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 }; 1573 1574This ioctl would copy current vcpu's xsave struct to the userspace. 1575 1576 15774.43 KVM_SET_XSAVE 1578------------------ 1579 1580:Capability: KVM_CAP_XSAVE 1581:Architectures: x86 1582:Type: vcpu ioctl 1583:Parameters: struct kvm_xsave (in) 1584:Returns: 0 on success, -1 on error 1585 1586:: 1587 1588 1589 struct kvm_xsave { 1590 __u32 region[1024]; 1591 }; 1592 1593This ioctl would copy userspace's xsave struct to the kernel. 1594 1595 15964.44 KVM_GET_XCRS 1597----------------- 1598 1599:Capability: KVM_CAP_XCRS 1600:Architectures: x86 1601:Type: vcpu ioctl 1602:Parameters: struct kvm_xcrs (out) 1603:Returns: 0 on success, -1 on error 1604 1605:: 1606 1607 struct kvm_xcr { 1608 __u32 xcr; 1609 __u32 reserved; 1610 __u64 value; 1611 }; 1612 1613 struct kvm_xcrs { 1614 __u32 nr_xcrs; 1615 __u32 flags; 1616 struct kvm_xcr xcrs[KVM_MAX_XCRS]; 1617 __u64 padding[16]; 1618 }; 1619 1620This ioctl would copy current vcpu's xcrs to the userspace. 1621 1622 16234.45 KVM_SET_XCRS 1624----------------- 1625 1626:Capability: KVM_CAP_XCRS 1627:Architectures: x86 1628:Type: vcpu ioctl 1629:Parameters: struct kvm_xcrs (in) 1630:Returns: 0 on success, -1 on error 1631 1632:: 1633 1634 struct kvm_xcr { 1635 __u32 xcr; 1636 __u32 reserved; 1637 __u64 value; 1638 }; 1639 1640 struct kvm_xcrs { 1641 __u32 nr_xcrs; 1642 __u32 flags; 1643 struct kvm_xcr xcrs[KVM_MAX_XCRS]; 1644 __u64 padding[16]; 1645 }; 1646 1647This ioctl would set vcpu's xcr to the value userspace specified. 1648 1649 16504.46 KVM_GET_SUPPORTED_CPUID 1651---------------------------- 1652 1653:Capability: KVM_CAP_EXT_CPUID 1654:Architectures: x86 1655:Type: system ioctl 1656:Parameters: struct kvm_cpuid2 (in/out) 1657:Returns: 0 on success, -1 on error 1658 1659:: 1660 1661 struct kvm_cpuid2 { 1662 __u32 nent; 1663 __u32 padding; 1664 struct kvm_cpuid_entry2 entries[0]; 1665 }; 1666 1667 #define KVM_CPUID_FLAG_SIGNIFCANT_INDEX BIT(0) 1668 #define KVM_CPUID_FLAG_STATEFUL_FUNC BIT(1) /* deprecated */ 1669 #define KVM_CPUID_FLAG_STATE_READ_NEXT BIT(2) /* deprecated */ 1670 1671 struct kvm_cpuid_entry2 { 1672 __u32 function; 1673 __u32 index; 1674 __u32 flags; 1675 __u32 eax; 1676 __u32 ebx; 1677 __u32 ecx; 1678 __u32 edx; 1679 __u32 padding[3]; 1680 }; 1681 1682This ioctl returns x86 cpuid features which are supported by both the 1683hardware and kvm in its default configuration. Userspace can use the 1684information returned by this ioctl to construct cpuid information (for 1685KVM_SET_CPUID2) that is consistent with hardware, kernel, and 1686userspace capabilities, and with user requirements (for example, the 1687user may wish to constrain cpuid to emulate older hardware, or for 1688feature consistency across a cluster). 1689 1690Note that certain capabilities, such as KVM_CAP_X86_DISABLE_EXITS, may 1691expose cpuid features (e.g. MONITOR) which are not supported by kvm in 1692its default configuration. If userspace enables such capabilities, it 1693is responsible for modifying the results of this ioctl appropriately. 1694 1695Userspace invokes KVM_GET_SUPPORTED_CPUID by passing a kvm_cpuid2 structure 1696with the 'nent' field indicating the number of entries in the variable-size 1697array 'entries'. If the number of entries is too low to describe the cpu 1698capabilities, an error (E2BIG) is returned. If the number is too high, 1699the 'nent' field is adjusted and an error (ENOMEM) is returned. If the 1700number is just right, the 'nent' field is adjusted to the number of valid 1701entries in the 'entries' array, which is then filled. 1702 1703The entries returned are the host cpuid as returned by the cpuid instruction, 1704with unknown or unsupported features masked out. Some features (for example, 1705x2apic), may not be present in the host cpu, but are exposed by kvm if it can 1706emulate them efficiently. The fields in each entry are defined as follows: 1707 1708 function: 1709 the eax value used to obtain the entry 1710 1711 index: 1712 the ecx value used to obtain the entry (for entries that are 1713 affected by ecx) 1714 1715 flags: 1716 an OR of zero or more of the following: 1717 1718 KVM_CPUID_FLAG_SIGNIFCANT_INDEX: 1719 if the index field is valid 1720 1721 eax, ebx, ecx, edx: 1722 the values returned by the cpuid instruction for 1723 this function/index combination 1724 1725The TSC deadline timer feature (CPUID leaf 1, ecx[24]) is always returned 1726as false, since the feature depends on KVM_CREATE_IRQCHIP for local APIC 1727support. Instead it is reported via:: 1728 1729 ioctl(KVM_CHECK_EXTENSION, KVM_CAP_TSC_DEADLINE_TIMER) 1730 1731if that returns true and you use KVM_CREATE_IRQCHIP, or if you emulate the 1732feature in userspace, then you can enable the feature for KVM_SET_CPUID2. 1733 1734 17354.47 KVM_PPC_GET_PVINFO 1736----------------------- 1737 1738:Capability: KVM_CAP_PPC_GET_PVINFO 1739:Architectures: ppc 1740:Type: vm ioctl 1741:Parameters: struct kvm_ppc_pvinfo (out) 1742:Returns: 0 on success, !0 on error 1743 1744:: 1745 1746 struct kvm_ppc_pvinfo { 1747 __u32 flags; 1748 __u32 hcall[4]; 1749 __u8 pad[108]; 1750 }; 1751 1752This ioctl fetches PV specific information that need to be passed to the guest 1753using the device tree or other means from vm context. 1754 1755The hcall array defines 4 instructions that make up a hypercall. 1756 1757If any additional field gets added to this structure later on, a bit for that 1758additional piece of information will be set in the flags bitmap. 1759 1760The flags bitmap is defined as:: 1761 1762 /* the host supports the ePAPR idle hcall 1763 #define KVM_PPC_PVINFO_FLAGS_EV_IDLE (1<<0) 1764 17654.52 KVM_SET_GSI_ROUTING 1766------------------------ 1767 1768:Capability: KVM_CAP_IRQ_ROUTING 1769:Architectures: x86 s390 arm arm64 1770:Type: vm ioctl 1771:Parameters: struct kvm_irq_routing (in) 1772:Returns: 0 on success, -1 on error 1773 1774Sets the GSI routing table entries, overwriting any previously set entries. 1775 1776On arm/arm64, GSI routing has the following limitation: 1777 1778- GSI routing does not apply to KVM_IRQ_LINE but only to KVM_IRQFD. 1779 1780:: 1781 1782 struct kvm_irq_routing { 1783 __u32 nr; 1784 __u32 flags; 1785 struct kvm_irq_routing_entry entries[0]; 1786 }; 1787 1788No flags are specified so far, the corresponding field must be set to zero. 1789 1790:: 1791 1792 struct kvm_irq_routing_entry { 1793 __u32 gsi; 1794 __u32 type; 1795 __u32 flags; 1796 __u32 pad; 1797 union { 1798 struct kvm_irq_routing_irqchip irqchip; 1799 struct kvm_irq_routing_msi msi; 1800 struct kvm_irq_routing_s390_adapter adapter; 1801 struct kvm_irq_routing_hv_sint hv_sint; 1802 __u32 pad[8]; 1803 } u; 1804 }; 1805 1806 /* gsi routing entry types */ 1807 #define KVM_IRQ_ROUTING_IRQCHIP 1 1808 #define KVM_IRQ_ROUTING_MSI 2 1809 #define KVM_IRQ_ROUTING_S390_ADAPTER 3 1810 #define KVM_IRQ_ROUTING_HV_SINT 4 1811 1812flags: 1813 1814- KVM_MSI_VALID_DEVID: used along with KVM_IRQ_ROUTING_MSI routing entry 1815 type, specifies that the devid field contains a valid value. The per-VM 1816 KVM_CAP_MSI_DEVID capability advertises the requirement to provide 1817 the device ID. If this capability is not available, userspace should 1818 never set the KVM_MSI_VALID_DEVID flag as the ioctl might fail. 1819- zero otherwise 1820 1821:: 1822 1823 struct kvm_irq_routing_irqchip { 1824 __u32 irqchip; 1825 __u32 pin; 1826 }; 1827 1828 struct kvm_irq_routing_msi { 1829 __u32 address_lo; 1830 __u32 address_hi; 1831 __u32 data; 1832 union { 1833 __u32 pad; 1834 __u32 devid; 1835 }; 1836 }; 1837 1838If KVM_MSI_VALID_DEVID is set, devid contains a unique device identifier 1839for the device that wrote the MSI message. For PCI, this is usually a 1840BFD identifier in the lower 16 bits. 1841 1842On x86, address_hi is ignored unless the KVM_X2APIC_API_USE_32BIT_IDS 1843feature of KVM_CAP_X2APIC_API capability is enabled. If it is enabled, 1844address_hi bits 31-8 provide bits 31-8 of the destination id. Bits 7-0 of 1845address_hi must be zero. 1846 1847:: 1848 1849 struct kvm_irq_routing_s390_adapter { 1850 __u64 ind_addr; 1851 __u64 summary_addr; 1852 __u64 ind_offset; 1853 __u32 summary_offset; 1854 __u32 adapter_id; 1855 }; 1856 1857 struct kvm_irq_routing_hv_sint { 1858 __u32 vcpu; 1859 __u32 sint; 1860 }; 1861 1862 18634.55 KVM_SET_TSC_KHZ 1864-------------------- 1865 1866:Capability: KVM_CAP_TSC_CONTROL 1867:Architectures: x86 1868:Type: vcpu ioctl 1869:Parameters: virtual tsc_khz 1870:Returns: 0 on success, -1 on error 1871 1872Specifies the tsc frequency for the virtual machine. The unit of the 1873frequency is KHz. 1874 1875 18764.56 KVM_GET_TSC_KHZ 1877-------------------- 1878 1879:Capability: KVM_CAP_GET_TSC_KHZ 1880:Architectures: x86 1881:Type: vcpu ioctl 1882:Parameters: none 1883:Returns: virtual tsc-khz on success, negative value on error 1884 1885Returns the tsc frequency of the guest. The unit of the return value is 1886KHz. If the host has unstable tsc this ioctl returns -EIO instead as an 1887error. 1888 1889 18904.57 KVM_GET_LAPIC 1891------------------ 1892 1893:Capability: KVM_CAP_IRQCHIP 1894:Architectures: x86 1895:Type: vcpu ioctl 1896:Parameters: struct kvm_lapic_state (out) 1897:Returns: 0 on success, -1 on error 1898 1899:: 1900 1901 #define KVM_APIC_REG_SIZE 0x400 1902 struct kvm_lapic_state { 1903 char regs[KVM_APIC_REG_SIZE]; 1904 }; 1905 1906Reads the Local APIC registers and copies them into the input argument. The 1907data format and layout are the same as documented in the architecture manual. 1908 1909If KVM_X2APIC_API_USE_32BIT_IDS feature of KVM_CAP_X2APIC_API is 1910enabled, then the format of APIC_ID register depends on the APIC mode 1911(reported by MSR_IA32_APICBASE) of its VCPU. x2APIC stores APIC ID in 1912the APIC_ID register (bytes 32-35). xAPIC only allows an 8-bit APIC ID 1913which is stored in bits 31-24 of the APIC register, or equivalently in 1914byte 35 of struct kvm_lapic_state's regs field. KVM_GET_LAPIC must then 1915be called after MSR_IA32_APICBASE has been set with KVM_SET_MSR. 1916 1917If KVM_X2APIC_API_USE_32BIT_IDS feature is disabled, struct kvm_lapic_state 1918always uses xAPIC format. 1919 1920 19214.58 KVM_SET_LAPIC 1922------------------ 1923 1924:Capability: KVM_CAP_IRQCHIP 1925:Architectures: x86 1926:Type: vcpu ioctl 1927:Parameters: struct kvm_lapic_state (in) 1928:Returns: 0 on success, -1 on error 1929 1930:: 1931 1932 #define KVM_APIC_REG_SIZE 0x400 1933 struct kvm_lapic_state { 1934 char regs[KVM_APIC_REG_SIZE]; 1935 }; 1936 1937Copies the input argument into the Local APIC registers. The data format 1938and layout are the same as documented in the architecture manual. 1939 1940The format of the APIC ID register (bytes 32-35 of struct kvm_lapic_state's 1941regs field) depends on the state of the KVM_CAP_X2APIC_API capability. 1942See the note in KVM_GET_LAPIC. 1943 1944 19454.59 KVM_IOEVENTFD 1946------------------ 1947 1948:Capability: KVM_CAP_IOEVENTFD 1949:Architectures: all 1950:Type: vm ioctl 1951:Parameters: struct kvm_ioeventfd (in) 1952:Returns: 0 on success, !0 on error 1953 1954This ioctl attaches or detaches an ioeventfd to a legal pio/mmio address 1955within the guest. A guest write in the registered address will signal the 1956provided event instead of triggering an exit. 1957 1958:: 1959 1960 struct kvm_ioeventfd { 1961 __u64 datamatch; 1962 __u64 addr; /* legal pio/mmio address */ 1963 __u32 len; /* 0, 1, 2, 4, or 8 bytes */ 1964 __s32 fd; 1965 __u32 flags; 1966 __u8 pad[36]; 1967 }; 1968 1969For the special case of virtio-ccw devices on s390, the ioevent is matched 1970to a subchannel/virtqueue tuple instead. 1971 1972The following flags are defined:: 1973 1974 #define KVM_IOEVENTFD_FLAG_DATAMATCH (1 << kvm_ioeventfd_flag_nr_datamatch) 1975 #define KVM_IOEVENTFD_FLAG_PIO (1 << kvm_ioeventfd_flag_nr_pio) 1976 #define KVM_IOEVENTFD_FLAG_DEASSIGN (1 << kvm_ioeventfd_flag_nr_deassign) 1977 #define KVM_IOEVENTFD_FLAG_VIRTIO_CCW_NOTIFY \ 1978 (1 << kvm_ioeventfd_flag_nr_virtio_ccw_notify) 1979 1980If datamatch flag is set, the event will be signaled only if the written value 1981to the registered address is equal to datamatch in struct kvm_ioeventfd. 1982 1983For virtio-ccw devices, addr contains the subchannel id and datamatch the 1984virtqueue index. 1985 1986With KVM_CAP_IOEVENTFD_ANY_LENGTH, a zero length ioeventfd is allowed, and 1987the kernel will ignore the length of guest write and may get a faster vmexit. 1988The speedup may only apply to specific architectures, but the ioeventfd will 1989work anyway. 1990 19914.60 KVM_DIRTY_TLB 1992------------------ 1993 1994:Capability: KVM_CAP_SW_TLB 1995:Architectures: ppc 1996:Type: vcpu ioctl 1997:Parameters: struct kvm_dirty_tlb (in) 1998:Returns: 0 on success, -1 on error 1999 2000:: 2001 2002 struct kvm_dirty_tlb { 2003 __u64 bitmap; 2004 __u32 num_dirty; 2005 }; 2006 2007This must be called whenever userspace has changed an entry in the shared 2008TLB, prior to calling KVM_RUN on the associated vcpu. 2009 2010The "bitmap" field is the userspace address of an array. This array 2011consists of a number of bits, equal to the total number of TLB entries as 2012determined by the last successful call to KVM_CONFIG_TLB, rounded up to the 2013nearest multiple of 64. 2014 2015Each bit corresponds to one TLB entry, ordered the same as in the shared TLB 2016array. 2017 2018The array is little-endian: the bit 0 is the least significant bit of the 2019first byte, bit 8 is the least significant bit of the second byte, etc. 2020This avoids any complications with differing word sizes. 2021 2022The "num_dirty" field is a performance hint for KVM to determine whether it 2023should skip processing the bitmap and just invalidate everything. It must 2024be set to the number of set bits in the bitmap. 2025 2026 20274.62 KVM_CREATE_SPAPR_TCE 2028------------------------- 2029 2030:Capability: KVM_CAP_SPAPR_TCE 2031:Architectures: powerpc 2032:Type: vm ioctl 2033:Parameters: struct kvm_create_spapr_tce (in) 2034:Returns: file descriptor for manipulating the created TCE table 2035 2036This creates a virtual TCE (translation control entry) table, which 2037is an IOMMU for PAPR-style virtual I/O. It is used to translate 2038logical addresses used in virtual I/O into guest physical addresses, 2039and provides a scatter/gather capability for PAPR virtual I/O. 2040 2041:: 2042 2043 /* for KVM_CAP_SPAPR_TCE */ 2044 struct kvm_create_spapr_tce { 2045 __u64 liobn; 2046 __u32 window_size; 2047 }; 2048 2049The liobn field gives the logical IO bus number for which to create a 2050TCE table. The window_size field specifies the size of the DMA window 2051which this TCE table will translate - the table will contain one 64 2052bit TCE entry for every 4kiB of the DMA window. 2053 2054When the guest issues an H_PUT_TCE hcall on a liobn for which a TCE 2055table has been created using this ioctl(), the kernel will handle it 2056in real mode, updating the TCE table. H_PUT_TCE calls for other 2057liobns will cause a vm exit and must be handled by userspace. 2058 2059The return value is a file descriptor which can be passed to mmap(2) 2060to map the created TCE table into userspace. This lets userspace read 2061the entries written by kernel-handled H_PUT_TCE calls, and also lets 2062userspace update the TCE table directly which is useful in some 2063circumstances. 2064 2065 20664.63 KVM_ALLOCATE_RMA 2067--------------------- 2068 2069:Capability: KVM_CAP_PPC_RMA 2070:Architectures: powerpc 2071:Type: vm ioctl 2072:Parameters: struct kvm_allocate_rma (out) 2073:Returns: file descriptor for mapping the allocated RMA 2074 2075This allocates a Real Mode Area (RMA) from the pool allocated at boot 2076time by the kernel. An RMA is a physically-contiguous, aligned region 2077of memory used on older POWER processors to provide the memory which 2078will be accessed by real-mode (MMU off) accesses in a KVM guest. 2079POWER processors support a set of sizes for the RMA that usually 2080includes 64MB, 128MB, 256MB and some larger powers of two. 2081 2082:: 2083 2084 /* for KVM_ALLOCATE_RMA */ 2085 struct kvm_allocate_rma { 2086 __u64 rma_size; 2087 }; 2088 2089The return value is a file descriptor which can be passed to mmap(2) 2090to map the allocated RMA into userspace. The mapped area can then be 2091passed to the KVM_SET_USER_MEMORY_REGION ioctl to establish it as the 2092RMA for a virtual machine. The size of the RMA in bytes (which is 2093fixed at host kernel boot time) is returned in the rma_size field of 2094the argument structure. 2095 2096The KVM_CAP_PPC_RMA capability is 1 or 2 if the KVM_ALLOCATE_RMA ioctl 2097is supported; 2 if the processor requires all virtual machines to have 2098an RMA, or 1 if the processor can use an RMA but doesn't require it, 2099because it supports the Virtual RMA (VRMA) facility. 2100 2101 21024.64 KVM_NMI 2103------------ 2104 2105:Capability: KVM_CAP_USER_NMI 2106:Architectures: x86 2107:Type: vcpu ioctl 2108:Parameters: none 2109:Returns: 0 on success, -1 on error 2110 2111Queues an NMI on the thread's vcpu. Note this is well defined only 2112when KVM_CREATE_IRQCHIP has not been called, since this is an interface 2113between the virtual cpu core and virtual local APIC. After KVM_CREATE_IRQCHIP 2114has been called, this interface is completely emulated within the kernel. 2115 2116To use this to emulate the LINT1 input with KVM_CREATE_IRQCHIP, use the 2117following algorithm: 2118 2119 - pause the vcpu 2120 - read the local APIC's state (KVM_GET_LAPIC) 2121 - check whether changing LINT1 will queue an NMI (see the LVT entry for LINT1) 2122 - if so, issue KVM_NMI 2123 - resume the vcpu 2124 2125Some guests configure the LINT1 NMI input to cause a panic, aiding in 2126debugging. 2127 2128 21294.65 KVM_S390_UCAS_MAP 2130---------------------- 2131 2132:Capability: KVM_CAP_S390_UCONTROL 2133:Architectures: s390 2134:Type: vcpu ioctl 2135:Parameters: struct kvm_s390_ucas_mapping (in) 2136:Returns: 0 in case of success 2137 2138The parameter is defined like this:: 2139 2140 struct kvm_s390_ucas_mapping { 2141 __u64 user_addr; 2142 __u64 vcpu_addr; 2143 __u64 length; 2144 }; 2145 2146This ioctl maps the memory at "user_addr" with the length "length" to 2147the vcpu's address space starting at "vcpu_addr". All parameters need to 2148be aligned by 1 megabyte. 2149 2150 21514.66 KVM_S390_UCAS_UNMAP 2152------------------------ 2153 2154:Capability: KVM_CAP_S390_UCONTROL 2155:Architectures: s390 2156:Type: vcpu ioctl 2157:Parameters: struct kvm_s390_ucas_mapping (in) 2158:Returns: 0 in case of success 2159 2160The parameter is defined like this:: 2161 2162 struct kvm_s390_ucas_mapping { 2163 __u64 user_addr; 2164 __u64 vcpu_addr; 2165 __u64 length; 2166 }; 2167 2168This ioctl unmaps the memory in the vcpu's address space starting at 2169"vcpu_addr" with the length "length". The field "user_addr" is ignored. 2170All parameters need to be aligned by 1 megabyte. 2171 2172 21734.67 KVM_S390_VCPU_FAULT 2174------------------------ 2175 2176:Capability: KVM_CAP_S390_UCONTROL 2177:Architectures: s390 2178:Type: vcpu ioctl 2179:Parameters: vcpu absolute address (in) 2180:Returns: 0 in case of success 2181 2182This call creates a page table entry on the virtual cpu's address space 2183(for user controlled virtual machines) or the virtual machine's address 2184space (for regular virtual machines). This only works for minor faults, 2185thus it's recommended to access subject memory page via the user page 2186table upfront. This is useful to handle validity intercepts for user 2187controlled virtual machines to fault in the virtual cpu's lowcore pages 2188prior to calling the KVM_RUN ioctl. 2189 2190 21914.68 KVM_SET_ONE_REG 2192-------------------- 2193 2194:Capability: KVM_CAP_ONE_REG 2195:Architectures: all 2196:Type: vcpu ioctl 2197:Parameters: struct kvm_one_reg (in) 2198:Returns: 0 on success, negative value on failure 2199 2200Errors: 2201 2202 ====== ============================================================ 2203 ENOENT no such register 2204 EINVAL invalid register ID, or no such register or used with VMs in 2205 protected virtualization mode on s390 2206 EPERM (arm64) register access not allowed before vcpu finalization 2207 ====== ============================================================ 2208 2209(These error codes are indicative only: do not rely on a specific error 2210code being returned in a specific situation.) 2211 2212:: 2213 2214 struct kvm_one_reg { 2215 __u64 id; 2216 __u64 addr; 2217 }; 2218 2219Using this ioctl, a single vcpu register can be set to a specific value 2220defined by user space with the passed in struct kvm_one_reg, where id 2221refers to the register identifier as described below and addr is a pointer 2222to a variable with the respective size. There can be architecture agnostic 2223and architecture specific registers. Each have their own range of operation 2224and their own constants and width. To keep track of the implemented 2225registers, find a list below: 2226 2227 ======= =============================== ============ 2228 Arch Register Width (bits) 2229 ======= =============================== ============ 2230 PPC KVM_REG_PPC_HIOR 64 2231 PPC KVM_REG_PPC_IAC1 64 2232 PPC KVM_REG_PPC_IAC2 64 2233 PPC KVM_REG_PPC_IAC3 64 2234 PPC KVM_REG_PPC_IAC4 64 2235 PPC KVM_REG_PPC_DAC1 64 2236 PPC KVM_REG_PPC_DAC2 64 2237 PPC KVM_REG_PPC_DABR 64 2238 PPC KVM_REG_PPC_DSCR 64 2239 PPC KVM_REG_PPC_PURR 64 2240 PPC KVM_REG_PPC_SPURR 64 2241 PPC KVM_REG_PPC_DAR 64 2242 PPC KVM_REG_PPC_DSISR 32 2243 PPC KVM_REG_PPC_AMR 64 2244 PPC KVM_REG_PPC_UAMOR 64 2245 PPC KVM_REG_PPC_MMCR0 64 2246 PPC KVM_REG_PPC_MMCR1 64 2247 PPC KVM_REG_PPC_MMCRA 64 2248 PPC KVM_REG_PPC_MMCR2 64 2249 PPC KVM_REG_PPC_MMCRS 64 2250 PPC KVM_REG_PPC_MMCR3 64 2251 PPC KVM_REG_PPC_SIAR 64 2252 PPC KVM_REG_PPC_SDAR 64 2253 PPC KVM_REG_PPC_SIER 64 2254 PPC KVM_REG_PPC_SIER2 64 2255 PPC KVM_REG_PPC_SIER3 64 2256 PPC KVM_REG_PPC_PMC1 32 2257 PPC KVM_REG_PPC_PMC2 32 2258 PPC KVM_REG_PPC_PMC3 32 2259 PPC KVM_REG_PPC_PMC4 32 2260 PPC KVM_REG_PPC_PMC5 32 2261 PPC KVM_REG_PPC_PMC6 32 2262 PPC KVM_REG_PPC_PMC7 32 2263 PPC KVM_REG_PPC_PMC8 32 2264 PPC KVM_REG_PPC_FPR0 64 2265 ... 2266 PPC KVM_REG_PPC_FPR31 64 2267 PPC KVM_REG_PPC_VR0 128 2268 ... 2269 PPC KVM_REG_PPC_VR31 128 2270 PPC KVM_REG_PPC_VSR0 128 2271 ... 2272 PPC KVM_REG_PPC_VSR31 128 2273 PPC KVM_REG_PPC_FPSCR 64 2274 PPC KVM_REG_PPC_VSCR 32 2275 PPC KVM_REG_PPC_VPA_ADDR 64 2276 PPC KVM_REG_PPC_VPA_SLB 128 2277 PPC KVM_REG_PPC_VPA_DTL 128 2278 PPC KVM_REG_PPC_EPCR 32 2279 PPC KVM_REG_PPC_EPR 32 2280 PPC KVM_REG_PPC_TCR 32 2281 PPC KVM_REG_PPC_TSR 32 2282 PPC KVM_REG_PPC_OR_TSR 32 2283 PPC KVM_REG_PPC_CLEAR_TSR 32 2284 PPC KVM_REG_PPC_MAS0 32 2285 PPC KVM_REG_PPC_MAS1 32 2286 PPC KVM_REG_PPC_MAS2 64 2287 PPC KVM_REG_PPC_MAS7_3 64 2288 PPC KVM_REG_PPC_MAS4 32 2289 PPC KVM_REG_PPC_MAS6 32 2290 PPC KVM_REG_PPC_MMUCFG 32 2291 PPC KVM_REG_PPC_TLB0CFG 32 2292 PPC KVM_REG_PPC_TLB1CFG 32 2293 PPC KVM_REG_PPC_TLB2CFG 32 2294 PPC KVM_REG_PPC_TLB3CFG 32 2295 PPC KVM_REG_PPC_TLB0PS 32 2296 PPC KVM_REG_PPC_TLB1PS 32 2297 PPC KVM_REG_PPC_TLB2PS 32 2298 PPC KVM_REG_PPC_TLB3PS 32 2299 PPC KVM_REG_PPC_EPTCFG 32 2300 PPC KVM_REG_PPC_ICP_STATE 64 2301 PPC KVM_REG_PPC_VP_STATE 128 2302 PPC KVM_REG_PPC_TB_OFFSET 64 2303 PPC KVM_REG_PPC_SPMC1 32 2304 PPC KVM_REG_PPC_SPMC2 32 2305 PPC KVM_REG_PPC_IAMR 64 2306 PPC KVM_REG_PPC_TFHAR 64 2307 PPC KVM_REG_PPC_TFIAR 64 2308 PPC KVM_REG_PPC_TEXASR 64 2309 PPC KVM_REG_PPC_FSCR 64 2310 PPC KVM_REG_PPC_PSPB 32 2311 PPC KVM_REG_PPC_EBBHR 64 2312 PPC KVM_REG_PPC_EBBRR 64 2313 PPC KVM_REG_PPC_BESCR 64 2314 PPC KVM_REG_PPC_TAR 64 2315 PPC KVM_REG_PPC_DPDES 64 2316 PPC KVM_REG_PPC_DAWR 64 2317 PPC KVM_REG_PPC_DAWRX 64 2318 PPC KVM_REG_PPC_CIABR 64 2319 PPC KVM_REG_PPC_IC 64 2320 PPC KVM_REG_PPC_VTB 64 2321 PPC KVM_REG_PPC_CSIGR 64 2322 PPC KVM_REG_PPC_TACR 64 2323 PPC KVM_REG_PPC_TCSCR 64 2324 PPC KVM_REG_PPC_PID 64 2325 PPC KVM_REG_PPC_ACOP 64 2326 PPC KVM_REG_PPC_VRSAVE 32 2327 PPC KVM_REG_PPC_LPCR 32 2328 PPC KVM_REG_PPC_LPCR_64 64 2329 PPC KVM_REG_PPC_PPR 64 2330 PPC KVM_REG_PPC_ARCH_COMPAT 32 2331 PPC KVM_REG_PPC_DABRX 32 2332 PPC KVM_REG_PPC_WORT 64 2333 PPC KVM_REG_PPC_SPRG9 64 2334 PPC KVM_REG_PPC_DBSR 32 2335 PPC KVM_REG_PPC_TIDR 64 2336 PPC KVM_REG_PPC_PSSCR 64 2337 PPC KVM_REG_PPC_DEC_EXPIRY 64 2338 PPC KVM_REG_PPC_PTCR 64 2339 PPC KVM_REG_PPC_DAWR1 64 2340 PPC KVM_REG_PPC_DAWRX1 64 2341 PPC KVM_REG_PPC_TM_GPR0 64 2342 ... 2343 PPC KVM_REG_PPC_TM_GPR31 64 2344 PPC KVM_REG_PPC_TM_VSR0 128 2345 ... 2346 PPC KVM_REG_PPC_TM_VSR63 128 2347 PPC KVM_REG_PPC_TM_CR 64 2348 PPC KVM_REG_PPC_TM_LR 64 2349 PPC KVM_REG_PPC_TM_CTR 64 2350 PPC KVM_REG_PPC_TM_FPSCR 64 2351 PPC KVM_REG_PPC_TM_AMR 64 2352 PPC KVM_REG_PPC_TM_PPR 64 2353 PPC KVM_REG_PPC_TM_VRSAVE 64 2354 PPC KVM_REG_PPC_TM_VSCR 32 2355 PPC KVM_REG_PPC_TM_DSCR 64 2356 PPC KVM_REG_PPC_TM_TAR 64 2357 PPC KVM_REG_PPC_TM_XER 64 2358 2359 MIPS KVM_REG_MIPS_R0 64 2360 ... 2361 MIPS KVM_REG_MIPS_R31 64 2362 MIPS KVM_REG_MIPS_HI 64 2363 MIPS KVM_REG_MIPS_LO 64 2364 MIPS KVM_REG_MIPS_PC 64 2365 MIPS KVM_REG_MIPS_CP0_INDEX 32 2366 MIPS KVM_REG_MIPS_CP0_ENTRYLO0 64 2367 MIPS KVM_REG_MIPS_CP0_ENTRYLO1 64 2368 MIPS KVM_REG_MIPS_CP0_CONTEXT 64 2369 MIPS KVM_REG_MIPS_CP0_CONTEXTCONFIG 32 2370 MIPS KVM_REG_MIPS_CP0_USERLOCAL 64 2371 MIPS KVM_REG_MIPS_CP0_XCONTEXTCONFIG 64 2372 MIPS KVM_REG_MIPS_CP0_PAGEMASK 32 2373 MIPS KVM_REG_MIPS_CP0_PAGEGRAIN 32 2374 MIPS KVM_REG_MIPS_CP0_SEGCTL0 64 2375 MIPS KVM_REG_MIPS_CP0_SEGCTL1 64 2376 MIPS KVM_REG_MIPS_CP0_SEGCTL2 64 2377 MIPS KVM_REG_MIPS_CP0_PWBASE 64 2378 MIPS KVM_REG_MIPS_CP0_PWFIELD 64 2379 MIPS KVM_REG_MIPS_CP0_PWSIZE 64 2380 MIPS KVM_REG_MIPS_CP0_WIRED 32 2381 MIPS KVM_REG_MIPS_CP0_PWCTL 32 2382 MIPS KVM_REG_MIPS_CP0_HWRENA 32 2383 MIPS KVM_REG_MIPS_CP0_BADVADDR 64 2384 MIPS KVM_REG_MIPS_CP0_BADINSTR 32 2385 MIPS KVM_REG_MIPS_CP0_BADINSTRP 32 2386 MIPS KVM_REG_MIPS_CP0_COUNT 32 2387 MIPS KVM_REG_MIPS_CP0_ENTRYHI 64 2388 MIPS KVM_REG_MIPS_CP0_COMPARE 32 2389 MIPS KVM_REG_MIPS_CP0_STATUS 32 2390 MIPS KVM_REG_MIPS_CP0_INTCTL 32 2391 MIPS KVM_REG_MIPS_CP0_CAUSE 32 2392 MIPS KVM_REG_MIPS_CP0_EPC 64 2393 MIPS KVM_REG_MIPS_CP0_PRID 32 2394 MIPS KVM_REG_MIPS_CP0_EBASE 64 2395 MIPS KVM_REG_MIPS_CP0_CONFIG 32 2396 MIPS KVM_REG_MIPS_CP0_CONFIG1 32 2397 MIPS KVM_REG_MIPS_CP0_CONFIG2 32 2398 MIPS KVM_REG_MIPS_CP0_CONFIG3 32 2399 MIPS KVM_REG_MIPS_CP0_CONFIG4 32 2400 MIPS KVM_REG_MIPS_CP0_CONFIG5 32 2401 MIPS KVM_REG_MIPS_CP0_CONFIG7 32 2402 MIPS KVM_REG_MIPS_CP0_XCONTEXT 64 2403 MIPS KVM_REG_MIPS_CP0_ERROREPC 64 2404 MIPS KVM_REG_MIPS_CP0_KSCRATCH1 64 2405 MIPS KVM_REG_MIPS_CP0_KSCRATCH2 64 2406 MIPS KVM_REG_MIPS_CP0_KSCRATCH3 64 2407 MIPS KVM_REG_MIPS_CP0_KSCRATCH4 64 2408 MIPS KVM_REG_MIPS_CP0_KSCRATCH5 64 2409 MIPS KVM_REG_MIPS_CP0_KSCRATCH6 64 2410 MIPS KVM_REG_MIPS_CP0_MAAR(0..63) 64 2411 MIPS KVM_REG_MIPS_COUNT_CTL 64 2412 MIPS KVM_REG_MIPS_COUNT_RESUME 64 2413 MIPS KVM_REG_MIPS_COUNT_HZ 64 2414 MIPS KVM_REG_MIPS_FPR_32(0..31) 32 2415 MIPS KVM_REG_MIPS_FPR_64(0..31) 64 2416 MIPS KVM_REG_MIPS_VEC_128(0..31) 128 2417 MIPS KVM_REG_MIPS_FCR_IR 32 2418 MIPS KVM_REG_MIPS_FCR_CSR 32 2419 MIPS KVM_REG_MIPS_MSA_IR 32 2420 MIPS KVM_REG_MIPS_MSA_CSR 32 2421 ======= =============================== ============ 2422 2423ARM registers are mapped using the lower 32 bits. The upper 16 of that 2424is the register group type, or coprocessor number: 2425 2426ARM core registers have the following id bit patterns:: 2427 2428 0x4020 0000 0010 <index into the kvm_regs struct:16> 2429 2430ARM 32-bit CP15 registers have the following id bit patterns:: 2431 2432 0x4020 0000 000F <zero:1> <crn:4> <crm:4> <opc1:4> <opc2:3> 2433 2434ARM 64-bit CP15 registers have the following id bit patterns:: 2435 2436 0x4030 0000 000F <zero:1> <zero:4> <crm:4> <opc1:4> <zero:3> 2437 2438ARM CCSIDR registers are demultiplexed by CSSELR value:: 2439 2440 0x4020 0000 0011 00 <csselr:8> 2441 2442ARM 32-bit VFP control registers have the following id bit patterns:: 2443 2444 0x4020 0000 0012 1 <regno:12> 2445 2446ARM 64-bit FP registers have the following id bit patterns:: 2447 2448 0x4030 0000 0012 0 <regno:12> 2449 2450ARM firmware pseudo-registers have the following bit pattern:: 2451 2452 0x4030 0000 0014 <regno:16> 2453 2454 2455arm64 registers are mapped using the lower 32 bits. The upper 16 of 2456that is the register group type, or coprocessor number: 2457 2458arm64 core/FP-SIMD registers have the following id bit patterns. Note 2459that the size of the access is variable, as the kvm_regs structure 2460contains elements ranging from 32 to 128 bits. The index is a 32bit 2461value in the kvm_regs structure seen as a 32bit array:: 2462 2463 0x60x0 0000 0010 <index into the kvm_regs struct:16> 2464 2465Specifically: 2466 2467======================= ========= ===== ======================================= 2468 Encoding Register Bits kvm_regs member 2469======================= ========= ===== ======================================= 2470 0x6030 0000 0010 0000 X0 64 regs.regs[0] 2471 0x6030 0000 0010 0002 X1 64 regs.regs[1] 2472 ... 