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 centered around different kinds of file descriptors 11and ioctls that can be issued to these file descriptors. An initial 12open("/dev/kvm") obtains a handle to the kvm subsystem; this handle 13can be used to issue system ioctls. A KVM_CREATE_VM ioctl on this 14handle will create a VM file descriptor which can be used to issue VM 15ioctls. A KVM_CREATE_VCPU or KVM_CREATE_DEVICE ioctl on a VM fd will 16create a virtual cpu or device and return a file descriptor pointing to 17the new resource. 18 19In other words, the kvm API is a set of ioctls that are issued to 20different kinds of file descriptor in order to control various aspects of 21a virtual machine. Depending on the file descriptor that accepts them, 22ioctls belong to the following classes: 23 24 - System ioctls: These query and set global attributes which affect the 25 whole kvm subsystem. In addition a system ioctl is used to create 26 virtual machines. 27 28 - VM ioctls: These query and set attributes that affect an entire virtual 29 machine, for example memory layout. In addition a VM ioctl is used to 30 create virtual cpus (vcpus) and devices. 31 32 VM ioctls must be issued from the same process (address space) that was 33 used to create the VM. 34 35 - vcpu ioctls: These query and set attributes that control the operation 36 of a single virtual cpu. 37 38 vcpu ioctls should be issued from the same thread that was used to create 39 the vcpu, except for asynchronous vcpu ioctl that are marked as such in 40 the documentation. Otherwise, the first ioctl after switching threads 41 could see a performance impact. 42 43 - device ioctls: These query and set attributes that control the operation 44 of a single device. 45 46 device ioctls must be issued from the same process (address space) that 47 was used to create the VM. 48 49While most ioctls are specific to one kind of file descriptor, in some 50cases the same ioctl can belong to more than one class. 51 52The KVM API grew over time. For this reason, KVM defines many constants 53of the form ``KVM_CAP_*``, each corresponding to a set of functionality 54provided by one or more ioctls. Availability of these "capabilities" can 55be checked with :ref:`KVM_CHECK_EXTENSION <KVM_CHECK_EXTENSION>`. Some 56capabilities also need to be enabled for VMs or VCPUs where their 57functionality is desired (see :ref:`cap_enable` and :ref:`cap_enable_vm`). 58 59 602. Restrictions 61=============== 62 63In general file descriptors can be migrated among processes by means 64of fork() and the SCM_RIGHTS facility of unix domain socket. These 65kinds of tricks are explicitly not supported by kvm. While they will 66not cause harm to the host, their actual behavior is not guaranteed by 67the API. See "General description" for details on the ioctl usage 68model that is supported by KVM. 69 70It is important to note that although VM ioctls may only be issued from 71the process that created the VM, a VM's lifecycle is associated with its 72file descriptor, not its creator (process). In other words, the VM and 73its resources, *including the associated address space*, are not freed 74until the last reference to the VM's file descriptor has been released. 75For example, if fork() is issued after ioctl(KVM_CREATE_VM), the VM will 76not be freed until both the parent (original) process and its child have 77put their references to the VM's file descriptor. 78 79Because a VM's resources are not freed until the last reference to its 80file descriptor is released, creating additional references to a VM 81via fork(), dup(), etc... without careful consideration is strongly 82discouraged and may have unwanted side effects, e.g. memory allocated 83by and on behalf of the VM's process may not be freed/unaccounted when 84the VM is shut down. 85 86 873. Extensions 88============= 89 90As of Linux 2.6.22, the KVM ABI has been stabilized: no backward 91incompatible change are allowed. However, there is an extension 92facility that allows backward-compatible extensions to the API to be 93queried and used. 94 95The extension mechanism is not based on the Linux version number. 96Instead, kvm defines extension identifiers and a facility to query 97whether a particular extension identifier is available. If it is, a 98set of ioctls is available for application use. 99 100 1014. API description 102================== 103 104This section describes ioctls that can be used to control kvm guests. 105For each ioctl, the following information is provided along with a 106description: 107 108 Capability: 109 which KVM extension provides this ioctl. Can be 'basic', 110 which means that is will be provided by any kernel that supports 111 API version 12 (see :ref:`KVM_GET_API_VERSION <KVM_GET_API_VERSION>`), 112 or a KVM_CAP_xyz constant that can be checked with 113 :ref:`KVM_CHECK_EXTENSION <KVM_CHECK_EXTENSION>`. 