM_CAP_SET_GUEST_DEBUG2 capability indicates the 3600supported KVM_GUESTDBG_* bits in the control field. 3601 3602When debug events exit the main run loop with the reason 3603KVM_EXIT_DEBUG with the kvm_debug_exit_arch part of the kvm_run 3604structure containing architecture specific debug information. 3605 36064.88 KVM_GET_EMULATED_CPUID 3607--------------------------- 3608 3609:Capability: KVM_CAP_EXT_EMUL_CPUID 3610:Architectures: x86 3611:Type: system ioctl 3612:Parameters: struct kvm_cpuid2 (in/out) 3613:Returns: 0 on success, -1 on error 3614 3615:: 3616 3617 struct kvm_cpuid2 { 3618 __u32 nent; 3619 __u32 flags; 3620 struct kvm_cpuid_entry2 entries[0]; 3621 }; 3622 3623The member 'flags' is used for passing flags from userspace. 3624 3625:: 3626 3627 #define KVM_CPUID_FLAG_SIGNIFCANT_INDEX BIT(0) 3628 #define KVM_CPUID_FLAG_STATEFUL_FUNC BIT(1) /* deprecated */ 3629 #define KVM_CPUID_FLAG_STATE_READ_NEXT BIT(2) /* deprecated */ 3630 3631 struct kvm_cpuid_entry2 { 3632 __u32 function; 3633 __u32 index; 3634 __u32 flags; 3635 __u32 eax; 3636 __u32 ebx; 3637 __u32 ecx; 3638 __u32 edx; 3639 __u32 padding[3]; 3640 }; 3641 3642This ioctl returns x86 cpuid features which are emulated by 3643kvm.Userspace can use the information returned by this ioctl to query 3644which features are emulated by kvm instead of being present natively. 3645 3646Userspace invokes KVM_GET_EMULATED_CPUID by passing a kvm_cpuid2 3647structure with the 'nent' field indicating the number of entries in 3648the variable-size array 'entries'. If the number of entries is too low 3649to describe the cpu capabilities, an error (E2BIG) is returned. If the 3650number is too high, the 'nent' field is adjusted and an error (ENOMEM) 3651is returned. If the number is just right, the 'nent' field is adjusted 3652to the number of valid entries in the 'entries' array, which is then 3653filled. 3654 3655The entries returned are the set CPUID bits of the respective features 3656which kvm emulates, as returned by the CPUID instruction, with unknown 3657or unsupported feature bits cleared. 3658 3659Features like x2apic, for example, may not be present in the host cpu 3660but are exposed by kvm in KVM_GET_SUPPORTED_CPUID because they can be 3661emulated efficiently and thus not included here. 3662 3663The fields in each entry are defined as follows: 3664 3665 function: 3666 the eax value used to obtain the entry 3667 index: 3668 the ecx value used to obtain the entry (for entries that are 3669 affected by ecx) 3670 flags: 3671 an OR of zero or more of the following: 3672 3673 KVM_CPUID_FLAG_SIGNIFCANT_INDEX: 3674 if the index field is valid 3675 3676 eax, ebx, ecx, edx: 3677 3678 the values returned by the cpuid instruction for 3679 this function/index combination 3680 36814.89 KVM_S390_MEM_OP 3682-------------------- 3683 3684:Capability: KVM_CAP_S390_MEM_OP 3685:Architectures: s390 3686:Type: vcpu ioctl 3687:Parameters: struct kvm_s390_mem_op (in) 3688:Returns: = 0 on success, 3689 < 0 on generic error (e.g. -EFAULT or -ENOMEM), 3690 > 0 if an exception occurred while walking the page tables 3691 3692Read or write data from/to the logical (virtual) memory of a VCPU. 3693 3694Parameters are specified via the following structure:: 3695 3696 struct kvm_s390_mem_op { 3697 __u64 gaddr; /* the guest address */ 3698 __u64 flags; /* flags */ 3699 __u32 size; /* amount of bytes */ 3700 __u32 op; /* type of operation */ 3701 __u64 buf; /* buffer in userspace */ 3702 __u8 ar; /* the access register number */ 3703 __u8 reserved[31]; /* should be set to 0 */ 3704 }; 3705 3706The type of operation is specified in the "op" field. It is either 3707KVM_S390_MEMOP_LOGICAL_READ for reading from logical memory space or 3708KVM_S390_MEMOP_LOGICAL_WRITE for writing to logical memory space. The 3709KVM_S390_MEMOP_F_CHECK_ONLY flag can be set in the "flags" field to check 3710whether the corresponding memory access would create an access exception 3711(without touching the data in the memory at the destination). In case an 3712access exception occurred while walking the MMU tables of the guest, the 3713ioctl returns a positive error number to indicate the type of exception. 3714This exception is also raised directly at the corresponding VCPU if the 3715flag KVM_S390_MEMOP_F_INJECT_EXCEPTION is set in the "flags" field. 3716 3717The start address of the memory region has to be specified in the "gaddr" 3718field, and the length of the region in the "size" field (which must not 3719be 0). The maximum value for "size" can be obtained by checking the 3720KVM_CAP_S390_MEM_OP capability. "buf" is the buffer supplied by the 3721userspace application where the read data should be written to for 3722KVM_S390_MEMOP_LOGICAL_READ, or where the data that should be written is 3723stored for a KVM_S390_MEMOP_LOGICAL_WRITE. When KVM_S390_MEMOP_F_CHECK_ONLY 3724is specified, "buf" is unused and can be NULL. "ar" designates the access 3725register number to be used; the valid range is 0..15. 3726 3727The "reserved" field is meant for future extensions. It is not used by 3728KVM with the currently defined set of flags. 3729 37304.90 KVM_S390_GET_SKEYS 3731----------------------- 3732 3733:Capability: KVM_CAP_S390_SKEYS 3734:Architectures: s390 3735:Type: vm ioctl 3736:Parameters: struct kvm_s390_skeys 3737:Returns: 0 on success, KVM_S390_GET_SKEYS_NONE if guest is not using storage 3738 keys, negative value on error 3739 3740This ioctl is used to get guest storage key values on the s390 3741architecture. The ioctl takes parameters via the kvm_s390_skeys struct:: 3742 3743 struct kvm_s390_skeys { 3744 __u64 start_gfn; 3745 __u64 count; 3746 __u64 skeydata_addr; 3747 __u32 flags; 3748 __u32 reserved[9]; 3749 }; 3750 3751The start_gfn field is the number of the first guest frame whose storage keys 3752you want to get. 3753 3754The count field is the number of consecutive frames (starting from start_gfn) 3755whose storage keys to get. The count field must be at least 1 and the maximum 3756allowed value is defined as KVM_S390_SKEYS_MAX. Values outside this range 3757will cause the ioctl to return -EINVAL. 3758 3759The skeydata_addr field is the address to a buffer large enough to hold count 3760bytes. This buffer will be filled with storage key data by the ioctl. 3761 37624.91 KVM_S390_SET_SKEYS 3763----------------------- 3764 3765:Capability: KVM_CAP_S390_SKEYS 3766:Architectures: s390 3767:Type: vm ioctl 3768:Parameters: struct kvm_s390_skeys 3769:Returns: 0 on success, negative value on error 3770 3771This ioctl is used to set guest storage key values on the s390 3772architecture. The ioctl takes parameters via the kvm_s390_skeys struct. 3773See section on KVM_S390_GET_SKEYS for struct definition. 3774 3775The start_gfn field is the number of the first guest frame whose storage keys 3776you want to set. 3777 3778The count field is the number of consecutive frames (starting from start_gfn) 3779whose storage keys to get. The count field must be at least 1 and the maximum 3780allowed value is defined as KVM_S390_SKEYS_MAX. Values outside this range 3781will cause the ioctl to return -EINVAL. 3782 3783The skeydata_addr field is the address to a buffer containing count bytes of 3784storage keys. Each byte in the buffer will be set as the storage key for a 3785single frame starting at start_gfn for count frames. 3786 3787Note: If any architecturally invalid key value is found in the given data then 3788the ioctl will return -EINVAL. 3789 37904.92 KVM_S390_IRQ 3791----------------- 3792 3793:Capability: KVM_CAP_S390_INJECT_IRQ 3794:Architectures: s390 3795:Type: vcpu ioctl 3796:Parameters: struct kvm_s390_irq (in) 3797:Returns: 0 on success, -1 on error 3798 3799Errors: 3800 3801 3802 ====== ================================================================= 3803 EINVAL interrupt type is invalid 3804 type is KVM_S390_SIGP_STOP and flag parameter is invalid value, 3805 type is KVM_S390_INT_EXTERNAL_CALL and code is bigger 3806 than the maximum of VCPUs 3807 EBUSY type is KVM_S390_SIGP_SET_PREFIX and vcpu is not stopped, 3808 type is KVM_S390_SIGP_STOP and a stop irq is already pending, 3809 type is KVM_S390_INT_EXTERNAL_CALL and an external call interrupt 3810 is already pending 3811 ====== ================================================================= 3812 3813Allows to inject an interrupt to the guest. 3814 3815Using struct kvm_s390_irq as a parameter allows 3816to inject additional payload which is not 3817possible via KVM_S390_INTERRUPT. 3818 3819Interrupt parameters are passed via kvm_s390_irq:: 3820 3821 struct kvm_s390_irq { 3822 __u64 type; 3823 union { 3824 struct kvm_s390_io_info io; 3825 struct kvm_s390_ext_info ext; 3826 struct kvm_s390_pgm_info pgm; 3827 struct kvm_s390_emerg_info emerg; 3828 struct kvm_s390_extcall_info extcall; 3829 struct kvm_s390_prefix_info prefix; 3830 struct kvm_s390_stop_info stop; 3831 struct kvm_s390_mchk_info mchk; 3832 char reserved[64]; 3833 } u; 3834 }; 3835 3836type can be one of the following: 3837 3838- KVM_S390_SIGP_STOP - sigp stop; parameter in .stop 3839- KVM_S390_PROGRAM_INT - program check; parameters in .pgm 3840- KVM_S390_SIGP_SET_PREFIX - sigp set prefix; parameters in .prefix 3841- KVM_S390_RESTART - restart; no parameters 3842- KVM_S390_INT_CLOCK_COMP - clock comparator interrupt; no parameters 3843- KVM_S390_INT_CPU_TIMER - CPU timer interrupt; no parameters 3844- KVM_S390_INT_EMERGENCY - sigp emergency; parameters in .emerg 3845- KVM_S390_INT_EXTERNAL_CALL - sigp external call; parameters in .extcall 3846- KVM_S390_MCHK - machine check interrupt; parameters in .mchk 3847 3848This is an asynchronous vcpu ioctl and can be invoked from any thread. 3849 38504.94 KVM_S390_GET_IRQ_STATE 3851--------------------------- 3852 3853:Capability: KVM_CAP_S390_IRQ_STATE 3854:Architectures: s390 3855:Type: vcpu ioctl 3856:Parameters: struct kvm_s390_irq_state (out) 3857:Returns: >= number of bytes copied into buffer, 3858 -EINVAL if buffer size is 0, 3859 -ENOBUFS if buffer size is too small to fit all pending interrupts, 3860 -EFAULT if the buffer address was invalid 3861 3862This ioctl allows userspace to retrieve the complete state of all currently 3863pending interrupts in a single buffer. Use cases include migration 3864and introspection. The parameter structure contains the address of a 3865userspace buffer and its length:: 3866 3867 struct kvm_s390_irq_state { 3868 __u64 buf; 3869 __u32 flags; /* will stay unused for compatibility reasons */ 3870 __u32 len; 3871 __u32 reserved[4]; /* will stay unused for compatibility reasons */ 3872 }; 3873 3874Userspace passes in the above struct and for each pending interrupt a 3875struct kvm_s390_irq is copied to the provided buffer. 3876 3877The structure contains a flags and a reserved field for future extensions. As 3878the kernel never checked for flags == 0 and QEMU never pre-zeroed flags and 3879reserved, these fields can not be used in the future without breaking 3880compatibility. 3881 3882If -ENOBUFS is returned the buffer provided was too small and userspace 3883may retry with a bigger buffer. 3884 38854.95 KVM_S390_SET_IRQ_STATE 3886--------------------------- 3887 3888:Capability: KVM_CAP_S390_IRQ_STATE 3889:Architectures: s390 3890:Type: vcpu ioctl 3891:Parameters: struct kvm_s390_irq_state (in) 3892:Returns: 0 on success, 3893 -EFAULT if the buffer address was invalid, 3894 -EINVAL for an invalid buffer length (see below), 3895 -EBUSY if there were already interrupts pending, 3896 errors occurring when actually injecting the 3897 interrupt. See KVM_S390_IRQ. 3898 3899This ioctl allows userspace to set the complete state of all cpu-local 3900interrupts currently pending for the vcpu. It is intended for restoring 3901interrupt state after a migration. The input parameter is a userspace buffer 3902containing a struct kvm_s390_irq_state:: 3903 3904 struct kvm_s390_irq_state { 3905 __u64 buf; 3906 __u32 flags; /* will stay unused for compatibility reasons */ 3907 __u32 len; 3908 __u32 reserved[4]; /* will stay unused for compatibility reasons */ 3909 }; 3910 3911The restrictions for flags and reserved apply as well. 3912(see KVM_S390_GET_IRQ_STATE) 3913 3914The userspace memory referenced by buf contains a struct kvm_s390_irq 3915for each interrupt to be injected into the guest. 3916If one of the interrupts could not be injected for some reason the 3917ioctl aborts. 3918 3919len must be a multiple of sizeof(struct kvm_s390_irq). It must be > 0 3920and it must not exceed (max_vcpus + 32) * sizeof(struct kvm_s390_irq), 3921which is the maximum number of possibly pending cpu-local interrupts. 3922 39234.96 KVM_SMI 3924------------ 3925 3926:Capability: KVM_CAP_X86_SMM 3927:Architectures: x86 3928:Type: vcpu ioctl 3929:Parameters: none 3930:Returns: 0 on success, -1 on error 3931 3932Queues an SMI on the thread's vcpu. 3933 39344.97 KVM_X86_SET_MSR_FILTER 3935---------------------------- 3936 3937:Capability: KVM_X86_SET_MSR_FILTER 3938:Architectures: x86 3939:Type: vm ioctl 3940:Parameters: struct kvm_msr_filter 3941:Returns: 0 on success, < 0 on error 3942 3943:: 3944 3945 struct kvm_msr_filter_range { 3946 #define KVM_MSR_FILTER_READ (1 << 0) 3947 #define KVM_MSR_FILTER_WRITE (1 << 1) 3948 __u32 flags; 3949 __u32 nmsrs; /* number of msrs in bitmap */ 3950 __u32 base; /* MSR index the bitmap starts at */ 3951 __u8 *bitmap; /* a 1 bit allows the operations in flags, 0 denies */ 3952 }; 3953 3954 #define KVM_MSR_FILTER_MAX_RANGES 16 3955 struct kvm_msr_filter { 3956 #define KVM_MSR_FILTER_DEFAULT_ALLOW (0 << 0) 3957 #define KVM_MSR_FILTER_DEFAULT_DENY (1 << 0) 3958 __u32 flags; 3959 struct kvm_msr_filter_range ranges[KVM_MSR_FILTER_MAX_RANGES]; 3960 }; 3961 3962flags values for ``struct kvm_msr_filter_range``: 3963 3964``KVM_MSR_FILTER_READ`` 3965 3966 Filter read accesses to MSRs using the given bitmap. A 0 in the bitmap 3967 indicates that a read should immediately fail, while a 1 indicates that 3968 a read for a particular MSR should be handled regardless of the default 3969 filter action. 3970 3971``KVM_MSR_FILTER_WRITE`` 3972 3973 Filter write accesses to MSRs using the given bitmap. A 0 in the bitmap 3974 indicates that a write should immediately fail, while a 1 indicates that 3975 a write for a particular MSR should be handled regardless of the default 3976 filter action. 3977 3978``KVM_MSR_FILTER_READ | KVM_MSR_FILTER_WRITE`` 3979 3980 Filter both read and write accesses to MSRs using the given bitmap. A 0 3981 in the bitmap indicates that both reads and writes should immediately fail, 3982 while a 1 indicates that reads and writes for a particular MSR are not 3983 filtered by this range. 3984 3985flags values for ``struct kvm_msr_filter``: 3986 3987``KVM_MSR_FILTER_DEFAULT_ALLOW`` 3988 3989 If no filter range matches an MSR index that is getting accessed, KVM will 3990 fall back to allowing access to the MSR. 3991 3992``KVM_MSR_FILTER_DEFAULT_DENY`` 3993 3994 If no filter range matches an MSR index that is getting accessed, KVM will 3995 fall back to rejecting access to the MSR. In this mode, all MSRs that should 3996 be processed by KVM need to explicitly be marked as allowed in the bitmaps. 3997 3998This ioctl allows user space to define up to 16 bitmaps of MSR ranges to 3999specify whether a certain MSR access should be explicitly filtered for or not. 4000 4001If this ioctl has never been invoked, MSR accesses are not guarded and the 4002default KVM in-kernel emulation behavior is fully preserved. 4003 4004Calling this ioctl with an empty set of ranges (all nmsrs == 0) disables MSR 4005filtering. In that mode, ``KVM_MSR_FILTER_DEFAULT_DENY`` is invalid and causes 4006an error. 4007 4008As soon as the filtering is in place, every MSR access is processed through 4009the filtering except for accesses to the x2APIC MSRs (from 0x800 to 0x8ff); 4010x2APIC MSRs are always allowed, independent of the ``default_allow`` setting, 4011and their behavior depends on the ``X2APIC_ENABLE`` bit of the APIC base 4012register. 4013 4014If a bit is within one of the defined ranges, read and write accesses are 4015guarded by the bitmap's value for the MSR index if the kind of access 4016is included in the ``struct kvm_msr_filter_range`` flags. If no range 4017cover this particular access, the behavior is determined by the flags 4018field in the kvm_msr_filter struct: ``KVM_MSR_FILTER_DEFAULT_ALLOW`` 4019and ``KVM_MSR_FILTER_DEFAULT_DENY``. 4020 4021Each bitmap range specifies a range of MSRs to potentially allow access on. 4022The range goes from MSR index [base .. base+nmsrs]. The flags field 4023indicates whether reads, writes or both reads and writes are filtered 4024by setting a 1 bit in the bitmap for the corresponding MSR index. 4025 4026If an MSR access is not permitted through the filtering, it generates a 4027#GP inside the guest. When combined with KVM_CAP_X86_USER_SPACE_MSR, that 4028allows user space to deflect and potentially handle various MSR accesses 4029into user space. 4030 4031If a vCPU is in running state while this ioctl is invoked, the vCPU may 4032experience inconsistent filtering behavior on MSR accesses. 4033 40344.98 KVM_CREATE_SPAPR_TCE_64 4035---------------------------- 4036 4037:Capability: KVM_CAP_SPAPR_TCE_64 4038:Architectures: powerpc 4039:Type: vm ioctl 4040:Parameters: struct kvm_create_spapr_tce_64 (in) 4041:Returns: file descriptor for manipulating the created TCE table 4042 4043This is an extension for KVM_CAP_SPAPR_TCE which only supports 32bit 4044windows, described in 4.62 KVM_CREATE_SPAPR_TCE 4045 4046This capability uses extended struct in ioctl interface:: 4047 4048 /* for KVM_CAP_SPAPR_TCE_64 */ 4049 struct kvm_create_spapr_tce_64 { 4050 __u64 liobn; 4051 __u32 page_shift; 4052 __u32 flags; 4053 __u64 offset; /* in pages */ 4054 __u64 size; /* in pages */ 4055 }; 4056 4057The aim of extension is to support an additional bigger DMA window with 4058a variable page size. 4059KVM_CREATE_SPAPR_TCE_64 receives a 64bit window size, an IOMMU page shift and 4060a bus offset of the corresponding DMA window, @size and @offset are numbers 4061of IOMMU pages. 4062 4063@flags are not used at the moment. 4064 4065The rest of functionality is identical to KVM_CREATE_SPAPR_TCE. 4066 40674.99 KVM_REINJECT_CONTROL 4068------------------------- 4069 4070:Capability: KVM_CAP_REINJECT_CONTROL 4071:Architectures: x86 4072:Type: vm ioctl 4073:Parameters: struct kvm_reinject_control (in) 4074:Returns: 0 on success, 4075 -EFAULT if struct kvm_reinject_control cannot be read, 4076 -ENXIO if KVM_CREATE_PIT or KVM_CREATE_PIT2 didn't succeed earlier. 4077 4078i8254 (PIT) has two modes, reinject and !reinject. The default is reinject, 4079where KVM queues elapsed i8254 ticks and monitors completion of interrupt from 4080vector(s) that i8254 injects. Reinject mode dequeues a tick and injects its 4081interrupt whenever there isn't a pending interrupt from i8254. 4082!reinject mode injects an interrupt as soon as a tick arrives. 4083 4084:: 4085 4086 struct kvm_reinject_control { 4087 __u8 pit_reinject; 4088 __u8 reserved[31]; 4089 }; 4090 4091pit_reinject = 0 (!reinject mode) is recommended, unless running an old 4092operating system that uses the PIT for timing (e.g. Linux 2.4.x). 