2473 0x6030 0000 0010 003c X30 64 regs.regs[30] 2474 0x6030 0000 0010 003e SP 64 regs.sp 2475 0x6030 0000 0010 0040 PC 64 regs.pc 2476 0x6030 0000 0010 0042 PSTATE 64 regs.pstate 2477 0x6030 0000 0010 0044 SP_EL1 64 sp_el1 2478 0x6030 0000 0010 0046 ELR_EL1 64 elr_el1 2479 0x6030 0000 0010 0048 SPSR_EL1 64 spsr[KVM_SPSR_EL1] (alias SPSR_SVC) 2480 0x6030 0000 0010 004a SPSR_ABT 64 spsr[KVM_SPSR_ABT] 2481 0x6030 0000 0010 004c SPSR_UND 64 spsr[KVM_SPSR_UND] 2482 0x6030 0000 0010 004e SPSR_IRQ 64 spsr[KVM_SPSR_IRQ] 2483 0x6060 0000 0010 0050 SPSR_FIQ 64 spsr[KVM_SPSR_FIQ] 2484 0x6040 0000 0010 0054 V0 128 fp_regs.vregs[0] [1]_ 2485 0x6040 0000 0010 0058 V1 128 fp_regs.vregs[1] [1]_ 2486 ... 2487 0x6040 0000 0010 00d0 V31 128 fp_regs.vregs[31] [1]_ 2488 0x6020 0000 0010 00d4 FPSR 32 fp_regs.fpsr 2489 0x6020 0000 0010 00d5 FPCR 32 fp_regs.fpcr 2490======================= ========= ===== ======================================= 2491 2492.. [1] These encodings are not accepted for SVE-enabled vcpus. See 2493 KVM_ARM_VCPU_INIT. 2494 2495 The equivalent register content can be accessed via bits [127:0] of 2496 the corresponding SVE Zn registers instead for vcpus that have SVE 2497 enabled (see below). 2498 2499arm64 CCSIDR registers are demultiplexed by CSSELR value:: 2500 2501 0x6020 0000 0011 00 <csselr:8> 2502 2503arm64 system registers have the following id bit patterns:: 2504 2505 0x6030 0000 0013 <op0:2> <op1:3> <crn:4> <crm:4> <op2:3> 2506 2507.. warning:: 2508 2509 Two system register IDs do not follow the specified pattern. These 2510 are KVM_REG_ARM_TIMER_CVAL and KVM_REG_ARM_TIMER_CNT, which map to 2511 system registers CNTV_CVAL_EL0 and CNTVCT_EL0 respectively. These 2512 two had their values accidentally swapped, which means TIMER_CVAL is 2513 derived from the register encoding for CNTVCT_EL0 and TIMER_CNT is 2514 derived from the register encoding for CNTV_CVAL_EL0. As this is 2515 API, it must remain this way. 2516 2517arm64 firmware pseudo-registers have the following bit pattern:: 2518 2519 0x6030 0000 0014 <regno:16> 2520 2521arm64 SVE registers have the following bit patterns:: 2522 2523 0x6080 0000 0015 00 <n:5> <slice:5> Zn bits[2048*slice + 2047 : 2048*slice] 2524 0x6050 0000 0015 04 <n:4> <slice:5> Pn bits[256*slice + 255 : 256*slice] 2525 0x6050 0000 0015 060 <slice:5> FFR bits[256*slice + 255 : 256*slice] 2526 0x6060 0000 0015 ffff KVM_REG_ARM64_SVE_VLS pseudo-register 2527 2528Access to register IDs where 2048 * slice >= 128 * max_vq will fail with 2529ENOENT. max_vq is the vcpu's maximum supported vector length in 128-bit 2530quadwords: see [2]_ below. 2531 2532These registers are only accessible on vcpus for which SVE is enabled. 2533See KVM_ARM_VCPU_INIT for details. 2534 2535In addition, except for KVM_REG_ARM64_SVE_VLS, these registers are not 2536accessible until the vcpu's SVE configuration has been finalized 2537using KVM_ARM_VCPU_FINALIZE(KVM_ARM_VCPU_SVE). See KVM_ARM_VCPU_INIT 2538and KVM_ARM_VCPU_FINALIZE for more information about this procedure. 2539 2540KVM_REG_ARM64_SVE_VLS is a pseudo-register that allows the set of vector 2541lengths supported by the vcpu to be discovered and configured by 2542userspace. When transferred to or from user memory via KVM_GET_ONE_REG 2543or KVM_SET_ONE_REG, the value of this register is of type 2544__u64[KVM_ARM64_SVE_VLS_WORDS], and encodes the set of vector lengths as 2545follows:: 2546 2547 __u64 vector_lengths[KVM_ARM64_SVE_VLS_WORDS]; 2548 2549 if (vq >= SVE_VQ_MIN && vq <= SVE_VQ_MAX && 2550 ((vector_lengths[(vq - KVM_ARM64_SVE_VQ_MIN) / 64] >> 2551 ((vq - KVM_ARM64_SVE_VQ_MIN) % 64)) & 1)) 2552 /* Vector length vq * 16 bytes supported */ 2553 else 2554 /* Vector length vq * 16 bytes not supported */ 2555 2556.. [2] The maximum value vq for which the above condition is true is 2557 max_vq. This is the maximum vector length available to the guest on 2558 this vcpu, and determines which register slices are visible through 2559 this ioctl interface. 2560 2561(See Documentation/arm64/sve.rst for an explanation of the "vq" 2562nomenclature.) 2563 2564KVM_REG_ARM64_SVE_VLS is only accessible after KVM_ARM_VCPU_INIT. 2565KVM_ARM_VCPU_INIT initialises it to the best set of vector lengths that 2566the host supports. 2567 2568Userspace may subsequently modify it if desired until the vcpu's SVE 2569configuration is finalized using KVM_ARM_VCPU_FINALIZE(KVM_ARM_VCPU_SVE). 2570 2571Apart from simply removing all vector lengths from the host set that 2572exceed some value, support for arbitrarily chosen sets of vector lengths 2573is hardware-dependent and may not be available. Attempting to configure 2574an invalid set of vector lengths via KVM_SET_ONE_REG will fail with 2575EINVAL. 2576 2577After the vcpu's SVE configuration is finalized, further attempts to 2578write this register will fail with EPERM. 2579 2580 2581MIPS registers are mapped using the lower 32 bits. The upper 16 of that is 2582the register group type: 2583 2584MIPS core registers (see above) have the following id bit patterns:: 2585 2586 0x7030 0000 0000 <reg:16> 2587 2588MIPS CP0 registers (see KVM_REG_MIPS_CP0_* above) have the following id bit 2589patterns depending on whether they're 32-bit or 64-bit registers:: 2590 2591 0x7020 0000 0001 00 <reg:5> <sel:3> (32-bit) 2592 0x7030 0000 0001 00 <reg:5> <sel:3> (64-bit) 2593 2594Note: KVM_REG_MIPS_CP0_ENTRYLO0 and KVM_REG_MIPS_CP0_ENTRYLO1 are the MIPS64 2595versions of the EntryLo registers regardless of the word size of the host 2596hardware, host kernel, guest, and whether XPA is present in the guest, i.e. 2597with the RI and XI bits (if they exist) in bits 63 and 62 respectively, and 2598the PFNX field starting at bit 30. 2599 2600MIPS MAARs (see KVM_REG_MIPS_CP0_MAAR(*) above) have the following id bit 2601patterns:: 2602 2603 0x7030 0000 0001 01 <reg:8> 2604 2605MIPS KVM control registers (see above) have the following id bit patterns:: 2606 2607 0x7030 0000 0002 <reg:16> 2608 2609MIPS FPU registers (see KVM_REG_MIPS_FPR_{32,64}() above) have the following 2610id bit patterns depending on the size of the register being accessed. They are 2611always accessed according to the current guest FPU mode (Status.FR and 2612Config5.FRE), i.e. as the guest would see them, and they become unpredictable 2613if the guest FPU mode is changed. MIPS SIMD Architecture (MSA) vector 2614registers (see KVM_REG_MIPS_VEC_128() above) have similar patterns as they 2615overlap the FPU registers:: 2616 2617 0x7020 0000 0003 00 <0:3> <reg:5> (32-bit FPU registers) 2618 0x7030 0000 0003 00 <0:3> <reg:5> (64-bit FPU registers) 2619 0x7040 0000 0003 00 <0:3> <reg:5> (128-bit MSA vector registers) 2620 2621MIPS FPU control registers (see KVM_REG_MIPS_FCR_{IR,CSR} above) have the 2622following id bit patterns:: 2623 2624 0x7020 0000 0003 01 <0:3> <reg:5> 2625 2626MIPS MSA control registers (see KVM_REG_MIPS_MSA_{IR,CSR} above) have the 2627following id bit patterns:: 2628 2629 0x7020 0000 0003 02 <0:3> <reg:5> 2630 2631RISC-V registers are mapped using the lower 32 bits. The upper 8 bits of 2632that is the register group type. 2633 2634RISC-V config registers are meant for configuring a Guest VCPU and it has 2635the following id bit patterns:: 2636 2637 0x8020 0000 01 <index into the kvm_riscv_config struct:24> (32bit Host) 2638 0x8030 0000 01 <index into the kvm_riscv_config struct:24> (64bit Host) 2639 2640Following are the RISC-V config registers: 2641 2642======================= ========= ============================================= 2643 Encoding Register Description 2644======================= ========= ============================================= 2645 0x80x0 0000 0100 0000 isa ISA feature bitmap of Guest VCPU 2646======================= ========= ============================================= 2647 2648The isa config register can be read anytime but can only be written before 2649a Guest VCPU runs. It will have ISA feature bits matching underlying host 2650set by default. 2651 2652RISC-V core registers represent the general excution state of a Guest VCPU 2653and it has the following id bit patterns:: 2654 2655 0x8020 0000 02 <index into the kvm_riscv_core struct:24> (32bit Host) 2656 0x8030 0000 02 <index into the kvm_riscv_core struct:24> (64bit Host) 2657 2658Following are the RISC-V core registers: 2659 2660======================= ========= ============================================= 2661 Encoding Register Description 2662======================= ========= ============================================= 2663 0x80x0 0000 0200 0000 regs.pc Program counter 2664 0x80x0 0000 0200 0001 regs.ra Return address 2665 0x80x0 0000 0200 0002 regs.sp Stack pointer 2666 0x80x0 0000 0200 0003 regs.gp Global pointer 2667 0x80x0 0000 0200 0004 regs.tp Task pointer 2668 0x80x0 0000 0200 0005 regs.t0 Caller saved register 0 2669 0x80x0 0000 0200 0006 regs.t1 Caller saved register 1 2670 0x80x0 0000 0200 0007 regs.t2 Caller saved register 2 2671 0x80x0 0000 0200 0008 regs.s0 Callee saved register 0 2672 0x80x0 0000 0200 0009 regs.s1 Callee saved register 1 2673 0x80x0 0000 0200 000a regs.a0 Function argument (or return value) 0 2674 0x80x0 0000 0200 000b regs.a1 Function argument (or return value) 1 2675 0x80x0 0000 0200 000c regs.a2 Function argument 2 2676 0x80x0 0000 0200 000d regs.a3 Function argument 3 2677 0x80x0 0000 0200 000e regs.a4 Function argument 4 2678 0x80x0 0000 0200 000f regs.a5 Function argument 5 2679 0x80x0 0000 0200 0010 regs.a6 Function argument 6 2680 0x80x0 0000 0200 0011 regs.a7 Function argument 7 2681 0x80x0 0000 0200 0012 regs.s2 Callee saved register 2 2682 0x80x0 0000 0200 0013 regs.s3 Callee saved register 3 2683 0x80x0 0000 0200 0014 regs.s4 Callee saved register 4 2684 0x80x0 0000 0200 0015 regs.s5 Callee saved register 5 2685 0x80x0 0000 0200 0016 regs.s6 Callee saved register 6 2686 0x80x0 0000 0200 0017 regs.s7 Callee saved register 7 2687 0x80x0 0000 0200 0018 regs.s8 Callee saved register 8 2688 0x80x0 0000 0200 0019 regs.s9 Callee saved register 9 2689 0x80x0 0000 0200 001a regs.s10 Callee saved register 10 2690 0x80x0 0000 0200 001b regs.s11 Callee saved register 11 2691 0x80x0 0000 0200 001c regs.t3 Caller saved register 3 2692 0x80x0 0000 0200 001d regs.t4 Caller saved register 4 2693 0x80x0 0000 0200 001e regs.t5 Caller saved register 5 2694 0x80x0 0000 0200 001f regs.t6 Caller saved register 6 2695 0x80x0 0000 0200 0020 mode Privilege mode (1 = S-mode or 0 = U-mode) 2696======================= ========= ============================================= 2697 2698RISC-V csr registers represent the supervisor mode control/status registers 2699of a Guest VCPU and it has the following id bit patterns:: 2700 2701 0x8020 0000 03 <index into the kvm_riscv_csr struct:24> (32bit Host) 2702 0x8030 0000 03 <index into the kvm_riscv_csr struct:24> (64bit Host) 2703 2704Following are the RISC-V csr registers: 2705 2706======================= ========= ============================================= 2707 Encoding Register Description 2708======================= ========= ============================================= 2709 0x80x0 0000 0300 0000 sstatus Supervisor status 2710 0x80x0 0000 0300 0001 sie Supervisor interrupt enable 2711 0x80x0 0000 0300 0002 stvec Supervisor trap vector base 2712 0x80x0 0000 0300 0003 sscratch Supervisor scratch register 2713 0x80x0 0000 0300 0004 sepc Supervisor exception program counter 2714 0x80x0 0000 0300 0005 scause Supervisor trap cause 2715 0x80x0 0000 0300 0006 stval Supervisor bad address or instruction 2716 0x80x0 0000 0300 0007 sip Supervisor interrupt pending 2717 0x80x0 0000 0300 0008 satp Supervisor address translation and protection 2718======================= ========= ============================================= 2719 2720RISC-V timer registers represent the timer state of a Guest VCPU and it has 2721the following id bit patterns:: 2722 2723 0x8030 0000 04 <index into the kvm_riscv_timer struct:24> 2724 2725Following are the RISC-V timer registers: 2726 2727======================= ========= ============================================= 2728 Encoding Register Description 2729======================= ========= ============================================= 2730 0x8030 0000 0400 0000 frequency Time base frequency (read-only) 2731 0x8030 0000 0400 0001 time Time value visible to Guest 2732 0x8030 0000 0400 0002 compare Time compare programmed by Guest 2733 0x8030 0000 0400 0003 state Time compare state (1 = ON or 0 = OFF) 2734======================= ========= ============================================= 2735 2736RISC-V F-extension registers represent the single precision floating point 2737state of a Guest VCPU and it has the following id bit patterns:: 2738 2739 0x8020 0000 05 <index into the __riscv_f_ext_state struct:24> 2740 2741Following are the RISC-V F-extension registers: 2742 2743======================= ========= ============================================= 2744 Encoding Register Description 2745======================= ========= ============================================= 2746 0x8020 0000 0500 0000 f[0] Floating point register 0 2747 ... 2748 0x8020 0000 0500 001f f[31] Floating point register 31 2749 0x8020 0000 0500 0020 fcsr Floating point control and status register 2750======================= ========= ============================================= 2751 2752RISC-V D-extension registers represent the double precision floating point 2753state of a Guest VCPU and it has the following id bit patterns:: 2754 2755 0x8020 0000 06 <index into the __riscv_d_ext_state struct:24> (fcsr) 2756 0x8030 0000 06 <index into the __riscv_d_ext_state struct:24> (non-fcsr) 2757 2758Following are the RISC-V D-extension registers: 2759 2760======================= ========= ============================================= 2761 Encoding Register Description 2762======================= ========= ============================================= 2763 0x8030 0000 0600 0000 f[0] Floating point register 0 2764 ... 2765 0x8030 0000 0600 001f f[31] Floating point register 31 2766 0x8020 0000 0600 0020 fcsr Floating point control and status register 2767======================= ========= ============================================= 2768 2769 27704.69 KVM_GET_ONE_REG 2771-------------------- 2772 2773:Capability: KVM_CAP_ONE_REG 2774:Architectures: all 2775:Type: vcpu ioctl 2776:Parameters: struct kvm_one_reg (in and out) 2777:Returns: 0 on success, negative value on failure 2778 2779Errors include: 2780 2781 ======== ============================================================ 2782 ENOENT no such register 2783 EINVAL invalid register ID, or no such register or used with VMs in 2784 protected virtualization mode on s390 2785 EPERM (arm64) register access not allowed before vcpu finalization 2786 ======== ============================================================ 2787 2788(These error codes are indicative only: do not rely on a specific error 2789code being returned in a specific situation.) 2790 2791This ioctl allows to receive the value of a single register implemented 2792in a vcpu. The register to read is indicated by the "id" field of the 2793kvm_one_reg struct passed in. On success, the register value can be found 2794at the memory location pointed to by "addr". 2795 2796The list of registers accessible using this interface is identical to the 2797list in 4.68. 2798 2799 28004.70 KVM_KVMCLOCK_CTRL 2801---------------------- 2802 2803:Capability: KVM_CAP_KVMCLOCK_CTRL 2804:Architectures: Any that implement pvclocks (currently x86 only) 2805:Type: vcpu ioctl 2806:Parameters: None 2807:Returns: 0 on success, -1 on error 2808 2809This ioctl sets a flag accessible to the guest indicating that the specified 2810vCPU has been paused by the host userspace. 2811 2812The host will set a flag in the pvclock structure that is checked from the 2813soft lockup watchdog. The flag is part of the pvclock structure that is 2814shared between guest and host, specifically the second bit of the flags 2815field of the pvclock_vcpu_time_info structure. It will be set exclusively by 2816the host and read/cleared exclusively by the guest. The guest operation of 2817checking and clearing the flag must be an atomic operation so 2818load-link/store-conditional, or equivalent must be used. There are two cases 2819where the guest will clear the flag: when the soft lockup watchdog timer resets 2820itself or when a soft lockup is detected. This ioctl can be called any time 2821after pausing the vcpu, but before it is resumed. 2822 2823 28244.71 KVM_SIGNAL_MSI 2825------------------- 2826 2827:Capability: KVM_CAP_SIGNAL_MSI 2828:Architectures: x86 arm arm64 2829:Type: vm ioctl 2830:Parameters: struct kvm_msi (in) 2831:Returns: >0 on delivery, 0 if guest blocked the MSI, and -1 on error 2832 2833Directly inject a MSI message. Only valid with in-kernel irqchip that handles 2834MSI messages. 2835 2836:: 2837 2838 struct kvm_msi { 2839 __u32 address_lo; 2840 __u32 address_hi; 2841 __u32 data; 2842 __u32 flags; 2843 __u32 devid; 2844 __u8 pad[12]; 2845 }; 2846 2847flags: 2848 KVM_MSI_VALID_DEVID: devid contains a valid value. The per-VM 2849 KVM_CAP_MSI_DEVID capability advertises the requirement to provide 2850 the device ID. If this capability is not available, userspace 2851 should never set the KVM_MSI_VALID_DEVID flag as the ioctl might fail. 2852 2853If KVM_MSI_VALID_DEVID is set, devid contains a unique device identifier 2854for the device that wrote the MSI message. For PCI, this is usually a 2855BFD identifier in the lower 16 bits. 2856 2857On x86, address_hi is ignored unless the KVM_X2APIC_API_USE_32BIT_IDS 2858feature of KVM_CAP_X2APIC_API capability is enabled. If it is enabled, 2859address_hi bits 31-8 provide bits 31-8 of the destination id. Bits 7-0 of 2860address_hi must be zero. 2861 2862 28634.71 KVM_CREATE_PIT2 2864-------------------- 2865 2866:Capability: KVM_CAP_PIT2 2867:Architectures: x86 2868:Type: vm ioctl 2869:Parameters: struct kvm_pit_config (in) 2870:Returns: 0 on success, -1 on error 2871 2872Creates an in-kernel device model for the i8254 PIT. This call is only valid 2873after enabling in-kernel irqchip support via KVM_CREATE_IRQCHIP. The following 2874parameters have to be passed:: 2875 2876 struct kvm_pit_config { 2877 __u32 flags; 2878 __u32 pad[15]; 2879 }; 2880 2881Valid flags are:: 2882 2883 #define KVM_PIT_SPEAKER_DUMMY 1 /* emulate speaker port stub */ 2884 2885PIT timer interrupts may use a per-VM kernel thread for injection. If it 2886exists, this thread will have a name of the following pattern:: 2887 2888 kvm-pit/<owner-process-pid> 2889 2890When running a guest with elevated priorities, the scheduling parameters of 2891this thread may have to be adjusted accordingly. 2892 2893This IOCTL replaces the obsolete KVM_CREATE_PIT. 2894 2895 28964.72 KVM_GET_PIT2 2897----------------- 2898 2899:Capability: KVM_CAP_PIT_STATE2 2900:Architectures: x86 2901:Type: vm ioctl 2902:Parameters: struct kvm_pit_state2 (out) 2903:Returns: 0 on success, -1 on error 2904 2905Retrieves the state of the in-kernel PIT model. Only valid after 2906KVM_CREATE_PIT2. The state is returned in the following structure:: 2907 2908 struct kvm_pit_state2 { 2909 struct kvm_pit_channel_state channels[3]; 2910 __u32 flags; 2911 __u32 reserved[9]; 2912 }; 2913 2914Valid flags are:: 2915 2916 /* disable PIT in HPET legacy mode */ 2917 #define KVM_PIT_FLAGS_HPET_LEGACY 0x00000001 2918 2919This IOCTL replaces the obsolete KVM_GET_PIT. 2920 2921 29224.73 KVM_SET_PIT2 2923----------------- 2924 2925:Capability: KVM_CAP_PIT_STATE2 2926:Architectures: x86 2927:Type: vm ioctl 2928:Parameters: struct kvm_pit_state2 (in) 2929:Returns: 0 on success, -1 on error 2930 2931Sets the state of the in-kernel PIT model. Only valid after KVM_CREATE_PIT2. 2932See KVM_GET_PIT2 for details on struct kvm_pit_state2. 2933 2934This IOCTL replaces the obsolete KVM_SET_PIT. 2935 2936 29374.74 KVM_PPC_GET_SMMU_INFO 2938-------------------------- 2939 2940:Capability: KVM_CAP_PPC_GET_SMMU_INFO 2941:Architectures: powerpc 2942:Type: vm ioctl 2943:Parameters: None 2944:Returns: 0 on success, -1 on error 2945 2946This populates and returns a structure describing the features of 2947the "Server" class MMU emulation supported by KVM. 2948This can in turn be used by userspace to generate the appropriate 2949device-tree properties for the guest operating system. 2950 2951The structure contains some global information, followed by an 2952array of supported segment page sizes:: 2953 2954 struct kvm_ppc_smmu_info { 2955 __u64 flags; 2956 __u32 slb_size; 2957 __u32 pad; 2958 struct kvm_ppc_one_seg_page_size sps[KVM_PPC_PAGE_SIZES_MAX_SZ]; 2959 }; 2960 2961The supported flags are: 2962 2963 - KVM_PPC_PAGE_SIZES_REAL: 2964 When that flag is set, guest page sizes must "fit" the backing 2965 store page sizes. When not set, any page size in the list can 2966 be used regardless of how they are backed by userspace. 2967 2968 - KVM_PPC_1T_SEGMENTS 2969 The emulated MMU supports 1T segments in addition to the 2970 standard 256M ones. 2971 2972 - KVM_PPC_NO_HASH 2973 This flag indicates that HPT guests are not supported by KVM, 2974 thus all guests must use radix MMU mode. 2975 2976The "slb_size" field indicates how many SLB entries are supported 2977 2978The "sps" array contains 8 entries indicating the supported base 2979page sizes for a segment in increasing order. Each entry is defined 2980as follow:: 2981 2982 struct kvm_ppc_one_seg_page_size { 2983 __u32 page_shift; /* Base page shift of segment (or 0) */ 2984 __u32 slb_enc; /* SLB encoding for BookS */ 2985 struct kvm_ppc_one_page_size enc[KVM_PPC_PAGE_SIZES_MAX_SZ]; 2986 }; 2987 2988An entry with a "page_shift" of 0 is unused. Because the array is 2989organized in increasing order, a lookup can stop when encoutering 2990such an entry. 2991 2992The "slb_enc" field provides the encoding to use in the SLB for the 2993page size. The bits are in positions such as the value can directly 2994be OR'ed into the "vsid" argument of the slbmte instruction. 2995 2996The "enc" array is a list which for each of those segment base page 2997size provides the list of supported actual page sizes (which can be 2998only larger or equal to the base page size), along with the 2999corresponding encoding in the hash PTE. Similarly, the array is 30008 entries sorted by increasing sizes and an entry with a "0" shift 3001is an empty entry and a terminator:: 3002 3003 struct kvm_ppc_one_page_size { 3004 __u32 page_shift; /* Page shift (or 0) */ 3005 __u32 pte_enc; /* Encoding in the HPTE (>>12) */ 3006 }; 3007 3008The "pte_enc" field provides a value that can OR'ed into the hash 3009PTE's RPN field (ie, it needs to be shifted left by 12 to OR it 3010into the hash PTE second double word). 3011 30124.75 KVM_IRQFD 3013-------------- 3014 3015:Capability: KVM_CAP_IRQFD 3016:Architectures: x86 s390 arm arm64 3017:Type: vm ioctl 3018:Parameters: struct kvm_irqfd (in) 3019:Returns: 0 on success, -1 on error 3020 3021Allows setting an eventfd to directly trigger a guest interrupt. 3022kvm_irqfd.fd specifies the file descriptor to use as the eventfd and 3023kvm_irqfd.gsi specifies the irqchip pin toggled by this event. When 3024an event is triggered on the eventfd, an interrupt is injected into 3025the guest using the specified gsi pin. The irqfd is removed using 3026the KVM_IRQFD_FLAG_DEASSIGN flag, specifying both kvm_irqfd.fd 3027and kvm_irqfd.gsi. 3028 3029With KVM_CAP_IRQFD_RESAMPLE, KVM_IRQFD supports a de-assert and notify 3030mechanism allowing emulation of level-triggered, irqfd-based 3031interrupts. When KVM_IRQFD_FLAG_RESAMPLE is set the user must pass an 3032additional eventfd in the kvm_irqfd.resamplefd field. When operating 3033in resample mode, posting of an interrupt through kvm_irq.fd asserts 3034the specified gsi in the irqchip. When the irqchip is resampled, such 3035as from an EOI, the gsi is de-asserted and the user is notified via 3036kvm_irqfd.resamplefd. It is the user's responsibility to re-queue 3037the interrupt if the device making use of it still requires service. 3038Note that closing the resamplefd is not sufficient to disable the 3039irqfd. The KVM_IRQFD_FLAG_RESAMPLE is only necessary on assignment 3040and need not be specified with KVM_IRQFD_FLAG_DEASSIGN. 3041 3042On arm/arm64, gsi routing being supported, the following can happen: 3043 3044- in case no routing entry is associated to this gsi, injection fails 3045- in case the gsi is associated to an irqchip routing entry, 3046 irqchip.pin + 32 corresponds to the injected SPI ID. 3047- in case the gsi is associated to an MSI routing entry, the MSI 3048 message and device ID are translated into an LPI (support restricted 3049 to GICv3 ITS in-kernel emulation). 3050 30514.76 KVM_PPC_ALLOCATE_HTAB 3052-------------------------- 3053 3054:Capability: KVM_CAP_PPC_ALLOC_HTAB 3055:Architectures: powerpc 3056:Type: vm ioctl 3057:Parameters: Pointer to u32 containing hash table order (in/out) 3058:Returns: 0 on success, -1 on error 3059 3060This requests the host kernel to allocate an MMU hash table for a 3061guest using the PAPR paravirtualization interface. This only does 3062anything if the kernel is configured to use the Book 3S HV style of 3063virtualization. Otherwise the capability doesn't exist and the ioctl 3064returns an ENOTTY error. The rest of this description assumes Book 3S 3065HV. 3066 3067There must be no vcpus running when this ioctl is called; if there 3068are, it will do nothing and return an EBUSY error. 3069 3070The parameter is a pointer to a 32-bit unsigned integer variable 3071containing the order (log base 2) of the desired size of the hash 3072table, which must be between 18 and 46. On successful return from the 3073ioctl, the value will not be changed by the kernel. 3074 3075If no hash table has been allocated when any vcpu is asked to run 3076(with the KVM_RUN ioctl), the host kernel will allocate a 3077default-sized hash table (16 MB). 3078 3079If this ioctl is called when a hash table has already been allocated, 3080with a different order from the existing hash table, the existing hash 3081table will be freed and a new one allocated. If this is ioctl is 3082called when a hash table has already been allocated of the same order 3083as specified, the kernel will clear out the existing hash table (zero 3084all HPTEs). In either case, if the guest is using the virtualized 3085real-mode area (VRMA) facility, the kernel will re-create the VMRA 3086HPTEs on the next KVM_RUN of any vcpu. 3087 30884.77 KVM_S390_INTERRUPT 3089----------------------- 3090 3091:Capability: basic 3092:Architectures: s390 3093:Type: vm ioctl, vcpu ioctl 3094:Parameters: struct kvm_s390_interrupt (in) 3095:Returns: 0 on success, -1 on error 3096 3097Allows to inject an interrupt to the guest. Interrupts can be floating 3098(vm ioctl) or per cpu (vcpu ioctl), depending on the interrupt type. 3099 3100Interrupt parameters are passed via kvm_s390_interrupt:: 3101 3102 struct kvm_s390_interrupt { 3103 __u32 type; 3104 __u32 parm; 3105 __u64 parm64; 3106 }; 3107 3108type can be one of the following: 3109 3110KVM_S390_SIGP_STOP (vcpu) 3111 - sigp stop; optional flags in parm 3112KVM_S390_PROGRAM_INT (vcpu) 3113 - program check; code in parm 3114KVM_S390_SIGP_SET_PREFIX (vcpu) 3115 - sigp set prefix; prefix address in parm 3116KVM_S390_RESTART (vcpu) 3117 - restart 3118KVM_S390_INT_CLOCK_COMP (vcpu) 3119 - clock comparator interrupt 3120KVM_S390_INT_CPU_TIMER (vcpu) 3121 - CPU timer interrupt 3122KVM_S390_INT_VIRTIO (vm) 3123 - virtio external interrupt; external interrupt 3124 parameters in parm and parm64 3125KVM_S390_INT_SERVICE (vm) 3126 - sclp external interrupt; sclp parameter in parm 3127KVM_S390_INT_EMERGENCY (vcpu) 3128 - sigp emergency; source cpu in parm 3129KVM_S390_INT_EXTERNAL_CALL (vcpu) 3130 - sigp external call; source cpu in parm 3131KVM_S390_INT_IO(ai,cssid,ssid,schid) (vm) 3132 - compound value to indicate an 3133 I/O interrupt (ai - adapter interrupt; cssid,ssid,schid - subchannel); 3134 I/O interruption parameters in parm (subchannel) and parm64 (intparm, 3135 interruption subclass) 3136KVM_S390_MCHK (vm, vcpu) 3137 - machine check interrupt; cr 14 bits in parm, machine check interrupt 3138 code in parm64 (note that machine checks needing further payload are not 3139 supported by this ioctl) 3140 3141This is an asynchronous vcpu ioctl and can be invoked from any thread. 