114 115 Architectures: 116 which instruction set architectures provide this ioctl. 117 x86 includes both i386 and x86_64. 118 119 Type: 120 system, vm, or vcpu. 121 122 Parameters: 123 what parameters are accepted by the ioctl. 124 125 Returns: 126 the return value. General error numbers (EBADF, ENOMEM, EINVAL) 127 are not detailed, but errors with specific meanings are. 128 129 130.. _KVM_GET_API_VERSION: 131 1324.1 KVM_GET_API_VERSION 133----------------------- 134 135:Capability: basic 136:Architectures: all 137:Type: system ioctl 138:Parameters: none 139:Returns: the constant KVM_API_VERSION (=12) 140 141This identifies the API version as the stable kvm API. It is not 142expected that this number will change. However, Linux 2.6.20 and 1432.6.21 report earlier versions; these are not documented and not 144supported. Applications should refuse to run if KVM_GET_API_VERSION 145returns a value other than 12. If this check passes, all ioctls 146described as 'basic' will be available. 147 148 1494.2 KVM_CREATE_VM 150----------------- 151 152:Capability: basic 153:Architectures: all 154:Type: system ioctl 155:Parameters: machine type identifier (KVM_VM_*) 156:Returns: a VM fd that can be used to control the new virtual machine. 157 158The new VM has no virtual cpus and no memory. 159You probably want to use 0 as machine type. 160 161X86: 162^^^^ 163 164Supported X86 VM types can be queried via KVM_CAP_VM_TYPES. 165 166S390: 167^^^^^ 168 169In order to create user controlled virtual machines on S390, check 170KVM_CAP_S390_UCONTROL and use the flag KVM_VM_S390_UCONTROL as 171privileged user (CAP_SYS_ADMIN). 172 173MIPS: 174^^^^^ 175 176To use hardware assisted virtualization on MIPS (VZ ASE) rather than 177the default trap & emulate implementation (which changes the virtual 178memory layout to fit in user mode), check KVM_CAP_MIPS_VZ and use the 179flag KVM_VM_MIPS_VZ. 180 181ARM64: 182^^^^^^ 183 184On arm64, the physical address size for a VM (IPA Size limit) is limited 185to 40bits by default. The limit can be configured if the host supports the 186extension KVM_CAP_ARM_VM_IPA_SIZE. When supported, use 187KVM_VM_TYPE_ARM_IPA_SIZE(IPA_Bits) to set the size in the machine type 188identifier, where IPA_Bits is the maximum width of any physical 189address used by the VM. The IPA_Bits is encoded in bits[7-0] of the 190machine type identifier. 191 192e.g, to configure a guest to use 48bit physical address size:: 193 194 vm_fd = ioctl(dev_fd, KVM_CREATE_VM, KVM_VM_TYPE_ARM_IPA_SIZE(48)); 195 196The requested size (IPA_Bits) must be: 197 198 == ========================================================= 199 0 Implies default size, 40bits (for backward compatibility) 200 N Implies N bits, where N is a positive integer such that, 201 32 <= N <= Host_IPA_Limit 202 == ========================================================= 203 204Host_IPA_Limit is the maximum possible value for IPA_Bits on the host and 205is dependent on the CPU capability and the kernel configuration. The limit can 206be retrieved using KVM_CAP_ARM_VM_IPA_SIZE of the KVM_CHECK_EXTENSION 207ioctl() at run-time. 208 209Creation of the VM will fail if the requested IPA size (whether it is 210implicit or explicit) is unsupported on the host. 211 212Please note that configuring the IPA size does not affect the capability 213exposed by the guest CPUs in ID_AA64MMFR0_EL1[PARange]. It only affects 214size of the address translated by the stage2 level (guest physical to 215host physical address translations). 216 217 2184.3 KVM_GET_MSR_INDEX_LIST, KVM_GET_MSR_FEATURE_INDEX_LIST 219---------------------------------------------------------- 220 221:Capability: basic, KVM_CAP_GET_MSR_FEATURES for KVM_GET_MSR_FEATURE_INDEX_LIST 222:Architectures: x86 223:Type: system ioctl 224:Parameters: struct kvm_msr_list (in/out) 225:Returns: 0 on success; -1 on error 226 227Errors: 228 229 ====== ============================================================ 230 EFAULT the msr index list cannot be read from or written to 231 E2BIG the msr index list is too big to fit in the array specified by 232 the user. 233 ====== ============================================================ 234 235:: 236 237 struct kvm_msr_list { 238 __u32 nmsrs; /* number of msrs in entries */ 239 __u32 indices[0]; 240 }; 241 242The user fills in the size of the indices array in nmsrs, and in return 243kvm adjusts nmsrs to reflect the actual number of msrs and fills in the 244indices array with their numbers. 245 246KVM_GET_MSR_INDEX_LIST returns the guest msrs that are supported. The list 247varies by kvm version and host processor, but does not change otherwise. 248 249Note: if kvm indicates supports MCE (KVM_CAP_MCE), then the MCE bank MSRs are 250not returned in the MSR list, as different vcpus can have a different number 251of banks, as set via the KVM_X86_SETUP_MCE ioctl. 