4093 40944.100 KVM_PPC_CONFIGURE_V3_MMU 4095------------------------------ 4096 4097:Capability: KVM_CAP_PPC_RADIX_MMU or KVM_CAP_PPC_HASH_MMU_V3 4098:Architectures: ppc 4099:Type: vm ioctl 4100:Parameters: struct kvm_ppc_mmuv3_cfg (in) 4101:Returns: 0 on success, 4102 -EFAULT if struct kvm_ppc_mmuv3_cfg cannot be read, 4103 -EINVAL if the configuration is invalid 4104 4105This ioctl controls whether the guest will use radix or HPT (hashed 4106page table) translation, and sets the pointer to the process table for 4107the guest. 4108 4109:: 4110 4111 struct kvm_ppc_mmuv3_cfg { 4112 __u64 flags; 4113 __u64 process_table; 4114 }; 4115 4116There are two bits that can be set in flags; KVM_PPC_MMUV3_RADIX and 4117KVM_PPC_MMUV3_GTSE. KVM_PPC_MMUV3_RADIX, if set, configures the guest 4118to use radix tree translation, and if clear, to use HPT translation. 4119KVM_PPC_MMUV3_GTSE, if set and if KVM permits it, configures the guest 4120to be able to use the global TLB and SLB invalidation instructions; 4121if clear, the guest may not use these instructions. 4122 4123The process_table field specifies the address and size of the guest 4124process table, which is in the guest's space. This field is formatted 4125as the second doubleword of the partition table entry, as defined in 4126the Power ISA V3.00, Book III section 5.7.6.1. 4127 41284.101 KVM_PPC_GET_RMMU_INFO 4129--------------------------- 4130 4131:Capability: KVM_CAP_PPC_RADIX_MMU 4132:Architectures: ppc 4133:Type: vm ioctl 4134:Parameters: struct kvm_ppc_rmmu_info (out) 4135:Returns: 0 on success, 4136 -EFAULT if struct kvm_ppc_rmmu_info cannot be written, 4137 -EINVAL if no useful information can be returned 4138 4139This ioctl returns a structure containing two things: (a) a list 4140containing supported radix tree geometries, and (b) a list that maps 4141page sizes to put in the "AP" (actual page size) field for the tlbie 4142(TLB invalidate entry) instruction. 4143 4144:: 4145 4146 struct kvm_ppc_rmmu_info { 4147 struct kvm_ppc_radix_geom { 4148 __u8 page_shift; 4149 __u8 level_bits[4]; 4150 __u8 pad[3]; 4151 } geometries[8]; 4152 __u32 ap_encodings[8]; 4153 }; 4154 4155The geometries[] field gives up to 8 supported geometries for the 4156radix page table, in terms of the log base 2 of the smallest page 4157size, and the number of bits indexed at each level of the tree, from 4158the PTE level up to the PGD level in that order. Any unused entries 4159will have 0 in the page_shift field. 4160 4161The ap_encodings gives the supported page sizes and their AP field 4162encodings, encoded with the AP value in the top 3 bits and the log 4163base 2 of the page size in the bottom 6 bits. 4164 41654.102 KVM_PPC_RESIZE_HPT_PREPARE 4166-------------------------------- 4167 4168:Capability: KVM_CAP_SPAPR_RESIZE_HPT 4169:Architectures: powerpc 4170:Type: vm ioctl 4171:Parameters: struct kvm_ppc_resize_hpt (in) 4172:Returns: 0 on successful completion, 4173 >0 if a new HPT is being prepared, the value is an estimated 4174 number of milliseconds until preparation is complete, 4175 -EFAULT if struct kvm_reinject_control cannot be read, 4176 -EINVAL if the supplied shift or flags are invalid, 4177 -ENOMEM if unable to allocate the new HPT, 4178 4179Used to implement the PAPR extension for runtime resizing of a guest's 4180Hashed Page Table (HPT). Specifically this starts, stops or monitors 4181the preparation of a new potential HPT for the guest, essentially 4182implementing the H_RESIZE_HPT_PREPARE hypercall. 4183 4184:: 4185 4186 struct kvm_ppc_resize_hpt { 4187 __u64 flags; 4188 __u32 shift; 4189 __u32 pad; 4190 }; 4191 4192If called with shift > 0 when there is no pending HPT for the guest, 4193this begins preparation of a new pending HPT of size 2^(shift) bytes. 4194It then returns a positive integer with the estimated number of 4195milliseconds until preparation is complete. 4196 4197If called when there is a pending HPT whose size does not match that 4198requested in the parameters, discards the existing pending HPT and 4199creates a new one as above. 4200 4201If called when there is a pending HPT of the size requested, will: 4202 4203 * If preparation of the pending HPT is already complete, return 0 4204 * If preparation of the pending HPT has failed, return an error 4205 code, then discard the pending HPT. 4206 * If preparation of the pending HPT is still in progress, return an 4207 estimated number of milliseconds until preparation is complete. 4208 4209If called with shift == 0, discards any currently pending HPT and 4210returns 0 (i.e. cancels any in-progress preparation). 4211 4212flags is reserved for future expansion, currently setting any bits in 4213flags will result in an -EINVAL. 4214 4215Normally this will be called repeatedly with the same parameters until 4216it returns <= 0. The first call will initiate preparation, subsequent 4217ones will monitor preparation until it completes or fails. 4218 42194.103 KVM_PPC_RESIZE_HPT_COMMIT 4220------------------------------- 4221 4222:Capability: KVM_CAP_SPAPR_RESIZE_HPT 4223:Architectures: powerpc 4224:Type: vm ioctl 4225:Parameters: struct kvm_ppc_resize_hpt (in) 4226:Returns: 0 on successful completion, 4227 -EFAULT if struct kvm_reinject_control cannot be read, 4228 -EINVAL if the supplied shift or flags are invalid, 4229 -ENXIO is there is no pending HPT, or the pending HPT doesn't 4230 have the requested size, 4231 -EBUSY if the pending HPT is not fully prepared, 4232 -ENOSPC if there was a hash collision when moving existing 4233 HPT entries to the new HPT, 4234 -EIO on other error conditions 4235 4236Used to implement the PAPR extension for runtime resizing of a guest's 4237Hashed Page Table (HPT). Specifically this requests that the guest be 4238transferred to working with the new HPT, essentially implementing the 4239H_RESIZE_HPT_COMMIT hypercall. 4240 4241:: 4242 4243 struct kvm_ppc_resize_hpt { 4244 __u64 flags; 4245 __u32 shift; 4246 __u32 pad; 4247 }; 4248 4249This should only be called after KVM_PPC_RESIZE_HPT_PREPARE has 4250returned 0 with the same parameters. In other cases 4251KVM_PPC_RESIZE_HPT_COMMIT will return an error (usually -ENXIO or 4252-EBUSY, though others may be possible if the preparation was started, 4253but failed). 4254 4255This will have undefined effects on the guest if it has not already 4256placed itself in a quiescent state where no vcpu will make MMU enabled 4257memory accesses. 4258 4259On succsful completion, the pending HPT will become the guest's active 4260HPT and the previous HPT will be discarded. 4261 4262On failure, the guest will still be operating on its previous HPT. 4263 42644.104 KVM_X86_GET_MCE_CAP_SUPPORTED 4265----------------------------------- 4266 4267:Capability: KVM_CAP_MCE 4268:Architectures: x86 4269:Type: system ioctl 4270:Parameters: u64 mce_cap (out) 4271:Returns: 0 on success, -1 on error 4272 4273Returns supported MCE capabilities. The u64 mce_cap parameter 4274has the same format as the MSR_IA32_MCG_CAP register. Supported 4275capabilities will have the corresponding bits set. 4276 42774.105 KVM_X86_SETUP_MCE 4278----------------------- 4279 4280:Capability: KVM_CAP_MCE 4281:Architectures: x86 4282:Type: vcpu ioctl 4283:Parameters: u64 mcg_cap (in) 4284:Returns: 0 on success, 4285 -EFAULT if u64 mcg_cap cannot be read, 4286 -EINVAL if the requested number of banks is invalid, 4287 -EINVAL if requested MCE capability is not supported. 4288 4289Initializes MCE support for use. The u64 mcg_cap parameter 4290has the same format as the MSR_IA32_MCG_CAP register and 4291specifies which capabilities should be enabled. The maximum 4292supported number of error-reporting banks can be retrieved when 4293checking for KVM_CAP_MCE. The supported capabilities can be 4294retrieved with KVM_X86_GET_MCE_CAP_SUPPORTED. 4295 42964.106 KVM_X86_SET_MCE 4297--------------------- 4298 4299:Capability: KVM_CAP_MCE 4300:Architectures: x86 4301:Type: vcpu ioctl 4302:Parameters: struct kvm_x86_mce (in) 4303:Returns: 0 on success, 4304 -EFAULT if struct kvm_x86_mce cannot be read, 4305 -EINVAL if the bank number is invalid, 4306 -EINVAL if VAL bit is not set in status field. 4307 4308Inject a machine check error (MCE) into the guest. The input 4309parameter is:: 4310 4311 struct kvm_x86_mce { 4312 __u64 status; 4313 __u64 addr; 4314 __u64 misc; 4315 __u64 mcg_status; 4316 __u8 bank; 4317 __u8 pad1[7]; 4318 __u64 pad2[3]; 4319 }; 4320 4321If the MCE being reported is an uncorrected error, KVM will 4322inject it as an MCE exception into the guest. If the guest 4323MCG_STATUS register reports that an MCE is in progress, KVM 4324causes an KVM_EXIT_SHUTDOWN vmexit. 4325 4326Otherwise, if the MCE is a corrected error, KVM will just 4327store it in the corresponding bank (provided this bank is 4328not holding a previously reported uncorrected error). 4329 43304.107 KVM_S390_GET_CMMA_BITS 4331---------------------------- 4332 4333:Capability: KVM_CAP_S390_CMMA_MIGRATION 4334:Architectures: s390 4335:Type: vm ioctl 4336:Parameters: struct kvm_s390_cmma_log (in, out) 4337:Returns: 0 on success, a negative value on error 4338 4339This ioctl is used to get the values of the CMMA bits on the s390 4340architecture. It is meant to be used in two scenarios: 4341 4342- During live migration to save the CMMA values. Live migration needs 4343 to be enabled via the KVM_REQ_START_MIGRATION VM property. 4344- To non-destructively peek at the CMMA values, with the flag 4345 KVM_S390_CMMA_PEEK set. 4346 4347The ioctl takes parameters via the kvm_s390_cmma_log struct. The desired 4348values are written to a buffer whose location is indicated via the "values" 4349member in the kvm_s390_cmma_log struct. The values in the input struct are 4350also updated as needed. 4351 4352Each CMMA value takes up one byte. 4353 4354:: 4355 4356 struct kvm_s390_cmma_log { 4357 __u64 start_gfn; 4358 __u32 count; 4359 __u32 flags; 4360 union { 4361 __u64 remaining; 4362 __u64 mask; 4363 }; 4364 __u64 values; 4365 }; 4366 4367start_gfn is the number of the first guest frame whose CMMA values are 4368to be retrieved, 4369 4370count is the length of the buffer in bytes, 4371 4372values points to the buffer where the result will be written to. 4373 4374If count is greater than KVM_S390_SKEYS_MAX, then it is considered to be 4375KVM_S390_SKEYS_MAX. KVM_S390_SKEYS_MAX is re-used for consistency with 4376other ioctls. 4377 4378The result is written in the buffer pointed to by the field values, and 4379the values of the input parameter are updated as follows. 4380 4381Depending on the flags, different actions are performed. The only 4382supported flag so far is KVM_S390_CMMA_PEEK. 4383 4384The default behaviour if KVM_S390_CMMA_PEEK is not set is: 4385start_gfn will indicate the first page frame whose CMMA bits were dirty. 4386It is not necessarily the same as the one passed as input, as clean pages 4387are skipped. 4388 4389count will indicate the number of bytes actually written in the buffer. 4390It can (and very often will) be smaller than the input value, since the 4391buffer is only filled until 16 bytes of clean values are found (which 4392are then not copied in the buffer). Since a CMMA migration block needs 4393the base address and the length, for a total of 16 bytes, we will send 4394back some clean data if there is some dirty data afterwards, as long as 4395the size of the clean data does not exceed the size of the header. This 4396allows to minimize the amount of data to be saved or transferred over 4397the network at the expense of more roundtrips to userspace. The next 4398invocation of the ioctl will skip over all the clean values, saving 4399potentially more than just the 16 bytes we found. 4400 4401If KVM_S390_CMMA_PEEK is set: 4402the existing storage attributes are read even when not in migration 4403mode, and no other action is performed; 4404 4405the output start_gfn will be equal to the input start_gfn, 4406 4407the output count will be equal to the input count, except if the end of 4408memory has been reached. 4409 4410In both cases: 4411the field "remaining" will indicate the total number of dirty CMMA values 4412still remaining, or 0 if KVM_S390_CMMA_PEEK is set and migration mode is 4413not enabled. 4414 4415mask is unused. 4416 4417values points to the userspace buffer where the result will be stored. 4418 4419This ioctl can fail with -ENOMEM if not enough memory can be allocated to 4420complete the task, with -ENXIO if CMMA is not enabled, with -EINVAL if 4421KVM_S390_CMMA_PEEK is not set but migration mode was not enabled, with 4422-EFAULT if the userspace address is invalid or if no page table is 4423present for the addresses (e.g. when using hugepages). 4424 44254.108 KVM_S390_SET_CMMA_BITS 4426---------------------------- 4427 4428:Capability: KVM_CAP_S390_CMMA_MIGRATION 4429:Architectures: s390 4430:Type: vm ioctl 4431:Parameters: struct kvm_s390_cmma_log (in) 4432:Returns: 0 on success, a negative value on error 4433 4434This ioctl is used to set the values of the CMMA bits on the s390 4435architecture. It is meant to be used during live migration to restore 4436the CMMA values, but there are no restrictions on its use. 4437The ioctl takes parameters via the kvm_s390_cmma_values struct. 4438Each CMMA value takes up one byte. 4439 4440:: 4441 4442 struct kvm_s390_cmma_log { 4443 __u64 start_gfn; 4444 __u32 count; 4445 __u32 flags; 4446 union { 4447 __u64 remaining; 4448 __u64 mask; 4449 }; 4450 __u64 values; 4451 }; 4452 4453start_gfn indicates the starting guest frame number, 4454 4455count indicates how many values are to be considered in the buffer, 4456 4457flags is not used and must be 0. 4458 4459mask indicates which PGSTE bits are to be considered. 4460 4461remaining is not used. 4462 4463values points to the buffer in userspace where to store the values. 4464 4465This ioctl can fail with -ENOMEM if not enough memory can be allocated to 4466complete the task, with -ENXIO if CMMA is not enabled, with -EINVAL if 4467the count field is too large (e.g. more than KVM_S390_CMMA_SIZE_MAX) or 4468if the flags field was not 0, with -EFAULT if the userspace address is 4469invalid, if invalid pages are written to (e.g. after the end of memory) 4470or if no page table is present for the addresses (e.g. when using 4471hugepages). 4472 44734.109 KVM_PPC_GET_CPU_CHAR 4474-------------------------- 4475 4476:Capability: KVM_CAP_PPC_GET_CPU_CHAR 4477:Architectures: powerpc 4478:Type: vm ioctl 4479:Parameters: struct kvm_ppc_cpu_char (out) 4480:Returns: 0 on successful completion, 4481 -EFAULT if struct kvm_ppc_cpu_char cannot be written 4482 4483This ioctl gives userspace information about certain characteristics 4484of the CPU relating to speculative execution of instructions and 4485possible information leakage resulting from speculative execution (see 4486CVE-2017-5715, CVE-2017-5753 and CVE-2017-5754). The information is 4487returned in struct kvm_ppc_cpu_char, which looks like this:: 4488 4489 struct kvm_ppc_cpu_char { 4490 __u64 character; /* characteristics of the CPU */ 4491 __u64 behaviour; /* recommended software behaviour */ 4492 __u64 character_mask; /* valid bits in character */ 4493 __u64 behaviour_mask; /* valid bits in behaviour */ 4494 }; 4495 4496For extensibility, the character_mask and behaviour_mask fields 4497indicate which bits of character and behaviour have been filled in by 4498the kernel. If the set of defined bits is extended in future then 4499userspace will be able to tell whether it is running on a kernel that 4500knows about the new bits. 4501 4502The character field describes attributes of the CPU which can help 4503with preventing inadvertent information disclosure - specifically, 4504whether there is an instruction to flash-invalidate the L1 data cache 4505(ori 30,30,0 or mtspr SPRN_TRIG2,rN), whether the L1 data cache is set 4506to a mode where entries can only be used by the thread that created 4507them, whether the bcctr[l] instruction prevents speculation, and 4508whether a speculation barrier instruction (ori 31,31,0) is provided. 4509 4510The behaviour field describes actions that software should take to 4511prevent inadvertent information disclosure, and thus describes which 4512vulnerabilities the hardware is subject to; specifically whether the 4513L1 data cache should be flushed when returning to user mode from the 4514kernel, and whether a speculation barrier should be placed between an 4515array bounds check and the array access. 4516 4517These fields use the same bit definitions as the new 4518H_GET_CPU_CHARACTERISTICS hypercall. 4519 45204.110 KVM_MEMORY_ENCRYPT_OP 4521--------------------------- 4522 4523:Capability: basic 4524:Architectures: x86 4525:Type: vm 4526:Parameters: an opaque platform specific structure (in/out) 4527:Returns: 0 on success; -1 on error 4528 4529If the platform supports creating encrypted VMs then this ioctl can be used 4530for issuing platform-specific memory encryption commands to manage those 4531encrypted VMs. 4532 4533Currently, this ioctl is used for issuing Secure Encrypted Virtualization 4534(SEV) commands on AMD Processors. The SEV commands are defined in 4535Documentation/virt/kvm/amd-memory-encryption.rst. 4536 45374.111 KVM_MEMORY_ENCRYPT_REG_REGION 4538----------------------------------- 4539 4540:Capability: basic 4541:Architectures: x86 4542:Type: system 4543:Parameters: struct kvm_enc_region (in) 4544:Returns: 0 on success; -1 on error 4545 4546This ioctl can be used to register a guest memory region which may 4547contain encrypted data (e.g. guest RAM, SMRAM etc). 4548 4549It is used in the SEV-enabled guest. When encryption is enabled, a guest 4550memory region may contain encrypted data. The SEV memory encryption 4551engine uses a tweak such that two identical plaintext pages, each at 4552different locations will have differing ciphertexts. So swapping or 4553moving ciphertext of those pages will not result in plaintext being 4554swapped. So relocating (or migrating) physical backing pages for the SEV 4555guest will require some additional steps. 4556 4557Note: The current SEV key management spec does not provide commands to 4558swap or migrate (move) ciphertext pages. Hence, for now we pin the guest 4559memory region registered with the ioctl. 4560 45614.112 KVM_MEMORY_ENCRYPT_UNREG_REGION 4562------------------------------------- 4563 4564:Capability: basic 4565:Architectures: x86 4566:Type: system 4567:Parameters: struct kvm_enc_region (in) 4568:Returns: 0 on success; -1 on error 4569 4570This ioctl can be used to unregister the guest memory region registered 4571with KVM_MEMORY_ENCRYPT_REG_REGION ioctl above. 