3142 31434.78 KVM_PPC_GET_HTAB_FD 3144------------------------ 3145 3146:Capability: KVM_CAP_PPC_HTAB_FD 3147:Architectures: powerpc 3148:Type: vm ioctl 3149:Parameters: Pointer to struct kvm_get_htab_fd (in) 3150:Returns: file descriptor number (>= 0) on success, -1 on error 3151 3152This returns a file descriptor that can be used either to read out the 3153entries in the guest's hashed page table (HPT), or to write entries to 3154initialize the HPT. The returned fd can only be written to if the 3155KVM_GET_HTAB_WRITE bit is set in the flags field of the argument, and 3156can only be read if that bit is clear. The argument struct looks like 3157this:: 3158 3159 /* For KVM_PPC_GET_HTAB_FD */ 3160 struct kvm_get_htab_fd { 3161 __u64 flags; 3162 __u64 start_index; 3163 __u64 reserved[2]; 3164 }; 3165 3166 /* Values for kvm_get_htab_fd.flags */ 3167 #define KVM_GET_HTAB_BOLTED_ONLY ((__u64)0x1) 3168 #define KVM_GET_HTAB_WRITE ((__u64)0x2) 3169 3170The 'start_index' field gives the index in the HPT of the entry at 3171which to start reading. It is ignored when writing. 3172 3173Reads on the fd will initially supply information about all 3174"interesting" HPT entries. Interesting entries are those with the 3175bolted bit set, if the KVM_GET_HTAB_BOLTED_ONLY bit is set, otherwise 3176all entries. When the end of the HPT is reached, the read() will 3177return. If read() is called again on the fd, it will start again from 3178the beginning of the HPT, but will only return HPT entries that have 3179changed since they were last read. 3180 3181Data read or written is structured as a header (8 bytes) followed by a 3182series of valid HPT entries (16 bytes) each. The header indicates how 3183many valid HPT entries there are and how many invalid entries follow 3184the valid entries. The invalid entries are not represented explicitly 3185in the stream. The header format is:: 3186 3187 struct kvm_get_htab_header { 3188 __u32 index; 3189 __u16 n_valid; 3190 __u16 n_invalid; 3191 }; 3192 3193Writes to the fd create HPT entries starting at the index given in the 3194header; first 'n_valid' valid entries with contents from the data 3195written, then 'n_invalid' invalid entries, invalidating any previously 3196valid entries found. 3197 31984.79 KVM_CREATE_DEVICE 3199---------------------- 3200 3201:Capability: KVM_CAP_DEVICE_CTRL 3202:Type: vm ioctl 3203:Parameters: struct kvm_create_device (in/out) 3204:Returns: 0 on success, -1 on error 3205 3206Errors: 3207 3208 ====== ======================================================= 3209 ENODEV The device type is unknown or unsupported 3210 EEXIST Device already created, and this type of device may not 3211 be instantiated multiple times 3212 ====== ======================================================= 3213 3214 Other error conditions may be defined by individual device types or 3215 have their standard meanings. 3216 3217Creates an emulated device in the kernel. The file descriptor returned 3218in fd can be used with KVM_SET/GET/HAS_DEVICE_ATTR. 3219 3220If the KVM_CREATE_DEVICE_TEST flag is set, only test whether the 3221device type is supported (not necessarily whether it can be created 3222in the current vm). 3223 3224Individual devices should not define flags. Attributes should be used 3225for specifying any behavior that is not implied by the device type 3226number. 3227 3228:: 3229 3230 struct kvm_create_device { 3231 __u32 type; /* in: KVM_DEV_TYPE_xxx */ 3232 __u32 fd; /* out: device handle */ 3233 __u32 flags; /* in: KVM_CREATE_DEVICE_xxx */ 3234 }; 3235 32364.80 KVM_SET_DEVICE_ATTR/KVM_GET_DEVICE_ATTR 3237-------------------------------------------- 3238 3239:Capability: KVM_CAP_DEVICE_CTRL, KVM_CAP_VM_ATTRIBUTES for vm device, 3240 KVM_CAP_VCPU_ATTRIBUTES for vcpu device 3241:Type: device ioctl, vm ioctl, vcpu ioctl 3242:Parameters: struct kvm_device_attr 3243:Returns: 0 on success, -1 on error 3244 3245Errors: 3246 3247 ===== ============================================================= 3248 ENXIO The group or attribute is unknown/unsupported for this device 3249 or hardware support is missing. 3250 EPERM The attribute cannot (currently) be accessed this way 3251 (e.g. read-only attribute, or attribute that only makes 3252 sense when the device is in a different state) 3253 ===== ============================================================= 3254 3255 Other error conditions may be defined by individual device types. 3256 3257Gets/sets a specified piece of device configuration and/or state. The 3258semantics are device-specific. See individual device documentation in 3259the "devices" directory. As with ONE_REG, the size of the data 3260transferred is defined by the particular attribute. 3261 3262:: 3263 3264 struct kvm_device_attr { 3265 __u32 flags; /* no flags currently defined */ 3266 __u32 group; /* device-defined */ 3267 __u64 attr; /* group-defined */ 3268 __u64 addr; /* userspace address of attr data */ 3269 }; 3270 32714.81 KVM_HAS_DEVICE_ATTR 3272------------------------ 3273 3274:Capability: KVM_CAP_DEVICE_CTRL, KVM_CAP_VM_ATTRIBUTES for vm device, 3275 KVM_CAP_VCPU_ATTRIBUTES for vcpu device 3276:Type: device ioctl, vm ioctl, vcpu ioctl 3277:Parameters: struct kvm_device_attr 3278:Returns: 0 on success, -1 on error 3279 3280Errors: 3281 3282 ===== ============================================================= 3283 ENXIO The group or attribute is unknown/unsupported for this device 3284 or hardware support is missing. 3285 ===== ============================================================= 3286 3287Tests whether a device supports a particular attribute. A successful 3288return indicates the attribute is implemented. It does not necessarily 3289indicate that the attribute can be read or written in the device's 3290current state. "addr" is ignored. 3291 32924.82 KVM_ARM_VCPU_INIT 3293---------------------- 3294 3295:Capability: basic 3296:Architectures: arm, arm64 3297:Type: vcpu ioctl 3298:Parameters: struct kvm_vcpu_init (in) 3299:Returns: 0 on success; -1 on error 3300 3301Errors: 3302 3303 ====== ================================================================= 3304 EINVAL the target is unknown, or the combination of features is invalid. 3305 ENOENT a features bit specified is unknown. 3306 ====== ================================================================= 3307 3308This tells KVM what type of CPU to present to the guest, and what 3309optional features it should have. This will cause a reset of the cpu 3310registers to their initial values. If this is not called, KVM_RUN will 3311return ENOEXEC for that vcpu. 3312 3313The initial values are defined as: 3314 - Processor state: 3315 * AArch64: EL1h, D, A, I and F bits set. All other bits 3316 are cleared. 3317 * AArch32: SVC, A, I and F bits set. All other bits are 3318 cleared. 3319 - General Purpose registers, including PC and SP: set to 0 3320 - FPSIMD/NEON registers: set to 0 3321 - SVE registers: set to 0 3322 - System registers: Reset to their architecturally defined 3323 values as for a warm reset to EL1 (resp. SVC) 3324 3325Note that because some registers reflect machine topology, all vcpus 3326should be created before this ioctl is invoked. 3327 3328Userspace can call this function multiple times for a given vcpu, including 3329after the vcpu has been run. This will reset the vcpu to its initial 3330state. All calls to this function after the initial call must use the same 3331target and same set of feature flags, otherwise EINVAL will be returned. 3332 3333Possible features: 3334 3335 - KVM_ARM_VCPU_POWER_OFF: Starts the CPU in a power-off state. 3336 Depends on KVM_CAP_ARM_PSCI. If not set, the CPU will be powered on 3337 and execute guest code when KVM_RUN is called. 3338 - KVM_ARM_VCPU_EL1_32BIT: Starts the CPU in a 32bit mode. 3339 Depends on KVM_CAP_ARM_EL1_32BIT (arm64 only). 3340 - KVM_ARM_VCPU_PSCI_0_2: Emulate PSCI v0.2 (or a future revision 3341 backward compatible with v0.2) for the CPU. 3342 Depends on KVM_CAP_ARM_PSCI_0_2. 3343 - KVM_ARM_VCPU_PMU_V3: Emulate PMUv3 for the CPU. 3344 Depends on KVM_CAP_ARM_PMU_V3. 3345 3346 - KVM_ARM_VCPU_PTRAUTH_ADDRESS: Enables Address Pointer authentication 3347 for arm64 only. 3348 Depends on KVM_CAP_ARM_PTRAUTH_ADDRESS. 3349 If KVM_CAP_ARM_PTRAUTH_ADDRESS and KVM_CAP_ARM_PTRAUTH_GENERIC are 3350 both present, then both KVM_ARM_VCPU_PTRAUTH_ADDRESS and 3351 KVM_ARM_VCPU_PTRAUTH_GENERIC must be requested or neither must be 3352 requested. 3353 3354 - KVM_ARM_VCPU_PTRAUTH_GENERIC: Enables Generic Pointer authentication 3355 for arm64 only. 3356 Depends on KVM_CAP_ARM_PTRAUTH_GENERIC. 3357 If KVM_CAP_ARM_PTRAUTH_ADDRESS and KVM_CAP_ARM_PTRAUTH_GENERIC are 3358 both present, then both KVM_ARM_VCPU_PTRAUTH_ADDRESS and 3359 KVM_ARM_VCPU_PTRAUTH_GENERIC must be requested or neither must be 3360 requested. 3361 3362 - KVM_ARM_VCPU_SVE: Enables SVE for the CPU (arm64 only). 3363 Depends on KVM_CAP_ARM_SVE. 3364 Requires KVM_ARM_VCPU_FINALIZE(KVM_ARM_VCPU_SVE): 3365 3366 * After KVM_ARM_VCPU_INIT: 3367 3368 - KVM_REG_ARM64_SVE_VLS may be read using KVM_GET_ONE_REG: the 3369 initial value of this pseudo-register indicates the best set of 3370 vector lengths possible for a vcpu on this host. 3371 3372 * Before KVM_ARM_VCPU_FINALIZE(KVM_ARM_VCPU_SVE): 3373 3374 - KVM_RUN and KVM_GET_REG_LIST are not available; 3375 3376 - KVM_GET_ONE_REG and KVM_SET_ONE_REG cannot be used to access 3377 the scalable archietctural SVE registers 3378 KVM_REG_ARM64_SVE_ZREG(), KVM_REG_ARM64_SVE_PREG() or 3379 KVM_REG_ARM64_SVE_FFR; 3380 3381 - KVM_REG_ARM64_SVE_VLS may optionally be written using 3382 KVM_SET_ONE_REG, to modify the set of vector lengths available 3383 for the vcpu. 3384 3385 * After KVM_ARM_VCPU_FINALIZE(KVM_ARM_VCPU_SVE): 3386 3387 - the KVM_REG_ARM64_SVE_VLS pseudo-register is immutable, and can 3388 no longer be written using KVM_SET_ONE_REG. 3389 33904.83 KVM_ARM_PREFERRED_TARGET 3391----------------------------- 3392 3393:Capability: basic 3394:Architectures: arm, arm64 3395:Type: vm ioctl 3396:Parameters: struct kvm_vcpu_init (out) 3397:Returns: 0 on success; -1 on error 3398 3399Errors: 3400 3401 ====== ========================================== 3402 ENODEV no preferred target available for the host 3403 ====== ========================================== 3404 3405This queries KVM for preferred CPU target type which can be emulated 3406by KVM on underlying host. 3407 3408The ioctl returns struct kvm_vcpu_init instance containing information 3409about preferred CPU target type and recommended features for it. The 3410kvm_vcpu_init->features bitmap returned will have feature bits set if 3411the preferred target recommends setting these features, but this is 3412not mandatory. 3413 3414The information returned by this ioctl can be used to prepare an instance 3415of struct kvm_vcpu_init for KVM_ARM_VCPU_INIT ioctl which will result in 3416VCPU matching underlying host. 3417 3418 34194.84 KVM_GET_REG_LIST 3420--------------------- 3421 3422:Capability: basic 3423:Architectures: arm, arm64, mips 3424:Type: vcpu ioctl 3425:Parameters: struct kvm_reg_list (in/out) 3426:Returns: 0 on success; -1 on error 3427 3428Errors: 3429 3430 ===== ============================================================== 3431 E2BIG the reg index list is too big to fit in the array specified by 3432 the user (the number required will be written into n). 3433 ===== ============================================================== 3434 3435:: 3436 3437 struct kvm_reg_list { 3438 __u64 n; /* number of registers in reg[] */ 3439 __u64 reg[0]; 3440 }; 3441 3442This ioctl returns the guest registers that are supported for the 3443KVM_GET_ONE_REG/KVM_SET_ONE_REG calls. 3444 3445 34464.85 KVM_ARM_SET_DEVICE_ADDR (deprecated) 3447----------------------------------------- 3448 3449:Capability: KVM_CAP_ARM_SET_DEVICE_ADDR 3450:Architectures: arm, arm64 3451:Type: vm ioctl 3452:Parameters: struct kvm_arm_device_address (in) 3453:Returns: 0 on success, -1 on error 3454 3455Errors: 3456 3457 ====== ============================================ 3458 ENODEV The device id is unknown 3459 ENXIO Device not supported on current system 3460 EEXIST Address already set 3461 E2BIG Address outside guest physical address space 3462 EBUSY Address overlaps with other device range 3463 ====== ============================================ 3464 3465:: 3466 3467 struct kvm_arm_device_addr { 3468 __u64 id; 3469 __u64 addr; 3470 }; 3471 3472Specify a device address in the guest's physical address space where guests 3473can access emulated or directly exposed devices, which the host kernel needs 3474to know about. The id field is an architecture specific identifier for a 3475specific device. 3476 3477ARM/arm64 divides the id field into two parts, a device id and an 3478address type id specific to the individual device:: 3479 3480 bits: | 63 ... 32 | 31 ... 16 | 15 ... 0 | 3481 field: | 0x00000000 | device id | addr type id | 3482 3483ARM/arm64 currently only require this when using the in-kernel GIC 3484support for the hardware VGIC features, using KVM_ARM_DEVICE_VGIC_V2 3485as the device id. When setting the base address for the guest's 3486mapping of the VGIC virtual CPU and distributor interface, the ioctl 3487must be called after calling KVM_CREATE_IRQCHIP, but before calling 3488KVM_RUN on any of the VCPUs. Calling this ioctl twice for any of the 3489base addresses will return -EEXIST. 3490 3491Note, this IOCTL is deprecated and the more flexible SET/GET_DEVICE_ATTR API 3492should be used instead. 3493 3494 34954.86 KVM_PPC_RTAS_DEFINE_TOKEN 3496------------------------------ 3497 3498:Capability: KVM_CAP_PPC_RTAS 3499:Architectures: ppc 3500:Type: vm ioctl 3501:Parameters: struct kvm_rtas_token_args 3502:Returns: 0 on success, -1 on error 3503 3504Defines a token value for a RTAS (Run Time Abstraction Services) 3505service in order to allow it to be handled in the kernel. The 3506argument struct gives the name of the service, which must be the name 3507of a service that has a kernel-side implementation. If the token 3508value is non-zero, it will be associated with that service, and 3509subsequent RTAS calls by the guest specifying that token will be 3510handled by the kernel. If the token value is 0, then any token 3511associated with the service will be forgotten, and subsequent RTAS 3512calls by the guest for that service will be passed to userspace to be 3513handled. 3514 35154.87 KVM_SET_GUEST_DEBUG 3516------------------------ 3517 3518:Capability: KVM_CAP_SET_GUEST_DEBUG 3519:Architectures: x86, s390, ppc, arm64 3520:Type: vcpu ioctl 3521:Parameters: struct kvm_guest_debug (in) 3522:Returns: 0 on success; -1 on error 3523 3524:: 3525 3526 struct kvm_guest_debug { 3527 __u32 control; 3528 __u32 pad; 3529 struct kvm_guest_debug_arch arch; 3530 }; 3531 3532Set up the processor specific debug registers and configure vcpu for 3533handling guest debug events. There are two parts to the structure, the 3534first a control bitfield indicates the type of debug events to handle 3535when running. Common control bits are: 3536 3537 - KVM_GUESTDBG_ENABLE: guest debugging is enabled 3538 - KVM_GUESTDBG_SINGLESTEP: the next run should single-step 3539 3540The top 16 bits of the control field are architecture specific control 3541flags which can include the following: 3542 3543 - KVM_GUESTDBG_USE_SW_BP: using software breakpoints [x86, arm64] 3544 - KVM_GUESTDBG_USE_HW_BP: using hardware breakpoints [x86, s390] 3545 - KVM_GUESTDBG_USE_HW: using hardware debug events [arm64] 3546 - KVM_GUESTDBG_INJECT_DB: inject DB type exception [x86] 3547 - KVM_GUESTDBG_INJECT_BP: inject BP type exception [x86] 3548 - KVM_GUESTDBG_EXIT_PENDING: trigger an immediate guest exit [s390] 3549 - KVM_GUESTDBG_BLOCKIRQ: avoid injecting interrupts/NMI/SMI [x86] 3550 3551For example KVM_GUESTDBG_USE_SW_BP indicates that software breakpoints 3552are enabled in memory so we need to ensure breakpoint exceptions are 3553correctly trapped and the KVM run loop exits at the breakpoint and not 3554running off into the normal guest vector. For KVM_GUESTDBG_USE_HW_BP 3555we need to ensure the guest vCPUs architecture specific registers are 3556updated to the correct (supplied) values. 3557 3558The second part of the structure is architecture specific and 3559typically contains a set of debug registers. 3560 3561For arm64 the number of debug registers is implementation defined and 3562can be determined by querying the KVM_CAP_GUEST_DEBUG_HW_BPS and 3563KVM_CAP_GUEST_DEBUG_HW_WPS capabilities which return a positive number 3564indicating the number of supported registers. 3565 3566For ppc, the KVM_CAP_PPC_GUEST_DEBUG_SSTEP capability indicates whether 3567the single-step debug event (KVM_GUESTDBG_SINGLESTEP) is supported. 3568 3569Also when supported, KVM_CAP_SET_GUEST_DEBUG2 capability indicates the 3570supported KVM_GUESTDBG_* bits in the control field. 3571 3572When debug events exit the main run loop with the reason 3573KVM_EXIT_DEBUG with the kvm_debug_exit_arch part of the kvm_run 3574structure containing architecture specific debug information. 3575 35764.88 KVM_GET_EMULATED_CPUID 3577--------------------------- 3578 3579:Capability: KVM_CAP_EXT_EMUL_CPUID 3580:Architectures: x86 3581:Type: system ioctl 3582:Parameters: struct kvm_cpuid2 (in/out) 3583:Returns: 0 on success, -1 on error 3584 3585:: 3586 3587 struct kvm_cpuid2 { 3588 __u32 nent; 3589 __u32 flags; 3590 struct kvm_cpuid_entry2 entries[0]; 3591 }; 3592 3593The member 'flags' is used for passing flags from userspace. 3594 3595:: 3596 3597 #define KVM_CPUID_FLAG_SIGNIFCANT_INDEX BIT(0) 3598 #define KVM_CPUID_FLAG_STATEFUL_FUNC BIT(1) /* deprecated */ 3599 #define KVM_CPUID_FLAG_STATE_READ_NEXT BIT(2) /* deprecated */ 3600 3601 struct kvm_cpuid_entry2 { 3602 __u32 function; 3603 __u32 index; 3604 __u32 flags; 3605 __u32 eax; 3606 __u32 ebx; 3607 __u32 ecx; 3608 __u32 edx; 3609 __u32 padding[3]; 3610 }; 3611 3612This ioctl returns x86 cpuid features which are emulated by 3613kvm.Userspace can use the information returned by this ioctl to query 3614which features are emulated by kvm instead of being present natively. 3615 3616Userspace invokes KVM_GET_EMULATED_CPUID by passing a kvm_cpuid2 3617structure with the 'nent' field indicating the number of entries in 3618the variable-size array 'entries'. If the number of entries is too low 3619to describe the cpu capabilities, an error (E2BIG) is returned. If the 3620number is too high, the 'nent' field is adjusted and an error (ENOMEM) 3621is returned. If the number is just right, the 'nent' field is adjusted 3622to the number of valid entries in the 'entries' array, which is then 3623filled. 3624 3625The entries returned are the set CPUID bits of the respective features 3626which kvm emulates, as returned by the CPUID instruction, with unknown 3627or unsupported feature bits cleared. 3628 3629Features like x2apic, for example, may not be present in the host cpu 3630but are exposed by kvm in KVM_GET_SUPPORTED_CPUID because they can be 3631emulated efficiently and thus not included here. 3632 3633The fields in each entry are defined as follows: 3634 3635 function: 3636 the eax value used to obtain the entry 3637 index: 3638 the ecx value used to obtain the entry (for entries that are 3639 affected by ecx) 3640 flags: 3641 an OR of zero or more of the following: 3642 3643 KVM_CPUID_FLAG_SIGNIFCANT_INDEX: 3644 if the index field is valid 3645 3646 eax, ebx, ecx, edx: 3647 3648 the values returned by the cpuid instruction for 3649 this function/index combination 3650 36514.89 KVM_S390_MEM_OP 3652-------------------- 3653 3654:Capability: KVM_CAP_S390_MEM_OP 3655:Architectures: s390 3656:Type: vcpu ioctl 3657:Parameters: struct kvm_s390_mem_op (in) 3658:Returns: = 0 on success, 3659 < 0 on generic error (e.g. -EFAULT or -ENOMEM), 3660 > 0 if an exception occurred while walking the page tables 3661 3662Read or write data from/to the logical (virtual) memory of a VCPU. 3663 3664Parameters are specified via the following structure:: 3665 3666 struct kvm_s390_mem_op { 3667 __u64 gaddr; /* the guest address */ 3668 __u64 flags; /* flags */ 3669 __u32 size; /* amount of bytes */ 3670 __u32 op; /* type of operation */ 3671 __u64 buf; /* buffer in userspace */ 3672 __u8 ar; /* the access register number */ 3673 __u8 reserved[31]; /* should be set to 0 */ 3674 }; 3675 3676The type of operation is specified in the "op" field. It is either 3677KVM_S390_MEMOP_LOGICAL_READ for reading from logical memory space or 3678KVM_S390_MEMOP_LOGICAL_WRITE for writing to logical memory space. The 3679KVM_S390_MEMOP_F_CHECK_ONLY flag can be set in the "flags" field to check 3680whether the corresponding memory access would create an access exception 3681(without touching the data in the memory at the destination). In case an 3682access exception occurred while walking the MMU tables of the guest, the 3683ioctl returns a positive error number to indicate the type of exception. 3684This exception is also raised directly at the corresponding VCPU if the 3685flag KVM_S390_MEMOP_F_INJECT_EXCEPTION is set in the "flags" field. 3686 3687The start address of the memory region has to be specified in the "gaddr" 3688field, and the length of the region in the "size" field (which must not 3689be 0). The maximum value for "size" can be obtained by checking the 3690KVM_CAP_S390_MEM_OP capability. "buf" is the buffer supplied by the 3691userspace application where the read data should be written to for 3692KVM_S390_MEMOP_LOGICAL_READ, or where the data that should be written is 3693stored for a KVM_S390_MEMOP_LOGICAL_WRITE. When KVM_S390_MEMOP_F_CHECK_ONLY 3694is specified, "buf" is unused and can be NULL. "ar" designates the access 3695register number to be used; the valid range is 0..15. 3696 3697The "reserved" field is meant for future extensions. It is not used by 3698KVM with the currently defined set of flags. 3699 37004.90 KVM_S390_GET_SKEYS 3701----------------------- 3702 3703:Capability: KVM_CAP_S390_SKEYS 3704:Architectures: s390 3705:Type: vm ioctl 3706:Parameters: struct kvm_s390_skeys 3707:Returns: 0 on success, KVM_S390_GET_SKEYS_NONE if guest is not using storage 3708 keys, negative value on error 3709 3710This ioctl is used to get guest storage key values on the s390 3711architecture. The ioctl takes parameters via the kvm_s390_skeys struct:: 3712 3713 struct kvm_s390_skeys { 3714 __u64 start_gfn; 3715 __u64 count; 3716 __u64 skeydata_addr; 3717 __u32 flags; 3718 __u32 reserved[9]; 3719 }; 3720 3721The start_gfn field is the number of the first guest frame whose storage keys 3722you want to get. 3723 3724The count field is the number of consecutive frames (starting from start_gfn) 3725whose storage keys to get. The count field must be at least 1 and the maximum 3726allowed value is defined as KVM_S390_SKEYS_MAX. Values outside this range 3727will cause the ioctl to return -EINVAL. 3728 3729The skeydata_addr field is the address to a buffer large enough to hold count 3730bytes. This buffer will be filled with storage key data by the ioctl. 3731 37324.91 KVM_S390_SET_SKEYS 3733----------------------- 3734 3735:Capability: KVM_CAP_S390_SKEYS 3736:Architectures: s390 3737:Type: vm ioctl 3738:Parameters: struct kvm_s390_skeys 3739:Returns: 0 on success, negative value on error 3740 3741This ioctl is used to set guest storage key values on the s390 3742architecture. The ioctl takes parameters via the kvm_s390_skeys struct. 3743See section on KVM_S390_GET_SKEYS for struct definition. 3744 3745The start_gfn field is the number of the first guest frame whose storage keys 3746you want to set. 3747 3748The count field is the number of consecutive frames (starting from start_gfn) 3749whose storage keys to get. The count field must be at least 1 and the maximum 3750allowed value is defined as KVM_S390_SKEYS_MAX. Values outside this range 3751will cause the ioctl to return -EINVAL. 3752 3753The skeydata_addr field is the address to a buffer containing count bytes of 3754storage keys. Each byte in the buffer will be set as the storage key for a 3755single frame starting at start_gfn for count frames. 3756 3757Note: If any architecturally invalid key value is found in the given data then 3758the ioctl will return -EINVAL. 3759 37604.92 KVM_S390_IRQ 3761----------------- 3762 3763:Capability: KVM_CAP_S390_INJECT_IRQ 3764:Architectures: s390 3765:Type: vcpu ioctl 3766:Parameters: struct kvm_s390_irq (in) 3767:Returns: 0 on success, -1 on error 3768 3769Errors: 3770 3771 3772 ====== ================================================================= 3773 EINVAL interrupt type is invalid 3774 type is KVM_S390_SIGP_STOP and flag parameter is invalid value, 3775 type is KVM_S390_INT_EXTERNAL_CALL and code is bigger 3776 than the maximum of VCPUs 3777 EBUSY type is KVM_S390_SIGP_SET_PREFIX and vcpu is not stopped, 3778 type is KVM_S390_SIGP_STOP and a stop irq is already pending, 3779 type is KVM_S390_INT_EXTERNAL_CALL and an external call interrupt 3780 is already pending 3781 ====== ================================================================= 3782 3783Allows to inject an interrupt to the guest. 3784 3785Using struct kvm_s390_irq as a parameter allows 3786to inject additional payload which is not 3787possible via KVM_S390_INTERRUPT. 3788 3789Interrupt parameters are passed via kvm_s390_irq:: 3790 3791 struct kvm_s390_irq { 3792 __u64 type; 3793 union { 3794 struct kvm_s390_io_info io; 3795 struct kvm_s390_ext_info ext; 3796 struct kvm_s390_pgm_info pgm; 3797 struct kvm_s390_emerg_info emerg; 3798 struct kvm_s390_extcall_info extcall; 3799 struct kvm_s390_prefix_info prefix; 3800 struct kvm_s390_stop_info stop; 3801 struct kvm_s390_mchk_info mchk; 3802 char reserved[64]; 3803 } u; 3804 }; 3805 3806type can be one of the following: 3807 3808- KVM_S390_SIGP_STOP - sigp stop; parameter in .stop 3809- KVM_S390_PROGRAM_INT - program check; parameters in .pgm 3810- KVM_S390_SIGP_SET_PREFIX - sigp set prefix; parameters in .prefix 3811- KVM_S390_RESTART - restart; no parameters 3812- KVM_S390_INT_CLOCK_COMP - clock comparator interrupt; no parameters 3813- KVM_S390_INT_CPU_TIMER - CPU timer interrupt; no parameters 3814- KVM_S390_INT_EMERGENCY - sigp emergency; parameters in .emerg 3815- KVM_S390_INT_EXTERNAL_CALL - sigp external call; parameters in .extcall 3816- KVM_S390_MCHK - machine check interrupt; parameters in .mchk 3817 3818This is an asynchronous vcpu ioctl and can be invoked from any thread. 3819 38204.94 KVM_S390_GET_IRQ_STATE 3821--------------------------- 3822 3823:Capability: KVM_CAP_S390_IRQ_STATE 3824:Architectures: s390 3825:Type: vcpu ioctl 3826:Parameters: struct kvm_s390_irq_state (out) 3827:Returns: >= number of bytes copied into buffer, 3828 -EINVAL if buffer size is 0, 3829 -ENOBUFS if buffer size is too small to fit all pending interrupts, 3830 -EFAULT if the buffer address was invalid 3831 3832This ioctl allows userspace to retrieve the complete state of all currently 3833pending interrupts in a single buffer. Use cases include migration 3834and introspection. The parameter structure contains the address of a 3835userspace buffer and its length:: 3836 3837 struct kvm_s390_irq_state { 3838 __u64 buf; 3839 __u32 flags; /* will stay unused for compatibility reasons */ 3840 __u32 len; 3841 __u32 reserved[4]; /* will stay unused for compatibility reasons */ 3842 }; 3843 3844Userspace passes in the above struct and for each pending interrupt a 3845struct kvm_s390_irq is copied to the provided buffer. 