252 253KVM_GET_MSR_FEATURE_INDEX_LIST returns the list of MSRs that can be passed 254to the KVM_GET_MSRS system ioctl. This lets userspace probe host capabilities 255and processor features that are exposed via MSRs (e.g., VMX capabilities). 256This list also varies by kvm version and host processor, but does not change 257otherwise. 258 259 260.. _KVM_CHECK_EXTENSION: 261 2624.4 KVM_CHECK_EXTENSION 263----------------------- 264 265:Capability: basic, KVM_CAP_CHECK_EXTENSION_VM for vm ioctl 266:Architectures: all 267:Type: system ioctl, vm ioctl 268:Parameters: extension identifier (KVM_CAP_*) 269:Returns: 0 if unsupported; 1 (or some other positive integer) if supported 270 271The API allows the application to query about extensions to the core 272kvm API. Userspace passes an extension identifier (an integer) and 273receives an integer that describes the extension availability. 274Generally 0 means no and 1 means yes, but some extensions may report 275additional information in the integer return value. 276 277Based on their initialization different VMs may have different capabilities. 278It is thus encouraged to use the vm ioctl to query for capabilities (available 279with KVM_CAP_CHECK_EXTENSION_VM on the vm fd) 280 2814.5 KVM_GET_VCPU_MMAP_SIZE 282-------------------------- 283 284:Capability: basic 285:Architectures: all 286:Type: system ioctl 287:Parameters: none 288:Returns: size of vcpu mmap area, in bytes 289 290The KVM_RUN ioctl (cf.) communicates with userspace via a shared 291memory region. This ioctl returns the size of that region. See the 292KVM_RUN documentation for details. 293 294Besides the size of the KVM_RUN communication region, other areas of 295the VCPU file descriptor can be mmap-ed, including: 296 297- if KVM_CAP_COALESCED_MMIO is available, a page at 298 KVM_COALESCED_MMIO_PAGE_OFFSET * PAGE_SIZE; for historical reasons, 299 this page is included in the result of KVM_GET_VCPU_MMAP_SIZE. 300 KVM_CAP_COALESCED_MMIO is not documented yet. 301 302- if KVM_CAP_DIRTY_LOG_RING is available, a number of pages at 303 KVM_DIRTY_LOG_PAGE_OFFSET * PAGE_SIZE. For more information on 304 KVM_CAP_DIRTY_LOG_RING, see :ref:`KVM_CAP_DIRTY_LOG_RING`. 305 306 3074.7 KVM_CREATE_VCPU 308------------------- 309 310:Capability: basic 311:Architectures: all 312:Type: vm ioctl 313:Parameters: vcpu id (apic id on x86) 314:Returns: vcpu fd on success, -1 on error 315 316This API adds a vcpu to a virtual machine. No more than max_vcpus may be added. 317The vcpu id is an integer in the range [0, max_vcpu_id). 318 319The recommended max_vcpus value can be retrieved using the KVM_CAP_NR_VCPUS of 320the KVM_CHECK_EXTENSION ioctl() at run-time. 321The maximum possible value for max_vcpus can be retrieved using the 322KVM_CAP_MAX_VCPUS of the KVM_CHECK_EXTENSION ioctl() at run-time. 323 324If the KVM_CAP_NR_VCPUS does not exist, you should assume that max_vcpus is 4 325cpus max. 326If the KVM_CAP_MAX_VCPUS does not exist, you should assume that max_vcpus is 327same as the value returned from KVM_CAP_NR_VCPUS. 328 329The maximum possible value for max_vcpu_id can be retrieved using the 330KVM_CAP_MAX_VCPU_ID of the KVM_CHECK_EXTENSION ioctl() at run-time. 331 332If the KVM_CAP_MAX_VCPU_ID does not exist, you should assume that max_vcpu_id 333is the same as the value returned from KVM_CAP_MAX_VCPUS. 334 335On powerpc using book3s_hv mode, the vcpus are mapped onto virtual 336threads in one or more virtual CPU cores. (This is because the 337hardware requires all the hardware threads in a CPU core to be in the 338same partition.) The KVM_CAP_PPC_SMT capability indicates the number 339of vcpus per virtual core (vcore). The vcore id is obtained by 340dividing the vcpu id by the number of vcpus per vcore. The vcpus in a 341given vcore will always be in the same physical core as each other 342(though that might be a different physical core from time to time). 343Userspace can control the threading (SMT) mode of the guest by its 344allocation of vcpu ids. For example, if userspace wants 345single-threaded guest vcpus, it should make all vcpu ids be a multiple 346of the number of vcpus per vcore. 347 348For virtual cpus that have been created with S390 user controlled virtual 349machines, the resulting vcpu fd can be memory mapped at page offset 350KVM_S390_SIE_PAGE_OFFSET in order to obtain a memory map of the virtual 351cpu's hardware control block. 352 353 3544.8 KVM_GET_DIRTY_LOG 355--------------------- 356 357:Capability: basic 358:Architectures: all 359:Type: vm ioctl 360:Parameters: struct kvm_dirty_log (in/out) 361:Returns: 0 on success, -1 on error 362 363:: 364 365 /* for KVM_GET_DIRTY_LOG */ 366 struct kvm_dirty_log { 367 __u32 slot; 368 __u32 padding; 369 union { 370 void __user *dirty_bitmap; /* one bit per page */ 371 __u64 padding; 372 }; 373 }; 374 375Given a memory slot, return a bitmap containing any pages dirtied 376since the last call to this ioctl. Bit 0 is the first page in the 377memory slot. Ensure the entire structure is cleared to avoid padding 378issues. 