4572 45734.113 KVM_HYPERV_EVENTFD 4574------------------------ 4575 4576:Capability: KVM_CAP_HYPERV_EVENTFD 4577:Architectures: x86 4578:Type: vm ioctl 4579:Parameters: struct kvm_hyperv_eventfd (in) 4580 4581This ioctl (un)registers an eventfd to receive notifications from the guest on 4582the specified Hyper-V connection id through the SIGNAL_EVENT hypercall, without 4583causing a user exit. SIGNAL_EVENT hypercall with non-zero event flag number 4584(bits 24-31) still triggers a KVM_EXIT_HYPERV_HCALL user exit. 4585 4586:: 4587 4588 struct kvm_hyperv_eventfd { 4589 __u32 conn_id; 4590 __s32 fd; 4591 __u32 flags; 4592 __u32 padding[3]; 4593 }; 4594 4595The conn_id field should fit within 24 bits:: 4596 4597 #define KVM_HYPERV_CONN_ID_MASK 0x00ffffff 4598 4599The acceptable values for the flags field are:: 4600 4601 #define KVM_HYPERV_EVENTFD_DEASSIGN (1 << 0) 4602 4603:Returns: 0 on success, 4604 -EINVAL if conn_id or flags is outside the allowed range, 4605 -ENOENT on deassign if the conn_id isn't registered, 4606 -EEXIST on assign if the conn_id is already registered 4607 46084.114 KVM_GET_NESTED_STATE 4609-------------------------- 4610 4611:Capability: KVM_CAP_NESTED_STATE 4612:Architectures: x86 4613:Type: vcpu ioctl 4614:Parameters: struct kvm_nested_state (in/out) 4615:Returns: 0 on success, -1 on error 4616 4617Errors: 4618 4619 ===== ============================================================= 4620 E2BIG the total state size exceeds the value of 'size' specified by 4621 the user; the size required will be written into size. 4622 ===== ============================================================= 4623 4624:: 4625 4626 struct kvm_nested_state { 4627 __u16 flags; 4628 __u16 format; 4629 __u32 size; 4630 4631 union { 4632 struct kvm_vmx_nested_state_hdr vmx; 4633 struct kvm_svm_nested_state_hdr svm; 4634 4635 /* Pad the header to 128 bytes. */ 4636 __u8 pad[120]; 4637 } hdr; 4638 4639 union { 4640 struct kvm_vmx_nested_state_data vmx[0]; 4641 struct kvm_svm_nested_state_data svm[0]; 4642 } data; 4643 }; 4644 4645 #define KVM_STATE_NESTED_GUEST_MODE 0x00000001 4646 #define KVM_STATE_NESTED_RUN_PENDING 0x00000002 4647 #define KVM_STATE_NESTED_EVMCS 0x00000004 4648 4649 #define KVM_STATE_NESTED_FORMAT_VMX 0 4650 #define KVM_STATE_NESTED_FORMAT_SVM 1 4651 4652 #define KVM_STATE_NESTED_VMX_VMCS_SIZE 0x1000 4653 4654 #define KVM_STATE_NESTED_VMX_SMM_GUEST_MODE 0x00000001 4655 #define KVM_STATE_NESTED_VMX_SMM_VMXON 0x00000002 4656 4657 #define KVM_STATE_VMX_PREEMPTION_TIMER_DEADLINE 0x00000001 4658 4659 struct kvm_vmx_nested_state_hdr { 4660 __u64 vmxon_pa; 4661 __u64 vmcs12_pa; 4662 4663 struct { 4664 __u16 flags; 4665 } smm; 4666 4667 __u32 flags; 4668 __u64 preemption_timer_deadline; 4669 }; 4670 4671 struct kvm_vmx_nested_state_data { 4672 __u8 vmcs12[KVM_STATE_NESTED_VMX_VMCS_SIZE]; 4673 __u8 shadow_vmcs12[KVM_STATE_NESTED_VMX_VMCS_SIZE]; 4674 }; 4675 4676This ioctl copies the vcpu's nested virtualization state from the kernel to 4677userspace. 4678 4679The maximum size of the state can be retrieved by passing KVM_CAP_NESTED_STATE 4680to the KVM_CHECK_EXTENSION ioctl(). 4681 46824.115 KVM_SET_NESTED_STATE 4683-------------------------- 4684 4685:Capability: KVM_CAP_NESTED_STATE 4686:Architectures: x86 4687:Type: vcpu ioctl 4688:Parameters: struct kvm_nested_state (in) 4689:Returns: 0 on success, -1 on error 4690 4691This copies the vcpu's kvm_nested_state struct from userspace to the kernel. 4692For the definition of struct kvm_nested_state, see KVM_GET_NESTED_STATE. 4693 46944.116 KVM_(UN)REGISTER_COALESCED_MMIO 4695------------------------------------- 4696 4697:Capability: KVM_CAP_COALESCED_MMIO (for coalesced mmio) 4698 KVM_CAP_COALESCED_PIO (for coalesced pio) 4699:Architectures: all 4700:Type: vm ioctl 4701:Parameters: struct kvm_coalesced_mmio_zone 4702:Returns: 0 on success, < 0 on error 4703 4704Coalesced I/O is a performance optimization that defers hardware 4705register write emulation so that userspace exits are avoided. It is 4706typically used to reduce the overhead of emulating frequently accessed 4707hardware registers. 4708 4709When a hardware register is configured for coalesced I/O, write accesses 4710do not exit to userspace and their value is recorded in a ring buffer 4711that is shared between kernel and userspace. 4712 4713Coalesced I/O is used if one or more write accesses to a hardware 4714register can be deferred until a read or a write to another hardware 4715register on the same device. This last access will cause a vmexit and 4716userspace will process accesses from the ring buffer before emulating 4717it. That will avoid exiting to userspace on repeated writes. 4718 4719Coalesced pio is based on coalesced mmio. There is little difference 4720between coalesced mmio and pio except that coalesced pio records accesses 4721to I/O ports. 4722 47234.117 KVM_CLEAR_DIRTY_LOG (vm ioctl) 4724------------------------------------ 4725 4726:Capability: KVM_CAP_MANUAL_DIRTY_LOG_PROTECT2 4727:Architectures: x86, arm, arm64, mips 4728:Type: vm ioctl 4729:Parameters: struct kvm_clear_dirty_log (in) 4730:Returns: 0 on success, -1 on error 4731 4732:: 4733 4734 /* for KVM_CLEAR_DIRTY_LOG */ 4735 struct kvm_clear_dirty_log { 4736 __u32 slot; 4737 __u32 num_pages; 4738 __u64 first_page; 4739 union { 4740 void __user *dirty_bitmap; /* one bit per page */ 4741 __u64 padding; 4742 }; 4743 }; 4744 4745The ioctl clears the dirty status of pages in a memory slot, according to 4746the bitmap that is passed in struct kvm_clear_dirty_log's dirty_bitmap 4747field. Bit 0 of the bitmap corresponds to page "first_page" in the 4748memory slot, and num_pages is the size in bits of the input bitmap. 4749first_page must be a multiple of 64; num_pages must also be a multiple of 475064 unless first_page + num_pages is the size of the memory slot. For each 4751bit that is set in the input bitmap, the corresponding page is marked "clean" 4752in KVM's dirty bitmap, and dirty tracking is re-enabled for that page 4753(for example via write-protection, or by clearing the dirty bit in 4754a page table entry). 4755 4756If KVM_CAP_MULTI_ADDRESS_SPACE is available, bits 16-31 of slot field specifies 4757the address space for which you want to clear the dirty status. See 4758KVM_SET_USER_MEMORY_REGION for details on the usage of slot field. 4759 4760This ioctl is mostly useful when KVM_CAP_MANUAL_DIRTY_LOG_PROTECT2 4761is enabled; for more information, see the description of the capability. 4762However, it can always be used as long as KVM_CHECK_EXTENSION confirms 4763that KVM_CAP_MANUAL_DIRTY_LOG_PROTECT2 is present. 4764 47654.118 KVM_GET_SUPPORTED_HV_CPUID 4766-------------------------------- 4767 4768:Capability: KVM_CAP_HYPERV_CPUID (vcpu), KVM_CAP_SYS_HYPERV_CPUID (system) 4769:Architectures: x86 4770:Type: system ioctl, vcpu ioctl 4771:Parameters: struct kvm_cpuid2 (in/out) 4772:Returns: 0 on success, -1 on error 4773 4774:: 4775 4776 struct kvm_cpuid2 { 4777 __u32 nent; 4778 __u32 padding; 4779 struct kvm_cpuid_entry2 entries[0]; 4780 }; 4781 4782 struct kvm_cpuid_entry2 { 4783 __u32 function; 4784 __u32 index; 4785 __u32 flags; 4786 __u32 eax; 4787 __u32 ebx; 4788 __u32 ecx; 4789 __u32 edx; 4790 __u32 padding[3]; 4791 }; 4792 4793This ioctl returns x86 cpuid features leaves related to Hyper-V emulation in 4794KVM. Userspace can use the information returned by this ioctl to construct 4795cpuid information presented to guests consuming Hyper-V enlightenments (e.g. 4796Windows or Hyper-V guests). 4797 4798CPUID feature leaves returned by this ioctl are defined by Hyper-V Top Level 4799Functional Specification (TLFS). These leaves can't be obtained with 4800KVM_GET_SUPPORTED_CPUID ioctl because some of them intersect with KVM feature 4801leaves (0x40000000, 0x40000001). 4802 4803Currently, the following list of CPUID leaves are returned: 4804 4805 - HYPERV_CPUID_VENDOR_AND_MAX_FUNCTIONS 4806 - HYPERV_CPUID_INTERFACE 4807 - HYPERV_CPUID_VERSION 4808 - HYPERV_CPUID_FEATURES 4809 - HYPERV_CPUID_ENLIGHTMENT_INFO 4810 - HYPERV_CPUID_IMPLEMENT_LIMITS 4811 - HYPERV_CPUID_NESTED_FEATURES 4812 - HYPERV_CPUID_SYNDBG_VENDOR_AND_MAX_FUNCTIONS 4813 - HYPERV_CPUID_SYNDBG_INTERFACE 4814 - HYPERV_CPUID_SYNDBG_PLATFORM_CAPABILITIES 4815 4816Userspace invokes KVM_GET_SUPPORTED_HV_CPUID by passing a kvm_cpuid2 structure 4817with the 'nent' field indicating the number of entries in the variable-size 4818array 'entries'. If the number of entries is too low to describe all Hyper-V 4819feature leaves, an error (E2BIG) is returned. If the number is more or equal 4820to the number of Hyper-V feature leaves, the 'nent' field is adjusted to the 4821number of valid entries in the 'entries' array, which is then filled. 4822 4823'index' and 'flags' fields in 'struct kvm_cpuid_entry2' are currently reserved, 4824userspace should not expect to get any particular value there. 4825 4826Note, vcpu version of KVM_GET_SUPPORTED_HV_CPUID is currently deprecated. Unlike 4827system ioctl which exposes all supported feature bits unconditionally, vcpu 4828version has the following quirks: 4829 4830- HYPERV_CPUID_NESTED_FEATURES leaf and HV_X64_ENLIGHTENED_VMCS_RECOMMENDED 4831 feature bit are only exposed when Enlightened VMCS was previously enabled 4832 on the corresponding vCPU (KVM_CAP_HYPERV_ENLIGHTENED_VMCS). 4833- HV_STIMER_DIRECT_MODE_AVAILABLE bit is only exposed with in-kernel LAPIC. 4834 (presumes KVM_CREATE_IRQCHIP has already been called). 4835 48364.119 KVM_ARM_VCPU_FINALIZE 4837--------------------------- 4838 4839:Architectures: arm, arm64 4840:Type: vcpu ioctl 4841:Parameters: int feature (in) 4842:Returns: 0 on success, -1 on error 4843 4844Errors: 4845 4846 ====== ============================================================== 4847 EPERM feature not enabled, needs configuration, or already finalized 4848 EINVAL feature unknown or not present 4849 ====== ============================================================== 4850 4851Recognised values for feature: 4852 4853 ===== =========================================== 4854 arm64 KVM_ARM_VCPU_SVE (requires KVM_CAP_ARM_SVE) 4855 ===== =========================================== 4856 4857Finalizes the configuration of the specified vcpu feature. 4858 4859The vcpu must already have been initialised, enabling the affected feature, by 4860means of a successful KVM_ARM_VCPU_INIT call with the appropriate flag set in 4861features[]. 4862 4863For affected vcpu features, this is a mandatory step that must be performed 4864before the vcpu is fully usable. 4865 4866Between KVM_ARM_VCPU_INIT and KVM_ARM_VCPU_FINALIZE, the feature may be 4867configured by use of ioctls such as KVM_SET_ONE_REG. The exact configuration 4868that should be performaned and how to do it are feature-dependent. 4869 4870Other calls that depend on a particular feature being finalized, such as 4871KVM_RUN, KVM_GET_REG_LIST, KVM_GET_ONE_REG and KVM_SET_ONE_REG, will fail with 4872-EPERM unless the feature has already been finalized by means of a 4873KVM_ARM_VCPU_FINALIZE call. 4874 4875See KVM_ARM_VCPU_INIT for details of vcpu features that require finalization 4876using this ioctl. 4877 48784.120 KVM_SET_PMU_EVENT_FILTER 4879------------------------------ 4880 4881:Capability: KVM_CAP_PMU_EVENT_FILTER 4882:Architectures: x86 4883:Type: vm ioctl 4884:Parameters: struct kvm_pmu_event_filter (in) 4885:Returns: 0 on success, -1 on error 4886 4887:: 4888 4889 struct kvm_pmu_event_filter { 4890 __u32 action; 4891 __u32 nevents; 4892 __u32 fixed_counter_bitmap; 4893 __u32 flags; 4894 __u32 pad[4]; 4895 __u64 events[0]; 4896 }; 4897 4898This ioctl restricts the set of PMU events that the guest can program. 4899The argument holds a list of events which will be allowed or denied. 4900The eventsel+umask of each event the guest attempts to program is compared 4901against the events field to determine whether the guest should have access. 4902The events field only controls general purpose counters; fixed purpose 4903counters are controlled by the fixed_counter_bitmap. 4904 4905No flags are defined yet, the field must be zero. 4906 4907Valid values for 'action':: 4908 4909 #define KVM_PMU_EVENT_ALLOW 0 4910 #define KVM_PMU_EVENT_DENY 1 4911 49124.121 KVM_PPC_SVM_OFF 4913--------------------- 4914 4915:Capability: basic 4916:Architectures: powerpc 4917:Type: vm ioctl 4918:Parameters: none 4919:Returns: 0 on successful completion, 4920 4921Errors: 4922 4923 ====== ================================================================ 4924 EINVAL if ultravisor failed to terminate the secure guest 4925 ENOMEM if hypervisor failed to allocate new radix page tables for guest 4926 ====== ================================================================ 4927 4928This ioctl is used to turn off the secure mode of the guest or transition 4929the guest from secure mode to normal mode. This is invoked when the guest 4930is reset. This has no effect if called for a normal guest. 4931 4932This ioctl issues an ultravisor call to terminate the secure guest, 4933unpins the VPA pages and releases all the device pages that are used to 4934track the secure pages by hypervisor. 4935 49364.122 KVM_S390_NORMAL_RESET 4937--------------------------- 4938 4939:Capability: KVM_CAP_S390_VCPU_RESETS 4940:Architectures: s390 4941:Type: vcpu ioctl 4942:Parameters: none 4943:Returns: 0 4944 4945This ioctl resets VCPU registers and control structures according to 4946the cpu reset definition in the POP (Principles Of Operation). 4947 49484.123 KVM_S390_INITIAL_RESET 4949---------------------------- 4950 4951:Capability: none 4952:Architectures: s390 4953:Type: vcpu ioctl 4954:Parameters: none 4955:Returns: 0 4956 4957This ioctl resets VCPU registers and control structures according to 4958the initial cpu reset definition in the POP. However, the cpu is not 4959put into ESA mode. This reset is a superset of the normal reset. 4960 49614.124 KVM_S390_CLEAR_RESET 4962-------------------------- 4963 4964:Capability: KVM_CAP_S390_VCPU_RESETS 4965:Architectures: s390 4966:Type: vcpu ioctl 4967:Parameters: none 4968:Returns: 0 4969 4970This ioctl resets VCPU registers and control structures according to 4971the clear cpu reset definition in the POP. However, the cpu is not put 4972into ESA mode. This reset is a superset of the initial reset. 4973 4974 49754.125 KVM_S390_PV_COMMAND 4976------------------------- 4977 4978:Capability: KVM_CAP_S390_PROTECTED 4979:Architectures: s390 4980:Type: vm ioctl 4981:Parameters: struct kvm_pv_cmd 4982:Returns: 0 on success, < 0 on error 4983 4984:: 4985 4986 struct kvm_pv_cmd { 4987 __u32 cmd; /* Command to be executed */ 4988 __u16 rc; /* Ultravisor return code */ 4989 __u16 rrc; /* Ultravisor return reason code */ 4990 __u64 data; /* Data or address */ 4991 __u32 flags; /* flags for future extensions. Must be 0 for now */ 4992 __u32 reserved[3]; 4993 }; 4994 4995cmd values: 4996 4997KVM_PV_ENABLE 4998 Allocate memory and register the VM with the Ultravisor, thereby 4999 donating memory to the Ultravisor that will become inaccessible to 5000 KVM. All existing CPUs are converted to protected ones. After this 5001 command has succeeded, any CPU added via hotplug will become 5002 protected during its creation as well. 5003 5004 Errors: 5005 5006 ===== ============================= 5007 EINTR an unmasked signal is pending 5008 ===== ============================= 5009 5010KVM_PV_DISABLE 5011 5012 Deregister the VM from the Ultravisor and reclaim the memory that 5013 had been donated to the Ultravisor, making it usable by the kernel 5014 again. All registered VCPUs are converted back to non-protected 5015 ones. 5016 5017KVM_PV_VM_SET_SEC_PARMS 5018 Pass the image header from VM memory to the Ultravisor in 5019 preparation of image unpacking and verification. 5020 5021KVM_PV_VM_UNPACK 5022 Unpack (protect and decrypt) a page of the encrypted boot image. 5023 5024KVM_PV_VM_VERIFY 5025 Verify the integrity of the unpacked image. Only if this succeeds, 5026 KVM is allowed to start protected VCPUs. 5027 50284.126 KVM_X86_SET_MSR_FILTER 5029---------------------------- 5030 5031:Capability: KVM_CAP_X86_MSR_FILTER 5032:Architectures: x86 5033:Type: vm ioctl 5034:Parameters: struct kvm_msr_filter 5035:Returns: 0 on success, < 0 on error 5036 5037:: 5038 5039 struct kvm_msr_filter_range { 5040 #define KVM_MSR_FILTER_READ (1 << 0) 5041 #define KVM_MSR_FILTER_WRITE (1 << 1) 5042 __u32 flags; 5043 __u32 nmsrs; /* number of msrs in bitmap */ 5044 __u32 base; /* MSR index the bitmap starts at */ 5045 __u8 *bitmap; /* a 1 bit allows the operations in flags, 0 denies */ 5046 }; 5047 5048 #define KVM_MSR_FILTER_MAX_RANGES 16 5049 struct kvm_msr_filter { 5050 #define KVM_MSR_FILTER_DEFAULT_ALLOW (0 << 0) 5051 #define KVM_MSR_FILTER_DEFAULT_DENY (1 << 0) 5052 __u32 flags; 5053 struct kvm_msr_filter_range ranges[KVM_MSR_FILTER_MAX_RANGES]; 5054 }; 5055 5056flags values for ``struct kvm_msr_filter_range``: 5057 5058``KVM_MSR_FILTER_READ`` 5059 5060 Filter read accesses to MSRs using the given bitmap. A 0 in the bitmap 5061 indicates that a read should immediately fail, while a 1 indicates that 5062 a read for a particular MSR should be handled regardless of the default 5063 filter action. 5064 5065``KVM_MSR_FILTER_WRITE`` 5066 5067 Filter write accesses to MSRs using the given bitmap. A 0 in the bitmap 5068 indicates that a write should immediately fail, while a 1 indicates that 5069 a write for a particular MSR should be handled regardless of the default 5070 filter action. 5071 5072``KVM_MSR_FILTER_READ | KVM_MSR_FILTER_WRITE`` 5073 5074 Filter both read and write accesses to MSRs using the given bitmap. A 0 5075 in the bitmap indicates that both reads and writes should immediately fail, 5076 while a 1 indicates that reads and writes for a particular MSR are not 5077 filtered by this range. 5078 5079flags values for ``struct kvm_msr_filter``: 5080 5081``KVM_MSR_FILTER_DEFAULT_ALLOW`` 5082 5083 If no filter range matches an MSR index that is getting accessed, KVM will 5084 fall back to allowing access to the MSR. 5085 5086``KVM_MSR_FILTER_DEFAULT_DENY`` 5087 5088 If no filter range matches an MSR index that is getting accessed, KVM will 5089 fall back to rejecting access to the MSR. In this mode, all MSRs that should 5090 be processed by KVM need to explicitly be marked as allowed in the bitmaps. 5091 5092This ioctl allows user space to define up to 16 bitmaps of MSR ranges to 5093specify whether a certain MSR access should be explicitly filtered for or not. 5094 5095If this ioctl has never been invoked, MSR accesses are not guarded and the 5096default KVM in-kernel emulation behavior is fully preserved. 5097 5098Calling this ioctl with an empty set of ranges (all nmsrs == 0) disables MSR 5099filtering. In that mode, ``KVM_MSR_FILTER_DEFAULT_DENY`` is invalid and causes 5100an error. 5101 5102As soon as the filtering is in place, every MSR access is processed through 5103the filtering except for accesses to the x2APIC MSRs (from 0x800 to 0x8ff); 5104x2APIC MSRs are always allowed, independent of the ``default_allow`` setting, 5105and their behavior depends on the ``X2APIC_ENABLE`` bit of the APIC base 5106register. 5107 5108If a bit is within one of the defined ranges, read and write accesses are 5109guarded by the bitmap's value for the MSR index if the kind of access 5110is included in the ``struct kvm_msr_filter_range`` flags. If no range 5111cover this particular access, the behavior is determined by the flags 5112field in the kvm_msr_filter struct: ``KVM_MSR_FILTER_DEFAULT_ALLOW`` 5113and ``KVM_MSR_FILTER_DEFAULT_DENY``. 