3846 3847The structure contains a flags and a reserved field for future extensions. As 3848the kernel never checked for flags == 0 and QEMU never pre-zeroed flags and 3849reserved, these fields can not be used in the future without breaking 3850compatibility. 3851 3852If -ENOBUFS is returned the buffer provided was too small and userspace 3853may retry with a bigger buffer. 3854 38554.95 KVM_S390_SET_IRQ_STATE 3856--------------------------- 3857 3858:Capability: KVM_CAP_S390_IRQ_STATE 3859:Architectures: s390 3860:Type: vcpu ioctl 3861:Parameters: struct kvm_s390_irq_state (in) 3862:Returns: 0 on success, 3863 -EFAULT if the buffer address was invalid, 3864 -EINVAL for an invalid buffer length (see below), 3865 -EBUSY if there were already interrupts pending, 3866 errors occurring when actually injecting the 3867 interrupt. See KVM_S390_IRQ. 3868 3869This ioctl allows userspace to set the complete state of all cpu-local 3870interrupts currently pending for the vcpu. It is intended for restoring 3871interrupt state after a migration. The input parameter is a userspace buffer 3872containing a struct kvm_s390_irq_state:: 3873 3874 struct kvm_s390_irq_state { 3875 __u64 buf; 3876 __u32 flags; /* will stay unused for compatibility reasons */ 3877 __u32 len; 3878 __u32 reserved[4]; /* will stay unused for compatibility reasons */ 3879 }; 3880 3881The restrictions for flags and reserved apply as well. 3882(see KVM_S390_GET_IRQ_STATE) 3883 3884The userspace memory referenced by buf contains a struct kvm_s390_irq 3885for each interrupt to be injected into the guest. 3886If one of the interrupts could not be injected for some reason the 3887ioctl aborts. 3888 3889len must be a multiple of sizeof(struct kvm_s390_irq). It must be > 0 3890and it must not exceed (max_vcpus + 32) * sizeof(struct kvm_s390_irq), 3891which is the maximum number of possibly pending cpu-local interrupts. 3892 38934.96 KVM_SMI 3894------------ 3895 3896:Capability: KVM_CAP_X86_SMM 3897:Architectures: x86 3898:Type: vcpu ioctl 3899:Parameters: none 3900:Returns: 0 on success, -1 on error 3901 3902Queues an SMI on the thread's vcpu. 3903 39044.97 KVM_X86_SET_MSR_FILTER 3905---------------------------- 3906 3907:Capability: KVM_X86_SET_MSR_FILTER 3908:Architectures: x86 3909:Type: vm ioctl 3910:Parameters: struct kvm_msr_filter 3911:Returns: 0 on success, < 0 on error 3912 3913:: 3914 3915 struct kvm_msr_filter_range { 3916 #define KVM_MSR_FILTER_READ (1 << 0) 3917 #define KVM_MSR_FILTER_WRITE (1 << 1) 3918 __u32 flags; 3919 __u32 nmsrs; /* number of msrs in bitmap */ 3920 __u32 base; /* MSR index the bitmap starts at */ 3921 __u8 *bitmap; /* a 1 bit allows the operations in flags, 0 denies */ 3922 }; 3923 3924 #define KVM_MSR_FILTER_MAX_RANGES 16 3925 struct kvm_msr_filter { 3926 #define KVM_MSR_FILTER_DEFAULT_ALLOW (0 << 0) 3927 #define KVM_MSR_FILTER_DEFAULT_DENY (1 << 0) 3928 __u32 flags; 3929 struct kvm_msr_filter_range ranges[KVM_MSR_FILTER_MAX_RANGES]; 3930 }; 3931 3932flags values for ``struct kvm_msr_filter_range``: 3933 3934``KVM_MSR_FILTER_READ`` 3935 3936 Filter read accesses to MSRs using the given bitmap. A 0 in the bitmap 3937 indicates that a read should immediately fail, while a 1 indicates that 3938 a read for a particular MSR should be handled regardless of the default 3939 filter action. 3940 3941``KVM_MSR_FILTER_WRITE`` 3942 3943 Filter write accesses to MSRs using the given bitmap. A 0 in the bitmap 3944 indicates that a write should immediately fail, while a 1 indicates that 3945 a write for a particular MSR should be handled regardless of the default 3946 filter action. 3947 3948``KVM_MSR_FILTER_READ | KVM_MSR_FILTER_WRITE`` 3949 3950 Filter both read and write accesses to MSRs using the given bitmap. A 0 3951 in the bitmap indicates that both reads and writes should immediately fail, 3952 while a 1 indicates that reads and writes for a particular MSR are not 3953 filtered by this range. 3954 3955flags values for ``struct kvm_msr_filter``: 3956 3957``KVM_MSR_FILTER_DEFAULT_ALLOW`` 3958 3959 If no filter range matches an MSR index that is getting accessed, KVM will 3960 fall back to allowing access to the MSR. 3961 3962``KVM_MSR_FILTER_DEFAULT_DENY`` 3963 3964 If no filter range matches an MSR index that is getting accessed, KVM will 3965 fall back to rejecting access to the MSR. In this mode, all MSRs that should 3966 be processed by KVM need to explicitly be marked as allowed in the bitmaps. 3967 3968This ioctl allows user space to define up to 16 bitmaps of MSR ranges to 3969specify whether a certain MSR access should be explicitly filtered for or not. 3970 3971If this ioctl has never been invoked, MSR accesses are not guarded and the 3972default KVM in-kernel emulation behavior is fully preserved. 3973 3974Calling this ioctl with an empty set of ranges (all nmsrs == 0) disables MSR 3975filtering. In that mode, ``KVM_MSR_FILTER_DEFAULT_DENY`` is invalid and causes 3976an error. 3977 3978As soon as the filtering is in place, every MSR access is processed through 3979the filtering except for accesses to the x2APIC MSRs (from 0x800 to 0x8ff); 3980x2APIC MSRs are always allowed, independent of the ``default_allow`` setting, 3981and their behavior depends on the ``X2APIC_ENABLE`` bit of the APIC base 3982register. 3983 3984If a bit is within one of the defined ranges, read and write accesses are 3985guarded by the bitmap's value for the MSR index if the kind of access 3986is included in the ``struct kvm_msr_filter_range`` flags. If no range 3987cover this particular access, the behavior is determined by the flags 3988field in the kvm_msr_filter struct: ``KVM_MSR_FILTER_DEFAULT_ALLOW`` 3989and ``KVM_MSR_FILTER_DEFAULT_DENY``. 3990 3991Each bitmap range specifies a range of MSRs to potentially allow access on. 3992The range goes from MSR index [base .. base+nmsrs]. The flags field 3993indicates whether reads, writes or both reads and writes are filtered 3994by setting a 1 bit in the bitmap for the corresponding MSR index. 3995 3996If an MSR access is not permitted through the filtering, it generates a 3997#GP inside the guest. When combined with KVM_CAP_X86_USER_SPACE_MSR, that 3998allows user space to deflect and potentially handle various MSR accesses 3999into user space. 4000 4001If a vCPU is in running state while this ioctl is invoked, the vCPU may 4002experience inconsistent filtering behavior on MSR accesses. 4003 40044.98 KVM_CREATE_SPAPR_TCE_64 4005---------------------------- 4006 4007:Capability: KVM_CAP_SPAPR_TCE_64 4008:Architectures: powerpc 4009:Type: vm ioctl 4010:Parameters: struct kvm_create_spapr_tce_64 (in) 4011:Returns: file descriptor for manipulating the created TCE table 4012 4013This is an extension for KVM_CAP_SPAPR_TCE which only supports 32bit 4014windows, described in 4.62 KVM_CREATE_SPAPR_TCE 4015 4016This capability uses extended struct in ioctl interface:: 4017 4018 /* for KVM_CAP_SPAPR_TCE_64 */ 4019 struct kvm_create_spapr_tce_64 { 4020 __u64 liobn; 4021 __u32 page_shift; 4022 __u32 flags; 4023 __u64 offset; /* in pages */ 4024 __u64 size; /* in pages */ 4025 }; 4026 4027The aim of extension is to support an additional bigger DMA window with 4028a variable page size. 4029KVM_CREATE_SPAPR_TCE_64 receives a 64bit window size, an IOMMU page shift and 4030a bus offset of the corresponding DMA window, @size and @offset are numbers 4031of IOMMU pages. 4032 4033@flags are not used at the moment. 4034 4035The rest of functionality is identical to KVM_CREATE_SPAPR_TCE. 4036 40374.99 KVM_REINJECT_CONTROL 4038------------------------- 4039 4040:Capability: KVM_CAP_REINJECT_CONTROL 4041:Architectures: x86 4042:Type: vm ioctl 4043:Parameters: struct kvm_reinject_control (in) 4044:Returns: 0 on success, 4045 -EFAULT if struct kvm_reinject_control cannot be read, 4046 -ENXIO if KVM_CREATE_PIT or KVM_CREATE_PIT2 didn't succeed earlier. 4047 4048i8254 (PIT) has two modes, reinject and !reinject. The default is reinject, 4049where KVM queues elapsed i8254 ticks and monitors completion of interrupt from 4050vector(s) that i8254 injects. Reinject mode dequeues a tick and injects its 4051interrupt whenever there isn't a pending interrupt from i8254. 4052!reinject mode injects an interrupt as soon as a tick arrives. 4053 4054:: 4055 4056 struct kvm_reinject_control { 4057 __u8 pit_reinject; 4058 __u8 reserved[31]; 4059 }; 4060 4061pit_reinject = 0 (!reinject mode) is recommended, unless running an old 4062operating system that uses the PIT for timing (e.g. Linux 2.4.x). 4063 40644.100 KVM_PPC_CONFIGURE_V3_MMU 4065------------------------------ 4066 4067:Capability: KVM_CAP_PPC_RADIX_MMU or KVM_CAP_PPC_HASH_MMU_V3 4068:Architectures: ppc 4069:Type: vm ioctl 4070:Parameters: struct kvm_ppc_mmuv3_cfg (in) 4071:Returns: 0 on success, 4072 -EFAULT if struct kvm_ppc_mmuv3_cfg cannot be read, 4073 -EINVAL if the configuration is invalid 4074 4075This ioctl controls whether the guest will use radix or HPT (hashed 4076page table) translation, and sets the pointer to the process table for 4077the guest. 4078 4079:: 4080 4081 struct kvm_ppc_mmuv3_cfg { 4082 __u64 flags; 4083 __u64 process_table; 4084 }; 4085 4086There are two bits that can be set in flags; KVM_PPC_MMUV3_RADIX and 4087KVM_PPC_MMUV3_GTSE. KVM_PPC_MMUV3_RADIX, if set, configures the guest 4088to use radix tree translation, and if clear, to use HPT translation. 4089KVM_PPC_MMUV3_GTSE, if set and if KVM permits it, configures the guest 4090to be able to use the global TLB and SLB invalidation instructions; 4091if clear, the guest may not use these instructions. 4092 4093The process_table field specifies the address and size of the guest 4094process table, which is in the guest's space. This field is formatted 4095as the second doubleword of the partition table entry, as defined in 4096the Power ISA V3.00, Book III section 5.7.6.1. 4097 40984.101 KVM_PPC_GET_RMMU_INFO 4099--------------------------- 4100 4101:Capability: KVM_CAP_PPC_RADIX_MMU 4102:Architectures: ppc 4103:Type: vm ioctl 4104:Parameters: struct kvm_ppc_rmmu_info (out) 4105:Returns: 0 on success, 4106 -EFAULT if struct kvm_ppc_rmmu_info cannot be written, 4107 -EINVAL if no useful information can be returned 4108 4109This ioctl returns a structure containing two things: (a) a list 4110containing supported radix tree geometries, and (b) a list that maps 4111page sizes to put in the "AP" (actual page size) field for the tlbie 4112(TLB invalidate entry) instruction. 4113 4114:: 4115 4116 struct kvm_ppc_rmmu_info { 4117 struct kvm_ppc_radix_geom { 4118 __u8 page_shift; 4119 __u8 level_bits[4]; 4120 __u8 pad[3]; 4121 } geometries[8]; 4122 __u32 ap_encodings[8]; 4123 }; 4124 4125The geometries[] field gives up to 8 supported geometries for the 4126radix page table, in terms of the log base 2 of the smallest page 4127size, and the number of bits indexed at each level of the tree, from 4128the PTE level up to the PGD level in that order. Any unused entries 4129will have 0 in the page_shift field. 4130 4131The ap_encodings gives the supported page sizes and their AP field 4132encodings, encoded with the AP value in the top 3 bits and the log 4133base 2 of the page size in the bottom 6 bits. 4134 41354.102 KVM_PPC_RESIZE_HPT_PREPARE 4136-------------------------------- 4137 4138:Capability: KVM_CAP_SPAPR_RESIZE_HPT 4139:Architectures: powerpc 4140:Type: vm ioctl 4141:Parameters: struct kvm_ppc_resize_hpt (in) 4142:Returns: 0 on successful completion, 4143 >0 if a new HPT is being prepared, the value is an estimated 4144 number of milliseconds until preparation is complete, 4145 -EFAULT if struct kvm_reinject_control cannot be read, 4146 -EINVAL if the supplied shift or flags are invalid, 4147 -ENOMEM if unable to allocate the new HPT, 4148 4149Used to implement the PAPR extension for runtime resizing of a guest's 4150Hashed Page Table (HPT). Specifically this starts, stops or monitors 4151the preparation of a new potential HPT for the guest, essentially 4152implementing the H_RESIZE_HPT_PREPARE hypercall. 4153 4154:: 4155 4156 struct kvm_ppc_resize_hpt { 4157 __u64 flags; 4158 __u32 shift; 4159 __u32 pad; 4160 }; 4161 4162If called with shift > 0 when there is no pending HPT for the guest, 4163this begins preparation of a new pending HPT of size 2^(shift) bytes. 4164It then returns a positive integer with the estimated number of 4165milliseconds until preparation is complete. 4166 4167If called when there is a pending HPT whose size does not match that 4168requested in the parameters, discards the existing pending HPT and 4169creates a new one as above. 4170 4171If called when there is a pending HPT of the size requested, will: 4172 4173 * If preparation of the pending HPT is already complete, return 0 4174 * If preparation of the pending HPT has failed, return an error 4175 code, then discard the pending HPT. 4176 * If preparation of the pending HPT is still in progress, return an 4177 estimated number of milliseconds until preparation is complete. 4178 4179If called with shift == 0, discards any currently pending HPT and 4180returns 0 (i.e. cancels any in-progress preparation). 4181 4182flags is reserved for future expansion, currently setting any bits in 4183flags will result in an -EINVAL. 4184 4185Normally this will be called repeatedly with the same parameters until 4186it returns <= 0. The first call will initiate preparation, subsequent 4187ones will monitor preparation until it completes or fails. 4188 41894.103 KVM_PPC_RESIZE_HPT_COMMIT 4190------------------------------- 4191 4192:Capability: KVM_CAP_SPAPR_RESIZE_HPT 4193:Architectures: powerpc 4194:Type: vm ioctl 4195:Parameters: struct kvm_ppc_resize_hpt (in) 4196:Returns: 0 on successful completion, 4197 -EFAULT if struct kvm_reinject_control cannot be read, 4198 -EINVAL if the supplied shift or flags are invalid, 4199 -ENXIO is there is no pending HPT, or the pending HPT doesn't 4200 have the requested size, 4201 -EBUSY if the pending HPT is not fully prepared, 4202 -ENOSPC if there was a hash collision when moving existing 4203 HPT entries to the new HPT, 4204 -EIO on other error conditions 4205 4206Used to implement the PAPR extension for runtime resizing of a guest's 4207Hashed Page Table (HPT). Specifically this requests that the guest be 4208transferred to working with the new HPT, essentially implementing the 4209H_RESIZE_HPT_COMMIT hypercall. 4210 4211:: 4212 4213 struct kvm_ppc_resize_hpt { 4214 __u64 flags; 4215 __u32 shift; 4216 __u32 pad; 4217 }; 4218 4219This should only be called after KVM_PPC_RESIZE_HPT_PREPARE has 4220returned 0 with the same parameters. In other cases 4221KVM_PPC_RESIZE_HPT_COMMIT will return an error (usually -ENXIO or 4222-EBUSY, though others may be possible if the preparation was started, 4223but failed). 4224 4225This will have undefined effects on the guest if it has not already 4226placed itself in a quiescent state where no vcpu will make MMU enabled 4227memory accesses. 4228 4229On succsful completion, the pending HPT will become the guest's active 4230HPT and the previous HPT will be discarded. 4231 4232On failure, the guest will still be operating on its previous HPT. 4233 42344.104 KVM_X86_GET_MCE_CAP_SUPPORTED 4235----------------------------------- 4236 4237:Capability: KVM_CAP_MCE 4238:Architectures: x86 4239:Type: system ioctl 4240:Parameters: u64 mce_cap (out) 4241:Returns: 0 on success, -1 on error 4242 4243Returns supported MCE capabilities. The u64 mce_cap parameter 4244has the same format as the MSR_IA32_MCG_CAP register. Supported 4245capabilities will have the corresponding bits set. 4246 42474.105 KVM_X86_SETUP_MCE 4248----------------------- 4249 4250:Capability: KVM_CAP_MCE 4251:Architectures: x86 4252:Type: vcpu ioctl 4253:Parameters: u64 mcg_cap (in) 4254:Returns: 0 on success, 4255 -EFAULT if u64 mcg_cap cannot be read, 4256 -EINVAL if the requested number of banks is invalid, 4257 -EINVAL if requested MCE capability is not supported. 4258 4259Initializes MCE support for use. The u64 mcg_cap parameter 4260has the same format as the MSR_IA32_MCG_CAP register and 4261specifies which capabilities should be enabled. The maximum 4262supported number of error-reporting banks can be retrieved when 4263checking for KVM_CAP_MCE. The supported capabilities can be 4264retrieved with KVM_X86_GET_MCE_CAP_SUPPORTED. 4265 42664.106 KVM_X86_SET_MCE 4267--------------------- 4268 4269:Capability: KVM_CAP_MCE 4270:Architectures: x86 4271:Type: vcpu ioctl 4272:Parameters: struct kvm_x86_mce (in) 4273:Returns: 0 on success, 4274 -EFAULT if struct kvm_x86_mce cannot be read, 4275 -EINVAL if the bank number is invalid, 4276 -EINVAL if VAL bit is not set in status field. 4277 4278Inject a machine check error (MCE) into the guest. The input 4279parameter is:: 4280 4281 struct kvm_x86_mce { 4282 __u64 status; 4283 __u64 addr; 4284 __u64 misc; 4285 __u64 mcg_status; 4286 __u8 bank; 4287 __u8 pad1[7]; 4288 __u64 pad2[3]; 4289 }; 4290 4291If the MCE being reported is an uncorrected error, KVM will 4292inject it as an MCE exception into the guest. If the guest 4293MCG_STATUS register reports that an MCE is in progress, KVM 4294causes an KVM_EXIT_SHUTDOWN vmexit. 4295 4296Otherwise, if the MCE is a corrected error, KVM will just 4297store it in the corresponding bank (provided this bank is 4298not holding a previously reported uncorrected error). 4299 43004.107 KVM_S390_GET_CMMA_BITS 4301---------------------------- 4302 4303:Capability: KVM_CAP_S390_CMMA_MIGRATION 4304:Architectures: s390 4305:Type: vm ioctl 4306:Parameters: struct kvm_s390_cmma_log (in, out) 4307:Returns: 0 on success, a negative value on error 4308 4309This ioctl is used to get the values of the CMMA bits on the s390 4310architecture. It is meant to be used in two scenarios: 4311 4312- During live migration to save the CMMA values. Live migration needs 4313 to be enabled via the KVM_REQ_START_MIGRATION VM property. 4314- To non-destructively peek at the CMMA values, with the flag 4315 KVM_S390_CMMA_PEEK set. 4316 4317The ioctl takes parameters via the kvm_s390_cmma_log struct. The desired 4318values are written to a buffer whose location is indicated via the "values" 4319member in the kvm_s390_cmma_log struct. The values in the input struct are 4320also updated as needed. 4321 4322Each CMMA value takes up one byte. 4323 4324:: 4325 4326 struct kvm_s390_cmma_log { 4327 __u64 start_gfn; 4328 __u32 count; 4329 __u32 flags; 4330 union { 4331 __u64 remaining; 4332 __u64 mask; 4333 }; 4334 __u64 values; 4335 }; 4336 4337start_gfn is the number of the first guest frame whose CMMA values are 4338to be retrieved, 4339 4340count is the length of the buffer in bytes, 4341 4342values points to the buffer where the result will be written to. 4343 4344If count is greater than KVM_S390_SKEYS_MAX, then it is considered to be 4345KVM_S390_SKEYS_MAX. KVM_S390_SKEYS_MAX is re-used for consistency with 4346other ioctls. 4347 4348The result is written in the buffer pointed to by the field values, and 4349the values of the input parameter are updated as follows. 4350 4351Depending on the flags, different actions are performed. The only 4352supported flag so far is KVM_S390_CMMA_PEEK. 4353 4354The default behaviour if KVM_S390_CMMA_PEEK is not set is: 4355start_gfn will indicate the first page frame whose CMMA bits were dirty. 4356It is not necessarily the same as the one passed as input, as clean pages 4357are skipped. 4358 4359count will indicate the number of bytes actually written in the buffer. 4360It can (and very often will) be smaller than the input value, since the 4361buffer is only filled until 16 bytes of clean values are found (which 4362are then not copied in the buffer). Since a CMMA migration block needs 4363the base address and the length, for a total of 16 bytes, we will send 4364back some clean data if there is some dirty data afterwards, as long as 4365the size of the clean data does not exceed the size of the header. This 4366allows to minimize the amount of data to be saved or transferred over 4367the network at the expense of more roundtrips to userspace. The next 4368invocation of the ioctl will skip over all the clean values, saving 4369potentially more than just the 16 bytes we found. 4370 4371If KVM_S390_CMMA_PEEK is set: 4372the existing storage attributes are read even when not in migration 4373mode, and no other action is performed; 4374 4375the output start_gfn will be equal to the input start_gfn, 4376 4377the output count will be equal to the input count, except if the end of 4378memory has been reached. 4379 4380In both cases: 4381the field "remaining" will indicate the total number of dirty CMMA values 4382still remaining, or 0 if KVM_S390_CMMA_PEEK is set and migration mode is 4383not enabled. 4384 4385mask is unused. 4386 4387values points to the userspace buffer where the result will be stored. 4388 4389This ioctl can fail with -ENOMEM if not enough memory can be allocated to 4390complete the task, with -ENXIO if CMMA is not enabled, with -EINVAL if 4391KVM_S390_CMMA_PEEK is not set but migration mode was not enabled, with 4392-EFAULT if the userspace address is invalid or if no page table is 4393present for the addresses (e.g. when using hugepages). 4394 43954.108 KVM_S390_SET_CMMA_BITS 4396---------------------------- 4397 4398:Capability: KVM_CAP_S390_CMMA_MIGRATION 4399:Architectures: s390 4400:Type: vm ioctl 4401:Parameters: struct kvm_s390_cmma_log (in) 4402:Returns: 0 on success, a negative value on error 4403 4404This ioctl is used to set the values of the CMMA bits on the s390 4405architecture. It is meant to be used during live migration to restore 4406the CMMA values, but there are no restrictions on its use. 4407The ioctl takes parameters via the kvm_s390_cmma_values struct. 4408Each CMMA value takes up one byte. 4409 4410:: 4411 4412 struct kvm_s390_cmma_log { 4413 __u64 start_gfn; 4414 __u32 count; 4415 __u32 flags; 4416 union { 4417 __u64 remaining; 4418 __u64 mask; 4419 }; 4420 __u64 values; 4421 }; 4422 4423start_gfn indicates the starting guest frame number, 4424 4425count indicates how many values are to be considered in the buffer, 4426 4427flags is not used and must be 0. 4428 4429mask indicates which PGSTE bits are to be considered. 4430 4431remaining is not used. 4432 4433values points to the buffer in userspace where to store the values. 4434 4435This ioctl can fail with -ENOMEM if not enough memory can be allocated to 4436complete the task, with -ENXIO if CMMA is not enabled, with -EINVAL if 4437the count field is too large (e.g. more than KVM_S390_CMMA_SIZE_MAX) or 4438if the flags field was not 0, with -EFAULT if the userspace address is 4439invalid, if invalid pages are written to (e.g. after the end of memory) 4440or if no page table is present for the addresses (e.g. when using 4441hugepages). 4442 44434.109 KVM_PPC_GET_CPU_CHAR 4444-------------------------- 4445 4446:Capability: KVM_CAP_PPC_GET_CPU_CHAR 4447:Architectures: powerpc 4448:Type: vm ioctl 4449:Parameters: struct kvm_ppc_cpu_char (out) 4450:Returns: 0 on successful completion, 4451 -EFAULT if struct kvm_ppc_cpu_char cannot be written 4452 4453This ioctl gives userspace information about certain characteristics 4454of the CPU relating to speculative execution of instructions and 4455possible information leakage resulting from speculative execution (see 4456CVE-2017-5715, CVE-2017-5753 and CVE-2017-5754). The information is 4457returned in struct kvm_ppc_cpu_char, which looks like this:: 4458 4459 struct kvm_ppc_cpu_char { 4460 __u64 character; /* characteristics of the CPU */ 4461 __u64 behaviour; /* recommended software behaviour */ 4462 __u64 character_mask; /* valid bits in character */ 4463 __u64 behaviour_mask; /* valid bits in behaviour */ 4464 }; 4465 4466For extensibility, the character_mask and behaviour_mask fields 4467indicate which bits of character and behaviour have been filled in by 4468the kernel. If the set of defined bits is extended in future then 4469userspace will be able to tell whether it is running on a kernel that 4470knows about the new bits. 4471 4472The character field describes attributes of the CPU which can help 4473with preventing inadvertent information disclosure - specifically, 4474whether there is an instruction to flash-invalidate the L1 data cache 4475(ori 30,30,0 or mtspr SPRN_TRIG2,rN), whether the L1 data cache is set 4476to a mode where entries can only be used by the thread that created 4477them, whether the bcctr[l] instruction prevents speculation, and 4478whether a speculation barrier instruction (ori 31,31,0) is provided. 4479 4480The behaviour field describes actions that software should take to 4481prevent inadvertent information disclosure, and thus describes which 4482vulnerabilities the hardware is subject to; specifically whether the 4483L1 data cache should be flushed when returning to user mode from the 4484kernel, and whether a speculation barrier should be placed between an 4485array bounds check and the array access. 4486 4487These fields use the same bit definitions as the new 4488H_GET_CPU_CHARACTERISTICS hypercall. 4489 44904.110 KVM_MEMORY_ENCRYPT_OP 4491--------------------------- 4492 4493:Capability: basic 4494:Architectures: x86 4495:Type: vm 4496:Parameters: an opaque platform specific structure (in/out) 4497:Returns: 0 on success; -1 on error 4498 4499If the platform supports creating encrypted VMs then this ioctl can be used 4500for issuing platform-specific memory encryption commands to manage those 4501encrypted VMs. 4502 4503Currently, this ioctl is used for issuing Secure Encrypted Virtualization 4504(SEV) commands on AMD Processors. The SEV commands are defined in 4505Documentation/virt/kvm/amd-memory-encryption.rst. 4506 45074.111 KVM_MEMORY_ENCRYPT_REG_REGION 4508----------------------------------- 4509 4510:Capability: basic 4511:Architectures: x86 4512:Type: system 4513:Parameters: struct kvm_enc_region (in) 4514:Returns: 0 on success; -1 on error 4515 4516This ioctl can be used to register a guest memory region which may 4517contain encrypted data (e.g. guest RAM, SMRAM etc). 4518 4519It is used in the SEV-enabled guest. When encryption is enabled, a guest 4520memory region may contain encrypted data. The SEV memory encryption 4521engine uses a tweak such that two identical plaintext pages, each at 4522different locations will have differing ciphertexts. So swapping or 4523moving ciphertext of those pages will not result in plaintext being 4524swapped. So relocating (or migrating) physical backing pages for the SEV 4525guest will require some additional steps. 4526 4527Note: The current SEV key management spec does not provide commands to 4528swap or migrate (move) ciphertext pages. Hence, for now we pin the guest 4529memory region registered with the ioctl. 4530 45314.112 KVM_MEMORY_ENCRYPT_UNREG_REGION 4532------------------------------------- 4533 4534:Capability: basic 4535:Architectures: x86 4536:Type: system 4537:Parameters: struct kvm_enc_region (in) 4538:Returns: 0 on success; -1 on error 4539 4540This ioctl can be used to unregister the guest memory region registered 4541with KVM_MEMORY_ENCRYPT_REG_REGION ioctl above. 