379 380If KVM_CAP_MULTI_ADDRESS_SPACE is available, bits 16-31 of slot field specifies 381the address space for which you want to return the dirty bitmap. See 382KVM_SET_USER_MEMORY_REGION for details on the usage of slot field. 383 384The bits in the dirty bitmap are cleared before the ioctl returns, unless 385KVM_CAP_MANUAL_DIRTY_LOG_PROTECT2 is enabled. For more information, 386see the description of the capability. 387 388Note that the Xen shared_info page, if configured, shall always be assumed 389to be dirty. KVM will not explicitly mark it such. 390 391 3924.10 KVM_RUN 393------------ 394 395:Capability: basic 396:Architectures: all 397:Type: vcpu ioctl 398:Parameters: none 399:Returns: 0 on success, -1 on error 400 401Errors: 402 403 ======= ============================================================== 404 EINTR an unmasked signal is pending 405 ENOEXEC the vcpu hasn't been initialized or the guest tried to execute 406 instructions from device memory (arm64) 407 ENOSYS data abort outside memslots with no syndrome info and 408 KVM_CAP_ARM_NISV_TO_USER not enabled (arm64) 409 EPERM SVE feature set but not finalized (arm64) 410 ======= ============================================================== 411 412This ioctl is used to run a guest virtual cpu. While there are no 413explicit parameters, there is an implicit parameter block that can be 414obtained by mmap()ing the vcpu fd at offset 0, with the size given by 415KVM_GET_VCPU_MMAP_SIZE. The parameter block is formatted as a 'struct 416kvm_run' (see below). 417 418 4194.11 KVM_GET_REGS 420----------------- 421 422:Capability: basic 423:Architectures: all except arm64 424:Type: vcpu ioctl 425:Parameters: struct kvm_regs (out) 426:Returns: 0 on success, -1 on error 427 428Reads the general purpose registers from the vcpu. 429 430:: 431 432 /* x86 */ 433 struct kvm_regs { 434 /* out (KVM_GET_REGS) / in (KVM_SET_REGS) */ 435 __u64 rax, rbx, rcx, rdx; 436 __u64 rsi, rdi, rsp, rbp; 437 __u64 r8, r9, r10, r11; 438 __u64 r12, r13, r14, r15; 439 __u64 rip, rflags; 440 }; 441 442 /* mips */ 443 struct kvm_regs { 444 /* out (KVM_GET_REGS) / in (KVM_SET_REGS) */ 445 __u64 gpr[32]; 446 __u64 hi; 447 __u64 lo; 448 __u64 pc; 449 }; 450 451 /* LoongArch */ 452 struct kvm_regs { 453 /* out (KVM_GET_REGS) / in (KVM_SET_REGS) */ 454 unsigned long gpr[32]; 455 unsigned long pc; 456 }; 457 458 4594.12 KVM_SET_REGS 460----------------- 461 462:Capability: basic 463:Architectures: all except arm64 464:Type: vcpu ioctl 465:Parameters: struct kvm_regs (in) 466:Returns: 0 on success, -1 on error 467 468Writes the general purpose registers into the vcpu. 469 470See KVM_GET_REGS for the data structure. 471 472 4734.13 KVM_GET_SREGS 474------------------ 475 476:Capability: basic 477:Architectures: x86, ppc 478:Type: vcpu ioctl 479:Parameters: struct kvm_sregs (out) 480:Returns: 0 on success, -1 on error 481 482Reads special registers from the vcpu. 483 484:: 485 486 /* x86 */ 487 struct kvm_sregs { 488 struct kvm_segment cs, ds, es, fs, gs, ss; 489 struct kvm_segment tr, ldt; 490 struct kvm_dtable gdt, idt; 491 __u64 cr0, cr2, cr3, cr4, cr8; 492 __u64 efer; 493 __u64 apic_base; 494 __u64 interrupt_bitmap[(KVM_NR_INTERRUPTS + 63) / 64]; 495 }; 496 497 /* ppc -- see arch/powerpc/include/uapi/asm/kvm.h */ 498 499interrupt_bitmap is a bitmap of pending external interrupts. At most 500one bit may be set. This interrupt has been acknowledged by the APIC 501but not yet injected into the cpu core. 502 503 5044.14 KVM_SET_SREGS 505------------------ 506 507:Capability: basic 508:Architectures: x86, ppc 509:Type: vcpu ioctl 510:Parameters: struct kvm_sregs (in) 511:Returns: 0 on success, -1 on error 512 513Writes special registers into the vcpu. See KVM_GET_SREGS for the 514data structures. 515 516 5174.15 KVM_TRANSLATE 518------------------ 519 520:Capability: basic 521:Architectures: x86 522:Type: vcpu ioctl 523:Parameters: struct kvm_translation (in/out) 524:Returns: 0 on success, -1 on error 525 526Translates a virtual address according to the vcpu's current address 527translation mode. 528 529:: 530 531 struct kvm_translation { 532 /* in */ 533 __u64 linear_address; 534 535 /* out */ 536 __u64 physical_address; 537 __u8 valid; 538 __u8 writeable; 539 __u8 usermode; 540 __u8 pad[5]; 541 }; 542 543 5444.16 KVM_INTERRUPT 545------------------ 546 547:Capability: basic 548:Architectures: x86, ppc, mips, riscv, loongarch 549:Type: vcpu ioctl 550:Parameters: struct kvm_interrupt (in) 551:Returns: 0 on success, negative on failure. 552 553Queues a hardware interrupt vector to be injected. 