5114 5115Each bitmap range specifies a range of MSRs to potentially allow access on. 5116The range goes from MSR index [base .. base+nmsrs]. The flags field 5117indicates whether reads, writes or both reads and writes are filtered 5118by setting a 1 bit in the bitmap for the corresponding MSR index. 5119 5120If an MSR access is not permitted through the filtering, it generates a 5121#GP inside the guest. When combined with KVM_CAP_X86_USER_SPACE_MSR, that 5122allows user space to deflect and potentially handle various MSR accesses 5123into user space. 5124 5125Note, invoking this ioctl with a vCPU is running is inherently racy. However, 5126KVM does guarantee that vCPUs will see either the previous filter or the new 5127filter, e.g. MSRs with identical settings in both the old and new filter will 5128have deterministic behavior. 5129 51304.127 KVM_XEN_HVM_SET_ATTR 5131-------------------------- 5132 5133:Capability: KVM_CAP_XEN_HVM / KVM_XEN_HVM_CONFIG_SHARED_INFO 5134:Architectures: x86 5135:Type: vm ioctl 5136:Parameters: struct kvm_xen_hvm_attr 5137:Returns: 0 on success, < 0 on error 5138 5139:: 5140 5141 struct kvm_xen_hvm_attr { 5142 __u16 type; 5143 __u16 pad[3]; 5144 union { 5145 __u8 long_mode; 5146 __u8 vector; 5147 struct { 5148 __u64 gfn; 5149 } shared_info; 5150 __u64 pad[4]; 5151 } u; 5152 }; 5153 5154type values: 5155 5156KVM_XEN_ATTR_TYPE_LONG_MODE 5157 Sets the ABI mode of the VM to 32-bit or 64-bit (long mode). This 5158 determines the layout of the shared info pages exposed to the VM. 5159 5160KVM_XEN_ATTR_TYPE_SHARED_INFO 5161 Sets the guest physical frame number at which the Xen "shared info" 5162 page resides. Note that although Xen places vcpu_info for the first 5163 32 vCPUs in the shared_info page, KVM does not automatically do so 5164 and instead requires that KVM_XEN_VCPU_ATTR_TYPE_VCPU_INFO be used 5165 explicitly even when the vcpu_info for a given vCPU resides at the 5166 "default" location in the shared_info page. This is because KVM is 5167 not aware of the Xen CPU id which is used as the index into the 5168 vcpu_info[] array, so cannot know the correct default location. 5169 5170 Note that the shared info page may be constantly written to by KVM; 5171 it contains the event channel bitmap used to deliver interrupts to 5172 a Xen guest, amongst other things. It is exempt from dirty tracking 5173 mechanisms — KVM will not explicitly mark the page as dirty each 5174 time an event channel interrupt is delivered to the guest! Thus, 5175 userspace should always assume that the designated GFN is dirty if 5176 any vCPU has been running or any event channel interrupts can be 5177 routed to the guest. 5178 5179KVM_XEN_ATTR_TYPE_UPCALL_VECTOR 5180 Sets the exception vector used to deliver Xen event channel upcalls. 5181 51824.127 KVM_XEN_HVM_GET_ATTR 5183-------------------------- 5184 5185:Capability: KVM_CAP_XEN_HVM / KVM_XEN_HVM_CONFIG_SHARED_INFO 5186:Architectures: x86 5187:Type: vm ioctl 5188:Parameters: struct kvm_xen_hvm_attr 5189:Returns: 0 on success, < 0 on error 5190 5191Allows Xen VM attributes to be read. For the structure and types, 5192see KVM_XEN_HVM_SET_ATTR above. 5193 51944.128 KVM_XEN_VCPU_SET_ATTR 5195--------------------------- 5196 5197:Capability: KVM_CAP_XEN_HVM / KVM_XEN_HVM_CONFIG_SHARED_INFO 5198:Architectures: x86 5199:Type: vcpu ioctl 5200:Parameters: struct kvm_xen_vcpu_attr 5201:Returns: 0 on success, < 0 on error 5202 5203:: 5204 5205 struct kvm_xen_vcpu_attr { 5206 __u16 type; 5207 __u16 pad[3]; 5208 union { 5209 __u64 gpa; 5210 __u64 pad[4]; 5211 struct { 5212 __u64 state; 5213 __u64 state_entry_time; 5214 __u64 time_running; 5215 __u64 time_runnable; 5216 __u64 time_blocked; 5217 __u64 time_offline; 5218 } runstate; 5219 } u; 5220 }; 5221 5222type values: 5223 5224KVM_XEN_VCPU_ATTR_TYPE_VCPU_INFO 5225 Sets the guest physical address of the vcpu_info for a given vCPU. 5226 5227KVM_XEN_VCPU_ATTR_TYPE_VCPU_TIME_INFO 5228 Sets the guest physical address of an additional pvclock structure 5229 for a given vCPU. This is typically used for guest vsyscall support. 5230 5231KVM_XEN_VCPU_ATTR_TYPE_RUNSTATE_ADDR 5232 Sets the guest physical address of the vcpu_runstate_info for a given 5233 vCPU. This is how a Xen guest tracks CPU state such as steal time. 5234 5235KVM_XEN_VCPU_ATTR_TYPE_RUNSTATE_CURRENT 5236 Sets the runstate (RUNSTATE_running/_runnable/_blocked/_offline) of 5237 the given vCPU from the .u.runstate.state member of the structure. 5238 KVM automatically accounts running and runnable time but blocked 5239 and offline states are only entered explicitly. 5240 5241KVM_XEN_VCPU_ATTR_TYPE_RUNSTATE_DATA 5242 Sets all fields of the vCPU runstate data from the .u.runstate member 5243 of the structure, including the current runstate. The state_entry_time 5244 must equal the sum of the other four times. 5245 5246KVM_XEN_VCPU_ATTR_TYPE_RUNSTATE_ADJUST 5247 This *adds* the contents of the .u.runstate members of the structure 5248 to the corresponding members of the given vCPU's runstate data, thus 5249 permitting atomic adjustments to the runstate times. The adjustment 5250 to the state_entry_time must equal the sum of the adjustments to the 5251 other four times. The state field must be set to -1, or to a valid 5252 runstate value (RUNSTATE_running, RUNSTATE_runnable, RUNSTATE_blocked 5253 or RUNSTATE_offline) to set the current accounted state as of the 5254 adjusted state_entry_time. 5255 52564.129 KVM_XEN_VCPU_GET_ATTR 5257--------------------------- 5258 5259:Capability: KVM_CAP_XEN_HVM / KVM_XEN_HVM_CONFIG_SHARED_INFO 5260:Architectures: x86 5261:Type: vcpu ioctl 5262:Parameters: struct kvm_xen_vcpu_attr 5263:Returns: 0 on success, < 0 on error 5264 5265Allows Xen vCPU attributes to be read. For the structure and types, 5266see KVM_XEN_VCPU_SET_ATTR above. 5267 5268The KVM_XEN_VCPU_ATTR_TYPE_RUNSTATE_ADJUST type may not be used 5269with the KVM_XEN_VCPU_GET_ATTR ioctl. 5270 52714.130 KVM_ARM_MTE_COPY_TAGS 5272--------------------------- 5273 5274:Capability: KVM_CAP_ARM_MTE 5275:Architectures: arm64 5276:Type: vm ioctl 5277:Parameters: struct kvm_arm_copy_mte_tags 5278:Returns: number of bytes copied, < 0 on error (-EINVAL for incorrect 5279 arguments, -EFAULT if memory cannot be accessed). 5280 5281:: 5282 5283 struct kvm_arm_copy_mte_tags { 5284 __u64 guest_ipa; 5285 __u64 length; 5286 void __user *addr; 5287 __u64 flags; 5288 __u64 reserved[2]; 5289 }; 5290 5291Copies Memory Tagging Extension (MTE) tags to/from guest tag memory. The 5292``guest_ipa`` and ``length`` fields must be ``PAGE_SIZE`` aligned. The ``addr`` 5293field must point to a buffer which the tags will be copied to or from. 5294 5295``flags`` specifies the direction of copy, either ``KVM_ARM_TAGS_TO_GUEST`` or 5296``KVM_ARM_TAGS_FROM_GUEST``. 5297 5298The size of the buffer to store the tags is ``(length / 16)`` bytes 5299(granules in MTE are 16 bytes long). Each byte contains a single tag 5300value. This matches the format of ``PTRACE_PEEKMTETAGS`` and 5301``PTRACE_POKEMTETAGS``. 5302 5303If an error occurs before any data is copied then a negative error code is 5304returned. If some tags have been copied before an error occurs then the number 5305of bytes successfully copied is returned. If the call completes successfully 5306then ``length`` is returned. 5307 53084.131 KVM_GET_SREGS2 5309-------------------- 5310 5311:Capability: KVM_CAP_SREGS2 5312:Architectures: x86 5313:Type: vcpu ioctl 5314:Parameters: struct kvm_sregs2 (out) 5315:Returns: 0 on success, -1 on error 5316 5317Reads special registers from the vcpu. 5318This ioctl (when supported) replaces the KVM_GET_SREGS. 5319 5320:: 5321 5322 struct kvm_sregs2 { 5323 /* out (KVM_GET_SREGS2) / in (KVM_SET_SREGS2) */ 5324 struct kvm_segment cs, ds, es, fs, gs, ss; 5325 struct kvm_segment tr, ldt; 5326 struct kvm_dtable gdt, idt; 5327 __u64 cr0, cr2, cr3, cr4, cr8; 5328 __u64 efer; 5329 __u64 apic_base; 5330 __u64 flags; 5331 __u64 pdptrs[4]; 5332 }; 5333 5334flags values for ``kvm_sregs2``: 5335 5336``KVM_SREGS2_FLAGS_PDPTRS_VALID`` 5337 5338 Indicates thats the struct contain valid PDPTR values. 5339 5340 53414.132 KVM_SET_SREGS2 5342-------------------- 5343 5344:Capability: KVM_CAP_SREGS2 5345:Architectures: x86 5346:Type: vcpu ioctl 5347:Parameters: struct kvm_sregs2 (in) 5348:Returns: 0 on success, -1 on error 5349 5350Writes special registers into the vcpu. 5351See KVM_GET_SREGS2 for the data structures. 5352This ioctl (when supported) replaces the KVM_SET_SREGS. 5353 53544.133 KVM_GET_STATS_FD 5355---------------------- 5356 5357:Capability: KVM_CAP_STATS_BINARY_FD 5358:Architectures: all 5359:Type: vm ioctl, vcpu ioctl 5360:Parameters: none 5361:Returns: statistics file descriptor on success, < 0 on error 5362 5363Errors: 5364 5365 ====== ====================================================== 5366 ENOMEM if the fd could not be created due to lack of memory 5367 EMFILE if the number of opened files exceeds the limit 5368 ====== ====================================================== 5369 5370The returned file descriptor can be used to read VM/vCPU statistics data in 5371binary format. The data in the file descriptor consists of four blocks 5372organized as follows: 5373 5374+-------------+ 5375| Header | 5376+-------------+ 5377| id string | 5378+-------------+ 5379| Descriptors | 5380+-------------+ 5381| Stats Data | 5382+-------------+ 5383 5384Apart from the header starting at offset 0, please be aware that it is 5385not guaranteed that the four blocks are adjacent or in the above order; 5386the offsets of the id, descriptors and data blocks are found in the 5387header. However, all four blocks are aligned to 64 bit offsets in the 5388file and they do not overlap. 5389 5390All blocks except the data block are immutable. Userspace can read them 5391only one time after retrieving the file descriptor, and then use ``pread`` or 5392``lseek`` to read the statistics repeatedly. 5393 5394All data is in system endianness. 5395 5396The format of the header is as follows:: 5397 5398 struct kvm_stats_header { 5399 __u32 flags; 5400 __u32 name_size; 5401 __u32 num_desc; 5402 __u32 id_offset; 5403 __u32 desc_offset; 5404 __u32 data_offset; 5405 }; 5406 5407The ``flags`` field is not used at the moment. It is always read as 0. 5408 5409The ``name_size`` field is the size (in byte) of the statistics name string 5410(including trailing '\0') which is contained in the "id string" block and 5411appended at the end of every descriptor. 5412 5413The ``num_desc`` field is the number of descriptors that are included in the 5414descriptor block. (The actual number of values in the data block may be 5415larger, since each descriptor may comprise more than one value). 5416 5417The ``id_offset`` field is the offset of the id string from the start of the 5418file indicated by the file descriptor. It is a multiple of 8. 5419 5420The ``desc_offset`` field is the offset of the Descriptors block from the start 5421of the file indicated by the file descriptor. It is a multiple of 8. 5422 5423The ``data_offset`` field is the offset of the Stats Data block from the start 5424of the file indicated by the file descriptor. It is a multiple of 8. 5425 5426The id string block contains a string which identifies the file descriptor on 5427which KVM_GET_STATS_FD was invoked. The size of the block, including the 5428trailing ``'\0'``, is indicated by the ``name_size`` field in the header. 5429 5430The descriptors block is only needed to be read once for the lifetime of the 5431file descriptor contains a sequence of ``struct kvm_stats_desc``, each followed 5432by a string of size ``name_size``. 5433:: 5434 5435 #define KVM_STATS_TYPE_SHIFT 0 5436 #define KVM_STATS_TYPE_MASK (0xF << KVM_STATS_TYPE_SHIFT) 5437 #define KVM_STATS_TYPE_CUMULATIVE (0x0 << KVM_STATS_TYPE_SHIFT) 5438 #define KVM_STATS_TYPE_INSTANT (0x1 << KVM_STATS_TYPE_SHIFT) 5439 #define KVM_STATS_TYPE_PEAK (0x2 << KVM_STATS_TYPE_SHIFT) 5440 #define KVM_STATS_TYPE_LINEAR_HIST (0x3 << KVM_STATS_TYPE_SHIFT) 5441 #define KVM_STATS_TYPE_LOG_HIST (0x4 << KVM_STATS_TYPE_SHIFT) 5442 #define KVM_STATS_TYPE_MAX KVM_STATS_TYPE_LOG_HIST 5443 5444 #define KVM_STATS_UNIT_SHIFT 4 5445 #define KVM_STATS_UNIT_MASK (0xF << KVM_STATS_UNIT_SHIFT) 5446 #define KVM_STATS_UNIT_NONE (0x0 << KVM_STATS_UNIT_SHIFT) 5447 #define KVM_STATS_UNIT_BYTES (0x1 << KVM_STATS_UNIT_SHIFT) 5448 #define KVM_STATS_UNIT_SECONDS (0x2 << KVM_STATS_UNIT_SHIFT) 5449 #define KVM_STATS_UNIT_CYCLES (0x3 << KVM_STATS_UNIT_SHIFT) 5450 #define KVM_STATS_UNIT_MAX KVM_STATS_UNIT_CYCLES 5451 5452 #define KVM_STATS_BASE_SHIFT 8 5453 #define KVM_STATS_BASE_MASK (0xF << KVM_STATS_BASE_SHIFT) 5454 #define KVM_STATS_BASE_POW10 (0x0 << KVM_STATS_BASE_SHIFT) 5455 #define KVM_STATS_BASE_POW2 (0x1 << KVM_STATS_BASE_SHIFT) 5456 #define KVM_STATS_BASE_MAX KVM_STATS_BASE_POW2 5457 5458 struct kvm_stats_desc { 5459 __u32 flags; 5460 __s16 exponent; 5461 __u16 size; 5462 __u32 offset; 5463 __u32 bucket_size; 5464 char name[]; 5465 }; 5466 5467The ``flags`` field contains the type and unit of the statistics data described 5468by this descriptor. Its endianness is CPU native. 5469The following flags are supported: 5470 5471Bits 0-3 of ``flags`` encode the type: 5472 5473 * ``KVM_STATS_TYPE_CUMULATIVE`` 5474 The statistics reports a cumulative count. The value of data can only be increased. 5475 Most of the counters used in KVM are of this type. 5476 The corresponding ``size`` field for this type is always 1. 5477 All cumulative statistics data are read/write. 5478 * ``KVM_STATS_TYPE_INSTANT`` 5479 The statistics reports an instantaneous value. Its value can be increased or 5480 decreased. This type is usually used as a measurement of some resources, 5481 like the number of dirty pages, the number of large pages, etc. 5482 All instant statistics are read only. 5483 The corresponding ``size`` field for this type is always 1. 5484 * ``KVM_STATS_TYPE_PEAK`` 5485 The statistics data reports a peak value, for example the maximum number 5486 of items in a hash table bucket, the longest time waited and so on. 5487 The value of data can only be increased. 5488 The corresponding ``size`` field for this type is always 1. 5489 * ``KVM_STATS_TYPE_LINEAR_HIST`` 5490 The statistic is reported as a linear histogram. The number of 5491 buckets is specified by the ``size`` field. The size of buckets is specified 5492 by the ``hist_param`` field. The range of the Nth bucket (1 <= N < ``size``) 5493 is [``hist_param``*(N-1), ``hist_param``*N), while the range of the last 5494 bucket is [``hist_param``*(``size``-1), +INF). (+INF means positive infinity 5495 value.) The bucket value indicates how many samples fell in the bucket's range. 5496 * ``KVM_STATS_TYPE_LOG_HIST`` 5497 The statistic is reported as a logarithmic histogram. The number of 5498 buckets is specified by the ``size`` field. The range of the first bucket is 5499 [0, 1), while the range of the last bucket is [pow(2, ``size``-2), +INF). 5500 Otherwise, The Nth bucket (1 < N < ``size``) covers 5501 [pow(2, N-2), pow(2, N-1)). The bucket value indicates how many samples fell 5502 in the bucket's range. 5503 5504Bits 4-7 of ``flags`` encode the unit: 5505 5506 * ``KVM_STATS_UNIT_NONE`` 5507 There is no unit for the value of statistics data. This usually means that 5508 the value is a simple counter of an event. 5509 * ``KVM_STATS_UNIT_BYTES`` 5510 It indicates that the statistics data is used to measure memory size, in the 5511 unit of Byte, KiByte, MiByte, GiByte, etc. The unit of the data is 5512 determined by the ``exponent`` field in the descriptor. 5513 * ``KVM_STATS_UNIT_SECONDS`` 5514 It indicates that the statistics data is used to measure time or latency. 5515 * ``KVM_STATS_UNIT_CYCLES`` 5516 It indicates that the statistics data is used to measure CPU clock cycles. 5517 5518Bits 8-11 of ``flags``, together with ``exponent``, encode the scale of the 5519unit: 5520 5521 * ``KVM_STATS_BASE_POW10`` 5522 The scale is based on power of 10. It is used for measurement of time and 5523 CPU clock cycles. For example, an exponent of -9 can be used with 5524 ``KVM_STATS_UNIT_SECONDS`` to express that the unit is nanoseconds. 5525 * ``KVM_STATS_BASE_POW2`` 5526 The scale is based on power of 2. It is used for measurement of memory size. 5527 For example, an exponent of 20 can be used with ``KVM_STATS_UNIT_BYTES`` to 5528 express that the unit is MiB. 5529 5530The ``size`` field is the number of values of this statistics data. Its 5531value is usually 1 for most of simple statistics. 1 means it contains an 5532unsigned 64bit data. 5533 5534The ``offset`` field is the offset from the start of Data Block to the start of 5535the corresponding statistics data. 5536 5537The ``bucket_size`` field is used as a parameter for histogram statistics data. 5538It is only used by linear histogram statistics data, specifying the size of a 5539bucket. 5540 5541The ``name`` field is the name string of the statistics data. The name string 5542starts at the end of ``struct kvm_stats_desc``. The maximum length including 5543the trailing ``'\0'``, is indicated by ``name_size`` in the header. 5544 5545The Stats Data block contains an array of 64-bit values in the same order 5546as the descriptors in Descriptors block. 5547 55484.134 KVM_GET_XSAVE2 5549-------------------- 5550 5551:Capability: KVM_CAP_XSAVE2 5552:Architectures: x86 5553:Type: vcpu ioctl 5554:Parameters: struct kvm_xsave (out) 5555:Returns: 0 on success, -1 on error 5556 5557 5558:: 5559 5560 struct kvm_xsave { 5561 __u32 region[1024]; 5562 __u32 extra[0]; 5563 }; 5564 5565This ioctl would copy current vcpu's xsave struct to the userspace. It 5566copies as many bytes as are returned by KVM_CHECK_EXTENSION(KVM_CAP_XSAVE2) 5567when invoked on the vm file descriptor. The size value returned by 5568KVM_CHECK_EXTENSION(KVM_CAP_XSAVE2) will always be at least 4096. 5569Currently, it is only greater than 4096 if a dynamic feature has been 5570enabled with ``arch_prctl()``, but this may change in the future. 5571 5572The offsets of the state save areas in struct kvm_xsave follow the contents 5573of CPUID leaf 0xD on the host. 5574 5575 55765. The kvm_run structure 5577======================== 5578 5579Application code obtains a pointer to the kvm_run structure by 5580mmap()ing a vcpu fd. From that point, application code can control 5581execution by changing fields in kvm_run prior to calling the KVM_RUN 5582ioctl, and obtain information about the reason KVM_RUN returned by 5583looking up structure members. 5584 5585:: 5586 5587 struct kvm_run { 5588 /* in */ 5589 __u8 request_interrupt_window; 5590 5591Request that KVM_RUN return when it becomes possible to inject external 5592interrupts into the guest. Useful in conjunction with KVM_INTERRUPT. 5593 5594:: 5595 5596 __u8 immediate_exit; 5597 5598This field is polled once when KVM_RUN starts; if non-zero, KVM_RUN 5599exits immediately, returning -EINTR. In the common scenario where a 5600signal is used to "kick" a VCPU out of KVM_RUN, this field can be used 5601to avoid usage of KVM_SET_SIGNAL_MASK, which has worse scalability. 5602Rather than blocking the signal outside KVM_RUN, userspace can set up 5603a signal handler that sets run->immediate_exit to a non-zero value. 5604 5605This field is ignored if KVM_CAP_IMMEDIATE_EXIT is not available. 