4542 45434.113 KVM_HYPERV_EVENTFD 4544------------------------ 4545 4546:Capability: KVM_CAP_HYPERV_EVENTFD 4547:Architectures: x86 4548:Type: vm ioctl 4549:Parameters: struct kvm_hyperv_eventfd (in) 4550 4551This ioctl (un)registers an eventfd to receive notifications from the guest on 4552the specified Hyper-V connection id through the SIGNAL_EVENT hypercall, without 4553causing a user exit. SIGNAL_EVENT hypercall with non-zero event flag number 4554(bits 24-31) still triggers a KVM_EXIT_HYPERV_HCALL user exit. 4555 4556:: 4557 4558 struct kvm_hyperv_eventfd { 4559 __u32 conn_id; 4560 __s32 fd; 4561 __u32 flags; 4562 __u32 padding[3]; 4563 }; 4564 4565The conn_id field should fit within 24 bits:: 4566 4567 #define KVM_HYPERV_CONN_ID_MASK 0x00ffffff 4568 4569The acceptable values for the flags field are:: 4570 4571 #define KVM_HYPERV_EVENTFD_DEASSIGN (1 << 0) 4572 4573:Returns: 0 on success, 4574 -EINVAL if conn_id or flags is outside the allowed range, 4575 -ENOENT on deassign if the conn_id isn't registered, 4576 -EEXIST on assign if the conn_id is already registered 4577 45784.114 KVM_GET_NESTED_STATE 4579-------------------------- 4580 4581:Capability: KVM_CAP_NESTED_STATE 4582:Architectures: x86 4583:Type: vcpu ioctl 4584:Parameters: struct kvm_nested_state (in/out) 4585:Returns: 0 on success, -1 on error 4586 4587Errors: 4588 4589 ===== ============================================================= 4590 E2BIG the total state size exceeds the value of 'size' specified by 4591 the user; the size required will be written into size. 4592 ===== ============================================================= 4593 4594:: 4595 4596 struct kvm_nested_state { 4597 __u16 flags; 4598 __u16 format; 4599 __u32 size; 4600 4601 union { 4602 struct kvm_vmx_nested_state_hdr vmx; 4603 struct kvm_svm_nested_state_hdr svm; 4604 4605 /* Pad the header to 128 bytes. */ 4606 __u8 pad[120]; 4607 } hdr; 4608 4609 union { 4610 struct kvm_vmx_nested_state_data vmx[0]; 4611 struct kvm_svm_nested_state_data svm[0]; 4612 } data; 4613 }; 4614 4615 #define KVM_STATE_NESTED_GUEST_MODE 0x00000001 4616 #define KVM_STATE_NESTED_RUN_PENDING 0x00000002 4617 #define KVM_STATE_NESTED_EVMCS 0x00000004 4618 4619 #define KVM_STATE_NESTED_FORMAT_VMX 0 4620 #define KVM_STATE_NESTED_FORMAT_SVM 1 4621 4622 #define KVM_STATE_NESTED_VMX_VMCS_SIZE 0x1000 4623 4624 #define KVM_STATE_NESTED_VMX_SMM_GUEST_MODE 0x00000001 4625 #define KVM_STATE_NESTED_VMX_SMM_VMXON 0x00000002 4626 4627 #define KVM_STATE_VMX_PREEMPTION_TIMER_DEADLINE 0x00000001 4628 4629 struct kvm_vmx_nested_state_hdr { 4630 __u64 vmxon_pa; 4631 __u64 vmcs12_pa; 4632 4633 struct { 4634 __u16 flags; 4635 } smm; 4636 4637 __u32 flags; 4638 __u64 preemption_timer_deadline; 4639 }; 4640 4641 struct kvm_vmx_nested_state_data { 4642 __u8 vmcs12[KVM_STATE_NESTED_VMX_VMCS_SIZE]; 4643 __u8 shadow_vmcs12[KVM_STATE_NESTED_VMX_VMCS_SIZE]; 4644 }; 4645 4646This ioctl copies the vcpu's nested virtualization state from the kernel to 4647userspace. 4648 4649The maximum size of the state can be retrieved by passing KVM_CAP_NESTED_STATE 4650to the KVM_CHECK_EXTENSION ioctl(). 4651 46524.115 KVM_SET_NESTED_STATE 4653-------------------------- 4654 4655:Capability: KVM_CAP_NESTED_STATE 4656:Architectures: x86 4657:Type: vcpu ioctl 4658:Parameters: struct kvm_nested_state (in) 4659:Returns: 0 on success, -1 on error 4660 4661This copies the vcpu's kvm_nested_state struct from userspace to the kernel. 4662For the definition of struct kvm_nested_state, see KVM_GET_NESTED_STATE. 4663 46644.116 KVM_(UN)REGISTER_COALESCED_MMIO 4665------------------------------------- 4666 4667:Capability: KVM_CAP_COALESCED_MMIO (for coalesced mmio) 4668 KVM_CAP_COALESCED_PIO (for coalesced pio) 4669:Architectures: all 4670:Type: vm ioctl 4671:Parameters: struct kvm_coalesced_mmio_zone 4672:Returns: 0 on success, < 0 on error 4673 4674Coalesced I/O is a performance optimization that defers hardware 4675register write emulation so that userspace exits are avoided. It is 4676typically used to reduce the overhead of emulating frequently accessed 4677hardware registers. 4678 4679When a hardware register is configured for coalesced I/O, write accesses 4680do not exit to userspace and their value is recorded in a ring buffer 4681that is shared between kernel and userspace. 4682 4683Coalesced I/O is used if one or more write accesses to a hardware 4684register can be deferred until a read or a write to another hardware 4685register on the same device. This last access will cause a vmexit and 4686userspace will process accesses from the ring buffer before emulating 4687it. That will avoid exiting to userspace on repeated writes. 4688 4689Coalesced pio is based on coalesced mmio. There is little difference 4690between coalesced mmio and pio except that coalesced pio records accesses 4691to I/O ports. 4692 46934.117 KVM_CLEAR_DIRTY_LOG (vm ioctl) 4694------------------------------------ 4695 4696:Capability: KVM_CAP_MANUAL_DIRTY_LOG_PROTECT2 4697:Architectures: x86, arm, arm64, mips 4698:Type: vm ioctl 4699:Parameters: struct kvm_clear_dirty_log (in) 4700:Returns: 0 on success, -1 on error 4701 4702:: 4703 4704 /* for KVM_CLEAR_DIRTY_LOG */ 4705 struct kvm_clear_dirty_log { 4706 __u32 slot; 4707 __u32 num_pages; 4708 __u64 first_page; 4709 union { 4710 void __user *dirty_bitmap; /* one bit per page */ 4711 __u64 padding; 4712 }; 4713 }; 4714 4715The ioctl clears the dirty status of pages in a memory slot, according to 4716the bitmap that is passed in struct kvm_clear_dirty_log's dirty_bitmap 4717field. Bit 0 of the bitmap corresponds to page "first_page" in the 4718memory slot, and num_pages is the size in bits of the input bitmap. 4719first_page must be a multiple of 64; num_pages must also be a multiple of 472064 unless first_page + num_pages is the size of the memory slot. For each 4721bit that is set in the input bitmap, the corresponding page is marked "clean" 4722in KVM's dirty bitmap, and dirty tracking is re-enabled for that page 4723(for example via write-protection, or by clearing the dirty bit in 4724a page table entry). 4725 4726If KVM_CAP_MULTI_ADDRESS_SPACE is available, bits 16-31 of slot field specifies 4727the address space for which you want to clear the dirty status. See 4728KVM_SET_USER_MEMORY_REGION for details on the usage of slot field. 4729 4730This ioctl is mostly useful when KVM_CAP_MANUAL_DIRTY_LOG_PROTECT2 4731is enabled; for more information, see the description of the capability. 4732However, it can always be used as long as KVM_CHECK_EXTENSION confirms 4733that KVM_CAP_MANUAL_DIRTY_LOG_PROTECT2 is present. 4734 47354.118 KVM_GET_SUPPORTED_HV_CPUID 4736-------------------------------- 4737 4738:Capability: KVM_CAP_HYPERV_CPUID (vcpu), KVM_CAP_SYS_HYPERV_CPUID (system) 4739:Architectures: x86 4740:Type: system ioctl, vcpu ioctl 4741:Parameters: struct kvm_cpuid2 (in/out) 4742:Returns: 0 on success, -1 on error 4743 4744:: 4745 4746 struct kvm_cpuid2 { 4747 __u32 nent; 4748 __u32 padding; 4749 struct kvm_cpuid_entry2 entries[0]; 4750 }; 4751 4752 struct kvm_cpuid_entry2 { 4753 __u32 function; 4754 __u32 index; 4755 __u32 flags; 4756 __u32 eax; 4757 __u32 ebx; 4758 __u32 ecx; 4759 __u32 edx; 4760 __u32 padding[3]; 4761 }; 4762 4763This ioctl returns x86 cpuid features leaves related to Hyper-V emulation in 4764KVM. Userspace can use the information returned by this ioctl to construct 4765cpuid information presented to guests consuming Hyper-V enlightenments (e.g. 4766Windows or Hyper-V guests). 4767 4768CPUID feature leaves returned by this ioctl are defined by Hyper-V Top Level 4769Functional Specification (TLFS). These leaves can't be obtained with 4770KVM_GET_SUPPORTED_CPUID ioctl because some of them intersect with KVM feature 4771leaves (0x40000000, 0x40000001). 4772 4773Currently, the following list of CPUID leaves are returned: 4774 4775 - HYPERV_CPUID_VENDOR_AND_MAX_FUNCTIONS 4776 - HYPERV_CPUID_INTERFACE 4777 - HYPERV_CPUID_VERSION 4778 - HYPERV_CPUID_FEATURES 4779 - HYPERV_CPUID_ENLIGHTMENT_INFO 4780 - HYPERV_CPUID_IMPLEMENT_LIMITS 4781 - HYPERV_CPUID_NESTED_FEATURES 4782 - HYPERV_CPUID_SYNDBG_VENDOR_AND_MAX_FUNCTIONS 4783 - HYPERV_CPUID_SYNDBG_INTERFACE 4784 - HYPERV_CPUID_SYNDBG_PLATFORM_CAPABILITIES 4785 4786Userspace invokes KVM_GET_SUPPORTED_HV_CPUID by passing a kvm_cpuid2 structure 4787with the 'nent' field indicating the number of entries in the variable-size 4788array 'entries'. If the number of entries is too low to describe all Hyper-V 4789feature leaves, an error (E2BIG) is returned. If the number is more or equal 4790to the number of Hyper-V feature leaves, the 'nent' field is adjusted to the 4791number of valid entries in the 'entries' array, which is then filled. 4792 4793'index' and 'flags' fields in 'struct kvm_cpuid_entry2' are currently reserved, 4794userspace should not expect to get any particular value there. 4795 4796Note, vcpu version of KVM_GET_SUPPORTED_HV_CPUID is currently deprecated. Unlike 4797system ioctl which exposes all supported feature bits unconditionally, vcpu 4798version has the following quirks: 4799 4800- HYPERV_CPUID_NESTED_FEATURES leaf and HV_X64_ENLIGHTENED_VMCS_RECOMMENDED 4801 feature bit are only exposed when Enlightened VMCS was previously enabled 4802 on the corresponding vCPU (KVM_CAP_HYPERV_ENLIGHTENED_VMCS). 4803- HV_STIMER_DIRECT_MODE_AVAILABLE bit is only exposed with in-kernel LAPIC. 4804 (presumes KVM_CREATE_IRQCHIP has already been called). 4805 48064.119 KVM_ARM_VCPU_FINALIZE 4807--------------------------- 4808 4809:Architectures: arm, arm64 4810:Type: vcpu ioctl 4811:Parameters: int feature (in) 4812:Returns: 0 on success, -1 on error 4813 4814Errors: 4815 4816 ====== ============================================================== 4817 EPERM feature not enabled, needs configuration, or already finalized 4818 EINVAL feature unknown or not present 4819 ====== ============================================================== 4820 4821Recognised values for feature: 4822 4823 ===== =========================================== 4824 arm64 KVM_ARM_VCPU_SVE (requires KVM_CAP_ARM_SVE) 4825 ===== =========================================== 4826 4827Finalizes the configuration of the specified vcpu feature. 4828 4829The vcpu must already have been initialised, enabling the affected feature, by 4830means of a successful KVM_ARM_VCPU_INIT call with the appropriate flag set in 4831features[]. 4832 4833For affected vcpu features, this is a mandatory step that must be performed 4834before the vcpu is fully usable. 4835 4836Between KVM_ARM_VCPU_INIT and KVM_ARM_VCPU_FINALIZE, the feature may be 4837configured by use of ioctls such as KVM_SET_ONE_REG. The exact configuration 4838that should be performaned and how to do it are feature-dependent. 4839 4840Other calls that depend on a particular feature being finalized, such as 4841KVM_RUN, KVM_GET_REG_LIST, KVM_GET_ONE_REG and KVM_SET_ONE_REG, will fail with 4842-EPERM unless the feature has already been finalized by means of a 4843KVM_ARM_VCPU_FINALIZE call. 4844 4845See KVM_ARM_VCPU_INIT for details of vcpu features that require finalization 4846using this ioctl. 4847 48484.120 KVM_SET_PMU_EVENT_FILTER 4849------------------------------ 4850 4851:Capability: KVM_CAP_PMU_EVENT_FILTER 4852:Architectures: x86 4853:Type: vm ioctl 4854:Parameters: struct kvm_pmu_event_filter (in) 4855:Returns: 0 on success, -1 on error 4856 4857:: 4858 4859 struct kvm_pmu_event_filter { 4860 __u32 action; 4861 __u32 nevents; 4862 __u32 fixed_counter_bitmap; 4863 __u32 flags; 4864 __u32 pad[4]; 4865 __u64 events[0]; 4866 }; 4867 4868This ioctl restricts the set of PMU events that the guest can program. 4869The argument holds a list of events which will be allowed or denied. 4870The eventsel+umask of each event the guest attempts to program is compared 4871against the events field to determine whether the guest should have access. 4872The events field only controls general purpose counters; fixed purpose 4873counters are controlled by the fixed_counter_bitmap. 4874 4875No flags are defined yet, the field must be zero. 4876 4877Valid values for 'action':: 4878 4879 #define KVM_PMU_EVENT_ALLOW 0 4880 #define KVM_PMU_EVENT_DENY 1 4881 48824.121 KVM_PPC_SVM_OFF 4883--------------------- 4884 4885:Capability: basic 4886:Architectures: powerpc 4887:Type: vm ioctl 4888:Parameters: none 4889:Returns: 0 on successful completion, 4890 4891Errors: 4892 4893 ====== ================================================================ 4894 EINVAL if ultravisor failed to terminate the secure guest 4895 ENOMEM if hypervisor failed to allocate new radix page tables for guest 4896 ====== ================================================================ 4897 4898This ioctl is used to turn off the secure mode of the guest or transition 4899the guest from secure mode to normal mode. This is invoked when the guest 4900is reset. This has no effect if called for a normal guest. 4901 4902This ioctl issues an ultravisor call to terminate the secure guest, 4903unpins the VPA pages and releases all the device pages that are used to 4904track the secure pages by hypervisor. 4905 49064.122 KVM_S390_NORMAL_RESET 4907--------------------------- 4908 4909:Capability: KVM_CAP_S390_VCPU_RESETS 4910:Architectures: s390 4911:Type: vcpu ioctl 4912:Parameters: none 4913:Returns: 0 4914 4915This ioctl resets VCPU registers and control structures according to 4916the cpu reset definition in the POP (Principles Of Operation). 4917 49184.123 KVM_S390_INITIAL_RESET 4919---------------------------- 4920 4921:Capability: none 4922:Architectures: s390 4923:Type: vcpu ioctl 4924:Parameters: none 4925:Returns: 0 4926 4927This ioctl resets VCPU registers and control structures according to 4928the initial cpu reset definition in the POP. However, the cpu is not 4929put into ESA mode. This reset is a superset of the normal reset. 4930 49314.124 KVM_S390_CLEAR_RESET 4932-------------------------- 4933 4934:Capability: KVM_CAP_S390_VCPU_RESETS 4935:Architectures: s390 4936:Type: vcpu ioctl 4937:Parameters: none 4938:Returns: 0 4939 4940This ioctl resets VCPU registers and control structures according to 4941the clear cpu reset definition in the POP. However, the cpu is not put 4942into ESA mode. This reset is a superset of the initial reset. 4943 4944 49454.125 KVM_S390_PV_COMMAND 4946------------------------- 4947 4948:Capability: KVM_CAP_S390_PROTECTED 4949:Architectures: s390 4950:Type: vm ioctl 4951:Parameters: struct kvm_pv_cmd 4952:Returns: 0 on success, < 0 on error 4953 4954:: 4955 4956 struct kvm_pv_cmd { 4957 __u32 cmd; /* Command to be executed */ 4958 __u16 rc; /* Ultravisor return code */ 4959 __u16 rrc; /* Ultravisor return reason code */ 4960 __u64 data; /* Data or address */ 4961 __u32 flags; /* flags for future extensions. Must be 0 for now */ 4962 __u32 reserved[3]; 4963 }; 4964 4965cmd values: 4966 4967KVM_PV_ENABLE 4968 Allocate memory and register the VM with the Ultravisor, thereby 4969 donating memory to the Ultravisor that will become inaccessible to 4970 KVM. All existing CPUs are converted to protected ones. After this 4971 command has succeeded, any CPU added via hotplug will become 4972 protected during its creation as well. 4973 4974 Errors: 4975 4976 ===== ============================= 4977 EINTR an unmasked signal is pending 4978 ===== ============================= 4979 4980KVM_PV_DISABLE 4981 4982 Deregister the VM from the Ultravisor and reclaim the memory that 4983 had been donated to the Ultravisor, making it usable by the kernel 4984 again. All registered VCPUs are converted back to non-protected 4985 ones. 4986 4987KVM_PV_VM_SET_SEC_PARMS 4988 Pass the image header from VM memory to the Ultravisor in 4989 preparation of image unpacking and verification. 4990 4991KVM_PV_VM_UNPACK 4992 Unpack (protect and decrypt) a page of the encrypted boot image. 4993 4994KVM_PV_VM_VERIFY 4995 Verify the integrity of the unpacked image. Only if this succeeds, 4996 KVM is allowed to start protected VCPUs. 4997 49984.126 KVM_X86_SET_MSR_FILTER 4999---------------------------- 5000 5001:Capability: KVM_CAP_X86_MSR_FILTER 5002:Architectures: x86 5003:Type: vm ioctl 5004:Parameters: struct kvm_msr_filter 5005:Returns: 0 on success, < 0 on error 5006 5007:: 5008 5009 struct kvm_msr_filter_range { 5010 #define KVM_MSR_FILTER_READ (1 << 0) 5011 #define KVM_MSR_FILTER_WRITE (1 << 1) 5012 __u32 flags; 5013 __u32 nmsrs; /* number of msrs in bitmap */ 5014 __u32 base; /* MSR index the bitmap starts at */ 5015 __u8 *bitmap; /* a 1 bit allows the operations in flags, 0 denies */ 5016 }; 5017 5018 #define KVM_MSR_FILTER_MAX_RANGES 16 5019 struct kvm_msr_filter { 5020 #define KVM_MSR_FILTER_DEFAULT_ALLOW (0 << 0) 5021 #define KVM_MSR_FILTER_DEFAULT_DENY (1 << 0) 5022 __u32 flags; 5023 struct kvm_msr_filter_range ranges[KVM_MSR_FILTER_MAX_RANGES]; 5024 }; 5025 5026flags values for ``struct kvm_msr_filter_range``: 5027 5028``KVM_MSR_FILTER_READ`` 5029 5030 Filter read accesses to MSRs using the given bitmap. A 0 in the bitmap 5031 indicates that a read should immediately fail, while a 1 indicates that 5032 a read for a particular MSR should be handled regardless of the default 5033 filter action. 5034 5035``KVM_MSR_FILTER_WRITE`` 5036 5037 Filter write accesses to MSRs using the given bitmap. A 0 in the bitmap 5038 indicates that a write should immediately fail, while a 1 indicates that 5039 a write for a particular MSR should be handled regardless of the default 5040 filter action. 5041 5042``KVM_MSR_FILTER_READ | KVM_MSR_FILTER_WRITE`` 5043 5044 Filter both read and write accesses to MSRs using the given bitmap. A 0 5045 in the bitmap indicates that both reads and writes should immediately fail, 5046 while a 1 indicates that reads and writes for a particular MSR are not 5047 filtered by this range. 5048 5049flags values for ``struct kvm_msr_filter``: 5050 5051``KVM_MSR_FILTER_DEFAULT_ALLOW`` 5052 5053 If no filter range matches an MSR index that is getting accessed, KVM will 5054 fall back to allowing access to the MSR. 5055 5056``KVM_MSR_FILTER_DEFAULT_DENY`` 5057 5058 If no filter range matches an MSR index that is getting accessed, KVM will 5059 fall back to rejecting access to the MSR. In this mode, all MSRs that should 5060 be processed by KVM need to explicitly be marked as allowed in the bitmaps. 5061 5062This ioctl allows user space to define up to 16 bitmaps of MSR ranges to 5063specify whether a certain MSR access should be explicitly filtered for or not. 5064 5065If this ioctl has never been invoked, MSR accesses are not guarded and the 5066default KVM in-kernel emulation behavior is fully preserved. 5067 5068Calling this ioctl with an empty set of ranges (all nmsrs == 0) disables MSR 5069filtering. In that mode, ``KVM_MSR_FILTER_DEFAULT_DENY`` is invalid and causes 5070an error. 5071 5072As soon as the filtering is in place, every MSR access is processed through 5073the filtering except for accesses to the x2APIC MSRs (from 0x800 to 0x8ff); 5074x2APIC MSRs are always allowed, independent of the ``default_allow`` setting, 5075and their behavior depends on the ``X2APIC_ENABLE`` bit of the APIC base 5076register. 5077 5078If a bit is within one of the defined ranges, read and write accesses are 5079guarded by the bitmap's value for the MSR index if the kind of access 5080is included in the ``struct kvm_msr_filter_range`` flags. If no range 5081cover this particular access, the behavior is determined by the flags 5082field in the kvm_msr_filter struct: ``KVM_MSR_FILTER_DEFAULT_ALLOW`` 5083and ``KVM_MSR_FILTER_DEFAULT_DENY``. 5084 5085Each bitmap range specifies a range of MSRs to potentially allow access on. 5086The range goes from MSR index [base .. base+nmsrs]. The flags field 5087indicates whether reads, writes or both reads and writes are filtered 5088by setting a 1 bit in the bitmap for the corresponding MSR index. 5089 5090If an MSR access is not permitted through the filtering, it generates a 5091#GP inside the guest. When combined with KVM_CAP_X86_USER_SPACE_MSR, that 5092allows user space to deflect and potentially handle various MSR accesses 5093into user space. 5094 5095Note, invoking this ioctl with a vCPU is running is inherently racy. However, 5096KVM does guarantee that vCPUs will see either the previous filter or the new 5097filter, e.g. MSRs with identical settings in both the old and new filter will 5098have deterministic behavior. 5099 51004.127 KVM_XEN_HVM_SET_ATTR 5101-------------------------- 5102 5103:Capability: KVM_CAP_XEN_HVM / KVM_XEN_HVM_CONFIG_SHARED_INFO 5104:Architectures: x86 5105:Type: vm ioctl 5106:Parameters: struct kvm_xen_hvm_attr 5107:Returns: 0 on success, < 0 on error 5108 5109:: 5110 5111 struct kvm_xen_hvm_attr { 5112 __u16 type; 5113 __u16 pad[3]; 5114 union { 5115 __u8 long_mode; 5116 __u8 vector; 5117 struct { 5118 __u64 gfn; 5119 } shared_info; 5120 __u64 pad[4]; 5121 } u; 5122 }; 5123 5124type values: 5125 5126KVM_XEN_ATTR_TYPE_LONG_MODE 5127 Sets the ABI mode of the VM to 32-bit or 64-bit (long mode). This 5128 determines the layout of the shared info pages exposed to the VM. 5129 5130KVM_XEN_ATTR_TYPE_SHARED_INFO 5131 Sets the guest physical frame number at which the Xen "shared info" 5132 page resides. Note that although Xen places vcpu_info for the first 5133 32 vCPUs in the shared_info page, KVM does not automatically do so 5134 and instead requires that KVM_XEN_VCPU_ATTR_TYPE_VCPU_INFO be used 5135 explicitly even when the vcpu_info for a given vCPU resides at the 5136 "default" location in the shared_info page. This is because KVM is 5137 not aware of the Xen CPU id which is used as the index into the 5138 vcpu_info[] array, so cannot know the correct default location. 5139 5140 Note that the shared info page may be constantly written to by KVM; 5141 it contains the event channel bitmap used to deliver interrupts to 5142 a Xen guest, amongst other things. It is exempt from dirty tracking 5143 mechanisms — KVM will not explicitly mark the page as dirty each 5144 time an event channel interrupt is delivered to the guest! Thus, 5145 userspace should always assume that the designated GFN is dirty if 5146 any vCPU has been running or any event channel interrupts can be 5147 routed to the guest. 5148 5149KVM_XEN_ATTR_TYPE_UPCALL_VECTOR 5150 Sets the exception vector used to deliver Xen event channel upcalls. 5151 51524.127 KVM_XEN_HVM_GET_ATTR 5153-------------------------- 5154 5155:Capability: KVM_CAP_XEN_HVM / KVM_XEN_HVM_CONFIG_SHARED_INFO 5156:Architectures: x86 5157:Type: vm ioctl 5158:Parameters: struct kvm_xen_hvm_attr 5159:Returns: 0 on success, < 0 on error 5160 5161Allows Xen VM attributes to be read. For the structure and types, 5162see KVM_XEN_HVM_SET_ATTR above. 5163 51644.128 KVM_XEN_VCPU_SET_ATTR 5165--------------------------- 5166 5167:Capability: KVM_CAP_XEN_HVM / KVM_XEN_HVM_CONFIG_SHARED_INFO 5168:Architectures: x86 5169:Type: vcpu ioctl 5170:Parameters: struct kvm_xen_vcpu_attr 5171:Returns: 0 on success, < 0 on error 5172 5173:: 5174 5175 struct kvm_xen_vcpu_attr { 5176 __u16 type; 5177 __u16 pad[3]; 5178 union { 5179 __u64 gpa; 5180 __u64 pad[4]; 5181 struct { 5182 __u64 state; 5183 __u64 state_entry_time; 5184 __u64 time_running; 5185 __u64 time_runnable; 5186 __u64 time_blocked; 5187 __u64 time_offline; 5188 } runstate; 5189 } u; 5190 }; 5191 5192type values: 5193 5194KVM_XEN_VCPU_ATTR_TYPE_VCPU_INFO 5195 Sets the guest physical address of the vcpu_info for a given vCPU. 5196 5197KVM_XEN_VCPU_ATTR_TYPE_VCPU_TIME_INFO 5198 Sets the guest physical address of an additional pvclock structure 5199 for a given vCPU. This is typically used for guest vsyscall support. 5200 5201KVM_XEN_VCPU_ATTR_TYPE_RUNSTATE_ADDR 5202 Sets the guest physical address of the vcpu_runstate_info for a given 5203 vCPU. This is how a Xen guest tracks CPU state such as steal time. 5204 5205KVM_XEN_VCPU_ATTR_TYPE_RUNSTATE_CURRENT 5206 Sets the runstate (RUNSTATE_running/_runnable/_blocked/_offline) of 5207 the given vCPU from the .u.runstate.state member of the structure. 5208 KVM automatically accounts running and runnable time but blocked 5209 and offline states are only entered explicitly. 5210 5211KVM_XEN_VCPU_ATTR_TYPE_RUNSTATE_DATA 5212 Sets all fields of the vCPU runstate data from the .u.runstate member 5213 of the structure, including the current runstate. The state_entry_time 5214 must equal the sum of the other four times. 5215 5216KVM_XEN_VCPU_ATTR_TYPE_RUNSTATE_ADJUST 5217 This *adds* the contents of the .u.runstate members of the structure 5218 to the corresponding members of the given vCPU's runstate data, thus 5219 permitting atomic adjustments to the runstate times. The adjustment 5220 to the state_entry_time must equal the sum of the adjustments to the 5221 other four times. The state field must be set to -1, or to a valid 5222 runstate value (RUNSTATE_running, RUNSTATE_runnable, RUNSTATE_blocked 5223 or RUNSTATE_offline) to set the current accounted state as of the 5224 adjusted state_entry_time. 5225 52264.129 KVM_XEN_VCPU_GET_ATTR 5227--------------------------- 5228 5229:Capability: KVM_CAP_XEN_HVM / KVM_XEN_HVM_CONFIG_SHARED_INFO 5230:Architectures: x86 5231:Type: vcpu ioctl 5232:Parameters: struct kvm_xen_vcpu_attr 5233:Returns: 0 on success, < 0 on error 5234 5235Allows Xen vCPU attributes to be read. For the structure and types, 5236see KVM_XEN_VCPU_SET_ATTR above. 5237 5238The KVM_XEN_VCPU_ATTR_TYPE_RUNSTATE_ADJUST type may not be used 5239with the KVM_XEN_VCPU_GET_ATTR ioctl. 5240 52414.130 KVM_ARM_MTE_COPY_TAGS 5242--------------------------- 5243 5244:Capability: KVM_CAP_ARM_MTE 5245:Architectures: arm64 5246:Type: vm ioctl 5247:Parameters: struct kvm_arm_copy_mte_tags 5248:Returns: number of bytes copied, < 0 on error (-EINVAL for incorrect 5249 arguments, -EFAULT if memory cannot be accessed). 5250 5251:: 5252 5253 struct kvm_arm_copy_mte_tags { 5254 __u64 guest_ipa; 5255 __u64 length; 5256 void __user *addr; 5257 __u64 flags; 5258 __u64 reserved[2]; 5259 }; 5260 5261Copies Memory Tagging Extension (MTE) tags to/from guest tag memory. The 5262``guest_ipa`` and ``length`` fields must be ``PAGE_SIZE`` aligned. The ``addr`` 5263field must point to a buffer which the tags will be copied to or from. 5264 5265``flags`` specifies the direction of copy, either ``KVM_ARM_TAGS_TO_GUEST`` or 5266``KVM_ARM_TAGS_FROM_GUEST``. 5267 5268The size of the buffer to store the tags is ``(length / 16)`` bytes 5269(granules in MTE are 16 bytes long). Each byte contains a single tag 5270value. This matches the format of ``PTRACE_PEEKMTETAGS`` and 5271``PTRACE_POKEMTETAGS``. 5272 5273If an error occurs before any data is copied then a negative error code is 5274returned. If some tags have been copied before an error occurs then the number 5275of bytes successfully copied is returned. If the call completes successfully 5276then ``length`` is returned. 5277 52784.131 KVM_GET_SREGS2 5279-------------------- 5280 5281:Capability: KVM_CAP_SREGS2 5282:Architectures: x86 5283:Type: vcpu ioctl 5284:Parameters: struct kvm_sregs2 (out) 5285:Returns: 0 on success, -1 on error 5286 5287Reads special registers from the vcpu. 5288This ioctl (when supported) replaces the KVM_GET_SREGS. 5289 5290:: 5291 5292 struct kvm_sregs2 { 5293 /* out (KVM_GET_SREGS2) / in (KVM_SET_SREGS2) */ 5294 struct kvm_segment cs, ds, es, fs, gs, ss; 5295 struct kvm_segment tr, ldt; 5296 struct kvm_dtable gdt, idt; 5297 __u64 cr0, cr2, cr3, cr4, cr8; 5298 __u64 efer; 5299 __u64 apic_base; 5300 __u64 flags; 5301 __u64 pdptrs[4]; 5302 }; 5303 5304flags values for ``kvm_sregs2``: 5305 5306``KVM_SREGS2_FLAGS_PDPTRS_VALID`` 5307 5308 Indicates thats the struct contain valid PDPTR values. 5309 5310 53114.132 KVM_SET_SREGS2 5312-------------------- 5313 5314:Capability: KVM_CAP_SREGS2 5315:Architectures: x86 5316:Type: vcpu ioctl 5317:Parameters: struct kvm_sregs2 (in) 5318:Returns: 0 on success, -1 on error 5319 5320Writes special registers into the vcpu. 5321See KVM_GET_SREGS2 for the data structures. 5322This ioctl (when supported) replaces the KVM_SET_SREGS. 5323 53244.133 KVM_GET_STATS_FD 5325---------------------- 5326 5327:Capability: KVM_CAP_STATS_BINARY_FD 5328:Architectures: all 5329:Type: vm ioctl, vcpu ioctl 5330:Parameters: none 5331:Returns: statistics file descriptor on success, < 0 on error 5332 5333Errors: 5334 5335 ====== ====================================================== 5336 ENOMEM if the fd could not be created due to lack of memory 5337 EMFILE if the number of opened files exceeds the limit 5338 ====== ====================================================== 5339 5340The returned file descriptor can be used to read VM/vCPU statistics data in 5341binary format. The data in the file descriptor consists of four blocks 5342organized as follows: 5343 5344+-------------+ 5345| Header | 5346+-------------+ 5347| id string | 5348+-------------+ 5349| Descriptors | 5350+-------------+ 5351| Stats Data | 5352+-------------+ 5353 5354Apart from the header starting at offset 0, please be aware that it is 5355not guaranteed that the four blocks are adjacent or in the above order; 5356the offsets of the id, descriptors and data blocks are found in the 5357header. However, all four blocks are aligned to 64 bit offsets in the 5358file and they do not overlap. 5359 5360All blocks except the data block are immutable. Userspace can read them 5361only one time after retrieving the file descriptor, and then use ``pread`` or 5362``lseek`` to read the statistics repeatedly. 