554 555:: 556 557 /* for KVM_INTERRUPT */ 558 struct kvm_interrupt { 559 /* in */ 560 __u32 irq; 561 }; 562 563X86: 564^^^^ 565 566:Returns: 567 568 ========= =================================== 569 0 on success, 570 -EEXIST if an interrupt is already enqueued 571 -EINVAL the irq number is invalid 572 -ENXIO if the PIC is in the kernel 573 -EFAULT if the pointer is invalid 574 ========= =================================== 575 576Note 'irq' is an interrupt vector, not an interrupt pin or line. This 577ioctl is useful if the in-kernel PIC is not used. 578 579PPC: 580^^^^ 581 582Queues an external interrupt to be injected. This ioctl is overloaded 583with 3 different irq values: 584 585a) KVM_INTERRUPT_SET 586 587 This injects an edge type external interrupt into the guest once it's ready 588 to receive interrupts. When injected, the interrupt is done. 589 590b) KVM_INTERRUPT_UNSET 591 592 This unsets any pending interrupt. 593 594 Only available with KVM_CAP_PPC_UNSET_IRQ. 595 596c) KVM_INTERRUPT_SET_LEVEL 597 598 This injects a level type external interrupt into the guest context. The 599 interrupt stays pending until a specific ioctl with KVM_INTERRUPT_UNSET 600 is triggered. 601 602 Only available with KVM_CAP_PPC_IRQ_LEVEL. 603 604Note that any value for 'irq' other than the ones stated above is invalid 605and incurs unexpected behavior. 606 607This is an asynchronous vcpu ioctl and can be invoked from any thread. 608 609MIPS: 610^^^^^ 611 612Queues an external interrupt to be injected into the virtual CPU. A negative 613interrupt number dequeues the interrupt. 614 615This is an asynchronous vcpu ioctl and can be invoked from any thread. 616 617RISC-V: 618^^^^^^^ 619 620Queues an external interrupt to be injected into the virtual CPU. This ioctl 621is overloaded with 2 different irq values: 622 623a) KVM_INTERRUPT_SET 624 625 This sets external interrupt for a virtual CPU and it will receive 626 once it is ready. 627 628b) KVM_INTERRUPT_UNSET 629 630 This clears pending external interrupt for a virtual CPU. 631 632This is an asynchronous vcpu ioctl and can be invoked from any thread. 633 634LOONGARCH: 635^^^^^^^^^^ 636 637Queues an external interrupt to be injected into the virtual CPU. A negative 638interrupt number dequeues the interrupt. 639 640This is an asynchronous vcpu ioctl and can be invoked from any thread. 641 642 6434.18 KVM_GET_MSRS 644----------------- 645 646:Capability: basic (vcpu), KVM_CAP_GET_MSR_FEATURES (system) 647:Architectures: x86 648:Type: system ioctl, vcpu ioctl 649:Parameters: struct kvm_msrs (in/out) 650:Returns: number of msrs successfully returned; 651 -1 on error 652 653When used as a system ioctl: 654Reads the values of MSR-based features that are available for the VM. This 655is similar to KVM_GET_SUPPORTED_CPUID, but it returns MSR indices and values. 656The list of msr-based features can be obtained using KVM_GET_MSR_FEATURE_INDEX_LIST 657in a system ioctl. 658 659When used as a vcpu ioctl: 660Reads model-specific registers from the vcpu. Supported msr indices can 661be obtained using KVM_GET_MSR_INDEX_LIST in a system ioctl. 662 663:: 664 665 struct kvm_msrs { 666 __u32 nmsrs; /* number of msrs in entries */ 667 __u32 pad; 668 669 struct kvm_msr_entry entries[0]; 670 }; 671 672 struct kvm_msr_entry { 673 __u32 index; 674 __u32 reserved; 675 __u64 data; 676 }; 677 678Application code should set the 'nmsrs' member (which indicates the 679size of the entries array) and the 'index' member of each array entry. 680kvm will fill in the 'data' member. 681 682 6834.19 KVM_SET_MSRS 684----------------- 685 686:Capability: basic 687:Architectures: x86 688:Type: vcpu ioctl 689:Parameters: struct kvm_msrs (in) 690:Returns: number of msrs successfully set (see below), -1 on error 691 692Writes model-specific registers to the vcpu. See KVM_GET_MSRS for the 693data structures. 694 695Application code should set the 'nmsrs' member (which indicates the 696size of the entries array), and the 'index' and 'data' members of each 697array entry. 698 699It tries to set the MSRs in array entries[] one by one. If setting an MSR 700fails, e.g., due to setting reserved bits, the MSR isn't supported/emulated 701by KVM, etc..., it stops processing the MSR list and returns the number of 702MSRs that have been set successfully. 703 704 7054.20 KVM_SET_CPUID 706------------------ 707 708:Capability: basic 709:Architectures: x86 710:Type: vcpu ioctl 711:Parameters: struct kvm_cpuid (in) 712:Returns: 0 on success, -1 on error 713 714Defines the vcpu responses to the cpuid instruction. Applications 715should use the KVM_SET_CPUID2 ioctl if available. 716 717Caveat emptor: 718 - If this IOCTL fails, KVM gives no guarantees that previous valid CPUID 719 configuration (if there is) is not corrupted. Userspace can get a copy 720 of the resulting CPUID configuration through KVM_GET_CPUID2 in case. 721 - Using KVM_SET_CPUID{,2} after KVM_RUN, i.e. changing the guest vCPU model 722 after running the guest, may cause guest instability. 