5606 5607:: 5608 5609 __u8 padding1[6]; 5610 5611 /* out */ 5612 __u32 exit_reason; 5613 5614When KVM_RUN has returned successfully (return value 0), this informs 5615application code why KVM_RUN has returned. Allowable values for this 5616field are detailed below. 5617 5618:: 5619 5620 __u8 ready_for_interrupt_injection; 5621 5622If request_interrupt_window has been specified, this field indicates 5623an interrupt can be injected now with KVM_INTERRUPT. 5624 5625:: 5626 5627 __u8 if_flag; 5628 5629The value of the current interrupt flag. Only valid if in-kernel 5630local APIC is not used. 5631 5632:: 5633 5634 __u16 flags; 5635 5636More architecture-specific flags detailing state of the VCPU that may 5637affect the device's behavior. Current defined flags:: 5638 5639 /* x86, set if the VCPU is in system management mode */ 5640 #define KVM_RUN_X86_SMM (1 << 0) 5641 /* x86, set if bus lock detected in VM */ 5642 #define KVM_RUN_BUS_LOCK (1 << 1) 5643 5644:: 5645 5646 /* in (pre_kvm_run), out (post_kvm_run) */ 5647 __u64 cr8; 5648 5649The value of the cr8 register. Only valid if in-kernel local APIC is 5650not used. Both input and output. 5651 5652:: 5653 5654 __u64 apic_base; 5655 5656The value of the APIC BASE msr. Only valid if in-kernel local 5657APIC is not used. Both input and output. 5658 5659:: 5660 5661 union { 5662 /* KVM_EXIT_UNKNOWN */ 5663 struct { 5664 __u64 hardware_exit_reason; 5665 } hw; 5666 5667If exit_reason is KVM_EXIT_UNKNOWN, the vcpu has exited due to unknown 5668reasons. Further architecture-specific information is available in 5669hardware_exit_reason. 5670 5671:: 5672 5673 /* KVM_EXIT_FAIL_ENTRY */ 5674 struct { 5675 __u64 hardware_entry_failure_reason; 5676 __u32 cpu; /* if KVM_LAST_CPU */ 5677 } fail_entry; 5678 5679If exit_reason is KVM_EXIT_FAIL_ENTRY, the vcpu could not be run due 5680to unknown reasons. Further architecture-specific information is 5681available in hardware_entry_failure_reason. 5682 5683:: 5684 5685 /* KVM_EXIT_EXCEPTION */ 5686 struct { 5687 __u32 exception; 5688 __u32 error_code; 5689 } ex; 5690 5691Unused. 5692 5693:: 5694 5695 /* KVM_EXIT_IO */ 5696 struct { 5697 #define KVM_EXIT_IO_IN 0 5698 #define KVM_EXIT_IO_OUT 1 5699 __u8 direction; 5700 __u8 size; /* bytes */ 5701 __u16 port; 5702 __u32 count; 5703 __u64 data_offset; /* relative to kvm_run start */ 5704 } io; 5705 5706If exit_reason is KVM_EXIT_IO, then the vcpu has 5707executed a port I/O instruction which could not be satisfied by kvm. 5708data_offset describes where the data is located (KVM_EXIT_IO_OUT) or 5709where kvm expects application code to place the data for the next 5710KVM_RUN invocation (KVM_EXIT_IO_IN). Data format is a packed array. 5711 5712:: 5713 5714 /* KVM_EXIT_DEBUG */ 5715 struct { 5716 struct kvm_debug_exit_arch arch; 5717 } debug; 5718 5719If the exit_reason is KVM_EXIT_DEBUG, then a vcpu is processing a debug event 5720for which architecture specific information is returned. 5721 5722:: 5723 5724 /* KVM_EXIT_MMIO */ 5725 struct { 5726 __u64 phys_addr; 5727 __u8 data[8]; 5728 __u32 len; 5729 __u8 is_write; 5730 } mmio; 5731 5732If exit_reason is KVM_EXIT_MMIO, then the vcpu has 5733executed a memory-mapped I/O instruction which could not be satisfied 5734by kvm. The 'data' member contains the written data if 'is_write' is 5735true, and should be filled by application code otherwise. 5736 5737The 'data' member contains, in its first 'len' bytes, the value as it would 5738appear if the VCPU performed a load or store of the appropriate width directly 5739to the byte array. 5740 5741.. note:: 5742 5743 For KVM_EXIT_IO, KVM_EXIT_MMIO, KVM_EXIT_OSI, KVM_EXIT_PAPR, KVM_EXIT_XEN, 5744 KVM_EXIT_EPR, KVM_EXIT_X86_RDMSR and KVM_EXIT_X86_WRMSR the corresponding 5745 operations are complete (and guest state is consistent) only after userspace 5746 has re-entered the kernel with KVM_RUN. The kernel side will first finish 5747 incomplete operations and then check for pending signals. 5748 5749 The pending state of the operation is not preserved in state which is 5750 visible to userspace, thus userspace should ensure that the operation is 5751 completed before performing a live migration. Userspace can re-enter the 5752 guest with an unmasked signal pending or with the immediate_exit field set 5753 to complete pending operations without allowing any further instructions 5754 to be executed. 5755 5756:: 5757 5758 /* KVM_EXIT_HYPERCALL */ 5759 struct { 5760 __u64 nr; 5761 __u64 args[6]; 5762 __u64 ret; 5763 __u32 longmode; 5764 __u32 pad; 5765 } hypercall; 5766 5767Unused. This was once used for 'hypercall to userspace'. To implement 5768such functionality, use KVM_EXIT_IO (x86) or KVM_EXIT_MMIO (all except s390). 5769 5770.. note:: KVM_EXIT_IO is significantly faster than KVM_EXIT_MMIO. 5771 5772:: 5773 5774 /* KVM_EXIT_TPR_ACCESS */ 5775 struct { 5776 __u64 rip; 5777 __u32 is_write; 5778 __u32 pad; 5779 } tpr_access; 5780 5781To be documented (KVM_TPR_ACCESS_REPORTING). 5782 5783:: 5784 5785 /* KVM_EXIT_S390_SIEIC */ 5786 struct { 5787 __u8 icptcode; 5788 __u64 mask; /* psw upper half */ 5789 __u64 addr; /* psw lower half */ 5790 __u16 ipa; 5791 __u32 ipb; 5792 } s390_sieic; 5793 5794s390 specific. 5795 5796:: 5797 5798 /* KVM_EXIT_S390_RESET */ 5799 #define KVM_S390_RESET_POR 1 5800 #define KVM_S390_RESET_CLEAR 2 5801 #define KVM_S390_RESET_SUBSYSTEM 4 5802 #define KVM_S390_RESET_CPU_INIT 8 5803 #define KVM_S390_RESET_IPL 16 5804 __u64 s390_reset_flags; 5805 5806s390 specific. 5807 5808:: 5809 5810 /* KVM_EXIT_S390_UCONTROL */ 5811 struct { 5812 __u64 trans_exc_code; 5813 __u32 pgm_code; 5814 } s390_ucontrol; 5815 5816s390 specific. A page fault has occurred for a user controlled virtual 5817machine (KVM_VM_S390_UNCONTROL) on it's host page table that cannot be 5818resolved by the kernel. 5819The program code and the translation exception code that were placed 5820in the cpu's lowcore are presented here as defined by the z Architecture 5821Principles of Operation Book in the Chapter for Dynamic Address Translation 5822(DAT) 5823 5824:: 5825 5826 /* KVM_EXIT_DCR */ 5827 struct { 5828 __u32 dcrn; 5829 __u32 data; 5830 __u8 is_write; 5831 } dcr; 5832 5833Deprecated - was used for 440 KVM. 5834 5835:: 5836 5837 /* KVM_EXIT_OSI */ 5838 struct { 5839 __u64 gprs[32]; 5840 } osi; 5841 5842MOL uses a special hypercall interface it calls 'OSI'. To enable it, we catch 5843hypercalls and exit with this exit struct that contains all the guest gprs. 5844 5845If exit_reason is KVM_EXIT_OSI, then the vcpu has triggered such a hypercall. 5846Userspace can now handle the hypercall and when it's done modify the gprs as 5847necessary. Upon guest entry all guest GPRs will then be replaced by the values 5848in this struct. 5849 5850:: 5851 5852 /* KVM_EXIT_PAPR_HCALL */ 5853 struct { 5854 __u64 nr; 5855 __u64 ret; 5856 __u64 args[9]; 5857 } papr_hcall; 5858 5859This is used on 64-bit PowerPC when emulating a pSeries partition, 5860e.g. with the 'pseries' machine type in qemu. It occurs when the 5861guest does a hypercall using the 'sc 1' instruction. The 'nr' field 5862contains the hypercall number (from the guest R3), and 'args' contains 5863the arguments (from the guest R4 - R12). Userspace should put the 5864return code in 'ret' and any extra returned values in args[]. 5865The possible hypercalls are defined in the Power Architecture Platform 5866Requirements (PAPR) document available from www.power.org (free 5867developer registration required to access it). 5868 5869:: 5870 5871 /* KVM_EXIT_S390_TSCH */ 5872 struct { 5873 __u16 subchannel_id; 5874 __u16 subchannel_nr; 5875 __u32 io_int_parm; 5876 __u32 io_int_word; 5877 __u32 ipb; 5878 __u8 dequeued; 5879 } s390_tsch; 5880 5881s390 specific. This exit occurs when KVM_CAP_S390_CSS_SUPPORT has been enabled 5882and TEST SUBCHANNEL was intercepted. If dequeued is set, a pending I/O 5883interrupt for the target subchannel has been dequeued and subchannel_id, 5884subchannel_nr, io_int_parm and io_int_word contain the parameters for that 5885interrupt. ipb is needed for instruction parameter decoding. 5886 5887:: 5888 5889 /* KVM_EXIT_EPR */ 5890 struct { 5891 __u32 epr; 5892 } epr; 5893 5894On FSL BookE PowerPC chips, the interrupt controller has a fast patch 5895interrupt acknowledge path to the core. When the core successfully 5896delivers an interrupt, it automatically populates the EPR register with 5897the interrupt vector number and acknowledges the interrupt inside 5898the interrupt controller. 5899 5900In case the interrupt controller lives in user space, we need to do 5901the interrupt acknowledge cycle through it to fetch the next to be 5902delivered interrupt vector using this exit. 5903 5904It gets triggered whenever both KVM_CAP_PPC_EPR are enabled and an 5905external interrupt has just been delivered into the guest. User space 5906should put the acknowledged interrupt vector into the 'epr' field. 5907 5908:: 5909 5910 /* KVM_EXIT_SYSTEM_EVENT */ 5911 struct { 5912 #define KVM_SYSTEM_EVENT_SHUTDOWN 1 5913 #define KVM_SYSTEM_EVENT_RESET 2 5914 #define KVM_SYSTEM_EVENT_CRASH 3 5915 __u32 type; 5916 __u64 flags; 5917 } system_event; 5918 5919If exit_reason is KVM_EXIT_SYSTEM_EVENT then the vcpu has triggered 5920a system-level event using some architecture specific mechanism (hypercall 5921or some special instruction). In case of ARM/ARM64, this is triggered using 5922HVC instruction based PSCI call from the vcpu. The 'type' field describes 5923the system-level event type. The 'flags' field describes architecture 5924specific flags for the system-level event. 5925 5926Valid values for 'type' are: 5927 5928 - KVM_SYSTEM_EVENT_SHUTDOWN -- the guest has requested a shutdown of the 5929 VM. Userspace is not obliged to honour this, and if it does honour 5930 this does not need to destroy the VM synchronously (ie it may call 5931 KVM_RUN again before shutdown finally occurs). 5932 - KVM_SYSTEM_EVENT_RESET -- the guest has requested a reset of the VM. 5933 As with SHUTDOWN, userspace can choose to ignore the request, or 5934 to schedule the reset to occur in the future and may call KVM_RUN again. 5935 - KVM_SYSTEM_EVENT_CRASH -- the guest crash occurred and the guest 5936 has requested a crash condition maintenance. Userspace can choose 5937 to ignore the request, or to gather VM memory core dump and/or 5938 reset/shutdown of the VM. 5939 5940:: 5941 5942 /* KVM_EXIT_IOAPIC_EOI */ 5943 struct { 5944 __u8 vector; 5945 } eoi; 5946 5947Indicates that the VCPU's in-kernel local APIC received an EOI for a 5948level-triggered IOAPIC interrupt. This exit only triggers when the 5949IOAPIC is implemented in userspace (i.e. KVM_CAP_SPLIT_IRQCHIP is enabled); 5950the userspace IOAPIC should process the EOI and retrigger the interrupt if 5951it is still asserted. Vector is the LAPIC interrupt vector for which the 5952EOI was received. 5953 5954:: 5955 5956 struct kvm_hyperv_exit { 5957 #define KVM_EXIT_HYPERV_SYNIC 1 5958 #define KVM_EXIT_HYPERV_HCALL 2 5959 #define KVM_EXIT_HYPERV_SYNDBG 3 5960 __u32 type; 5961 __u32 pad1; 5962 union { 5963 struct { 5964 __u32 msr; 5965 __u32 pad2; 5966 __u64 control; 5967 __u64 evt_page; 5968 __u64 msg_page; 5969 } synic; 5970 struct { 5971 __u64 input; 5972 __u64 result; 5973 __u64 params[2]; 5974 } hcall; 5975 struct { 5976 __u32 msr; 5977 __u32 pad2; 5978 __u64 control; 5979 __u64 status; 5980 __u64 send_page; 5981 __u64 recv_page; 5982 __u64 pending_page; 5983 } syndbg; 5984 } u; 5985 }; 5986 /* KVM_EXIT_HYPERV */ 5987 struct kvm_hyperv_exit hyperv; 5988 5989Indicates that the VCPU exits into userspace to process some tasks 5990related to Hyper-V emulation. 5991 5992Valid values for 'type' are: 5993 5994 - KVM_EXIT_HYPERV_SYNIC -- synchronously notify user-space about 5995 5996Hyper-V SynIC state change. Notification is used to remap SynIC 5997event/message pages and to enable/disable SynIC messages/events processing 5998in userspace. 5999 6000 - KVM_EXIT_HYPERV_SYNDBG -- synchronously notify user-space about 6001 6002Hyper-V Synthetic debugger state change. Notification is used to either update 6003the pending_page location or to send a control command (send the buffer located 6004in send_page or recv a buffer to recv_page). 6005 6006:: 6007 6008 /* KVM_EXIT_ARM_NISV */ 6009 struct { 6010 __u64 esr_iss; 6011 __u64 fault_ipa; 6012 } arm_nisv; 6013 6014Used on arm and arm64 systems. If a guest accesses memory not in a memslot, 6015KVM will typically return to userspace and ask it to do MMIO emulation on its 6016behalf. However, for certain classes of instructions, no instruction decode 6017(direction, length of memory access) is provided, and fetching and decoding 6018the instruction from the VM is overly complicated to live in the kernel. 6019 6020Historically, when this situation occurred, KVM would print a warning and kill 6021the VM. KVM assumed that if the guest accessed non-memslot memory, it was 6022trying to do I/O, which just couldn't be emulated, and the warning message was 6023phrased accordingly. However, what happened more often was that a guest bug 6024caused access outside the guest memory areas which should lead to a more 6025meaningful warning message and an external abort in the guest, if the access 6026did not fall within an I/O window. 6027 6028Userspace implementations can query for KVM_CAP_ARM_NISV_TO_USER, and enable 6029this capability at VM creation. Once this is done, these types of errors will 6030instead return to userspace with KVM_EXIT_ARM_NISV, with the valid bits from 6031the HSR (arm) and ESR_EL2 (arm64) in the esr_iss field, and the faulting IPA 6032in the fault_ipa field. Userspace can either fix up the access if it's 6033actually an I/O access by decoding the instruction from guest memory (if it's 6034very brave) and continue executing the guest, or it can decide to suspend, 6035dump, or restart the guest. 6036 6037Note that KVM does not skip the faulting instruction as it does for 6038KVM_EXIT_MMIO, but userspace has to emulate any change to the processing state 6039if it decides to decode and emulate the instruction. 6040 6041:: 6042 6043 /* KVM_EXIT_X86_RDMSR / KVM_EXIT_X86_WRMSR */ 6044 struct { 6045 __u8 error; /* user -> kernel */ 6046 __u8 pad[7]; 6047 __u32 reason; /* kernel -> user */ 6048 __u32 index; /* kernel -> user */ 6049 __u64 data; /* kernel <-> user */ 6050 } msr; 6051 6052Used on x86 systems. When the VM capability KVM_CAP_X86_USER_SPACE_MSR is 6053enabled, MSR accesses to registers that would invoke a #GP by KVM kernel code 6054will instead trigger a KVM_EXIT_X86_RDMSR exit for reads and KVM_EXIT_X86_WRMSR 6055exit for writes. 6056 6057The "reason" field specifies why the MSR trap occurred. User space will only 6058receive MSR exit traps when a particular reason was requested during through 6059ENABLE_CAP. Currently valid exit reasons are: 6060 6061 KVM_MSR_EXIT_REASON_UNKNOWN - access to MSR that is unknown to KVM 6062 KVM_MSR_EXIT_REASON_INVAL - access to invalid MSRs or reserved bits 6063 KVM_MSR_EXIT_REASON_FILTER - access blocked by KVM_X86_SET_MSR_FILTER 6064 6065For KVM_EXIT_X86_RDMSR, the "index" field tells user space which MSR the guest 6066wants to read. To respond to this request with a successful read, user space 6067writes the respective data into the "data" field and must continue guest 6068execution to ensure the read data is transferred into guest register state. 6069 6070If the RDMSR request was unsuccessful, user space indicates that with a "1" in 6071the "error" field. This will inject a #GP into the guest when the VCPU is 6072executed again. 6073 6074For KVM_EXIT_X86_WRMSR, the "index" field tells user space which MSR the guest 6075wants to write. Once finished processing the event, user space must continue 6076vCPU execution. If the MSR write was unsuccessful, user space also sets the 6077"error" field to "1". 6078 6079:: 6080 6081 6082 struct kvm_xen_exit { 6083 #define KVM_EXIT_XEN_HCALL 1 6084 __u32 type; 6085 union { 6086 struct { 6087 __u32 longmode; 6088 __u32 cpl; 6089 __u64 input; 6090 __u64 result; 6091 __u64 params[6]; 6092 } hcall; 6093 } u; 6094 }; 6095 /* KVM_EXIT_XEN */ 6096 struct kvm_hyperv_exit xen; 6097 6098Indicates that the VCPU exits into userspace to process some tasks 6099related to Xen emulation. 6100 6101Valid values for 'type' are: 6102 6103 - KVM_EXIT_XEN_HCALL -- synchronously notify user-space about Xen hypercall. 6104 Userspace is expected to place the hypercall result into the appropriate 6105 field before invoking KVM_RUN again. 6106 6107:: 6108 6109 /* KVM_EXIT_RISCV_SBI */ 6110 struct { 6111 unsigned long extension_id; 6112 unsigned long function_id; 6113 unsigned long args[6]; 6114 unsigned long ret[2]; 6115 } riscv_sbi; 6116If exit reason is KVM_EXIT_RISCV_SBI then it indicates that the VCPU has 6117done a SBI call which is not handled by KVM RISC-V kernel module. The details 6118of the SBI call are available in 'riscv_sbi' member of kvm_run structure. The 6119'extension_id' field of 'riscv_sbi' represents SBI extension ID whereas the 6120'function_id' field represents function ID of given SBI extension. The 'args' 6121array field of 'riscv_sbi' represents parameters for the SBI call and 'ret' 6122array field represents return values. The userspace should update the return 6123values of SBI call before resuming the VCPU. For more details on RISC-V SBI 6124spec refer, https://github.com/riscv/riscv-sbi-doc. 6125 6126:: 6127 6128 /* Fix the size of the union. */ 6129 char padding[256]; 6130 }; 6131 6132 /* 6133 * shared registers between kvm and userspace. 6134 * kvm_valid_regs specifies the register classes set by the host 6135 * kvm_dirty_regs specified the register classes dirtied by userspace 6136 * struct kvm_sync_regs is architecture specific, as well as the 6137 * bits for kvm_valid_regs and kvm_dirty_regs 6138 */ 6139 __u64 kvm_valid_regs; 6140 __u64 kvm_dirty_regs; 6141 union { 6142 struct kvm_sync_regs regs; 6143 char padding[SYNC_REGS_SIZE_BYTES]; 6144 } s; 6145 6146If KVM_CAP_SYNC_REGS is defined, these fields allow userspace to access 6147certain guest registers without having to call SET/GET_*REGS. Thus we can 6148avoid some system call overhead if userspace has to handle the exit. 6149Userspace can query the validity of the structure by checking 6150kvm_valid_regs for specific bits. These bits are architecture specific 6151and usually define the validity of a groups of registers. (e.g. one bit 6152for general purpose registers) 6153 6154Please note that the kernel is allowed to use the kvm_run structure as the 6155primary storage for certain register types. Therefore, the kernel may use the 6156values in kvm_run even if the corresponding bit in kvm_dirty_regs is not set. 6157 6158:: 6159 6160 }; 6161 6162 6163 61646. Capabilities that can be enabled on vCPUs 6165============================================ 6166 6167There are certain capabilities that change the behavior of the virtual CPU or 6168the virtual machine when enabled. To enable them, please see section 4.37. 6169Below you can find a list of capabilities and what their effect on the vCPU or 6170the virtual machine is when enabling them. 6171 6172The following information is provided along with the description: 6173 6174 Architectures: 6175 which instruction set architectures provide this ioctl. 