5363 5364All data is in system endianness. 5365 5366The format of the header is as follows:: 5367 5368 struct kvm_stats_header { 5369 __u32 flags; 5370 __u32 name_size; 5371 __u32 num_desc; 5372 __u32 id_offset; 5373 __u32 desc_offset; 5374 __u32 data_offset; 5375 }; 5376 5377The ``flags`` field is not used at the moment. It is always read as 0. 5378 5379The ``name_size`` field is the size (in byte) of the statistics name string 5380(including trailing '\0') which is contained in the "id string" block and 5381appended at the end of every descriptor. 5382 5383The ``num_desc`` field is the number of descriptors that are included in the 5384descriptor block. (The actual number of values in the data block may be 5385larger, since each descriptor may comprise more than one value). 5386 5387The ``id_offset`` field is the offset of the id string from the start of the 5388file indicated by the file descriptor. It is a multiple of 8. 5389 5390The ``desc_offset`` field is the offset of the Descriptors block from the start 5391of the file indicated by the file descriptor. It is a multiple of 8. 5392 5393The ``data_offset`` field is the offset of the Stats Data block from the start 5394of the file indicated by the file descriptor. It is a multiple of 8. 5395 5396The id string block contains a string which identifies the file descriptor on 5397which KVM_GET_STATS_FD was invoked. The size of the block, including the 5398trailing ``'\0'``, is indicated by the ``name_size`` field in the header. 5399 5400The descriptors block is only needed to be read once for the lifetime of the 5401file descriptor contains a sequence of ``struct kvm_stats_desc``, each followed 5402by a string of size ``name_size``. 5403:: 5404 5405 #define KVM_STATS_TYPE_SHIFT 0 5406 #define KVM_STATS_TYPE_MASK (0xF << KVM_STATS_TYPE_SHIFT) 5407 #define KVM_STATS_TYPE_CUMULATIVE (0x0 << KVM_STATS_TYPE_SHIFT) 5408 #define KVM_STATS_TYPE_INSTANT (0x1 << KVM_STATS_TYPE_SHIFT) 5409 #define KVM_STATS_TYPE_PEAK (0x2 << KVM_STATS_TYPE_SHIFT) 5410 #define KVM_STATS_TYPE_LINEAR_HIST (0x3 << KVM_STATS_TYPE_SHIFT) 5411 #define KVM_STATS_TYPE_LOG_HIST (0x4 << KVM_STATS_TYPE_SHIFT) 5412 #define KVM_STATS_TYPE_MAX KVM_STATS_TYPE_LOG_HIST 5413 5414 #define KVM_STATS_UNIT_SHIFT 4 5415 #define KVM_STATS_UNIT_MASK (0xF << KVM_STATS_UNIT_SHIFT) 5416 #define KVM_STATS_UNIT_NONE (0x0 << KVM_STATS_UNIT_SHIFT) 5417 #define KVM_STATS_UNIT_BYTES (0x1 << KVM_STATS_UNIT_SHIFT) 5418 #define KVM_STATS_UNIT_SECONDS (0x2 << KVM_STATS_UNIT_SHIFT) 5419 #define KVM_STATS_UNIT_CYCLES (0x3 << KVM_STATS_UNIT_SHIFT) 5420 #define KVM_STATS_UNIT_MAX KVM_STATS_UNIT_CYCLES 5421 5422 #define KVM_STATS_BASE_SHIFT 8 5423 #define KVM_STATS_BASE_MASK (0xF << KVM_STATS_BASE_SHIFT) 5424 #define KVM_STATS_BASE_POW10 (0x0 << KVM_STATS_BASE_SHIFT) 5425 #define KVM_STATS_BASE_POW2 (0x1 << KVM_STATS_BASE_SHIFT) 5426 #define KVM_STATS_BASE_MAX KVM_STATS_BASE_POW2 5427 5428 struct kvm_stats_desc { 5429 __u32 flags; 5430 __s16 exponent; 5431 __u16 size; 5432 __u32 offset; 5433 __u32 bucket_size; 5434 char name[]; 5435 }; 5436 5437The ``flags`` field contains the type and unit of the statistics data described 5438by this descriptor. Its endianness is CPU native. 5439The following flags are supported: 5440 5441Bits 0-3 of ``flags`` encode the type: 5442 5443 * ``KVM_STATS_TYPE_CUMULATIVE`` 5444 The statistics reports a cumulative count. The value of data can only be increased. 5445 Most of the counters used in KVM are of this type. 5446 The corresponding ``size`` field for this type is always 1. 5447 All cumulative statistics data are read/write. 5448 * ``KVM_STATS_TYPE_INSTANT`` 5449 The statistics reports an instantaneous value. Its value can be increased or 5450 decreased. This type is usually used as a measurement of some resources, 5451 like the number of dirty pages, the number of large pages, etc. 5452 All instant statistics are read only. 5453 The corresponding ``size`` field for this type is always 1. 5454 * ``KVM_STATS_TYPE_PEAK`` 5455 The statistics data reports a peak value, for example the maximum number 5456 of items in a hash table bucket, the longest time waited and so on. 5457 The value of data can only be increased. 5458 The corresponding ``size`` field for this type is always 1. 5459 * ``KVM_STATS_TYPE_LINEAR_HIST`` 5460 The statistic is reported as a linear histogram. The number of 5461 buckets is specified by the ``size`` field. The size of buckets is specified 5462 by the ``hist_param`` field. The range of the Nth bucket (1 <= N < ``size``) 5463 is [``hist_param``*(N-1), ``hist_param``*N), while the range of the last 5464 bucket is [``hist_param``*(``size``-1), +INF). (+INF means positive infinity 5465 value.) The bucket value indicates how many samples fell in the bucket's range. 5466 * ``KVM_STATS_TYPE_LOG_HIST`` 5467 The statistic is reported as a logarithmic histogram. The number of 5468 buckets is specified by the ``size`` field. The range of the first bucket is 5469 [0, 1), while the range of the last bucket is [pow(2, ``size``-2), +INF). 5470 Otherwise, The Nth bucket (1 < N < ``size``) covers 5471 [pow(2, N-2), pow(2, N-1)). The bucket value indicates how many samples fell 5472 in the bucket's range. 5473 5474Bits 4-7 of ``flags`` encode the unit: 5475 5476 * ``KVM_STATS_UNIT_NONE`` 5477 There is no unit for the value of statistics data. This usually means that 5478 the value is a simple counter of an event. 5479 * ``KVM_STATS_UNIT_BYTES`` 5480 It indicates that the statistics data is used to measure memory size, in the 5481 unit of Byte, KiByte, MiByte, GiByte, etc. The unit of the data is 5482 determined by the ``exponent`` field in the descriptor. 5483 * ``KVM_STATS_UNIT_SECONDS`` 5484 It indicates that the statistics data is used to measure time or latency. 5485 * ``KVM_STATS_UNIT_CYCLES`` 5486 It indicates that the statistics data is used to measure CPU clock cycles. 5487 5488Bits 8-11 of ``flags``, together with ``exponent``, encode the scale of the 5489unit: 5490 5491 * ``KVM_STATS_BASE_POW10`` 5492 The scale is based on power of 10. It is used for measurement of time and 5493 CPU clock cycles. For example, an exponent of -9 can be used with 5494 ``KVM_STATS_UNIT_SECONDS`` to express that the unit is nanoseconds. 5495 * ``KVM_STATS_BASE_POW2`` 5496 The scale is based on power of 2. It is used for measurement of memory size. 5497 For example, an exponent of 20 can be used with ``KVM_STATS_UNIT_BYTES`` to 5498 express that the unit is MiB. 5499 5500The ``size`` field is the number of values of this statistics data. Its 5501value is usually 1 for most of simple statistics. 1 means it contains an 5502unsigned 64bit data. 5503 5504The ``offset`` field is the offset from the start of Data Block to the start of 5505the corresponding statistics data. 5506 5507The ``bucket_size`` field is used as a parameter for histogram statistics data. 5508It is only used by linear histogram statistics data, specifying the size of a 5509bucket. 5510 5511The ``name`` field is the name string of the statistics data. The name string 5512starts at the end of ``struct kvm_stats_desc``. The maximum length including 5513the trailing ``'\0'``, is indicated by ``name_size`` in the header. 5514 5515The Stats Data block contains an array of 64-bit values in the same order 5516as the descriptors in Descriptors block. 5517 55185. The kvm_run structure 5519======================== 5520 5521Application code obtains a pointer to the kvm_run structure by 5522mmap()ing a vcpu fd. From that point, application code can control 5523execution by changing fields in kvm_run prior to calling the KVM_RUN 5524ioctl, and obtain information about the reason KVM_RUN returned by 5525looking up structure members. 5526 5527:: 5528 5529 struct kvm_run { 5530 /* in */ 5531 __u8 request_interrupt_window; 5532 5533Request that KVM_RUN return when it becomes possible to inject external 5534interrupts into the guest. Useful in conjunction with KVM_INTERRUPT. 5535 5536:: 5537 5538 __u8 immediate_exit; 5539 5540This field is polled once when KVM_RUN starts; if non-zero, KVM_RUN 5541exits immediately, returning -EINTR. In the common scenario where a 5542signal is used to "kick" a VCPU out of KVM_RUN, this field can be used 5543to avoid usage of KVM_SET_SIGNAL_MASK, which has worse scalability. 5544Rather than blocking the signal outside KVM_RUN, userspace can set up 5545a signal handler that sets run->immediate_exit to a non-zero value. 5546 5547This field is ignored if KVM_CAP_IMMEDIATE_EXIT is not available. 5548 5549:: 5550 5551 __u8 padding1[6]; 5552 5553 /* out */ 5554 __u32 exit_reason; 5555 5556When KVM_RUN has returned successfully (return value 0), this informs 5557application code why KVM_RUN has returned. Allowable values for this 5558field are detailed below. 5559 5560:: 5561 5562 __u8 ready_for_interrupt_injection; 5563 5564If request_interrupt_window has been specified, this field indicates 5565an interrupt can be injected now with KVM_INTERRUPT. 5566 5567:: 5568 5569 __u8 if_flag; 5570 5571The value of the current interrupt flag. Only valid if in-kernel 5572local APIC is not used. 5573 5574:: 5575 5576 __u16 flags; 5577 5578More architecture-specific flags detailing state of the VCPU that may 5579affect the device's behavior. Current defined flags:: 5580 5581 /* x86, set if the VCPU is in system management mode */ 5582 #define KVM_RUN_X86_SMM (1 << 0) 5583 /* x86, set if bus lock detected in VM */ 5584 #define KVM_RUN_BUS_LOCK (1 << 1) 5585 5586:: 5587 5588 /* in (pre_kvm_run), out (post_kvm_run) */ 5589 __u64 cr8; 5590 5591The value of the cr8 register. Only valid if in-kernel local APIC is 5592not used. Both input and output. 5593 5594:: 5595 5596 __u64 apic_base; 5597 5598The value of the APIC BASE msr. Only valid if in-kernel local 5599APIC is not used. Both input and output. 5600 5601:: 5602 5603 union { 5604 /* KVM_EXIT_UNKNOWN */ 5605 struct { 5606 __u64 hardware_exit_reason; 5607 } hw; 5608 5609If exit_reason is KVM_EXIT_UNKNOWN, the vcpu has exited due to unknown 5610reasons. Further architecture-specific information is available in 5611hardware_exit_reason. 5612 5613:: 5614 5615 /* KVM_EXIT_FAIL_ENTRY */ 5616 struct { 5617 __u64 hardware_entry_failure_reason; 5618 __u32 cpu; /* if KVM_LAST_CPU */ 5619 } fail_entry; 5620 5621If exit_reason is KVM_EXIT_FAIL_ENTRY, the vcpu could not be run due 5622to unknown reasons. Further architecture-specific information is 5623available in hardware_entry_failure_reason. 5624 5625:: 5626 5627 /* KVM_EXIT_EXCEPTION */ 5628 struct { 5629 __u32 exception; 5630 __u32 error_code; 5631 } ex; 5632 5633Unused. 5634 5635:: 5636 5637 /* KVM_EXIT_IO */ 5638 struct { 5639 #define KVM_EXIT_IO_IN 0 5640 #define KVM_EXIT_IO_OUT 1 5641 __u8 direction; 5642 __u8 size; /* bytes */ 5643 __u16 port; 5644 __u32 count; 5645 __u64 data_offset; /* relative to kvm_run start */ 5646 } io; 5647 5648If exit_reason is KVM_EXIT_IO, then the vcpu has 5649executed a port I/O instruction which could not be satisfied by kvm. 5650data_offset describes where the data is located (KVM_EXIT_IO_OUT) or 5651where kvm expects application code to place the data for the next 5652KVM_RUN invocation (KVM_EXIT_IO_IN). Data format is a packed array. 5653 5654:: 5655 5656 /* KVM_EXIT_DEBUG */ 5657 struct { 5658 struct kvm_debug_exit_arch arch; 5659 } debug; 5660 5661If the exit_reason is KVM_EXIT_DEBUG, then a vcpu is processing a debug event 5662for which architecture specific information is returned. 5663 5664:: 5665 5666 /* KVM_EXIT_MMIO */ 5667 struct { 5668 __u64 phys_addr; 5669 __u8 data[8]; 5670 __u32 len; 5671 __u8 is_write; 5672 } mmio; 5673 5674If exit_reason is KVM_EXIT_MMIO, then the vcpu has 5675executed a memory-mapped I/O instruction which could not be satisfied 5676by kvm. The 'data' member contains the written data if 'is_write' is 5677true, and should be filled by application code otherwise. 5678 5679The 'data' member contains, in its first 'len' bytes, the value as it would 5680appear if the VCPU performed a load or store of the appropriate width directly 5681to the byte array. 5682 5683.. note:: 5684 5685 For KVM_EXIT_IO, KVM_EXIT_MMIO, KVM_EXIT_OSI, KVM_EXIT_PAPR, KVM_EXIT_XEN, 5686 KVM_EXIT_EPR, KVM_EXIT_X86_RDMSR and KVM_EXIT_X86_WRMSR the corresponding 5687 operations are complete (and guest state is consistent) only after userspace 5688 has re-entered the kernel with KVM_RUN. The kernel side will first finish 5689 incomplete operations and then check for pending signals. 5690 5691 The pending state of the operation is not preserved in state which is 5692 visible to userspace, thus userspace should ensure that the operation is 5693 completed before performing a live migration. Userspace can re-enter the 5694 guest with an unmasked signal pending or with the immediate_exit field set 5695 to complete pending operations without allowing any further instructions 5696 to be executed. 5697 5698:: 5699 5700 /* KVM_EXIT_HYPERCALL */ 5701 struct { 5702 __u64 nr; 5703 __u64 args[6]; 5704 __u64 ret; 5705 __u32 longmode; 5706 __u32 pad; 5707 } hypercall; 5708 5709Unused. This was once used for 'hypercall to userspace'. To implement 5710such functionality, use KVM_EXIT_IO (x86) or KVM_EXIT_MMIO (all except s390). 5711 5712.. note:: KVM_EXIT_IO is significantly faster than KVM_EXIT_MMIO. 5713 5714:: 5715 5716 /* KVM_EXIT_TPR_ACCESS */ 5717 struct { 5718 __u64 rip; 5719 __u32 is_write; 5720 __u32 pad; 5721 } tpr_access; 5722 5723To be documented (KVM_TPR_ACCESS_REPORTING). 5724 5725:: 5726 5727 /* KVM_EXIT_S390_SIEIC */ 5728 struct { 5729 __u8 icptcode; 5730 __u64 mask; /* psw upper half */ 5731 __u64 addr; /* psw lower half */ 5732 __u16 ipa; 5733 __u32 ipb; 5734 } s390_sieic; 5735 5736s390 specific. 5737 5738:: 5739 5740 /* KVM_EXIT_S390_RESET */ 5741 #define KVM_S390_RESET_POR 1 5742 #define KVM_S390_RESET_CLEAR 2 5743 #define KVM_S390_RESET_SUBSYSTEM 4 5744 #define KVM_S390_RESET_CPU_INIT 8 5745 #define KVM_S390_RESET_IPL 16 5746 __u64 s390_reset_flags; 5747 5748s390 specific. 5749 5750:: 5751 5752 /* KVM_EXIT_S390_UCONTROL */ 5753 struct { 5754 __u64 trans_exc_code; 5755 __u32 pgm_code; 5756 } s390_ucontrol; 5757 5758s390 specific. A page fault has occurred for a user controlled virtual 5759machine (KVM_VM_S390_UNCONTROL) on it's host page table that cannot be 5760resolved by the kernel. 5761The program code and the translation exception code that were placed 5762in the cpu's lowcore are presented here as defined by the z Architecture 5763Principles of Operation Book in the Chapter for Dynamic Address Translation 5764(DAT) 5765 5766:: 5767 5768 /* KVM_EXIT_DCR */ 5769 struct { 5770 __u32 dcrn; 5771 __u32 data; 5772 __u8 is_write; 5773 } dcr; 5774 5775Deprecated - was used for 440 KVM. 5776 5777:: 5778 5779 /* KVM_EXIT_OSI */ 5780 struct { 5781 __u64 gprs[32]; 5782 } osi; 5783 5784MOL uses a special hypercall interface it calls 'OSI'. To enable it, we catch 5785hypercalls and exit with this exit struct that contains all the guest gprs. 5786 5787If exit_reason is KVM_EXIT_OSI, then the vcpu has triggered such a hypercall. 5788Userspace can now handle the hypercall and when it's done modify the gprs as 5789necessary. Upon guest entry all guest GPRs will then be replaced by the values 5790in this struct. 5791 5792:: 5793 5794 /* KVM_EXIT_PAPR_HCALL */ 5795 struct { 5796 __u64 nr; 5797 __u64 ret; 5798 __u64 args[9]; 5799 } papr_hcall; 5800 5801This is used on 64-bit PowerPC when emulating a pSeries partition, 5802e.g. with the 'pseries' machine type in qemu. It occurs when the 5803guest does a hypercall using the 'sc 1' instruction. The 'nr' field 5804contains the hypercall number (from the guest R3), and 'args' contains 5805the arguments (from the guest R4 - R12). Userspace should put the 5806return code in 'ret' and any extra returned values in args[]. 5807The possible hypercalls are defined in the Power Architecture Platform 5808Requirements (PAPR) document available from www.power.org (free 5809developer registration required to access it). 5810 5811:: 5812 5813 /* KVM_EXIT_S390_TSCH */ 5814 struct { 5815 __u16 subchannel_id; 5816 __u16 subchannel_nr; 5817 __u32 io_int_parm; 5818 __u32 io_int_word; 5819 __u32 ipb; 5820 __u8 dequeued; 5821 } s390_tsch; 5822 5823s390 specific. This exit occurs when KVM_CAP_S390_CSS_SUPPORT has been enabled 5824and TEST SUBCHANNEL was intercepted. If dequeued is set, a pending I/O 5825interrupt for the target subchannel has been dequeued and subchannel_id, 5826subchannel_nr, io_int_parm and io_int_word contain the parameters for that 5827interrupt. ipb is needed for instruction parameter decoding. 5828 5829:: 5830 5831 /* KVM_EXIT_EPR */ 5832 struct { 5833 __u32 epr; 5834 } epr; 5835 5836On FSL BookE PowerPC chips, the interrupt controller has a fast patch 5837interrupt acknowledge path to the core. When the core successfully 5838delivers an interrupt, it automatically populates the EPR register with 5839the interrupt vector number and acknowledges the interrupt inside 5840the interrupt controller. 5841 5842In case the interrupt controller lives in user space, we need to do 5843the interrupt acknowledge cycle through it to fetch the next to be 5844delivered interrupt vector using this exit. 5845 5846It gets triggered whenever both KVM_CAP_PPC_EPR are enabled and an 5847external interrupt has just been delivered into the guest. User space 5848should put the acknowledged interrupt vector into the 'epr' field. 5849 5850:: 5851 5852 /* KVM_EXIT_SYSTEM_EVENT */ 5853 struct { 5854 #define KVM_SYSTEM_EVENT_SHUTDOWN 1 5855 #define KVM_SYSTEM_EVENT_RESET 2 5856 #define KVM_SYSTEM_EVENT_CRASH 3 5857 __u32 type; 5858 __u64 flags; 5859 } system_event; 5860 5861If exit_reason is KVM_EXIT_SYSTEM_EVENT then the vcpu has triggered 5862a system-level event using some architecture specific mechanism (hypercall 5863or some special instruction). In case of ARM/ARM64, this is triggered using 5864HVC instruction based PSCI call from the vcpu. The 'type' field describes 5865the system-level event type. The 'flags' field describes architecture 5866specific flags for the system-level event. 5867 5868Valid values for 'type' are: 5869 5870 - KVM_SYSTEM_EVENT_SHUTDOWN -- the guest has requested a shutdown of the 5871 VM. Userspace is not obliged to honour this, and if it does honour 5872 this does not need to destroy the VM synchronously (ie it may call 5873 KVM_RUN again before shutdown finally occurs). 5874 - KVM_SYSTEM_EVENT_RESET -- the guest has requested a reset of the VM. 5875 As with SHUTDOWN, userspace can choose to ignore the request, or 5876 to schedule the reset to occur in the future and may call KVM_RUN again. 5877 - KVM_SYSTEM_EVENT_CRASH -- the guest crash occurred and the guest 5878 has requested a crash condition maintenance. Userspace can choose 5879 to ignore the request, or to gather VM memory core dump and/or 5880 reset/shutdown of the VM. 5881 5882:: 5883 5884 /* KVM_EXIT_IOAPIC_EOI */ 5885 struct { 5886 __u8 vector; 5887 } eoi; 5888 5889Indicates that the VCPU's in-kernel local APIC received an EOI for a 5890level-triggered IOAPIC interrupt. This exit only triggers when the 5891IOAPIC is implemented in userspace (i.e. KVM_CAP_SPLIT_IRQCHIP is enabled); 5892the userspace IOAPIC should process the EOI and retrigger the interrupt if 5893it is still asserted. Vector is the LAPIC interrupt vector for which the 5894EOI was received. 5895 5896:: 5897 5898 struct kvm_hyperv_exit { 5899 #define KVM_EXIT_HYPERV_SYNIC 1 5900 #define KVM_EXIT_HYPERV_HCALL 2 5901 #define KVM_EXIT_HYPERV_SYNDBG 3 5902 __u32 type; 5903 __u32 pad1; 5904 union { 5905 struct { 5906 __u32 msr; 5907 __u32 pad2; 5908 __u64 control; 5909 __u64 evt_page; 5910 __u64 msg_page; 5911 } synic; 5912 struct { 5913 __u64 input; 5914 __u64 result; 5915 __u64 params[2]; 5916 } hcall; 5917 struct { 5918 __u32 msr; 5919 __u32 pad2; 5920 __u64 control; 5921 __u64 status; 5922 __u64 send_page; 5923 __u64 recv_page; 5924 __u64 pending_page; 5925 } syndbg; 5926 } u; 5927 }; 5928 /* KVM_EXIT_HYPERV */ 5929 struct kvm_hyperv_exit hyperv; 5930 5931Indicates that the VCPU exits into userspace to process some tasks 5932related to Hyper-V emulation. 5933 5934Valid values for 'type' are: 5935 5936 - KVM_EXIT_HYPERV_SYNIC -- synchronously notify user-space about 5937 5938Hyper-V SynIC state change. Notification is used to remap SynIC 5939event/message pages and to enable/disable SynIC messages/events processing 5940in userspace. 5941 5942 - KVM_EXIT_HYPERV_SYNDBG -- synchronously notify user-space about 5943 5944Hyper-V Synthetic debugger state change. Notification is used to either update 5945the pending_page location or to send a control command (send the buffer located 5946in send_page or recv a buffer to recv_page). 5947 5948:: 5949 5950 /* KVM_EXIT_ARM_NISV */ 5951 struct { 5952 __u64 esr_iss; 5953 __u64 fault_ipa; 5954 } arm_nisv; 5955 5956Used on arm and arm64 systems. If a guest accesses memory not in a memslot, 5957KVM will typically return to userspace and ask it to do MMIO emulation on its 5958behalf. However, for certain classes of instructions, no instruction decode 5959(direction, length of memory access) is provided, and fetching and decoding 5960the instruction from the VM is overly complicated to live in the kernel. 5961 5962Historically, when this situation occurred, KVM would print a warning and kill 5963the VM. KVM assumed that if the guest accessed non-memslot memory, it was 5964trying to do I/O, which just couldn't be emulated, and the warning message was 5965phrased accordingly. However, what happened more often was that a guest bug 5966caused access outside the guest memory areas which should lead to a more 5967meaningful warning message and an external abort in the guest, if the access 5968did not fall within an I/O window. 5969 5970Userspace implementations can query for KVM_CAP_ARM_NISV_TO_USER, and enable 5971this capability at VM creation. Once this is done, these types of errors will 5972instead return to userspace with KVM_EXIT_ARM_NISV, with the valid bits from 5973the HSR (arm) and ESR_EL2 (arm64) in the esr_iss field, and the faulting IPA 5974in the fault_ipa field. Userspace can either fix up the access if it's 5975actually an I/O access by decoding the instruction from guest memory (if it's 5976very brave) and continue executing the guest, or it can decide to suspend, 5977dump, or restart the guest. 5978 5979Note that KVM does not skip the faulting instruction as it does for 5980KVM_EXIT_MMIO, but userspace has to emulate any change to the processing state 5981if it decides to decode and emulate the instruction. 5982 5983:: 5984 5985 /* KVM_EXIT_X86_RDMSR / KVM_EXIT_X86_WRMSR */ 5986 struct { 5987 __u8 error; /* user -> kernel */ 5988 __u8 pad[7]; 5989 __u32 reason; /* kernel -> user */ 5990 __u32 index; /* kernel -> user */ 5991 __u64 data; /* kernel <-> user */ 5992 } msr; 5993 5994Used on x86 systems. When the VM capability KVM_CAP_X86_USER_SPACE_MSR is 5995enabled, MSR accesses to registers that would invoke a #GP by KVM kernel code 5996will instead trigger a KVM_EXIT_X86_RDMSR exit for reads and KVM_EXIT_X86_WRMSR 5997exit for writes. 5998 5999The "reason" field specifies why the MSR trap occurred. User space will only 6000receive MSR exit traps when a particular reason was requested during through 6001ENABLE_CAP. Currently valid exit reasons are: 6002 6003 KVM_MSR_EXIT_REASON_UNKNOWN - access to MSR that is unknown to KVM 6004 KVM_MSR_EXIT_REASON_INVAL - access to invalid MSRs or reserved bits 6005 KVM_MSR_EXIT_REASON_FILTER - access blocked by KVM_X86_SET_MSR_FILTER 6006 6007For KVM_EXIT_X86_RDMSR, the "index" field tells user space which MSR the guest 6008wants to read. To respond to this request with a successful read, user space 6009writes the respective data into the "data" field and must continue guest 6010execution to ensure the read data is transferred into guest register state. 6011 6012If the RDMSR request was unsuccessful, user space indicates that with a "1" in 6013the "error" field. This will inject a #GP into the guest when the VCPU is 6014executed again. 6015 6016For KVM_EXIT_X86_WRMSR, the "index" field tells user space which MSR the guest 6017wants to write. Once finished processing the event, user space must continue 6018vCPU execution. If the MSR write was unsuccessful, user space also sets the 6019"error" field to "1". 6020 6021:: 6022 6023 6024 struct kvm_xen_exit { 6025 #define KVM_EXIT_XEN_HCALL 1 6026 __u32 type; 6027 union { 6028 struct { 6029 __u32 longmode; 6030 __u32 cpl; 6031 __u64 input; 6032 __u64 result; 6033 __u64 params[6]; 6034 } hcall; 6035 } u; 6036 }; 6037 /* KVM_EXIT_XEN */ 6038 struct kvm_hyperv_exit xen; 6039 6040Indicates that the VCPU exits into userspace to process some tasks 6041related to Xen emulation. 6042 6043Valid values for 'type' are: 6044 6045 - KVM_EXIT_XEN_HCALL -- synchronously notify user-space about Xen hypercall. 6046 Userspace is expected to place the hypercall result into the appropriate 6047 field before invoking KVM_RUN again. 6048 6049:: 6050 6051 /* KVM_EXIT_RISCV_SBI */ 6052 struct { 6053 unsigned long extension_id; 6054 unsigned long function_id; 6055 unsigned long args[6]; 6056 unsigned long ret[2]; 6057 } riscv_sbi; 6058If exit reason is KVM_EXIT_RISCV_SBI then it indicates that the VCPU has 6059done a SBI call which is not handled by KVM RISC-V kernel module. The details 6060of the SBI call are available in 'riscv_sbi' member of kvm_run structure. The 6061'extension_id' field of 'riscv_sbi' represents SBI extension ID whereas the 6062'function_id' field represents function ID of given SBI extension. The 'args' 6063array field of 'riscv_sbi' represents parameters for the SBI call and 'ret' 6064array field represents return values. The userspace should update the return 6065values of SBI call before resuming the VCPU. For more details on RISC-V SBI 6066spec refer, https://github.com/riscv/riscv-sbi-doc. 6067 6068:: 6069 6070 /* Fix the size of the union. */ 6071 char padding[256]; 6072 }; 6073 6074 /* 6075 * shared registers between kvm and userspace. 6076 * kvm_valid_regs specifies the register classes set by the host 6077 * kvm_dirty_regs specified the register classes dirtied by userspace 6078 * struct kvm_sync_regs is architecture specific, as well as the 6079 * bits for kvm_valid_regs and kvm_dirty_regs 6080 */ 6081 __u64 kvm_valid_regs; 6082 __u64 kvm_dirty_regs; 6083 union { 6084 struct kvm_sync_regs regs; 6085 char padding[SYNC_REGS_SIZE_BYTES]; 6086 } s; 6087 6088If KVM_CAP_SYNC_REGS is defined, these fields allow userspace to access 6089certain guest registers without having to call SET/GET_*REGS. Thus we can 6090avoid some system call overhead if userspace has to handle the exit. 6091Userspace can query the validity of the structure by checking 6092kvm_valid_regs for specific bits. These bits are architecture specific 6093and usually define the validity of a groups of registers. (e.g. one bit 6094for general purpose registers) 6095 6096Please note that the kernel is allowed to use the kvm_run structure as the 6097primary storage for certain register types. Therefore, the kernel may use the 6098values in kvm_run even if the corresponding bit in kvm_dirty_regs is not set. 6099 6100:: 6101 6102 }; 6103 6104 6105 61066. Capabilities that can be enabled on vCPUs 6107============================================ 6108 6109There are certain capabilities that change the behavior of the virtual CPU or 6110the virtual machine when enabled. To enable them, please see section 4.37. 6111Below you can find a list of capabilities and what their effect on the vCPU or 6112the virtual machine is when enabling them. 6113 6114The following information is provided along with the description: 6115 6116 Architectures: 6117 which instruction set architectures provide this ioctl. 6118 x86 includes both i386 and x86_64. 6119 6120 Target: 6121 whether this is a per-vcpu or per-vm capability. 6122 6123 Parameters: 6124 what parameters are accepted by the capability. 