723 - Using heterogeneous CPUID configurations, modulo APIC IDs, topology, etc... 724 may cause guest instability. 725 726:: 727 728 struct kvm_cpuid_entry { 729 __u32 function; 730 __u32 eax; 731 __u32 ebx; 732 __u32 ecx; 733 __u32 edx; 734 __u32 padding; 735 }; 736 737 /* for KVM_SET_CPUID */ 738 struct kvm_cpuid { 739 __u32 nent; 740 __u32 padding; 741 struct kvm_cpuid_entry entries[0]; 742 }; 743 744 7454.21 KVM_SET_SIGNAL_MASK 746------------------------ 747 748:Capability: basic 749:Architectures: all 750:Type: vcpu ioctl 751:Parameters: struct kvm_signal_mask (in) 752:Returns: 0 on success, -1 on error 753 754Defines which signals are blocked during execution of KVM_RUN. This 755signal mask temporarily overrides the threads signal mask. Any 756unblocked signal received (except SIGKILL and SIGSTOP, which retain 757their traditional behaviour) will cause KVM_RUN to return with -EINTR. 758 759Note the signal will only be delivered if not blocked by the original 760signal mask. 761 762:: 763 764 /* for KVM_SET_SIGNAL_MASK */ 765 struct kvm_signal_mask { 766 __u32 len; 767 __u8 sigset[0]; 768 }; 769 770 7714.22 KVM_GET_FPU 772---------------- 773 774:Capability: basic 775:Architectures: x86, loongarch 776:Type: vcpu ioctl 777:Parameters: struct kvm_fpu (out) 778:Returns: 0 on success, -1 on error 779 780Reads the floating point state from the vcpu. 781 782:: 783 784 /* x86: for KVM_GET_FPU and KVM_SET_FPU */ 785 struct kvm_fpu { 786 __u8 fpr[8][16]; 787 __u16 fcw; 788 __u16 fsw; 789 __u8 ftwx; /* in fxsave format */ 790 __u8 pad1; 791 __u16 last_opcode; 792 __u64 last_ip; 793 __u64 last_dp; 794 __u8 xmm[16][16]; 795 __u32 mxcsr; 796 __u32 pad2; 797 }; 798 799 /* LoongArch: for KVM_GET_FPU and KVM_SET_FPU */ 800 struct kvm_fpu { 801 __u32 fcsr; 802 __u64 fcc; 803 struct kvm_fpureg { 804 __u64 val64[4]; 805 }fpr[32]; 806 }; 807 808 8094.23 KVM_SET_FPU 810---------------- 811 812:Capability: basic 813:Architectures: x86, loongarch 814:Type: vcpu ioctl 815:Parameters: struct kvm_fpu (in) 816:Returns: 0 on success, -1 on error 817 818Writes the floating point state to the vcpu. 819 820:: 821 822 /* x86: for KVM_GET_FPU and KVM_SET_FPU */ 823 struct kvm_fpu { 824 __u8 fpr[8][16]; 825 __u16 fcw; 826 __u16 fsw; 827 __u8 ftwx; /* in fxsave format */ 828 __u8 pad1; 829 __u16 last_opcode; 830 __u64 last_ip; 831 __u64 last_dp; 832 __u8 xmm[16][16]; 833 __u32 mxcsr; 834 __u32 pad2; 835 }; 836 837 /* LoongArch: for KVM_GET_FPU and KVM_SET_FPU */ 838 struct kvm_fpu { 839 __u32 fcsr; 840 __u64 fcc; 841 struct kvm_fpureg { 842 __u64 val64[4]; 843 }fpr[32]; 844 }; 845 846 8474.24 KVM_CREATE_IRQCHIP 848----------------------- 849 850:Capability: KVM_CAP_IRQCHIP, KVM_CAP_S390_IRQCHIP (s390) 851:Architectures: x86, arm64, s390 852:Type: vm ioctl 853:Parameters: none 854:Returns: 0 on success, -1 on error 855 856Creates an interrupt controller model in the kernel. 857On x86, creates a virtual ioapic, a virtual PIC (two PICs, nested), and sets up 858future vcpus to have a local APIC. IRQ routing for GSIs 0-15 is set to both 859PIC and IOAPIC; GSI 16-23 only go to the IOAPIC. 860On arm64, a GICv2 is created. Any other GIC versions require the usage of 861KVM_CREATE_DEVICE, which also supports creating a GICv2. Using 862KVM_CREATE_DEVICE is preferred over KVM_CREATE_IRQCHIP for GICv2. 863On s390, a dummy irq routing table is created. 864 865Note that on s390 the KVM_CAP_S390_IRQCHIP vm capability needs to be enabled 866before KVM_CREATE_IRQCHIP can be used. 867 868 8694.25 KVM_IRQ_LINE 870----------------- 871 872:Capability: KVM_CAP_IRQCHIP 873:Architectures: x86, arm64 874:Type: vm ioctl 875:Parameters: struct kvm_irq_level 876:Returns: 0 on success, -1 on error 877 878Sets the level of a GSI input to the interrupt controller model in the kernel. 879On some architectures it is required that an interrupt controller model has 880been previously created with KVM_CREATE_IRQCHIP. Note that edge-triggered 881interrupts require the level to be set to 1 and then back to 0. 882 883On real hardware, interrupt pins can be active-low or active-high. This 884does not matter for the level field of struct kvm_irq_level: 1 always 885means active (asserted), 0 means inactive (deasserted). 886 887x86 allows the operating system to program the interrupt polarity 888(active-low/active-high) for level-triggered interrupts, and KVM used 889to consider the polarity. However, due to bitrot in the handling of 890active-low interrupts, the above convention is now valid on x86 too. 891This is signaled by KVM_CAP_X86_IOAPIC_POLARITY_IGNORED. Userspace 892should not present interrupts to the guest as active-low unless this 893capability is present (or unless it is not using the in-kernel irqchip, 894of course). 