6176 x86 includes both i386 and x86_64. 6177 6178 Target: 6179 whether this is a per-vcpu or per-vm capability. 6180 6181 Parameters: 6182 what parameters are accepted by the capability. 6183 6184 Returns: 6185 the return value. General error numbers (EBADF, ENOMEM, EINVAL) 6186 are not detailed, but errors with specific meanings are. 6187 6188 61896.1 KVM_CAP_PPC_OSI 6190------------------- 6191 6192:Architectures: ppc 6193:Target: vcpu 6194:Parameters: none 6195:Returns: 0 on success; -1 on error 6196 6197This capability enables interception of OSI hypercalls that otherwise would 6198be treated as normal system calls to be injected into the guest. OSI hypercalls 6199were invented by Mac-on-Linux to have a standardized communication mechanism 6200between the guest and the host. 6201 6202When this capability is enabled, KVM_EXIT_OSI can occur. 6203 6204 62056.2 KVM_CAP_PPC_PAPR 6206-------------------- 6207 6208:Architectures: ppc 6209:Target: vcpu 6210:Parameters: none 6211:Returns: 0 on success; -1 on error 6212 6213This capability enables interception of PAPR hypercalls. PAPR hypercalls are 6214done using the hypercall instruction "sc 1". 6215 6216It also sets the guest privilege level to "supervisor" mode. Usually the guest 6217runs in "hypervisor" privilege mode with a few missing features. 6218 6219In addition to the above, it changes the semantics of SDR1. In this mode, the 6220HTAB address part of SDR1 contains an HVA instead of a GPA, as PAPR keeps the 6221HTAB invisible to the guest. 6222 6223When this capability is enabled, KVM_EXIT_PAPR_HCALL can occur. 6224 6225 62266.3 KVM_CAP_SW_TLB 6227------------------ 6228 6229:Architectures: ppc 6230:Target: vcpu 6231:Parameters: args[0] is the address of a struct kvm_config_tlb 6232:Returns: 0 on success; -1 on error 6233 6234:: 6235 6236 struct kvm_config_tlb { 6237 __u64 params; 6238 __u64 array; 6239 __u32 mmu_type; 6240 __u32 array_len; 6241 }; 6242 6243Configures the virtual CPU's TLB array, establishing a shared memory area 6244between userspace and KVM. The "params" and "array" fields are userspace 6245addresses of mmu-type-specific data structures. The "array_len" field is an 6246safety mechanism, and should be set to the size in bytes of the memory that 6247userspace has reserved for the array. It must be at least the size dictated 6248by "mmu_type" and "params". 6249 6250While KVM_RUN is active, the shared region is under control of KVM. Its 6251contents are undefined, and any modification by userspace results in 6252boundedly undefined behavior. 6253 6254On return from KVM_RUN, the shared region will reflect the current state of 6255the guest's TLB. If userspace makes any changes, it must call KVM_DIRTY_TLB 6256to tell KVM which entries have been changed, prior to calling KVM_RUN again 6257on this vcpu. 6258 6259For mmu types KVM_MMU_FSL_BOOKE_NOHV and KVM_MMU_FSL_BOOKE_HV: 6260 6261 - The "params" field is of type "struct kvm_book3e_206_tlb_params". 6262 - The "array" field points to an array of type "struct 6263 kvm_book3e_206_tlb_entry". 6264 - The array consists of all entries in the first TLB, followed by all 6265 entries in the second TLB. 6266 - Within a TLB, entries are ordered first by increasing set number. Within a 6267 set, entries are ordered by way (increasing ESEL). 6268 - The hash for determining set number in TLB0 is: (MAS2 >> 12) & (num_sets - 1) 6269 where "num_sets" is the tlb_sizes[] value divided by the tlb_ways[] value. 6270 - The tsize field of mas1 shall be set to 4K on TLB0, even though the 6271 hardware ignores this value for TLB0. 6272 62736.4 KVM_CAP_S390_CSS_SUPPORT 6274---------------------------- 6275 6276:Architectures: s390 6277:Target: vcpu 6278:Parameters: none 6279:Returns: 0 on success; -1 on error 6280 6281This capability enables support for handling of channel I/O instructions. 6282 6283TEST PENDING INTERRUPTION and the interrupt portion of TEST SUBCHANNEL are 6284handled in-kernel, while the other I/O instructions are passed to userspace. 6285 6286When this capability is enabled, KVM_EXIT_S390_TSCH will occur on TEST 6287SUBCHANNEL intercepts. 6288 6289Note that even though this capability is enabled per-vcpu, the complete 6290virtual machine is affected. 6291 62926.5 KVM_CAP_PPC_EPR 6293------------------- 6294 6295:Architectures: ppc 6296:Target: vcpu 6297:Parameters: args[0] defines whether the proxy facility is active 6298:Returns: 0 on success; -1 on error 6299 6300This capability enables or disables the delivery of interrupts through the 6301external proxy facility. 6302 6303When enabled (args[0] != 0), every time the guest gets an external interrupt 6304delivered, it automatically exits into user space with a KVM_EXIT_EPR exit 6305to receive the topmost interrupt vector. 6306 6307When disabled (args[0] == 0), behavior is as if this facility is unsupported. 6308 6309When this capability is enabled, KVM_EXIT_EPR can occur. 6310 63116.6 KVM_CAP_IRQ_MPIC 6312-------------------- 6313 6314:Architectures: ppc 6315:Parameters: args[0] is the MPIC device fd; 6316 args[1] is the MPIC CPU number for this vcpu 6317 6318This capability connects the vcpu to an in-kernel MPIC device. 6319 63206.7 KVM_CAP_IRQ_XICS 6321-------------------- 6322 6323:Architectures: ppc 6324:Target: vcpu 6325:Parameters: args[0] is the XICS device fd; 6326 args[1] is the XICS CPU number (server ID) for this vcpu 6327 6328This capability connects the vcpu to an in-kernel XICS device. 6329 63306.8 KVM_CAP_S390_IRQCHIP 6331------------------------ 6332 6333:Architectures: s390 6334:Target: vm 6335:Parameters: none 6336 6337This capability enables the in-kernel irqchip for s390. Please refer to 6338"4.24 KVM_CREATE_IRQCHIP" for details. 6339 63406.9 KVM_CAP_MIPS_FPU 6341-------------------- 6342 6343:Architectures: mips 6344:Target: vcpu 6345:Parameters: args[0] is reserved for future use (should be 0). 6346 6347This capability allows the use of the host Floating Point Unit by the guest. It 6348allows the Config1.FP bit to be set to enable the FPU in the guest. Once this is 6349done the ``KVM_REG_MIPS_FPR_*`` and ``KVM_REG_MIPS_FCR_*`` registers can be 6350accessed (depending on the current guest FPU register mode), and the Status.FR, 6351Config5.FRE bits are accessible via the KVM API and also from the guest, 6352depending on them being supported by the FPU. 6353 63546.10 KVM_CAP_MIPS_MSA 6355--------------------- 6356 6357:Architectures: mips 6358:Target: vcpu 6359:Parameters: args[0] is reserved for future use (should be 0). 6360 6361This capability allows the use of the MIPS SIMD Architecture (MSA) by the guest. 6362It allows the Config3.MSAP bit to be set to enable the use of MSA by the guest. 6363Once this is done the ``KVM_REG_MIPS_VEC_*`` and ``KVM_REG_MIPS_MSA_*`` 6364registers can be accessed, and the Config5.MSAEn bit is accessible via the 6365KVM API and also from the guest. 6366 63676.74 KVM_CAP_SYNC_REGS 6368---------------------- 6369 6370:Architectures: s390, x86 6371:Target: s390: always enabled, x86: vcpu 6372:Parameters: none 6373:Returns: x86: KVM_CHECK_EXTENSION returns a bit-array indicating which register 6374 sets are supported 6375 (bitfields defined in arch/x86/include/uapi/asm/kvm.h). 6376 6377As described above in the kvm_sync_regs struct info in section 5 (kvm_run): 6378KVM_CAP_SYNC_REGS "allow[s] userspace to access certain guest registers 6379without having to call SET/GET_*REGS". This reduces overhead by eliminating 6380repeated ioctl calls for setting and/or getting register values. This is 6381particularly important when userspace is making synchronous guest state 6382modifications, e.g. when emulating and/or intercepting instructions in 6383userspace. 6384 6385For s390 specifics, please refer to the source code. 6386 6387For x86: 6388 6389- the register sets to be copied out to kvm_run are selectable 6390 by userspace (rather that all sets being copied out for every exit). 6391- vcpu_events are available in addition to regs and sregs. 6392 6393For x86, the 'kvm_valid_regs' field of struct kvm_run is overloaded to 6394function as an input bit-array field set by userspace to indicate the 6395specific register sets to be copied out on the next exit. 6396 6397To indicate when userspace has modified values that should be copied into 6398the vCPU, the all architecture bitarray field, 'kvm_dirty_regs' must be set. 6399This is done using the same bitflags as for the 'kvm_valid_regs' field. 6400If the dirty bit is not set, then the register set values will not be copied 6401into the vCPU even if they've been modified. 6402 6403Unused bitfields in the bitarrays must be set to zero. 6404 6405:: 6406 6407 struct kvm_sync_regs { 6408 struct kvm_regs regs; 6409 struct kvm_sregs sregs; 6410 struct kvm_vcpu_events events; 6411 }; 6412 64136.75 KVM_CAP_PPC_IRQ_XIVE 6414------------------------- 6415 6416:Architectures: ppc 6417:Target: vcpu 6418:Parameters: args[0] is the XIVE device fd; 6419 args[1] is the XIVE CPU number (server ID) for this vcpu 6420 6421This capability connects the vcpu to an in-kernel XIVE device. 6422 64237. Capabilities that can be enabled on VMs 6424========================================== 6425 6426There are certain capabilities that change the behavior of the virtual 6427machine when enabled. To enable them, please see section 4.37. Below 6428you can find a list of capabilities and what their effect on the VM 6429is when enabling them. 6430 6431The following information is provided along with the description: 6432 6433 Architectures: 6434 which instruction set architectures provide this ioctl. 6435 x86 includes both i386 and x86_64. 6436 6437 Parameters: 6438 what parameters are accepted by the capability. 6439 6440 Returns: 6441 the return value. General error numbers (EBADF, ENOMEM, EINVAL) 6442 are not detailed, but errors with specific meanings are. 6443 6444 64457.1 KVM_CAP_PPC_ENABLE_HCALL 6446---------------------------- 6447 6448:Architectures: ppc 6449:Parameters: args[0] is the sPAPR hcall number; 6450 args[1] is 0 to disable, 1 to enable in-kernel handling 6451 6452This capability controls whether individual sPAPR hypercalls (hcalls) 6453get handled by the kernel or not. Enabling or disabling in-kernel 6454handling of an hcall is effective across the VM. On creation, an 6455initial set of hcalls are enabled for in-kernel handling, which 6456consists of those hcalls for which in-kernel handlers were implemented 6457before this capability was implemented. If disabled, the kernel will 6458not to attempt to handle the hcall, but will always exit to userspace 6459to handle it. Note that it may not make sense to enable some and 6460disable others of a group of related hcalls, but KVM does not prevent 6461userspace from doing that. 6462 6463If the hcall number specified is not one that has an in-kernel 6464implementation, the KVM_ENABLE_CAP ioctl will fail with an EINVAL 6465error. 6466 64677.2 KVM_CAP_S390_USER_SIGP 6468-------------------------- 6469 6470:Architectures: s390 6471:Parameters: none 6472 6473This capability controls which SIGP orders will be handled completely in user 6474space. With this capability enabled, all fast orders will be handled completely 6475in the kernel: 6476 6477- SENSE 6478- SENSE RUNNING 6479- EXTERNAL CALL 6480- EMERGENCY SIGNAL 6481- CONDITIONAL EMERGENCY SIGNAL 6482 6483All other orders will be handled completely in user space. 6484 6485Only privileged operation exceptions will be checked for in the kernel (or even 6486in the hardware prior to interception). If this capability is not enabled, the 6487old way of handling SIGP orders is used (partially in kernel and user space). 6488 64897.3 KVM_CAP_S390_VECTOR_REGISTERS 6490--------------------------------- 6491 6492:Architectures: s390 6493:Parameters: none 6494:Returns: 0 on success, negative value on error 6495 6496Allows use of the vector registers introduced with z13 processor, and 6497provides for the synchronization between host and user space. Will 6498return -EINVAL if the machine does not support vectors. 6499 65007.4 KVM_CAP_S390_USER_STSI 6501-------------------------- 6502 6503:Architectures: s390 6504:Parameters: none 6505 6506This capability allows post-handlers for the STSI instruction. After 6507initial handling in the kernel, KVM exits to user space with 6508KVM_EXIT_S390_STSI to allow user space to insert further data. 6509 6510Before exiting to userspace, kvm handlers should fill in s390_stsi field of 6511vcpu->run:: 6512 6513 struct { 6514 __u64 addr; 6515 __u8 ar; 6516 __u8 reserved; 6517 __u8 fc; 6518 __u8 sel1; 6519 __u16 sel2; 6520 } s390_stsi; 6521 6522 @addr - guest address of STSI SYSIB 6523 @fc - function code 6524 @sel1 - selector 1 6525 @sel2 - selector 2 6526 @ar - access register number 6527 6528KVM handlers should exit to userspace with rc = -EREMOTE. 6529 65307.5 KVM_CAP_SPLIT_IRQCHIP 6531------------------------- 6532 6533:Architectures: x86 6534:Parameters: args[0] - number of routes reserved for userspace IOAPICs 6535:Returns: 0 on success, -1 on error 6536 6537Create a local apic for each processor in the kernel. This can be used 6538instead of KVM_CREATE_IRQCHIP if the userspace VMM wishes to emulate the 6539IOAPIC and PIC (and also the PIT, even though this has to be enabled 6540separately). 6541 6542This capability also enables in kernel routing of interrupt requests; 6543when KVM_CAP_SPLIT_IRQCHIP only routes of KVM_IRQ_ROUTING_MSI type are 6544used in the IRQ routing table. The first args[0] MSI routes are reserved 6545for the IOAPIC pins. Whenever the LAPIC receives an EOI for these routes, 6546a KVM_EXIT_IOAPIC_EOI vmexit will be reported to userspace. 6547 6548Fails if VCPU has already been created, or if the irqchip is already in the 6549kernel (i.e. KVM_CREATE_IRQCHIP has already been called). 6550 65517.6 KVM_CAP_S390_RI 6552------------------- 6553 6554:Architectures: s390 6555:Parameters: none 6556 6557Allows use of runtime-instrumentation introduced with zEC12 processor. 6558Will return -EINVAL if the machine does not support runtime-instrumentation. 6559Will return -EBUSY if a VCPU has already been created. 6560 65617.7 KVM_CAP_X2APIC_API 6562---------------------- 6563 6564:Architectures: x86 6565:Parameters: args[0] - features that should be enabled 6566:Returns: 0 on success, -EINVAL when args[0] contains invalid features 6567 6568Valid feature flags in args[0] are:: 6569 6570 #define KVM_X2APIC_API_USE_32BIT_IDS (1ULL << 0) 6571 #define KVM_X2APIC_API_DISABLE_BROADCAST_QUIRK (1ULL << 1) 6572 6573Enabling KVM_X2APIC_API_USE_32BIT_IDS changes the behavior of 6574KVM_SET_GSI_ROUTING, KVM_SIGNAL_MSI, KVM_SET_LAPIC, and KVM_GET_LAPIC, 6575allowing the use of 32-bit APIC IDs. See KVM_CAP_X2APIC_API in their 6576respective sections. 6577 6578KVM_X2APIC_API_DISABLE_BROADCAST_QUIRK must be enabled for x2APIC to work 6579in logical mode or with more than 255 VCPUs. Otherwise, KVM treats 0xff 6580as a broadcast even in x2APIC mode in order to support physical x2APIC 6581without interrupt remapping. This is undesirable in logical mode, 6582where 0xff represents CPUs 0-7 in cluster 0. 6583 65847.8 KVM_CAP_S390_USER_INSTR0 6585---------------------------- 6586 6587:Architectures: s390 6588:Parameters: none 6589 6590With this capability enabled, all illegal instructions 0x0000 (2 bytes) will 6591be intercepted and forwarded to user space. User space can use this 6592mechanism e.g. to realize 2-byte software breakpoints. The kernel will 6593not inject an operating exception for these instructions, user space has 6594to take care of that. 6595 6596This capability can be enabled dynamically even if VCPUs were already 6597created and are running. 6598 65997.9 KVM_CAP_S390_GS 6600------------------- 6601 6602:Architectures: s390 6603:Parameters: none 6604:Returns: 0 on success; -EINVAL if the machine does not support 6605 guarded storage; -EBUSY if a VCPU has already been created. 6606 6607Allows use of guarded storage for the KVM guest. 6608 66097.10 KVM_CAP_S390_AIS 6610--------------------- 6611 6612:Architectures: s390 6613:Parameters: none 6614 6615Allow use of adapter-interruption suppression. 6616:Returns: 0 on success; -EBUSY if a VCPU has already been created. 6617 66187.11 KVM_CAP_PPC_SMT 6619-------------------- 6620 6621:Architectures: ppc 6622:Parameters: vsmt_mode, flags 6623 6624Enabling this capability on a VM provides userspace with a way to set 6625the desired virtual SMT mode (i.e. the number of virtual CPUs per 6626virtual core). The virtual SMT mode, vsmt_mode, must be a power of 2 6627between 1 and 8. On POWER8, vsmt_mode must also be no greater than 6628the number of threads per subcore for the host. Currently flags must 6629be 0. A successful call to enable this capability will result in 6630vsmt_mode being returned when the KVM_CAP_PPC_SMT capability is 6631subsequently queried for the VM. This capability is only supported by 6632HV KVM, and can only be set before any VCPUs have been created. 6633The KVM_CAP_PPC_SMT_POSSIBLE capability indicates which virtual SMT 6634modes are available. 6635 66367.12 KVM_CAP_PPC_FWNMI 6637---------------------- 6638 6639:Architectures: ppc 6640:Parameters: none 6641 6642With this capability a machine check exception in the guest address 6643space will cause KVM to exit the guest with NMI exit reason. This 6644enables QEMU to build error log and branch to guest kernel registered 6645machine check handling routine. Without this capability KVM will 6646branch to guests' 0x200 interrupt vector. 6647 66487.13 KVM_CAP_X86_DISABLE_EXITS 6649------------------------------ 6650 6651:Architectures: x86 6652:Parameters: args[0] defines which exits are disabled 6653:Returns: 0 on success, -EINVAL when args[0] contains invalid exits 6654 6655Valid bits in args[0] are:: 6656 6657 #define KVM_X86_DISABLE_EXITS_MWAIT (1 << 0) 6658 #define KVM_X86_DISABLE_EXITS_HLT (1 << 1) 6659 #define KVM_X86_DISABLE_EXITS_PAUSE (1 << 2) 6660 #define KVM_X86_DISABLE_EXITS_CSTATE (1 << 3) 6661 6662Enabling this capability on a VM provides userspace with a way to no 6663longer intercept some instructions for improved latency in some 6664workloads, and is suggested when vCPUs are associated to dedicated 6665physical CPUs. More bits can be added in the future; userspace can 6666just pass the KVM_CHECK_EXTENSION result to KVM_ENABLE_CAP to disable 6667all such vmexits. 6668 6669Do not enable KVM_FEATURE_PV_UNHALT if you disable HLT exits. 6670 66717.14 KVM_CAP_S390_HPAGE_1M 6672-------------------------- 6673 6674:Architectures: s390 6675:Parameters: none 6676:Returns: 0 on success, -EINVAL if hpage module parameter was not set 6677 or cmma is enabled, or the VM has the KVM_VM_S390_UCONTROL 6678 flag set 6679 6680With this capability the KVM support for memory backing with 1m pages 6681through hugetlbfs can be enabled for a VM. After the capability is 6682enabled, cmma can't be enabled anymore and pfmfi and the storage key 6683interpretation are disabled. If cmma has already been enabled or the 6684hpage module parameter is not set to 1, -EINVAL is returned. 