6125 6126 Returns: 6127 the return value. General error numbers (EBADF, ENOMEM, EINVAL) 6128 are not detailed, but errors with specific meanings are. 6129 6130 61316.1 KVM_CAP_PPC_OSI 6132------------------- 6133 6134:Architectures: ppc 6135:Target: vcpu 6136:Parameters: none 6137:Returns: 0 on success; -1 on error 6138 6139This capability enables interception of OSI hypercalls that otherwise would 6140be treated as normal system calls to be injected into the guest. OSI hypercalls 6141were invented by Mac-on-Linux to have a standardized communication mechanism 6142between the guest and the host. 6143 6144When this capability is enabled, KVM_EXIT_OSI can occur. 6145 6146 61476.2 KVM_CAP_PPC_PAPR 6148-------------------- 6149 6150:Architectures: ppc 6151:Target: vcpu 6152:Parameters: none 6153:Returns: 0 on success; -1 on error 6154 6155This capability enables interception of PAPR hypercalls. PAPR hypercalls are 6156done using the hypercall instruction "sc 1". 6157 6158It also sets the guest privilege level to "supervisor" mode. Usually the guest 6159runs in "hypervisor" privilege mode with a few missing features. 6160 6161In addition to the above, it changes the semantics of SDR1. In this mode, the 6162HTAB address part of SDR1 contains an HVA instead of a GPA, as PAPR keeps the 6163HTAB invisible to the guest. 6164 6165When this capability is enabled, KVM_EXIT_PAPR_HCALL can occur. 6166 6167 61686.3 KVM_CAP_SW_TLB 6169------------------ 6170 6171:Architectures: ppc 6172:Target: vcpu 6173:Parameters: args[0] is the address of a struct kvm_config_tlb 6174:Returns: 0 on success; -1 on error 6175 6176:: 6177 6178 struct kvm_config_tlb { 6179 __u64 params; 6180 __u64 array; 6181 __u32 mmu_type; 6182 __u32 array_len; 6183 }; 6184 6185Configures the virtual CPU's TLB array, establishing a shared memory area 6186between userspace and KVM. The "params" and "array" fields are userspace 6187addresses of mmu-type-specific data structures. The "array_len" field is an 6188safety mechanism, and should be set to the size in bytes of the memory that 6189userspace has reserved for the array. It must be at least the size dictated 6190by "mmu_type" and "params". 6191 6192While KVM_RUN is active, the shared region is under control of KVM. Its 6193contents are undefined, and any modification by userspace results in 6194boundedly undefined behavior. 6195 6196On return from KVM_RUN, the shared region will reflect the current state of 6197the guest's TLB. If userspace makes any changes, it must call KVM_DIRTY_TLB 6198to tell KVM which entries have been changed, prior to calling KVM_RUN again 6199on this vcpu. 6200 6201For mmu types KVM_MMU_FSL_BOOKE_NOHV and KVM_MMU_FSL_BOOKE_HV: 6202 6203 - The "params" field is of type "struct kvm_book3e_206_tlb_params". 6204 - The "array" field points to an array of type "struct 6205 kvm_book3e_206_tlb_entry". 6206 - The array consists of all entries in the first TLB, followed by all 6207 entries in the second TLB. 6208 - Within a TLB, entries are ordered first by increasing set number. Within a 6209 set, entries are ordered by way (increasing ESEL). 6210 - The hash for determining set number in TLB0 is: (MAS2 >> 12) & (num_sets - 1) 6211 where "num_sets" is the tlb_sizes[] value divided by the tlb_ways[] value. 6212 - The tsize field of mas1 shall be set to 4K on TLB0, even though the 6213 hardware ignores this value for TLB0. 6214 62156.4 KVM_CAP_S390_CSS_SUPPORT 6216---------------------------- 6217 6218:Architectures: s390 6219:Target: vcpu 6220:Parameters: none 6221:Returns: 0 on success; -1 on error 6222 6223This capability enables support for handling of channel I/O instructions. 6224 6225TEST PENDING INTERRUPTION and the interrupt portion of TEST SUBCHANNEL are 6226handled in-kernel, while the other I/O instructions are passed to userspace. 6227 6228When this capability is enabled, KVM_EXIT_S390_TSCH will occur on TEST 6229SUBCHANNEL intercepts. 6230 6231Note that even though this capability is enabled per-vcpu, the complete 6232virtual machine is affected. 6233 62346.5 KVM_CAP_PPC_EPR 6235------------------- 6236 6237:Architectures: ppc 6238:Target: vcpu 6239:Parameters: args[0] defines whether the proxy facility is active 6240:Returns: 0 on success; -1 on error 6241 6242This capability enables or disables the delivery of interrupts through the 6243external proxy facility. 6244 6245When enabled (args[0] != 0), every time the guest gets an external interrupt 6246delivered, it automatically exits into user space with a KVM_EXIT_EPR exit 6247to receive the topmost interrupt vector. 6248 6249When disabled (args[0] == 0), behavior is as if this facility is unsupported. 6250 6251When this capability is enabled, KVM_EXIT_EPR can occur. 6252 62536.6 KVM_CAP_IRQ_MPIC 6254-------------------- 6255 6256:Architectures: ppc 6257:Parameters: args[0] is the MPIC device fd; 6258 args[1] is the MPIC CPU number for this vcpu 6259 6260This capability connects the vcpu to an in-kernel MPIC device. 6261 62626.7 KVM_CAP_IRQ_XICS 6263-------------------- 6264 6265:Architectures: ppc 6266:Target: vcpu 6267:Parameters: args[0] is the XICS device fd; 6268 args[1] is the XICS CPU number (server ID) for this vcpu 6269 6270This capability connects the vcpu to an in-kernel XICS device. 6271 62726.8 KVM_CAP_S390_IRQCHIP 6273------------------------ 6274 6275:Architectures: s390 6276:Target: vm 6277:Parameters: none 6278 6279This capability enables the in-kernel irqchip for s390. Please refer to 6280"4.24 KVM_CREATE_IRQCHIP" for details. 6281 62826.9 KVM_CAP_MIPS_FPU 6283-------------------- 6284 6285:Architectures: mips 6286:Target: vcpu 6287:Parameters: args[0] is reserved for future use (should be 0). 6288 6289This capability allows the use of the host Floating Point Unit by the guest. It 6290allows the Config1.FP bit to be set to enable the FPU in the guest. Once this is 6291done the ``KVM_REG_MIPS_FPR_*`` and ``KVM_REG_MIPS_FCR_*`` registers can be 6292accessed (depending on the current guest FPU register mode), and the Status.FR, 6293Config5.FRE bits are accessible via the KVM API and also from the guest, 6294depending on them being supported by the FPU. 6295 62966.10 KVM_CAP_MIPS_MSA 6297--------------------- 6298 6299:Architectures: mips 6300:Target: vcpu 6301:Parameters: args[0] is reserved for future use (should be 0). 6302 6303This capability allows the use of the MIPS SIMD Architecture (MSA) by the guest. 6304It allows the Config3.MSAP bit to be set to enable the use of MSA by the guest. 6305Once this is done the ``KVM_REG_MIPS_VEC_*`` and ``KVM_REG_MIPS_MSA_*`` 6306registers can be accessed, and the Config5.MSAEn bit is accessible via the 6307KVM API and also from the guest. 6308 63096.74 KVM_CAP_SYNC_REGS 6310---------------------- 6311 6312:Architectures: s390, x86 6313:Target: s390: always enabled, x86: vcpu 6314:Parameters: none 6315:Returns: x86: KVM_CHECK_EXTENSION returns a bit-array indicating which register 6316 sets are supported 6317 (bitfields defined in arch/x86/include/uapi/asm/kvm.h). 6318 6319As described above in the kvm_sync_regs struct info in section 5 (kvm_run): 6320KVM_CAP_SYNC_REGS "allow[s] userspace to access certain guest registers 6321without having to call SET/GET_*REGS". This reduces overhead by eliminating 6322repeated ioctl calls for setting and/or getting register values. This is 6323particularly important when userspace is making synchronous guest state 6324modifications, e.g. when emulating and/or intercepting instructions in 6325userspace. 6326 6327For s390 specifics, please refer to the source code. 6328 6329For x86: 6330 6331- the register sets to be copied out to kvm_run are selectable 6332 by userspace (rather that all sets being copied out for every exit). 6333- vcpu_events are available in addition to regs and sregs. 6334 6335For x86, the 'kvm_valid_regs' field of struct kvm_run is overloaded to 6336function as an input bit-array field set by userspace to indicate the 6337specific register sets to be copied out on the next exit. 6338 6339To indicate when userspace has modified values that should be copied into 6340the vCPU, the all architecture bitarray field, 'kvm_dirty_regs' must be set. 6341This is done using the same bitflags as for the 'kvm_valid_regs' field. 6342If the dirty bit is not set, then the register set values will not be copied 6343into the vCPU even if they've been modified. 6344 6345Unused bitfields in the bitarrays must be set to zero. 6346 6347:: 6348 6349 struct kvm_sync_regs { 6350 struct kvm_regs regs; 6351 struct kvm_sregs sregs; 6352 struct kvm_vcpu_events events; 6353 }; 6354 63556.75 KVM_CAP_PPC_IRQ_XIVE 6356------------------------- 6357 6358:Architectures: ppc 6359:Target: vcpu 6360:Parameters: args[0] is the XIVE device fd; 6361 args[1] is the XIVE CPU number (server ID) for this vcpu 6362 6363This capability connects the vcpu to an in-kernel XIVE device. 6364 63657. Capabilities that can be enabled on VMs 6366========================================== 6367 6368There are certain capabilities that change the behavior of the virtual 6369machine when enabled. To enable them, please see section 4.37. Below 6370you can find a list of capabilities and what their effect on the VM 6371is when enabling them. 6372 6373The following information is provided along with the description: 6374 6375 Architectures: 6376 which instruction set architectures provide this ioctl. 6377 x86 includes both i386 and x86_64. 6378 6379 Parameters: 6380 what parameters are accepted by the capability. 6381 6382 Returns: 6383 the return value. General error numbers (EBADF, ENOMEM, EINVAL) 6384 are not detailed, but errors with specific meanings are. 6385 6386 63877.1 KVM_CAP_PPC_ENABLE_HCALL 6388---------------------------- 6389 6390:Architectures: ppc 6391:Parameters: args[0] is the sPAPR hcall number; 6392 args[1] is 0 to disable, 1 to enable in-kernel handling 6393 6394This capability controls whether individual sPAPR hypercalls (hcalls) 6395get handled by the kernel or not. Enabling or disabling in-kernel 6396handling of an hcall is effective across the VM. On creation, an 6397initial set of hcalls are enabled for in-kernel handling, which 6398consists of those hcalls for which in-kernel handlers were implemented 6399before this capability was implemented. If disabled, the kernel will 6400not to attempt to handle the hcall, but will always exit to userspace 6401to handle it. Note that it may not make sense to enable some and 6402disable others of a group of related hcalls, but KVM does not prevent 6403userspace from doing that. 6404 6405If the hcall number specified is not one that has an in-kernel 6406implementation, the KVM_ENABLE_CAP ioctl will fail with an EINVAL 6407error. 6408 64097.2 KVM_CAP_S390_USER_SIGP 6410-------------------------- 6411 6412:Architectures: s390 6413:Parameters: none 6414 6415This capability controls which SIGP orders will be handled completely in user 6416space. With this capability enabled, all fast orders will be handled completely 6417in the kernel: 6418 6419- SENSE 6420- SENSE RUNNING 6421- EXTERNAL CALL 6422- EMERGENCY SIGNAL 6423- CONDITIONAL EMERGENCY SIGNAL 6424 6425All other orders will be handled completely in user space. 6426 6427Only privileged operation exceptions will be checked for in the kernel (or even 6428in the hardware prior to interception). If this capability is not enabled, the 6429old way of handling SIGP orders is used (partially in kernel and user space). 6430 64317.3 KVM_CAP_S390_VECTOR_REGISTERS 6432--------------------------------- 6433 6434:Architectures: s390 6435:Parameters: none 6436:Returns: 0 on success, negative value on error 6437 6438Allows use of the vector registers introduced with z13 processor, and 6439provides for the synchronization between host and user space. Will 6440return -EINVAL if the machine does not support vectors. 6441 64427.4 KVM_CAP_S390_USER_STSI 6443-------------------------- 6444 6445:Architectures: s390 6446:Parameters: none 6447 6448This capability allows post-handlers for the STSI instruction. After 6449initial handling in the kernel, KVM exits to user space with 6450KVM_EXIT_S390_STSI to allow user space to insert further data. 6451 6452Before exiting to userspace, kvm handlers should fill in s390_stsi field of 6453vcpu->run:: 6454 6455 struct { 6456 __u64 addr; 6457 __u8 ar; 6458 __u8 reserved; 6459 __u8 fc; 6460 __u8 sel1; 6461 __u16 sel2; 6462 } s390_stsi; 6463 6464 @addr - guest address of STSI SYSIB 6465 @fc - function code 6466 @sel1 - selector 1 6467 @sel2 - selector 2 6468 @ar - access register number 6469 6470KVM handlers should exit to userspace with rc = -EREMOTE. 6471 64727.5 KVM_CAP_SPLIT_IRQCHIP 6473------------------------- 6474 6475:Architectures: x86 6476:Parameters: args[0] - number of routes reserved for userspace IOAPICs 6477:Returns: 0 on success, -1 on error 6478 6479Create a local apic for each processor in the kernel. This can be used 6480instead of KVM_CREATE_IRQCHIP if the userspace VMM wishes to emulate the 6481IOAPIC and PIC (and also the PIT, even though this has to be enabled 6482separately). 6483 6484This capability also enables in kernel routing of interrupt requests; 6485when KVM_CAP_SPLIT_IRQCHIP only routes of KVM_IRQ_ROUTING_MSI type are 6486used in the IRQ routing table. The first args[0] MSI routes are reserved 6487for the IOAPIC pins. Whenever the LAPIC receives an EOI for these routes, 6488a KVM_EXIT_IOAPIC_EOI vmexit will be reported to userspace. 6489 6490Fails if VCPU has already been created, or if the irqchip is already in the 6491kernel (i.e. KVM_CREATE_IRQCHIP has already been called). 6492 64937.6 KVM_CAP_S390_RI 6494------------------- 6495 6496:Architectures: s390 6497:Parameters: none 6498 6499Allows use of runtime-instrumentation introduced with zEC12 processor. 6500Will return -EINVAL if the machine does not support runtime-instrumentation. 6501Will return -EBUSY if a VCPU has already been created. 6502 65037.7 KVM_CAP_X2APIC_API 6504---------------------- 6505 6506:Architectures: x86 6507:Parameters: args[0] - features that should be enabled 6508:Returns: 0 on success, -EINVAL when args[0] contains invalid features 6509 6510Valid feature flags in args[0] are:: 6511 6512 #define KVM_X2APIC_API_USE_32BIT_IDS (1ULL << 0) 6513 #define KVM_X2APIC_API_DISABLE_BROADCAST_QUIRK (1ULL << 1) 6514 6515Enabling KVM_X2APIC_API_USE_32BIT_IDS changes the behavior of 6516KVM_SET_GSI_ROUTING, KVM_SIGNAL_MSI, KVM_SET_LAPIC, and KVM_GET_LAPIC, 6517allowing the use of 32-bit APIC IDs. See KVM_CAP_X2APIC_API in their 6518respective sections. 6519 6520KVM_X2APIC_API_DISABLE_BROADCAST_QUIRK must be enabled for x2APIC to work 6521in logical mode or with more than 255 VCPUs. Otherwise, KVM treats 0xff 6522as a broadcast even in x2APIC mode in order to support physical x2APIC 6523without interrupt remapping. This is undesirable in logical mode, 6524where 0xff represents CPUs 0-7 in cluster 0. 6525 65267.8 KVM_CAP_S390_USER_INSTR0 6527---------------------------- 6528 6529:Architectures: s390 6530:Parameters: none 6531 6532With this capability enabled, all illegal instructions 0x0000 (2 bytes) will 6533be intercepted and forwarded to user space. User space can use this 6534mechanism e.g. to realize 2-byte software breakpoints. The kernel will 6535not inject an operating exception for these instructions, user space has 6536to take care of that. 6537 6538This capability can be enabled dynamically even if VCPUs were already 6539created and are running. 6540 65417.9 KVM_CAP_S390_GS 6542------------------- 6543 6544:Architectures: s390 6545:Parameters: none 6546:Returns: 0 on success; -EINVAL if the machine does not support 6547 guarded storage; -EBUSY if a VCPU has already been created. 6548 6549Allows use of guarded storage for the KVM guest. 6550 65517.10 KVM_CAP_S390_AIS 6552--------------------- 6553 6554:Architectures: s390 6555:Parameters: none 6556 6557Allow use of adapter-interruption suppression. 6558:Returns: 0 on success; -EBUSY if a VCPU has already been created. 6559 65607.11 KVM_CAP_PPC_SMT 6561-------------------- 6562 6563:Architectures: ppc 6564:Parameters: vsmt_mode, flags 6565 6566Enabling this capability on a VM provides userspace with a way to set 6567the desired virtual SMT mode (i.e. the number of virtual CPUs per 6568virtual core). The virtual SMT mode, vsmt_mode, must be a power of 2 6569between 1 and 8. On POWER8, vsmt_mode must also be no greater than 6570the number of threads per subcore for the host. Currently flags must 6571be 0. A successful call to enable this capability will result in 6572vsmt_mode being returned when the KVM_CAP_PPC_SMT capability is 6573subsequently queried for the VM. This capability is only supported by 6574HV KVM, and can only be set before any VCPUs have been created. 6575The KVM_CAP_PPC_SMT_POSSIBLE capability indicates which virtual SMT 6576modes are available. 6577 65787.12 KVM_CAP_PPC_FWNMI 6579---------------------- 6580 6581:Architectures: ppc 6582:Parameters: none 6583 6584With this capability a machine check exception in the guest address 6585space will cause KVM to exit the guest with NMI exit reason. This 6586enables QEMU to build error log and branch to guest kernel registered 6587machine check handling routine. Without this capability KVM will 6588branch to guests' 0x200 interrupt vector. 6589 65907.13 KVM_CAP_X86_DISABLE_EXITS 6591------------------------------ 6592 6593:Architectures: x86 6594:Parameters: args[0] defines which exits are disabled 6595:Returns: 0 on success, -EINVAL when args[0] contains invalid exits 6596 6597Valid bits in args[0] are:: 6598 6599 #define KVM_X86_DISABLE_EXITS_MWAIT (1 << 0) 6600 #define KVM_X86_DISABLE_EXITS_HLT (1 << 1) 6601 #define KVM_X86_DISABLE_EXITS_PAUSE (1 << 2) 6602 #define KVM_X86_DISABLE_EXITS_CSTATE (1 << 3) 6603 6604Enabling this capability on a VM provides userspace with a way to no 6605longer intercept some instructions for improved latency in some 6606workloads, and is suggested when vCPUs are associated to dedicated 6607physical CPUs. More bits can be added in the future; userspace can 6608just pass the KVM_CHECK_EXTENSION result to KVM_ENABLE_CAP to disable 6609all such vmexits. 6610 6611Do not enable KVM_FEATURE_PV_UNHALT if you disable HLT exits. 6612 66137.14 KVM_CAP_S390_HPAGE_1M 6614-------------------------- 6615 6616:Architectures: s390 6617:Parameters: none 6618:Returns: 0 on success, -EINVAL if hpage module parameter was not set 6619 or cmma is enabled, or the VM has the KVM_VM_S390_UCONTROL 6620 flag set 6621 6622With this capability the KVM support for memory backing with 1m pages 6623through hugetlbfs can be enabled for a VM. After the capability is 6624enabled, cmma can't be enabled anymore and pfmfi and the storage key 6625interpretation are disabled. If cmma has already been enabled or the 6626hpage module parameter is not set to 1, -EINVAL is returned. 6627 6628While it is generally possible to create a huge page backed VM without 6629this capability, the VM will not be able to run. 6630 66317.15 KVM_CAP_MSR_PLATFORM_INFO 6632------------------------------ 6633 6634:Architectures: x86 6635:Parameters: args[0] whether feature should be enabled or not 6636 6637With this capability, a guest may read the MSR_PLATFORM_INFO MSR. Otherwise, 6638a #GP would be raised when the guest tries to access. Currently, this 6639capability does not enable write permissions of this MSR for the guest. 6640 66417.16 KVM_CAP_PPC_NESTED_HV 6642-------------------------- 6643 6644:Architectures: ppc 6645:Parameters: none 6646:Returns: 0 on success, -EINVAL when the implementation doesn't support 6647 nested-HV virtualization. 6648 6649HV-KVM on POWER9 and later systems allows for "nested-HV" 6650virtualization, which provides a way for a guest VM to run guests that 6651can run using the CPU's supervisor mode (privileged non-hypervisor 6652state). Enabling this capability on a VM depends on the CPU having 6653the necessary functionality and on the facility being enabled with a 6654kvm-hv module parameter. 6655 66567.17 KVM_CAP_EXCEPTION_PAYLOAD 6657------------------------------ 6658 6659:Architectures: x86 6660:Parameters: args[0] whether feature should be enabled or not 6661 6662With this capability enabled, CR2 will not be modified prior to the 6663emulated VM-exit when L1 intercepts a #PF exception that occurs in 6664L2. Similarly, for kvm-intel only, DR6 will not be modified prior to 6665the emulated VM-exit when L1 intercepts a #DB exception that occurs in 6666L2. As a result, when KVM_GET_VCPU_EVENTS reports a pending #PF (or 6667#DB) exception for L2, exception.has_payload will be set and the 6668faulting address (or the new DR6 bits*) will be reported in the 6669exception_payload field. Similarly, when userspace injects a #PF (or 6670#DB) into L2 using KVM_SET_VCPU_EVENTS, it is expected to set 6671exception.has_payload and to put the faulting address - or the new DR6 6672bits\ [#]_ - in the exception_payload field. 6673 6674This capability also enables exception.pending in struct 6675kvm_vcpu_events, which allows userspace to distinguish between pending 6676and injected exceptions. 6677 6678 6679.. [#] For the new DR6 bits, note that bit 16 is set iff the #DB exception 6680 will clear DR6.RTM. 6681 66827.18 KVM_CAP_MANUAL_DIRTY_LOG_PROTECT2 6683 6684:Architectures: x86, arm, arm64, mips 6685:Parameters: args[0] whether feature should be enabled or not 6686 6687Valid flags are:: 6688 6689 #define KVM_DIRTY_LOG_MANUAL_PROTECT_ENABLE (1 << 0) 6690 #define KVM_DIRTY_LOG_INITIALLY_SET (1 << 1) 6691 6692With KVM_DIRTY_LOG_MANUAL_PROTECT_ENABLE is set, KVM_GET_DIRTY_LOG will not 6693automatically clear and write-protect all pages that are returned as dirty. 6694Rather, userspace will have to do this operation separately using 6695KVM_CLEAR_DIRTY_LOG. 6696 6697At the cost of a slightly more complicated operation, this provides better 6698scalability and responsiveness for two reasons. First, 6699KVM_CLEAR_DIRTY_LOG ioctl can operate on a 64-page granularity rather 6700than requiring to sync a full memslot; this ensures that KVM does not 6701take spinlocks for an extended period of time. Second, in some cases a 6702large amount of time can pass between a call to KVM_GET_DIRTY_LOG and 6703userspace actually using the data in the page. Pages can be modified 6704during this time, which is inefficient for both the guest and userspace: 6705the guest will incur a higher penalty due to write protection faults, 6706while userspace can see false reports of dirty pages. Manual reprotection 6707helps reducing this time, improving guest performance and reducing the 6708number of dirty log false positives. 6709 6710With KVM_DIRTY_LOG_INITIALLY_SET set, all the bits of the dirty bitmap 6711will be initialized to 1 when created. This also improves performance because 6712dirty logging can be enabled gradually in small chunks on the first call 6713to KVM_CLEAR_DIRTY_LOG. KVM_DIRTY_LOG_INITIALLY_SET depends on 6714KVM_DIRTY_LOG_MANUAL_PROTECT_ENABLE (it is also only available on 6715x86 and arm64 for now). 6716 6717KVM_CAP_MANUAL_DIRTY_LOG_PROTECT2 was previously available under the name 6718KVM_CAP_MANUAL_DIRTY_LOG_PROTECT, but the implementation had bugs that make 6719it hard or impossible to use it correctly. The availability of 6720KVM_CAP_MANUAL_DIRTY_LOG_PROTECT2 signals that those bugs are fixed. 6721Userspace should not try to use KVM_CAP_MANUAL_DIRTY_LOG_PROTECT. 6722 67237.19 KVM_CAP_PPC_SECURE_GUEST 6724------------------------------ 6725 6726:Architectures: ppc 6727 6728This capability indicates that KVM is running on a host that has 6729ultravisor firmware and thus can support a secure guest. On such a 6730system, a guest can ask the ultravisor to make it a secure guest, 6731one whose memory is inaccessible to the host except for pages which 6732are explicitly requested to be shared with the host. The ultravisor 6733notifies KVM when a guest requests to become a secure guest, and KVM 6734has the opportunity to veto the transition. 6735 6736If present, this capability can be enabled for a VM, meaning that KVM 6737will allow the transition to secure guest mode. Otherwise KVM will 6738veto the transition. 6739 67407.20 KVM_CAP_HALT_POLL 6741---------------------- 6742 6743:Architectures: all 6744:Target: VM 6745:Parameters: args[0] is the maximum poll time in nanoseconds 6746:Returns: 0 on success; -1 on error 6747 6748This capability overrides the kvm module parameter halt_poll_ns for the 6749target VM. 6750 6751VCPU polling allows a VCPU to poll for wakeup events instead of immediately 6752scheduling during guest halts. The maximum time a VCPU can spend polling is 6753controlled by the kvm module parameter halt_poll_ns. This capability allows 6754the maximum halt time to specified on a per-VM basis, effectively overriding 6755the module parameter for the target VM. 6756 67577.21 KVM_CAP_X86_USER_SPACE_MSR 6758------------------------------- 6759 6760:Architectures: x86 6761:Target: VM 6762:Parameters: args[0] contains the mask of KVM_MSR_EXIT_REASON_* events to report 6763:Returns: 0 on success; -1 on error 6764 6765This capability enables trapping of #GP invoking RDMSR and WRMSR instructions 6766into user space. 6767 6768When a guest requests to read or write an MSR, KVM may not implement all MSRs 6769that are relevant to a respective system. It also does not differentiate by 6770CPU type. 6771 6772To allow more fine grained control over MSR handling, user space may enable 6773this capability. With it enabled, MSR accesses that match the mask specified in 6774args[0] and trigger a #GP event inside the guest by KVM will instead trigger 6775KVM_EXIT_X86_RDMSR and KVM_EXIT_X86_WRMSR exit notifications which user space 6776can then handle to implement model specific MSR handling and/or user notifications 6777to inform a user that an MSR was not handled. 6778 67797.22 KVM_CAP_X86_BUS_LOCK_EXIT 6780------------------------------- 6781 6782:Architectures: x86 6783:Target: VM 6784:Parameters: args[0] defines the policy used when bus locks detected in guest 6785:Returns: 0 on success, -EINVAL when args[0] contains invalid bits 6786 6787Valid bits in args[0] are:: 6788 6789 #define KVM_BUS_LOCK_DETECTION_OFF (1 << 0) 6790 #define KVM_BUS_LOCK_DETECTION_EXIT (1 << 1) 6791 6792Enabling this capability on a VM provides userspace with a way to select 6793a policy to handle the bus locks detected in guest. Userspace can obtain 6794the supported modes from the result of KVM_CHECK_EXTENSION and define it 6795through the KVM_ENABLE_CAP. 6796 6797KVM_BUS_LOCK_DETECTION_OFF and KVM_BUS_LOCK_DETECTION_EXIT are supported 6798currently and mutually exclusive with each other. More bits can be added in 6799the future. 6800 6801With KVM_BUS_LOCK_DETECTION_OFF set, bus locks in guest will not cause vm exits 6802so that no additional actions are needed. This is the default mode. 6803 6804With KVM_BUS_LOCK_DETECTION_EXIT set, vm exits happen when bus lock detected 6805in VM. KVM just exits to userspace when handling them. Userspace can enforce 6806its own throttling or other policy based mitigations. 6807 6808This capability is aimed to address the thread that VM can exploit bus locks to 6809degree the performance of the whole system. Once the userspace enable this 6810capability and select the KVM_BUS_LOCK_DETECTION_EXIT mode, KVM will set the 6811KVM_RUN_BUS_LOCK flag in vcpu-run->flags field and exit to userspace. Concerning 6812the bus lock vm exit can be preempted by a higher priority VM exit, the exit 6813notifications to userspace can be KVM_EXIT_BUS_LOCK or other reasons. 6814KVM_RUN_BUS_LOCK flag is used to distinguish between them. 6815 68167.23 KVM_CAP_PPC_DAWR1 6817---------------------- 6818 6819:Architectures: ppc 6820:Parameters: none 6821:Returns: 0 on success, -EINVAL when CPU doesn't support 2nd DAWR 6822 6823This capability can be used to check / enable 2nd DAWR feature provided 6824by POWER10 processor. 6825 6826 68277.24 KVM_CAP_VM_COPY_ENC_CONTEXT_FROM 6828------------------------------------- 6829 6830Architectures: x86 SEV enabled 6831Type: vm 6832Parameters: args[0] is the fd of the source vm 6833Returns: 0 on success; ENOTTY on error 6834 6835This capability enables userspace to copy encryption context from the vm 6836indicated by the fd to the vm this is called on. 6837 6838This is intended to support in-guest workloads scheduled by the host. This 6839allows the in-guest workload to maintain its own NPTs and keeps the two vms 6840from accidentally clobbering each other with interrupts and the like (separate 6841APIC/MSRs/etc). 