895 896 897arm64 can signal an interrupt either at the CPU level, or at the 898in-kernel irqchip (GIC), and for in-kernel irqchip can tell the GIC to 899use PPIs designated for specific cpus. The irq field is interpreted 900like this:: 901 902 bits: | 31 ... 28 | 27 ... 24 | 23 ... 16 | 15 ... 0 | 903 field: | vcpu2_index | irq_type | vcpu_index | irq_id | 904 905The irq_type field has the following values: 906 907- KVM_ARM_IRQ_TYPE_CPU: 908 out-of-kernel GIC: irq_id 0 is IRQ, irq_id 1 is FIQ 909- KVM_ARM_IRQ_TYPE_SPI: 910 in-kernel GIC: SPI, irq_id between 32 and 1019 (incl.) 911 (the vcpu_index field is ignored) 912- KVM_ARM_IRQ_TYPE_PPI: 913 in-kernel GIC: PPI, irq_id between 16 and 31 (incl.) 914 915(The irq_id field thus corresponds nicely to the IRQ ID in the ARM GIC specs) 916 917In both cases, level is used to assert/deassert the line. 918 919When KVM_CAP_ARM_IRQ_LINE_LAYOUT_2 is supported, the target vcpu is 920identified as (256 * vcpu2_index + vcpu_index). Otherwise, vcpu2_index 921must be zero. 922 923Note that on arm64, the KVM_CAP_IRQCHIP capability only conditions 924injection of interrupts for the in-kernel irqchip. KVM_IRQ_LINE can always 925be used for a userspace interrupt controller. 926 927:: 928 929 struct kvm_irq_level { 930 union { 931 __u32 irq; /* GSI */ 932 __s32 status; /* not used for KVM_IRQ_LEVEL */ 933 }; 934 __u32 level; /* 0 or 1 */ 935 }; 936 937 9384.26 KVM_GET_IRQCHIP 939-------------------- 940 941:Capability: KVM_CAP_IRQCHIP 942:Architectures: x86 943:Type: vm ioctl 944:Parameters: struct kvm_irqchip (in/out) 945:Returns: 0 on success, -1 on error 946 947Reads the state of a kernel interrupt controller created with 948KVM_CREATE_IRQCHIP into a buffer provided by the caller. 949 950:: 951 952 struct kvm_irqchip { 953 __u32 chip_id; /* 0 = PIC1, 1 = PIC2, 2 = IOAPIC */ 954 __u32 pad; 955 union { 956 char dummy[512]; /* reserving space */ 957 struct kvm_pic_state pic; 958 struct kvm_ioapic_state ioapic; 959 } chip; 960 }; 961 962 9634.27 KVM_SET_IRQCHIP 964-------------------- 965 966:Capability: KVM_CAP_IRQCHIP 967:Architectures: x86 968:Type: vm ioctl 969:Parameters: struct kvm_irqchip (in) 970:Returns: 0 on success, -1 on error 971 972Sets the state of a kernel interrupt controller created with 973KVM_CREATE_IRQCHIP from a buffer provided by the caller. 974 975:: 976 977 struct kvm_irqchip { 978 __u32 chip_id; /* 0 = PIC1, 1 = PIC2, 2 = IOAPIC */ 979 __u32 pad; 980 union { 981 char dummy[512]; /* reserving space */ 982 struct kvm_pic_state pic; 983 struct kvm_ioapic_state ioapic; 984 } chip; 985 }; 986 987 9884.28 KVM_XEN_HVM_CONFIG 989----------------------- 990 991:Capability: KVM_CAP_XEN_HVM 992:Architectures: x86 993:Type: vm ioctl 994:Parameters: struct kvm_xen_hvm_config (in) 995:Returns: 0 on success, -1 on error 996 997Sets the MSR that the Xen HVM guest uses to initialize its hypercall 998page, and provides the starting address and size of the hypercall 999blobs in userspace. When the guest writes the MSR, kvm copies one 1000page of a blob (32- or 64-bit, depending on the vcpu mode) to guest 1001memory. 1002 1003The MSR index must be in the range [0x40000000, 0x4fffffff], i.e. must reside 1004in the range that is unofficially reserved for use by hypervisors. The min/max 1005values are enumerated via KVM_XEN_MSR_MIN_INDEX and KVM_XEN_MSR_MAX_INDEX. 1006 1007:: 1008 1009 struct kvm_xen_hvm_config { 1010 __u32 flags; 1011 __u32 msr; 1012 __u64 blob_addr_32; 1013 __u64 blob_addr_64; 1014 __u8 blob_size_32; 1015 __u8 blob_size_64; 1016 __u8 pad2[30]; 1017 }; 1018 1019If certain flags are returned from the KVM_CAP_XEN_HVM check, they may 1020be set in the flags field of this ioctl: 1021 1022The KVM_XEN_HVM_CONFIG_INTERCEPT_HCALL flag requests KVM to generate 1023the contents of the hypercall page automatically; hypercalls will be 1024intercepted and passed to userspace through KVM_EXIT_XEN. In this 1025case, all of the blob size and address fields must be zero. 1026 1027The KVM_XEN_HVM_CONFIG_EVTCHN_SEND flag indicates to KVM that userspace 1028will always use the KVM_XEN_HVM_EVTCHN_SEND ioctl to deliver event 1029channel interrupts rather than manipulating the guest's shared_info 1030structures directly. This, in turn, may allow KVM to enable features 1031such as intercepting the SCHEDOP_poll hypercall to accelerate PV 1032spinlock operation for the guest. Userspace may still use the ioctl 1033to deliver events if it was advertised, even if userspace does not 1034send this indication that it will always do so 1035 1036No other flags are currently valid in the struct kvm_xen_hvm_config. 1037 10384.29 KVM_GET_CLOCK 1039------------------ 1040 1041:Capability: KVM_CAP_ADJUST_CLOCK 1042:Architectures: x86 1043:Type: vm ioctl 1044:Parameters: struct kvm_clock_data (out) 1045:Returns: 0 on success, -1 on error 1046 1047Gets the current timestamp of kvmclock as seen by the current guest. In 1048conjunction with KVM_SET_CLOCK, it is used to ensure monotonicity on scenarios 1049such as migration. 