6685 6686While it is generally possible to create a huge page backed VM without 6687this capability, the VM will not be able to run. 6688 66897.15 KVM_CAP_MSR_PLATFORM_INFO 6690------------------------------ 6691 6692:Architectures: x86 6693:Parameters: args[0] whether feature should be enabled or not 6694 6695With this capability, a guest may read the MSR_PLATFORM_INFO MSR. Otherwise, 6696a #GP would be raised when the guest tries to access. Currently, this 6697capability does not enable write permissions of this MSR for the guest. 6698 66997.16 KVM_CAP_PPC_NESTED_HV 6700-------------------------- 6701 6702:Architectures: ppc 6703:Parameters: none 6704:Returns: 0 on success, -EINVAL when the implementation doesn't support 6705 nested-HV virtualization. 6706 6707HV-KVM on POWER9 and later systems allows for "nested-HV" 6708virtualization, which provides a way for a guest VM to run guests that 6709can run using the CPU's supervisor mode (privileged non-hypervisor 6710state). Enabling this capability on a VM depends on the CPU having 6711the necessary functionality and on the facility being enabled with a 6712kvm-hv module parameter. 6713 67147.17 KVM_CAP_EXCEPTION_PAYLOAD 6715------------------------------ 6716 6717:Architectures: x86 6718:Parameters: args[0] whether feature should be enabled or not 6719 6720With this capability enabled, CR2 will not be modified prior to the 6721emulated VM-exit when L1 intercepts a #PF exception that occurs in 6722L2. Similarly, for kvm-intel only, DR6 will not be modified prior to 6723the emulated VM-exit when L1 intercepts a #DB exception that occurs in 6724L2. As a result, when KVM_GET_VCPU_EVENTS reports a pending #PF (or 6725#DB) exception for L2, exception.has_payload will be set and the 6726faulting address (or the new DR6 bits*) will be reported in the 6727exception_payload field. Similarly, when userspace injects a #PF (or 6728#DB) into L2 using KVM_SET_VCPU_EVENTS, it is expected to set 6729exception.has_payload and to put the faulting address - or the new DR6 6730bits\ [#]_ - in the exception_payload field. 6731 6732This capability also enables exception.pending in struct 6733kvm_vcpu_events, which allows userspace to distinguish between pending 6734and injected exceptions. 6735 6736 6737.. [#] For the new DR6 bits, note that bit 16 is set iff the #DB exception 6738 will clear DR6.RTM. 6739 67407.18 KVM_CAP_MANUAL_DIRTY_LOG_PROTECT2 6741 6742:Architectures: x86, arm, arm64, mips 6743:Parameters: args[0] whether feature should be enabled or not 6744 6745Valid flags are:: 6746 6747 #define KVM_DIRTY_LOG_MANUAL_PROTECT_ENABLE (1 << 0) 6748 #define KVM_DIRTY_LOG_INITIALLY_SET (1 << 1) 6749 6750With KVM_DIRTY_LOG_MANUAL_PROTECT_ENABLE is set, KVM_GET_DIRTY_LOG will not 6751automatically clear and write-protect all pages that are returned as dirty. 6752Rather, userspace will have to do this operation separately using 6753KVM_CLEAR_DIRTY_LOG. 6754 6755At the cost of a slightly more complicated operation, this provides better 6756scalability and responsiveness for two reasons. First, 6757KVM_CLEAR_DIRTY_LOG ioctl can operate on a 64-page granularity rather 6758than requiring to sync a full memslot; this ensures that KVM does not 6759take spinlocks for an extended period of time. Second, in some cases a 6760large amount of time can pass between a call to KVM_GET_DIRTY_LOG and 6761userspace actually using the data in the page. Pages can be modified 6762during this time, which is inefficient for both the guest and userspace: 6763the guest will incur a higher penalty due to write protection faults, 6764while userspace can see false reports of dirty pages. Manual reprotection 6765helps reducing this time, improving guest performance and reducing the 6766number of dirty log false positives. 6767 6768With KVM_DIRTY_LOG_INITIALLY_SET set, all the bits of the dirty bitmap 6769will be initialized to 1 when created. This also improves performance because 6770dirty logging can be enabled gradually in small chunks on the first call 6771to KVM_CLEAR_DIRTY_LOG. KVM_DIRTY_LOG_INITIALLY_SET depends on 6772KVM_DIRTY_LOG_MANUAL_PROTECT_ENABLE (it is also only available on 6773x86 and arm64 for now). 6774 6775KVM_CAP_MANUAL_DIRTY_LOG_PROTECT2 was previously available under the name 6776KVM_CAP_MANUAL_DIRTY_LOG_PROTECT, but the implementation had bugs that make 6777it hard or impossible to use it correctly. The availability of 6778KVM_CAP_MANUAL_DIRTY_LOG_PROTECT2 signals that those bugs are fixed. 6779Userspace should not try to use KVM_CAP_MANUAL_DIRTY_LOG_PROTECT. 6780 67817.19 KVM_CAP_PPC_SECURE_GUEST 6782------------------------------ 6783 6784:Architectures: ppc 6785 6786This capability indicates that KVM is running on a host that has 6787ultravisor firmware and thus can support a secure guest. On such a 6788system, a guest can ask the ultravisor to make it a secure guest, 6789one whose memory is inaccessible to the host except for pages which 6790are explicitly requested to be shared with the host. The ultravisor 6791notifies KVM when a guest requests to become a secure guest, and KVM 6792has the opportunity to veto the transition. 6793 6794If present, this capability can be enabled for a VM, meaning that KVM 6795will allow the transition to secure guest mode. Otherwise KVM will 6796veto the transition. 6797 67987.20 KVM_CAP_HALT_POLL 6799---------------------- 6800 6801:Architectures: all 6802:Target: VM 6803:Parameters: args[0] is the maximum poll time in nanoseconds 6804:Returns: 0 on success; -1 on error 6805 6806This capability overrides the kvm module parameter halt_poll_ns for the 6807target VM. 6808 6809VCPU polling allows a VCPU to poll for wakeup events instead of immediately 6810scheduling during guest halts. The maximum time a VCPU can spend polling is 6811controlled by the kvm module parameter halt_poll_ns. This capability allows 6812the maximum halt time to specified on a per-VM basis, effectively overriding 6813the module parameter for the target VM. 6814 68157.21 KVM_CAP_X86_USER_SPACE_MSR 6816------------------------------- 6817 6818:Architectures: x86 6819:Target: VM 6820:Parameters: args[0] contains the mask of KVM_MSR_EXIT_REASON_* events to report 6821:Returns: 0 on success; -1 on error 6822 6823This capability enables trapping of #GP invoking RDMSR and WRMSR instructions 6824into user space. 6825 6826When a guest requests to read or write an MSR, KVM may not implement all MSRs 6827that are relevant to a respective system. It also does not differentiate by 6828CPU type. 6829 6830To allow more fine grained control over MSR handling, user space may enable 6831this capability. With it enabled, MSR accesses that match the mask specified in 6832args[0] and trigger a #GP event inside the guest by KVM will instead trigger 6833KVM_EXIT_X86_RDMSR and KVM_EXIT_X86_WRMSR exit notifications which user space 6834can then handle to implement model specific MSR handling and/or user notifications 6835to inform a user that an MSR was not handled. 6836 68377.22 KVM_CAP_X86_BUS_LOCK_EXIT 6838------------------------------- 6839 6840:Architectures: x86 6841:Target: VM 6842:Parameters: args[0] defines the policy used when bus locks detected in guest 6843:Returns: 0 on success, -EINVAL when args[0] contains invalid bits 6844 6845Valid bits in args[0] are:: 6846 6847 #define KVM_BUS_LOCK_DETECTION_OFF (1 << 0) 6848 #define KVM_BUS_LOCK_DETECTION_EXIT (1 << 1) 6849 6850Enabling this capability on a VM provides userspace with a way to select 6851a policy to handle the bus locks detected in guest. Userspace can obtain 6852the supported modes from the result of KVM_CHECK_EXTENSION and define it 6853through the KVM_ENABLE_CAP. 6854 6855KVM_BUS_LOCK_DETECTION_OFF and KVM_BUS_LOCK_DETECTION_EXIT are supported 6856currently and mutually exclusive with each other. More bits can be added in 6857the future. 6858 6859With KVM_BUS_LOCK_DETECTION_OFF set, bus locks in guest will not cause vm exits 6860so that no additional actions are needed. This is the default mode. 6861 6862With KVM_BUS_LOCK_DETECTION_EXIT set, vm exits happen when bus lock detected 6863in VM. KVM just exits to userspace when handling them. Userspace can enforce 6864its own throttling or other policy based mitigations. 6865 6866This capability is aimed to address the thread that VM can exploit bus locks to 6867degree the performance of the whole system. Once the userspace enable this 6868capability and select the KVM_BUS_LOCK_DETECTION_EXIT mode, KVM will set the 6869KVM_RUN_BUS_LOCK flag in vcpu-run->flags field and exit to userspace. Concerning 6870the bus lock vm exit can be preempted by a higher priority VM exit, the exit 6871notifications to userspace can be KVM_EXIT_BUS_LOCK or other reasons. 6872KVM_RUN_BUS_LOCK flag is used to distinguish between them. 6873 68747.23 KVM_CAP_PPC_DAWR1 6875---------------------- 6876 6877:Architectures: ppc 6878:Parameters: none 6879:Returns: 0 on success, -EINVAL when CPU doesn't support 2nd DAWR 6880 6881This capability can be used to check / enable 2nd DAWR feature provided 6882by POWER10 processor. 6883 6884 68857.24 KVM_CAP_VM_COPY_ENC_CONTEXT_FROM 6886------------------------------------- 6887 6888Architectures: x86 SEV enabled 6889Type: vm 6890Parameters: args[0] is the fd of the source vm 6891Returns: 0 on success; ENOTTY on error 6892 6893This capability enables userspace to copy encryption context from the vm 6894indicated by the fd to the vm this is called on. 6895 6896This is intended to support in-guest workloads scheduled by the host. This 6897allows the in-guest workload to maintain its own NPTs and keeps the two vms 6898from accidentally clobbering each other with interrupts and the like (separate 6899APIC/MSRs/etc). 6900 69017.25 KVM_CAP_SGX_ATTRIBUTE 6902-------------------------- 6903 6904:Architectures: x86 6905:Target: VM 6906:Parameters: args[0] is a file handle of a SGX attribute file in securityfs 6907:Returns: 0 on success, -EINVAL if the file handle is invalid or if a requested 6908 attribute is not supported by KVM. 6909 6910KVM_CAP_SGX_ATTRIBUTE enables a userspace VMM to grant a VM access to one or 6911more priveleged enclave attributes. args[0] must hold a file handle to a valid 6912SGX attribute file corresponding to an attribute that is supported/restricted 6913by KVM (currently only PROVISIONKEY). 6914 6915The SGX subsystem restricts access to a subset of enclave attributes to provide 6916additional security for an uncompromised kernel, e.g. use of the PROVISIONKEY 6917is restricted to deter malware from using the PROVISIONKEY to obtain a stable 6918system fingerprint. To prevent userspace from circumventing such restrictions 6919by running an enclave in a VM, KVM prevents access to privileged attributes by 6920default. 6921 6922See Documentation/x86/sgx.rst for more details. 6923 69247.26 KVM_CAP_PPC_RPT_INVALIDATE 6925------------------------------- 6926 6927:Capability: KVM_CAP_PPC_RPT_INVALIDATE 6928:Architectures: ppc 6929:Type: vm 6930 6931This capability indicates that the kernel is capable of handling 6932H_RPT_INVALIDATE hcall. 6933 6934In order to enable the use of H_RPT_INVALIDATE in the guest, 6935user space might have to advertise it for the guest. For example, 6936IBM pSeries (sPAPR) guest starts using it if "hcall-rpt-invalidate" is 6937present in the "ibm,hypertas-functions" device-tree property. 6938 6939This capability is enabled for hypervisors on platforms like POWER9 6940that support radix MMU. 6941 69427.27 KVM_CAP_EXIT_ON_EMULATION_FAILURE 6943-------------------------------------- 6944 6945:Architectures: x86 6946:Parameters: args[0] whether the feature should be enabled or not 6947 6948When this capability is enabled, an emulation failure will result in an exit 6949to userspace with KVM_INTERNAL_ERROR (except when the emulator was invoked 6950to handle a VMware backdoor instruction). Furthermore, KVM will now provide up 6951to 15 instruction bytes for any exit to userspace resulting from an emulation 6952failure. When these exits to userspace occur use the emulation_failure struct 6953instead of the internal struct. They both have the same layout, but the 6954emulation_failure struct matches the content better. It also explicitly 6955defines the 'flags' field which is used to describe the fields in the struct 6956that are valid (ie: if KVM_INTERNAL_ERROR_EMULATION_FLAG_INSTRUCTION_BYTES is 6957set in the 'flags' field then both 'insn_size' and 'insn_bytes' have valid data 6958in them.) 6959 69607.28 KVM_CAP_ARM_MTE 6961-------------------- 6962 6963:Architectures: arm64 6964:Parameters: none 6965 6966This capability indicates that KVM (and the hardware) supports exposing the 6967Memory Tagging Extensions (MTE) to the guest. It must also be enabled by the 6968VMM before creating any VCPUs to allow the guest access. Note that MTE is only 6969available to a guest running in AArch64 mode and enabling this capability will 6970cause attempts to create AArch32 VCPUs to fail. 6971 6972When enabled the guest is able to access tags associated with any memory given 6973to the guest. KVM will ensure that the tags are maintained during swap or 6974hibernation of the host; however the VMM needs to manually save/restore the 6975tags as appropriate if the VM is migrated. 6976 6977When this capability is enabled all memory in memslots must be mapped as 6978not-shareable (no MAP_SHARED), attempts to create a memslot with a 6979MAP_SHARED mmap will result in an -EINVAL return. 6980 6981When enabled the VMM may make use of the ``KVM_ARM_MTE_COPY_TAGS`` ioctl to 6982perform a bulk copy of tags to/from the guest. 6983 69847.29 KVM_CAP_VM_MOVE_ENC_CONTEXT_FROM 6985------------------------------------- 6986 6987Architectures: x86 SEV enabled 6988Type: vm 6989Parameters: args[0] is the fd of the source vm 6990Returns: 0 on success 6991 6992This capability enables userspace to migrate the encryption context from the VM 6993indicated by the fd to the VM this is called on. 6994 6995This is intended to support intra-host migration of VMs between userspace VMMs, 6996upgrading the VMM process without interrupting the guest. 6997 69988. Other capabilities. 6999====================== 7000 7001This section lists capabilities that give information about other 7002features of the KVM implementation. 7003 70048.1 KVM_CAP_PPC_HWRNG 7005--------------------- 7006 7007:Architectures: ppc 7008 7009This capability, if KVM_CHECK_EXTENSION indicates that it is 7010available, means that the kernel has an implementation of the 7011H_RANDOM hypercall backed by a hardware random-number generator. 7012If present, the kernel H_RANDOM handler can be enabled for guest use 7013with the KVM_CAP_PPC_ENABLE_HCALL capability. 7014 70158.2 KVM_CAP_HYPERV_SYNIC 7016------------------------ 7017 7018:Architectures: x86 7019 7020This capability, if KVM_CHECK_EXTENSION indicates that it is 7021available, means that the kernel has an implementation of the 7022Hyper-V Synthetic interrupt controller(SynIC). Hyper-V SynIC is 7023used to support Windows Hyper-V based guest paravirt drivers(VMBus). 7024 7025In order to use SynIC, it has to be activated by setting this 7026capability via KVM_ENABLE_CAP ioctl on the vcpu fd. Note that this 7027will disable the use of APIC hardware virtualization even if supported 7028by the CPU, as it's incompatible with SynIC auto-EOI behavior. 7029 70308.3 KVM_CAP_PPC_RADIX_MMU 7031------------------------- 7032 7033:Architectures: ppc 7034 7035This capability, if KVM_CHECK_EXTENSION indicates that it is 7036available, means that the kernel can support guests using the 7037radix MMU defined in Power ISA V3.00 (as implemented in the POWER9 7038processor). 7039 70408.4 KVM_CAP_PPC_HASH_MMU_V3 7041--------------------------- 7042 7043:Architectures: ppc 7044 7045This capability, if KVM_CHECK_EXTENSION indicates that it is 7046available, means that the kernel can support guests using the 7047hashed page table MMU defined in Power ISA V3.00 (as implemented in 7048the POWER9 processor), including in-memory segment tables. 7049 70508.5 KVM_CAP_MIPS_VZ 7051------------------- 7052 7053:Architectures: mips 7054 7055This capability, if KVM_CHECK_EXTENSION on the main kvm handle indicates that 7056it is available, means that full hardware assisted virtualization capabilities 7057of the hardware are available for use through KVM. An appropriate 7058KVM_VM_MIPS_* type must be passed to KVM_CREATE_VM to create a VM which 7059utilises it. 7060 7061If KVM_CHECK_EXTENSION on a kvm VM handle indicates that this capability is 7062available, it means that the VM is using full hardware assisted virtualization 7063capabilities of the hardware. This is useful to check after creating a VM with 7064KVM_VM_MIPS_DEFAULT. 7065 7066The value returned by KVM_CHECK_EXTENSION should be compared against known 7067values (see below). All other values are reserved. This is to allow for the 7068possibility of other hardware assisted virtualization implementations which 7069may be incompatible with the MIPS VZ ASE. 7070 7071== ========================================================================== 7072 0 The trap & emulate implementation is in use to run guest code in user 7073 mode. Guest virtual memory segments are rearranged to fit the guest in the 7074 user mode address space. 7075 7076 1 The MIPS VZ ASE is in use, providing full hardware assisted 7077 virtualization, including standard guest virtual memory segments. 7078== ========================================================================== 7079 70808.6 KVM_CAP_MIPS_TE 7081------------------- 7082 7083:Architectures: mips 7084 7085This capability, if KVM_CHECK_EXTENSION on the main kvm handle indicates that 7086it is available, means that the trap & emulate implementation is available to 7087run guest code in user mode, even if KVM_CAP_MIPS_VZ indicates that hardware 7088assisted virtualisation is also available. KVM_VM_MIPS_TE (0) must be passed 7089to KVM_CREATE_VM to create a VM which utilises it. 7090 7091If KVM_CHECK_EXTENSION on a kvm VM handle indicates that this capability is 7092available, it means that the VM is using trap & emulate. 7093 70948.7 KVM_CAP_MIPS_64BIT 7095---------------------- 7096 7097:Architectures: mips 7098 7099This capability indicates the supported architecture type of the guest, i.e. the 7100supported register and address width. 7101 7102The values returned when this capability is checked by KVM_CHECK_EXTENSION on a 7103kvm VM handle correspond roughly to the CP0_Config.AT register field, and should 7104be checked specifically against known values (see below). All other values are 7105reserved. 7106 7107== ======================================================================== 7108 0 MIPS32 or microMIPS32. 7109 Both registers and addresses are 32-bits wide. 7110 It will only be possible to run 32-bit guest code. 7111 7112 1 MIPS64 or microMIPS64 with access only to 32-bit compatibility segments. 7113 Registers are 64-bits wide, but addresses are 32-bits wide. 7114 64-bit guest code may run but cannot access MIPS64 memory segments. 7115 It will also be possible to run 32-bit guest code. 7116 7117 2 MIPS64 or microMIPS64 with access to all address segments. 