6842 68437.25 KVM_CAP_SGX_ATTRIBUTE 6844-------------------------- 6845 6846:Architectures: x86 6847:Target: VM 6848:Parameters: args[0] is a file handle of a SGX attribute file in securityfs 6849:Returns: 0 on success, -EINVAL if the file handle is invalid or if a requested 6850 attribute is not supported by KVM. 6851 6852KVM_CAP_SGX_ATTRIBUTE enables a userspace VMM to grant a VM access to one or 6853more priveleged enclave attributes. args[0] must hold a file handle to a valid 6854SGX attribute file corresponding to an attribute that is supported/restricted 6855by KVM (currently only PROVISIONKEY). 6856 6857The SGX subsystem restricts access to a subset of enclave attributes to provide 6858additional security for an uncompromised kernel, e.g. use of the PROVISIONKEY 6859is restricted to deter malware from using the PROVISIONKEY to obtain a stable 6860system fingerprint. To prevent userspace from circumventing such restrictions 6861by running an enclave in a VM, KVM prevents access to privileged attributes by 6862default. 6863 6864See Documentation/x86/sgx.rst for more details. 6865 68667.26 KVM_CAP_PPC_RPT_INVALIDATE 6867------------------------------- 6868 6869:Capability: KVM_CAP_PPC_RPT_INVALIDATE 6870:Architectures: ppc 6871:Type: vm 6872 6873This capability indicates that the kernel is capable of handling 6874H_RPT_INVALIDATE hcall. 6875 6876In order to enable the use of H_RPT_INVALIDATE in the guest, 6877user space might have to advertise it for the guest. For example, 6878IBM pSeries (sPAPR) guest starts using it if "hcall-rpt-invalidate" is 6879present in the "ibm,hypertas-functions" device-tree property. 6880 6881This capability is enabled for hypervisors on platforms like POWER9 6882that support radix MMU. 6883 68847.27 KVM_CAP_EXIT_ON_EMULATION_FAILURE 6885-------------------------------------- 6886 6887:Architectures: x86 6888:Parameters: args[0] whether the feature should be enabled or not 6889 6890When this capability is enabled, an emulation failure will result in an exit 6891to userspace with KVM_INTERNAL_ERROR (except when the emulator was invoked 6892to handle a VMware backdoor instruction). Furthermore, KVM will now provide up 6893to 15 instruction bytes for any exit to userspace resulting from an emulation 6894failure. When these exits to userspace occur use the emulation_failure struct 6895instead of the internal struct. They both have the same layout, but the 6896emulation_failure struct matches the content better. It also explicitly 6897defines the 'flags' field which is used to describe the fields in the struct 6898that are valid (ie: if KVM_INTERNAL_ERROR_EMULATION_FLAG_INSTRUCTION_BYTES is 6899set in the 'flags' field then both 'insn_size' and 'insn_bytes' have valid data 6900in them.) 6901 69027.28 KVM_CAP_ARM_MTE 6903-------------------- 6904 6905:Architectures: arm64 6906:Parameters: none 6907 6908This capability indicates that KVM (and the hardware) supports exposing the 6909Memory Tagging Extensions (MTE) to the guest. It must also be enabled by the 6910VMM before creating any VCPUs to allow the guest access. Note that MTE is only 6911available to a guest running in AArch64 mode and enabling this capability will 6912cause attempts to create AArch32 VCPUs to fail. 6913 6914When enabled the guest is able to access tags associated with any memory given 6915to the guest. KVM will ensure that the tags are maintained during swap or 6916hibernation of the host; however the VMM needs to manually save/restore the 6917tags as appropriate if the VM is migrated. 6918 6919When this capability is enabled all memory in memslots must be mapped as 6920not-shareable (no MAP_SHARED), attempts to create a memslot with a 6921MAP_SHARED mmap will result in an -EINVAL return. 6922 6923When enabled the VMM may make use of the ``KVM_ARM_MTE_COPY_TAGS`` ioctl to 6924perform a bulk copy of tags to/from the guest. 6925 69267.29 KVM_CAP_VM_MOVE_ENC_CONTEXT_FROM 6927------------------------------------- 6928 6929Architectures: x86 SEV enabled 6930Type: vm 6931Parameters: args[0] is the fd of the source vm 6932Returns: 0 on success 6933 6934This capability enables userspace to migrate the encryption context from the VM 6935indicated by the fd to the VM this is called on. 6936 6937This is intended to support intra-host migration of VMs between userspace VMMs, 6938upgrading the VMM process without interrupting the guest. 6939 69408. Other capabilities. 6941====================== 6942 6943This section lists capabilities that give information about other 6944features of the KVM implementation. 6945 69468.1 KVM_CAP_PPC_HWRNG 6947--------------------- 6948 6949:Architectures: ppc 6950 6951This capability, if KVM_CHECK_EXTENSION indicates that it is 6952available, means that the kernel has an implementation of the 6953H_RANDOM hypercall backed by a hardware random-number generator. 6954If present, the kernel H_RANDOM handler can be enabled for guest use 6955with the KVM_CAP_PPC_ENABLE_HCALL capability. 6956 69578.2 KVM_CAP_HYPERV_SYNIC 6958------------------------ 6959 6960:Architectures: x86 6961 6962This capability, if KVM_CHECK_EXTENSION indicates that it is 6963available, means that the kernel has an implementation of the 6964Hyper-V Synthetic interrupt controller(SynIC). Hyper-V SynIC is 6965used to support Windows Hyper-V based guest paravirt drivers(VMBus). 6966 6967In order to use SynIC, it has to be activated by setting this 6968capability via KVM_ENABLE_CAP ioctl on the vcpu fd. Note that this 6969will disable the use of APIC hardware virtualization even if supported 6970by the CPU, as it's incompatible with SynIC auto-EOI behavior. 6971 69728.3 KVM_CAP_PPC_RADIX_MMU 6973------------------------- 6974 6975:Architectures: ppc 6976 6977This capability, if KVM_CHECK_EXTENSION indicates that it is 6978available, means that the kernel can support guests using the 6979radix MMU defined in Power ISA V3.00 (as implemented in the POWER9 6980processor). 6981 69828.4 KVM_CAP_PPC_HASH_MMU_V3 6983--------------------------- 6984 6985:Architectures: ppc 6986 6987This capability, if KVM_CHECK_EXTENSION indicates that it is 6988available, means that the kernel can support guests using the 6989hashed page table MMU defined in Power ISA V3.00 (as implemented in 6990the POWER9 processor), including in-memory segment tables. 6991 69928.5 KVM_CAP_MIPS_VZ 6993------------------- 6994 6995:Architectures: mips 6996 6997This capability, if KVM_CHECK_EXTENSION on the main kvm handle indicates that 6998it is available, means that full hardware assisted virtualization capabilities 6999of the hardware are available for use through KVM. An appropriate 7000KVM_VM_MIPS_* type must be passed to KVM_CREATE_VM to create a VM which 7001utilises it. 7002 7003If KVM_CHECK_EXTENSION on a kvm VM handle indicates that this capability is 7004available, it means that the VM is using full hardware assisted virtualization 7005capabilities of the hardware. This is useful to check after creating a VM with 7006KVM_VM_MIPS_DEFAULT. 7007 7008The value returned by KVM_CHECK_EXTENSION should be compared against known 7009values (see below). All other values are reserved. This is to allow for the 7010possibility of other hardware assisted virtualization implementations which 7011may be incompatible with the MIPS VZ ASE. 7012 7013== ========================================================================== 7014 0 The trap & emulate implementation is in use to run guest code in user 7015 mode. Guest virtual memory segments are rearranged to fit the guest in the 7016 user mode address space. 7017 7018 1 The MIPS VZ ASE is in use, providing full hardware assisted 7019 virtualization, including standard guest virtual memory segments. 7020== ========================================================================== 7021 70228.6 KVM_CAP_MIPS_TE 7023------------------- 7024 7025:Architectures: mips 7026 7027This capability, if KVM_CHECK_EXTENSION on the main kvm handle indicates that 7028it is available, means that the trap & emulate implementation is available to 7029run guest code in user mode, even if KVM_CAP_MIPS_VZ indicates that hardware 7030assisted virtualisation is also available. KVM_VM_MIPS_TE (0) must be passed 7031to KVM_CREATE_VM to create a VM which utilises it. 7032 7033If KVM_CHECK_EXTENSION on a kvm VM handle indicates that this capability is 7034available, it means that the VM is using trap & emulate. 7035 70368.7 KVM_CAP_MIPS_64BIT 7037---------------------- 7038 7039:Architectures: mips 7040 7041This capability indicates the supported architecture type of the guest, i.e. the 7042supported register and address width. 7043 7044The values returned when this capability is checked by KVM_CHECK_EXTENSION on a 7045kvm VM handle correspond roughly to the CP0_Config.AT register field, and should 7046be checked specifically against known values (see below). All other values are 7047reserved. 7048 7049== ======================================================================== 7050 0 MIPS32 or microMIPS32. 7051 Both registers and addresses are 32-bits wide. 7052 It will only be possible to run 32-bit guest code. 7053 7054 1 MIPS64 or microMIPS64 with access only to 32-bit compatibility segments. 7055 Registers are 64-bits wide, but addresses are 32-bits wide. 7056 64-bit guest code may run but cannot access MIPS64 memory segments. 7057 It will also be possible to run 32-bit guest code. 7058 7059 2 MIPS64 or microMIPS64 with access to all address segments. 7060 Both registers and addresses are 64-bits wide. 7061 It will be possible to run 64-bit or 32-bit guest code. 7062== ======================================================================== 7063 70648.9 KVM_CAP_ARM_USER_IRQ 7065------------------------ 7066 7067:Architectures: arm, arm64 7068 7069This capability, if KVM_CHECK_EXTENSION indicates that it is available, means 7070that if userspace creates a VM without an in-kernel interrupt controller, it 7071will be notified of changes to the output level of in-kernel emulated devices, 7072which can generate virtual interrupts, presented to the VM. 7073For such VMs, on every return to userspace, the kernel 7074updates the vcpu's run->s.regs.device_irq_level field to represent the actual 7075output level of the device. 7076 7077Whenever kvm detects a change in the device output level, kvm guarantees at 7078least one return to userspace before running the VM. This exit could either 7079be a KVM_EXIT_INTR or any other exit event, like KVM_EXIT_MMIO. This way, 7080userspace can always sample the device output level and re-compute the state of 7081the userspace interrupt controller. Userspace should always check the state 7082of run->s.regs.device_irq_level on every kvm exit. 7083The value in run->s.regs.device_irq_level can represent both level and edge 7084triggered interrupt signals, depending on the device. Edge triggered interrupt 7085signals will exit to userspace with the bit in run->s.regs.device_irq_level 7086set exactly once per edge signal. 7087 7088The field run->s.regs.device_irq_level is available independent of 7089run->kvm_valid_regs or run->kvm_dirty_regs bits. 7090 7091If KVM_CAP_ARM_USER_IRQ is supported, the KVM_CHECK_EXTENSION ioctl returns a 7092number larger than 0 indicating the version of this capability is implemented 7093and thereby which bits in run->s.regs.device_irq_level can signal values. 7094 7095Currently the following bits are defined for the device_irq_level bitmap:: 7096 7097 KVM_CAP_ARM_USER_IRQ >= 1: 7098 7099 KVM_ARM_DEV_EL1_VTIMER - EL1 virtual timer 7100 KVM_ARM_DEV_EL1_PTIMER - EL1 physical timer 7101 KVM_ARM_DEV_PMU - ARM PMU overflow interrupt signal 7102 7103Future versions of kvm may implement additional events. These will get 7104indicated by returning a higher number from KVM_CHECK_EXTENSION and will be 7105listed above. 7106 71078.10 KVM_CAP_PPC_SMT_POSSIBLE 7108----------------------------- 7109 7110:Architectures: ppc 7111 7112Querying this capability returns a bitmap indicating the possible 7113virtual SMT modes that can be set using KVM_CAP_PPC_SMT. If bit N 7114(counting from the right) is set, then a virtual SMT mode of 2^N is 7115available. 7116 71178.11 KVM_CAP_HYPERV_SYNIC2 7118-------------------------- 7119 7120:Architectures: x86 7121 7122This capability enables a newer version of Hyper-V Synthetic interrupt 7123controller (SynIC). The only difference with KVM_CAP_HYPERV_SYNIC is that KVM 7124doesn't clear SynIC message and event flags pages when they are enabled by 7125writing to the respective MSRs. 7126 71278.12 KVM_CAP_HYPERV_VP_INDEX 7128---------------------------- 7129 7130:Architectures: x86 7131 7132This capability indicates that userspace can load HV_X64_MSR_VP_INDEX msr. Its 7133value is used to denote the target vcpu for a SynIC interrupt. For 7134compatibilty, KVM initializes this msr to KVM's internal vcpu index. When this 7135capability is absent, userspace can still query this msr's value. 7136 71378.13 KVM_CAP_S390_AIS_MIGRATION 7138------------------------------- 7139 7140:Architectures: s390 7141:Parameters: none 7142 7143This capability indicates if the flic device will be able to get/set the 7144AIS states for migration via the KVM_DEV_FLIC_AISM_ALL attribute and allows 7145to discover this without having to create a flic device. 7146 71478.14 KVM_CAP_S390_PSW 7148--------------------- 7149 7150:Architectures: s390 7151 7152This capability indicates that the PSW is exposed via the kvm_run structure. 7153 71548.15 KVM_CAP_S390_GMAP 7155---------------------- 7156 7157:Architectures: s390 7158 7159This capability indicates that the user space memory used as guest mapping can 7160be anywhere in the user memory address space, as long as the memory slots are 7161aligned and sized to a segment (1MB) boundary. 7162 71638.16 KVM_CAP_S390_COW 7164--------------------- 7165 7166:Architectures: s390 7167 7168This capability indicates that the user space memory used as guest mapping can 7169use copy-on-write semantics as well as dirty pages tracking via read-only page 7170tables. 7171 71728.17 KVM_CAP_S390_BPB 7173--------------------- 7174 7175:Architectures: s390 7176 7177This capability indicates that kvm will implement the interfaces to handle 7178reset, migration and nested KVM for branch prediction blocking. The stfle 7179facility 82 should not be provided to the guest without this capability. 7180 71818.18 KVM_CAP_HYPERV_TLBFLUSH 7182---------------------------- 7183 7184:Architectures: x86 7185 7186This capability indicates that KVM supports paravirtualized Hyper-V TLB Flush 7187hypercalls: 7188HvFlushVirtualAddressSpace, HvFlushVirtualAddressSpaceEx, 7189HvFlushVirtualAddressList, HvFlushVirtualAddressListEx. 7190 71918.19 KVM_CAP_ARM_INJECT_SERROR_ESR 7192---------------------------------- 7193 7194:Architectures: arm, arm64 7195 7196This capability indicates that userspace can specify (via the 7197KVM_SET_VCPU_EVENTS ioctl) the syndrome value reported to the guest when it 7198takes a virtual SError interrupt exception. 7199If KVM advertises this capability, userspace can only specify the ISS field for 7200the ESR syndrome. Other parts of the ESR, such as the EC are generated by the 7201CPU when the exception is taken. If this virtual SError is taken to EL1 using 7202AArch64, this value will be reported in the ISS field of ESR_ELx. 7203 7204See KVM_CAP_VCPU_EVENTS for more details. 7205 72068.20 KVM_CAP_HYPERV_SEND_IPI 7207---------------------------- 7208 7209:Architectures: x86 7210 7211This capability indicates that KVM supports paravirtualized Hyper-V IPI send 7212hypercalls: 7213HvCallSendSyntheticClusterIpi, HvCallSendSyntheticClusterIpiEx. 7214 72158.21 KVM_CAP_HYPERV_DIRECT_TLBFLUSH 7216----------------------------------- 7217 7218:Architectures: x86 7219 7220This capability indicates that KVM running on top of Hyper-V hypervisor 7221enables Direct TLB flush for its guests meaning that TLB flush 7222hypercalls are handled by Level 0 hypervisor (Hyper-V) bypassing KVM. 7223Due to the different ABI for hypercall parameters between Hyper-V and 7224KVM, enabling this capability effectively disables all hypercall 7225handling by KVM (as some KVM hypercall may be mistakenly treated as TLB 7226flush hypercalls by Hyper-V) so userspace should disable KVM identification 7227in CPUID and only exposes Hyper-V identification. In this case, guest 7228thinks it's running on Hyper-V and only use Hyper-V hypercalls. 7229 72308.22 KVM_CAP_S390_VCPU_RESETS 7231----------------------------- 7232 7233:Architectures: s390 7234 7235This capability indicates that the KVM_S390_NORMAL_RESET and 7236KVM_S390_CLEAR_RESET ioctls are available. 7237 72388.23 KVM_CAP_S390_PROTECTED 7239--------------------------- 7240 7241:Architectures: s390 7242 7243This capability indicates that the Ultravisor has been initialized and 7244KVM can therefore start protected VMs. 7245This capability governs the KVM_S390_PV_COMMAND ioctl and the 7246KVM_MP_STATE_LOAD MP_STATE. KVM_SET_MP_STATE can fail for protected 7247guests when the state change is invalid. 7248 72498.24 KVM_CAP_STEAL_TIME 7250----------------------- 7251 7252:Architectures: arm64, x86 7253 7254This capability indicates that KVM supports steal time accounting. 7255When steal time accounting is supported it may be enabled with 7256architecture-specific interfaces. This capability and the architecture- 7257specific interfaces must be consistent, i.e. if one says the feature 7258is supported, than the other should as well and vice versa. For arm64 7259see Documentation/virt/kvm/devices/vcpu.rst "KVM_ARM_VCPU_PVTIME_CTRL". 7260For x86 see Documentation/virt/kvm/msr.rst "MSR_KVM_STEAL_TIME". 7261 72628.25 KVM_CAP_S390_DIAG318 7263------------------------- 7264 7265:Architectures: s390 7266 7267This capability enables a guest to set information about its control program 7268(i.e. guest kernel type and version). The information is helpful during 7269system/firmware service events, providing additional data about the guest 7270environments running on the machine. 7271 7272The information is associated with the DIAGNOSE 0x318 instruction, which sets 7273an 8-byte value consisting of a one-byte Control Program Name Code (CPNC) and 7274a 7-byte Control Program Version Code (CPVC). The CPNC determines what 7275environment the control program is running in (e.g. Linux, z/VM...), and the 7276CPVC is used for information specific to OS (e.g. Linux version, Linux 7277distribution...) 7278 7279If this capability is available, then the CPNC and CPVC can be synchronized 7280between KVM and userspace via the sync regs mechanism (KVM_SYNC_DIAG318). 7281 72828.26 KVM_CAP_X86_USER_SPACE_MSR 7283------------------------------- 7284 7285:Architectures: x86 7286 7287This capability indicates that KVM supports deflection of MSR reads and 7288writes to user space. It can be enabled on a VM level. If enabled, MSR 7289accesses that would usually trigger a #GP by KVM into the guest will 7290instead get bounced to user space through the KVM_EXIT_X86_RDMSR and 7291KVM_EXIT_X86_WRMSR exit notifications. 7292 72938.27 KVM_CAP_X86_MSR_FILTER 7294--------------------------- 7295 7296:Architectures: x86 7297 7298This capability indicates that KVM supports that accesses to user defined MSRs 7299may be rejected. With this capability exposed, KVM exports new VM ioctl 7300KVM_X86_SET_MSR_FILTER which user space can call to specify bitmaps of MSR 7301ranges that KVM should reject access to. 7302 7303In combination with KVM_CAP_X86_USER_SPACE_MSR, this allows user space to 7304trap and emulate MSRs that are outside of the scope of KVM as well as 7305limit the attack surface on KVM's MSR emulation code. 7306 73078.28 KVM_CAP_ENFORCE_PV_FEATURE_CPUID 7308----------------------------- 7309 7310Architectures: x86 7311 7312When enabled, KVM will disable paravirtual features provided to the 7313guest according to the bits in the KVM_CPUID_FEATURES CPUID leaf 7314(0x40000001). Otherwise, a guest may use the paravirtual features 7315regardless of what has actually been exposed through the CPUID leaf. 7316 73178.29 KVM_CAP_DIRTY_LOG_RING 7318--------------------------- 7319 7320:Architectures: x86 7321:Parameters: args[0] - size of the dirty log ring 7322 7323KVM is capable of tracking dirty memory using ring buffers that are 7324mmaped into userspace; there is one dirty ring per vcpu. 7325 7326The dirty ring is available to userspace as an array of 7327``struct kvm_dirty_gfn``. Each dirty entry it's defined as:: 7328 7329 struct kvm_dirty_gfn { 7330 __u32 flags; 7331 __u32 slot; /* as_id | slot_id */ 7332 __u64 offset; 7333 }; 7334 7335The following values are defined for the flags field to define the 7336current state of the entry:: 7337 7338 #define KVM_DIRTY_GFN_F_DIRTY BIT(0) 7339 #define KVM_DIRTY_GFN_F_RESET BIT(1) 7340 #define KVM_DIRTY_GFN_F_MASK 0x3 7341 7342Userspace should call KVM_ENABLE_CAP ioctl right after KVM_CREATE_VM 7343ioctl to enable this capability for the new guest and set the size of 7344the rings. Enabling the capability is only allowed before creating any 7345vCPU, and the size of the ring must be a power of two. The larger the 7346ring buffer, the less likely the ring is full and the VM is forced to 7347exit to userspace. The optimal size depends on the workload, but it is 7348recommended that it be at least 64 KiB (4096 entries). 7349 7350Just like for dirty page bitmaps, the buffer tracks writes to 7351all user memory regions for which the KVM_MEM_LOG_DIRTY_PAGES flag was 7352set in KVM_SET_USER_MEMORY_REGION. Once a memory region is registered 7353with the flag set, userspace can start harvesting dirty pages from the 7354ring buffer. 7355 7356An entry in the ring buffer can be unused (flag bits ``00``), 7357dirty (flag bits ``01``) or harvested (flag bits ``1X``). The 7358state machine for the entry is as follows:: 7359 7360 dirtied harvested reset 7361 00 -----------> 01 -------------> 1X -------+ 7362 ^ | 7363 | | 7364 +------------------------------------------+ 7365 7366To harvest the dirty pages, userspace accesses the mmaped ring buffer 7367to read the dirty GFNs. If the flags has the DIRTY bit set (at this stage 7368the RESET bit must be cleared), then it means this GFN is a dirty GFN. 7369The userspace should harvest this GFN and mark the flags from state 7370``01b`` to ``1Xb`` (bit 0 will be ignored by KVM, but bit 1 must be set 7371to show that this GFN is harvested and waiting for a reset), and move 7372on to the next GFN. The userspace should continue to do this until the 7373flags of a GFN have the DIRTY bit cleared, meaning that it has harvested 7374all the dirty GFNs that were available. 7375 7376It's not necessary for userspace to harvest the all dirty GFNs at once. 7377However it must collect the dirty GFNs in sequence, i.e., the userspace 7378program cannot skip one dirty GFN to collect the one next to it. 7379 7380After processing one or more entries in the ring buffer, userspace 7381calls the VM ioctl KVM_RESET_DIRTY_RINGS to notify the kernel about 7382it, so that the kernel will reprotect those collected GFNs. 7383Therefore, the ioctl must be called *before* reading the content of 7384the dirty pages. 7385 7386The dirty ring can get full. When it happens, the KVM_RUN of the 7387vcpu will return with exit reason KVM_EXIT_DIRTY_LOG_FULL. 7388 7389The dirty ring interface has a major difference comparing to the 7390KVM_GET_DIRTY_LOG interface in that, when reading the dirty ring from 7391userspace, it's still possible that the kernel has not yet flushed the 7392processor's dirty page buffers into the kernel buffer (with dirty bitmaps, the 7393flushing is done by the KVM_GET_DIRTY_LOG ioctl). To achieve that, one 7394needs to kick the vcpu out of KVM_RUN using a signal. The resulting 7395vmexit ensures that all dirty GFNs are flushed to the dirty rings. 7396 7397NOTE: the capability KVM_CAP_DIRTY_LOG_RING and the corresponding 7398ioctl KVM_RESET_DIRTY_RINGS are mutual exclusive to the existing ioctls 7399KVM_GET_DIRTY_LOG and KVM_CLEAR_DIRTY_LOG. After enabling 7400KVM_CAP_DIRTY_LOG_RING with an acceptable dirty ring size, the virtual 7401machine will switch to ring-buffer dirty page tracking and further 7402KVM_GET_DIRTY_LOG or KVM_CLEAR_DIRTY_LOG ioctls will fail. 7403 74048.30 KVM_CAP_XEN_HVM 7405-------------------- 7406 7407:Architectures: x86 7408 7409This capability indicates the features that Xen supports for hosting Xen 7410PVHVM guests. Valid flags are:: 7411 7412 #define KVM_XEN_HVM_CONFIG_HYPERCALL_MSR (1 << 0) 7413 #define KVM_XEN_HVM_CONFIG_INTERCEPT_HCALL (1 << 1) 7414 #define KVM_XEN_HVM_CONFIG_SHARED_INFO (1 << 2) 7415 #define KVM_XEN_HVM_CONFIG_RUNSTATE (1 << 2) 7416 7417The KVM_XEN_HVM_CONFIG_HYPERCALL_MSR flag indicates that the KVM_XEN_HVM_CONFIG 7418ioctl is available, for the guest to set its hypercall page. 7419 7420If KVM_XEN_HVM_CONFIG_INTERCEPT_HCALL is also set, the same flag may also be 7421provided in the flags to KVM_XEN_HVM_CONFIG, without providing hypercall page 7422contents, to request that KVM generate hypercall page content automatically 7423and also enable interception of guest hypercalls with KVM_EXIT_XEN. 7424 7425The KVM_XEN_HVM_CONFIG_SHARED_INFO flag indicates the availability of the 7426KVM_XEN_HVM_SET_ATTR, KVM_XEN_HVM_GET_ATTR, KVM_XEN_VCPU_SET_ATTR and 7427KVM_XEN_VCPU_GET_ATTR ioctls, as well as the delivery of exception vectors 7428for event channel upcalls when the evtchn_upcall_pending field of a vcpu's 7429vcpu_info is set. 7430 7431The KVM_XEN_HVM_CONFIG_RUNSTATE flag indicates that the runstate-related 7432features KVM_XEN_VCPU_ATTR_TYPE_RUNSTATE_ADDR/_CURRENT/_DATA/_ADJUST are 7433supported by the KVM_XEN_VCPU_SET_ATTR/KVM_XEN_VCPU_GET_ATTR ioctls. 7434 74358.31 KVM_CAP_PPC_MULTITCE 7436------------------------- 7437 7438:Capability: KVM_CAP_PPC_MULTITCE 7439:Architectures: ppc 7440:Type: vm 7441 7442This capability means the kernel is capable of handling hypercalls 7443H_PUT_TCE_INDIRECT and H_STUFF_TCE without passing those into the user 7444space. This significantly accelerates DMA operations for PPC KVM guests. 7445User space should expect that its handlers for these hypercalls 7446are not going to be called if user space previously registered LIOBN 7447in KVM (via KVM_CREATE_SPAPR_TCE or similar calls). 7448 7449In order to enable H_PUT_TCE_INDIRECT and H_STUFF_TCE use in the guest, 7450user space might have to advertise it for the guest. For example, 7451IBM pSeries (sPAPR) guest starts using them if "hcall-multi-tce" is 7452present in the "ibm,hypertas-functions" device-tree property. 7453 7454The hypercalls mentioned above may or may not be processed successfully 7455in the kernel based fast path. If they can not be handled by the kernel, 7456they will get passed on to user space. So user space still has to have 7457an implementation for these despite the in kernel acceleration. 7458 7459This capability is always enabled. 7460 74618.32 KVM_CAP_PTP_KVM 7462-------------------- 7463 7464:Architectures: arm64 7465 7466This capability indicates that the KVM virtual PTP service is 7467supported in the host. A VMM can check whether the service is 7468available to the guest on migration. 7469 74708.33 KVM_CAP_HYPERV_ENFORCE_CPUID 7471--------------------------------- 7472 7473Architectures: x86 7474 7475When enabled, KVM will disable emulated Hyper-V features provided to the 7476guest according to the bits Hyper-V CPUID feature leaves. Otherwise, all 7477currently implmented Hyper-V features are provided unconditionally when 7478Hyper-V identification is set in the HYPERV_CPUID_INTERFACE (0x40000001) 7479leaf. 7480 74818.34 KVM_CAP_EXIT_HYPERCALL 7482--------------------------- 7483 7484:Capability: KVM_CAP_EXIT_HYPERCALL 7485:Architectures: x86 7486:Type: vm 7487 7488This capability, if enabled, will cause KVM to exit to userspace 7489with KVM_EXIT_HYPERCALL exit reason to process some hypercalls. 7490 7491Calling KVM_CHECK_EXTENSION for this capability will return a bitmask 7492of hypercalls that can be configured to exit to userspace. 7493Right now, the only such hypercall is KVM_HC_MAP_GPA_RANGE. 7494 7495The argument to KVM_ENABLE_CAP is also a bitmask, and must be a subset 7496of the result of KVM_CHECK_EXTENSION. KVM will forward to userspace 7497the hypercalls whose corresponding bit is in the argument, and return 7498ENOSYS for the others. 7499