7118 Both registers and addresses are 64-bits wide. 7119 It will be possible to run 64-bit or 32-bit guest code. 7120== ======================================================================== 7121 71228.9 KVM_CAP_ARM_USER_IRQ 7123------------------------ 7124 7125:Architectures: arm, arm64 7126 7127This capability, if KVM_CHECK_EXTENSION indicates that it is available, means 7128that if userspace creates a VM without an in-kernel interrupt controller, it 7129will be notified of changes to the output level of in-kernel emulated devices, 7130which can generate virtual interrupts, presented to the VM. 7131For such VMs, on every return to userspace, the kernel 7132updates the vcpu's run->s.regs.device_irq_level field to represent the actual 7133output level of the device. 7134 7135Whenever kvm detects a change in the device output level, kvm guarantees at 7136least one return to userspace before running the VM. This exit could either 7137be a KVM_EXIT_INTR or any other exit event, like KVM_EXIT_MMIO. This way, 7138userspace can always sample the device output level and re-compute the state of 7139the userspace interrupt controller. Userspace should always check the state 7140of run->s.regs.device_irq_level on every kvm exit. 7141The value in run->s.regs.device_irq_level can represent both level and edge 7142triggered interrupt signals, depending on the device. Edge triggered interrupt 7143signals will exit to userspace with the bit in run->s.regs.device_irq_level 7144set exactly once per edge signal. 7145 7146The field run->s.regs.device_irq_level is available independent of 7147run->kvm_valid_regs or run->kvm_dirty_regs bits. 7148 7149If KVM_CAP_ARM_USER_IRQ is supported, the KVM_CHECK_EXTENSION ioctl returns a 7150number larger than 0 indicating the version of this capability is implemented 7151and thereby which bits in run->s.regs.device_irq_level can signal values. 7152 7153Currently the following bits are defined for the device_irq_level bitmap:: 7154 7155 KVM_CAP_ARM_USER_IRQ >= 1: 7156 7157 KVM_ARM_DEV_EL1_VTIMER - EL1 virtual timer 7158 KVM_ARM_DEV_EL1_PTIMER - EL1 physical timer 7159 KVM_ARM_DEV_PMU - ARM PMU overflow interrupt signal 7160 7161Future versions of kvm may implement additional events. These will get 7162indicated by returning a higher number from KVM_CHECK_EXTENSION and will be 7163listed above. 7164 71658.10 KVM_CAP_PPC_SMT_POSSIBLE 7166----------------------------- 7167 7168:Architectures: ppc 7169 7170Querying this capability returns a bitmap indicating the possible 7171virtual SMT modes that can be set using KVM_CAP_PPC_SMT. If bit N 7172(counting from the right) is set, then a virtual SMT mode of 2^N is 7173available. 7174 71758.11 KVM_CAP_HYPERV_SYNIC2 7176-------------------------- 7177 7178:Architectures: x86 7179 7180This capability enables a newer version of Hyper-V Synthetic interrupt 7181controller (SynIC). The only difference with KVM_CAP_HYPERV_SYNIC is that KVM 7182doesn't clear SynIC message and event flags pages when they are enabled by 7183writing to the respective MSRs. 7184 71858.12 KVM_CAP_HYPERV_VP_INDEX 7186---------------------------- 7187 7188:Architectures: x86 7189 7190This capability indicates that userspace can load HV_X64_MSR_VP_INDEX msr. Its 7191value is used to denote the target vcpu for a SynIC interrupt. For 7192compatibilty, KVM initializes this msr to KVM's internal vcpu index. When this 7193capability is absent, userspace can still query this msr's value. 7194 71958.13 KVM_CAP_S390_AIS_MIGRATION 7196------------------------------- 7197 7198:Architectures: s390 7199:Parameters: none 7200 7201This capability indicates if the flic device will be able to get/set the 7202AIS states for migration via the KVM_DEV_FLIC_AISM_ALL attribute and allows 7203to discover this without having to create a flic device. 7204 72058.14 KVM_CAP_S390_PSW 7206--------------------- 7207 7208:Architectures: s390 7209 7210This capability indicates that the PSW is exposed via the kvm_run structure. 7211 72128.15 KVM_CAP_S390_GMAP 7213---------------------- 7214 7215:Architectures: s390 7216 7217This capability indicates that the user space memory used as guest mapping can 7218be anywhere in the user memory address space, as long as the memory slots are 7219aligned and sized to a segment (1MB) boundary. 7220 72218.16 KVM_CAP_S390_COW 7222--------------------- 7223 7224:Architectures: s390 7225 7226This capability indicates that the user space memory used as guest mapping can 7227use copy-on-write semantics as well as dirty pages tracking via read-only page 7228tables. 7229 72308.17 KVM_CAP_S390_BPB 7231--------------------- 7232 7233:Architectures: s390 7234 7235This capability indicates that kvm will implement the interfaces to handle 7236reset, migration and nested KVM for branch prediction blocking. The stfle 7237facility 82 should not be provided to the guest without this capability. 7238 72398.18 KVM_CAP_HYPERV_TLBFLUSH 7240---------------------------- 7241 7242:Architectures: x86 7243 7244This capability indicates that KVM supports paravirtualized Hyper-V TLB Flush 7245hypercalls: 7246HvFlushVirtualAddressSpace, HvFlushVirtualAddressSpaceEx, 7247HvFlushVirtualAddressList, HvFlushVirtualAddressListEx. 7248 72498.19 KVM_CAP_ARM_INJECT_SERROR_ESR 7250---------------------------------- 7251 7252:Architectures: arm, arm64 7253 7254This capability indicates that userspace can specify (via the 7255KVM_SET_VCPU_EVENTS ioctl) the syndrome value reported to the guest when it 7256takes a virtual SError interrupt exception. 7257If KVM advertises this capability, userspace can only specify the ISS field for 7258the ESR syndrome. Other parts of the ESR, such as the EC are generated by the 7259CPU when the exception is taken. If this virtual SError is taken to EL1 using 7260AArch64, this value will be reported in the ISS field of ESR_ELx. 7261 7262See KVM_CAP_VCPU_EVENTS for more details. 7263 72648.20 KVM_CAP_HYPERV_SEND_IPI 7265---------------------------- 7266 7267:Architectures: x86 7268 7269This capability indicates that KVM supports paravirtualized Hyper-V IPI send 7270hypercalls: 7271HvCallSendSyntheticClusterIpi, HvCallSendSyntheticClusterIpiEx. 7272 72738.21 KVM_CAP_HYPERV_DIRECT_TLBFLUSH 7274----------------------------------- 7275 7276:Architectures: x86 7277 7278This capability indicates that KVM running on top of Hyper-V hypervisor 7279enables Direct TLB flush for its guests meaning that TLB flush 7280hypercalls are handled by Level 0 hypervisor (Hyper-V) bypassing KVM. 7281Due to the different ABI for hypercall parameters between Hyper-V and 7282KVM, enabling this capability effectively disables all hypercall 7283handling by KVM (as some KVM hypercall may be mistakenly treated as TLB 7284flush hypercalls by Hyper-V) so userspace should disable KVM identification 7285in CPUID and only exposes Hyper-V identification. In this case, guest 7286thinks it's running on Hyper-V and only use Hyper-V hypercalls. 7287 72888.22 KVM_CAP_S390_VCPU_RESETS 7289----------------------------- 7290 7291:Architectures: s390 7292 7293This capability indicates that the KVM_S390_NORMAL_RESET and 7294KVM_S390_CLEAR_RESET ioctls are available. 7295 72968.23 KVM_CAP_S390_PROTECTED 7297--------------------------- 7298 7299:Architectures: s390 7300 7301This capability indicates that the Ultravisor has been initialized and 7302KVM can therefore start protected VMs. 7303This capability governs the KVM_S390_PV_COMMAND ioctl and the 7304KVM_MP_STATE_LOAD MP_STATE. KVM_SET_MP_STATE can fail for protected 7305guests when the state change is invalid. 7306 73078.24 KVM_CAP_STEAL_TIME 7308----------------------- 7309 7310:Architectures: arm64, x86 7311 7312This capability indicates that KVM supports steal time accounting. 7313When steal time accounting is supported it may be enabled with 7314architecture-specific interfaces. This capability and the architecture- 7315specific interfaces must be consistent, i.e. if one says the feature 7316is supported, than the other should as well and vice versa. For arm64 7317see Documentation/virt/kvm/devices/vcpu.rst "KVM_ARM_VCPU_PVTIME_CTRL". 7318For x86 see Documentation/virt/kvm/msr.rst "MSR_KVM_STEAL_TIME". 7319 73208.25 KVM_CAP_S390_DIAG318 7321------------------------- 7322 7323:Architectures: s390 7324 7325This capability enables a guest to set information about its control program 7326(i.e. guest kernel type and version). The information is helpful during 7327system/firmware service events, providing additional data about the guest 7328environments running on the machine. 7329 7330The information is associated with the DIAGNOSE 0x318 instruction, which sets 7331an 8-byte value consisting of a one-byte Control Program Name Code (CPNC) and 7332a 7-byte Control Program Version Code (CPVC). The CPNC determines what 7333environment the control program is running in (e.g. Linux, z/VM...), and the 7334CPVC is used for information specific to OS (e.g. Linux version, Linux 7335distribution...) 7336 7337If this capability is available, then the CPNC and CPVC can be synchronized 7338between KVM and userspace via the sync regs mechanism (KVM_SYNC_DIAG318). 7339 73408.26 KVM_CAP_X86_USER_SPACE_MSR 7341------------------------------- 7342 7343:Architectures: x86 7344 7345This capability indicates that KVM supports deflection of MSR reads and 7346writes to user space. It can be enabled on a VM level. If enabled, MSR 7347accesses that would usually trigger a #GP by KVM into the guest will 7348instead get bounced to user space through the KVM_EXIT_X86_RDMSR and 7349KVM_EXIT_X86_WRMSR exit notifications. 7350 73518.27 KVM_CAP_X86_MSR_FILTER 7352--------------------------- 7353 7354:Architectures: x86 7355 7356This capability indicates that KVM supports that accesses to user defined MSRs 7357may be rejected. With this capability exposed, KVM exports new VM ioctl 7358KVM_X86_SET_MSR_FILTER which user space can call to specify bitmaps of MSR 7359ranges that KVM should reject access to. 7360 7361In combination with KVM_CAP_X86_USER_SPACE_MSR, this allows user space to 7362trap and emulate MSRs that are outside of the scope of KVM as well as 7363limit the attack surface on KVM's MSR emulation code. 7364 73658.28 KVM_CAP_ENFORCE_PV_FEATURE_CPUID 7366------------------------------------- 7367 7368Architectures: x86 7369 7370When enabled, KVM will disable paravirtual features provided to the 7371guest according to the bits in the KVM_CPUID_FEATURES CPUID leaf 7372(0x40000001). Otherwise, a guest may use the paravirtual features 7373regardless of what has actually been exposed through the CPUID leaf. 7374 73758.29 KVM_CAP_DIRTY_LOG_RING 7376--------------------------- 7377 7378:Architectures: x86 7379:Parameters: args[0] - size of the dirty log ring 7380 7381KVM is capable of tracking dirty memory using ring buffers that are 7382mmaped into userspace; there is one dirty ring per vcpu. 7383 7384The dirty ring is available to userspace as an array of 7385``struct kvm_dirty_gfn``. Each dirty entry it's defined as:: 7386 7387 struct kvm_dirty_gfn { 7388 __u32 flags; 7389 __u32 slot; /* as_id | slot_id */ 7390 __u64 offset; 7391 }; 7392 7393The following values are defined for the flags field to define the 7394current state of the entry:: 7395 7396 #define KVM_DIRTY_GFN_F_DIRTY BIT(0) 7397 #define KVM_DIRTY_GFN_F_RESET BIT(1) 7398 #define KVM_DIRTY_GFN_F_MASK 0x3 7399 7400Userspace should call KVM_ENABLE_CAP ioctl right after KVM_CREATE_VM 7401ioctl to enable this capability for the new guest and set the size of 7402the rings. Enabling the capability is only allowed before creating any 7403vCPU, and the size of the ring must be a power of two. The larger the 7404ring buffer, the less likely the ring is full and the VM is forced to 7405exit to userspace. The optimal size depends on the workload, but it is 7406recommended that it be at least 64 KiB (4096 entries). 7407 7408Just like for dirty page bitmaps, the buffer tracks writes to 7409all user memory regions for which the KVM_MEM_LOG_DIRTY_PAGES flag was 7410set in KVM_SET_USER_MEMORY_REGION. Once a memory region is registered 7411with the flag set, userspace can start harvesting dirty pages from the 7412ring buffer. 7413 7414An entry in the ring buffer can be unused (flag bits ``00``), 7415dirty (flag bits ``01``) or harvested (flag bits ``1X``). The 7416state machine for the entry is as follows:: 7417 7418 dirtied harvested reset 7419 00 -----------> 01 -------------> 1X -------+ 7420 ^ | 7421 | | 7422 +------------------------------------------+ 7423 7424To harvest the dirty pages, userspace accesses the mmaped ring buffer 7425to read the dirty GFNs. If the flags has the DIRTY bit set (at this stage 7426the RESET bit must be cleared), then it means this GFN is a dirty GFN. 7427The userspace should harvest this GFN and mark the flags from state 7428``01b`` to ``1Xb`` (bit 0 will be ignored by KVM, but bit 1 must be set 7429to show that this GFN is harvested and waiting for a reset), and move 7430on to the next GFN. The userspace should continue to do this until the 7431flags of a GFN have the DIRTY bit cleared, meaning that it has harvested 7432all the dirty GFNs that were available. 7433 7434It's not necessary for userspace to harvest the all dirty GFNs at once. 7435However it must collect the dirty GFNs in sequence, i.e., the userspace 7436program cannot skip one dirty GFN to collect the one next to it. 7437 7438After processing one or more entries in the ring buffer, userspace 7439calls the VM ioctl KVM_RESET_DIRTY_RINGS to notify the kernel about 7440it, so that the kernel will reprotect those collected GFNs. 7441Therefore, the ioctl must be called *before* reading the content of 7442the dirty pages. 7443 7444The dirty ring can get full. When it happens, the KVM_RUN of the 7445vcpu will return with exit reason KVM_EXIT_DIRTY_LOG_FULL. 7446 7447The dirty ring interface has a major difference comparing to the 7448KVM_GET_DIRTY_LOG interface in that, when reading the dirty ring from 7449userspace, it's still possible that the kernel has not yet flushed the 7450processor's dirty page buffers into the kernel buffer (with dirty bitmaps, the 7451flushing is done by the KVM_GET_DIRTY_LOG ioctl). To achieve that, one 7452needs to kick the vcpu out of KVM_RUN using a signal. The resulting 7453vmexit ensures that all dirty GFNs are flushed to the dirty rings. 7454 7455NOTE: the capability KVM_CAP_DIRTY_LOG_RING and the corresponding 7456ioctl KVM_RESET_DIRTY_RINGS are mutual exclusive to the existing ioctls 7457KVM_GET_DIRTY_LOG and KVM_CLEAR_DIRTY_LOG. After enabling 7458KVM_CAP_DIRTY_LOG_RING with an acceptable dirty ring size, the virtual 7459machine will switch to ring-buffer dirty page tracking and further 7460KVM_GET_DIRTY_LOG or KVM_CLEAR_DIRTY_LOG ioctls will fail. 7461 74628.30 KVM_CAP_XEN_HVM 7463-------------------- 7464 7465:Architectures: x86 7466 7467This capability indicates the features that Xen supports for hosting Xen 7468PVHVM guests. Valid flags are:: 7469 7470 #define KVM_XEN_HVM_CONFIG_HYPERCALL_MSR (1 << 0) 7471 #define KVM_XEN_HVM_CONFIG_INTERCEPT_HCALL (1 << 1) 7472 #define KVM_XEN_HVM_CONFIG_SHARED_INFO (1 << 2) 7473 #define KVM_XEN_HVM_CONFIG_RUNSTATE (1 << 2) 7474 #define KVM_XEN_HVM_CONFIG_EVTCHN_2LEVEL (1 << 3) 7475 7476The KVM_XEN_HVM_CONFIG_HYPERCALL_MSR flag indicates that the KVM_XEN_HVM_CONFIG 7477ioctl is available, for the guest to set its hypercall page. 7478 7479If KVM_XEN_HVM_CONFIG_INTERCEPT_HCALL is also set, the same flag may also be 7480provided in the flags to KVM_XEN_HVM_CONFIG, without providing hypercall page 7481contents, to request that KVM generate hypercall page content automatically 7482and also enable interception of guest hypercalls with KVM_EXIT_XEN. 7483 7484The KVM_XEN_HVM_CONFIG_SHARED_INFO flag indicates the availability of the 7485KVM_XEN_HVM_SET_ATTR, KVM_XEN_HVM_GET_ATTR, KVM_XEN_VCPU_SET_ATTR and 7486KVM_XEN_VCPU_GET_ATTR ioctls, as well as the delivery of exception vectors 7487for event channel upcalls when the evtchn_upcall_pending field of a vcpu's 7488vcpu_info is set. 7489 7490The KVM_XEN_HVM_CONFIG_RUNSTATE flag indicates that the runstate-related 7491features KVM_XEN_VCPU_ATTR_TYPE_RUNSTATE_ADDR/_CURRENT/_DATA/_ADJUST are 7492supported by the KVM_XEN_VCPU_SET_ATTR/KVM_XEN_VCPU_GET_ATTR ioctls. 7493 7494The KVM_XEN_HVM_CONFIG_EVTCHN_2LEVEL flag indicates that IRQ routing entries 7495of the type KVM_IRQ_ROUTING_XEN_EVTCHN are supported, with the priority 7496field set to indicate 2 level event channel delivery. 7497 74988.31 KVM_CAP_PPC_MULTITCE 7499------------------------- 7500 7501:Capability: KVM_CAP_PPC_MULTITCE 7502:Architectures: ppc 7503:Type: vm 7504 7505This capability means the kernel is capable of handling hypercalls 7506H_PUT_TCE_INDIRECT and H_STUFF_TCE without passing those into the user 7507space. This significantly accelerates DMA operations for PPC KVM guests. 7508User space should expect that its handlers for these hypercalls 7509are not going to be called if user space previously registered LIOBN 7510in KVM (via KVM_CREATE_SPAPR_TCE or similar calls). 7511 7512In order to enable H_PUT_TCE_INDIRECT and H_STUFF_TCE use in the guest, 7513user space might have to advertise it for the guest. For example, 7514IBM pSeries (sPAPR) guest starts using them if "hcall-multi-tce" is 7515present in the "ibm,hypertas-functions" device-tree property. 7516 7517The hypercalls mentioned above may or may not be processed successfully 7518in the kernel based fast path. If they can not be handled by the kernel, 7519they will get passed on to user space. So user space still has to have 7520an implementation for these despite the in kernel acceleration. 7521 7522This capability is always enabled. 7523 75248.32 KVM_CAP_PTP_KVM 7525-------------------- 7526 7527:Architectures: arm64 7528 7529This capability indicates that the KVM virtual PTP service is 7530supported in the host. A VMM can check whether the service is 7531available to the guest on migration. 7532 75338.33 KVM_CAP_HYPERV_ENFORCE_CPUID 7534--------------------------------- 7535 7536Architectures: x86 7537 7538When enabled, KVM will disable emulated Hyper-V features provided to the 7539guest according to the bits Hyper-V CPUID feature leaves. Otherwise, all 7540currently implmented Hyper-V features are provided unconditionally when 7541Hyper-V identification is set in the HYPERV_CPUID_INTERFACE (0x40000001) 7542leaf. 7543 75448.34 KVM_CAP_EXIT_HYPERCALL 7545--------------------------- 7546 7547:Capability: KVM_CAP_EXIT_HYPERCALL 7548:Architectures: x86 7549:Type: vm 7550 7551This capability, if enabled, will cause KVM to exit to userspace 7552with KVM_EXIT_HYPERCALL exit reason to process some hypercalls. 7553 7554Calling KVM_CHECK_EXTENSION for this capability will return a bitmask 7555of hypercalls that can be configured to exit to userspace. 7556Right now, the only such hypercall is KVM_HC_MAP_GPA_RANGE. 7557 7558The argument to KVM_ENABLE_CAP is also a bitmask, and must be a subset 7559of the result of KVM_CHECK_EXTENSION. KVM will forward to userspace 7560the hypercalls whose corresponding bit is in the argument, and return 7561ENOSYS for the others. 7562