1050 1051When KVM_CAP_ADJUST_CLOCK is passed to KVM_CHECK_EXTENSION, it returns the 1052set of bits that KVM can return in struct kvm_clock_data's flag member. 1053 1054The following flags are defined: 1055 1056KVM_CLOCK_TSC_STABLE 1057 If set, the returned value is the exact kvmclock 1058 value seen by all VCPUs at the instant when KVM_GET_CLOCK was called. 1059 If clear, the returned value is simply CLOCK_MONOTONIC plus a constant 1060 offset; the offset can be modified with KVM_SET_CLOCK. KVM will try 1061 to make all VCPUs follow this clock, but the exact value read by each 1062 VCPU could differ, because the host TSC is not stable. 1063 1064KVM_CLOCK_REALTIME 1065 If set, the `realtime` field in the kvm_clock_data 1066 structure is populated with the value of the host's real time 1067 clocksource at the instant when KVM_GET_CLOCK was called. If clear, 1068 the `realtime` field does not contain a value. 1069 1070KVM_CLOCK_HOST_TSC 1071 If set, the `host_tsc` field in the kvm_clock_data 1072 structure is populated with the value of the host's timestamp counter (TSC) 1073 at the instant when KVM_GET_CLOCK was called. If clear, the `host_tsc` field 1074 does not contain a value. 1075 1076:: 1077 1078 struct kvm_clock_data { 1079 __u64 clock; /* kvmclock current value */ 1080 __u32 flags; 1081 __u32 pad0; 1082 __u64 realtime; 1083 __u64 host_tsc; 1084 __u32 pad[4]; 1085 }; 1086 1087 10884.30 KVM_SET_CLOCK 1089------------------ 1090 1091:Capability: KVM_CAP_ADJUST_CLOCK 1092:Architectures: x86 1093:Type: vm ioctl 1094:Parameters: struct kvm_clock_data (in) 1095:Returns: 0 on success, -1 on error 1096 1097Sets the current timestamp of kvmclock to the value specified in its parameter. 1098In conjunction with KVM_GET_CLOCK, it is used to ensure monotonicity on scenarios 1099such as migration. 1100 1101The following flags can be passed: 1102 1103KVM_CLOCK_REALTIME 1104 If set, KVM will compare the value of the `realtime` field 1105 with the value of the host's real time clocksource at the instant when 1106 KVM_SET_CLOCK was called. The difference in elapsed time is added to the final 1107 kvmclock value that will be provided to guests. 1108 1109Other flags returned by ``KVM_GET_CLOCK`` are accepted but ignored. 1110 1111:: 1112 1113 struct kvm_clock_data { 1114 __u64 clock; /* kvmclock current value */ 1115 __u32 flags; 1116 __u32 pad0; 1117 __u64 realtime; 1118 __u64 host_tsc; 1119 __u32 pad[4]; 1120 }; 1121 1122 11234.31 KVM_GET_VCPU_EVENTS 1124------------------------ 1125 1126:Capability: KVM_CAP_VCPU_EVENTS 1127:Extended by: KVM_CAP_INTR_SHADOW 1128:Architectures: x86, arm64 1129:Type: vcpu ioctl 1130:Parameters: struct kvm_vcpu_events (out) 1131:Returns: 0 on success, -1 on error 1132 1133X86: 1134^^^^ 1135 1136Gets currently pending exceptions, interrupts, and NMIs as well as related 1137states of the vcpu. 1138 1139:: 1140 1141 struct kvm_vcpu_events { 1142 struct { 1143 __u8 injected; 1144 __u8 nr; 1145 __u8 has_error_code; 1146 __u8 pending; 1147 __u32 error_code; 1148 } exception; 1149 struct { 1150 __u8 injected; 1151 __u8 nr; 1152 __u8 soft; 1153 __u8 shadow; 1154 } interrupt; 1155 struct { 1156 __u8 injected; 1157 __u8 pending; 1158 __u8 masked; 1159 __u8 pad; 1160 } nmi; 1161 __u32 sipi_vector; 1162 __u32 flags; 1163 struct { 1164 __u8 smm; 1165 __u8 pending; 1166 __u8 smm_inside_nmi; 1167 __u8 latched_init; 1168 } smi; 1169 __u8 reserved[27]; 1170 __u8 exception_has_payload; 1171 __u64 exception_payload; 1172 }; 1173 1174The following bits are defined in the flags field: 1175 1176- KVM_VCPUEVENT_VALID_SHADOW may be set to signal that 1177 interrupt.shadow contains a valid state. 1178 1179- KVM_VCPUEVENT_VALID_SMM may be set to signal that smi contains a 1180 valid state. 1181 1182- KVM_VCPUEVENT_VALID_PAYLOAD may be set to signal that the 1183 exception_has_payload, exception_payload, and exception.pending 1184 fields contain a valid state. This bit will be set whenever 1185 KVM_CAP_EXCEPTION_PAYLOAD is enabled. 1186 1187- KVM_VCPUEVENT_VALID_TRIPLE_FAULT may be set to signal that the 1188 triple_fault_pending field contains a valid state. This bit will 1189 be set whenever KVM_CAP_X86_TRIPLE_FAULT_EVENT is enabled. 1190 1191ARM64: 1192^^^^^^ 1193 1194If the guest accesses a device that is being emulated by the host kernel in 1195such a way that a real device would generate a physical SError, KVM may make 1196a virtual SError pending for that VCPU. This system error interrupt remains 1197pending until the guest takes the exception by unmasking PSTATE.A. 1198 1199Running the VCPU may cause it to take a pending SError, or make an access that 1200causes an SError to become pending. The event's description is only valid while 1201the VPCU is not running. 1202 1203This API provides a way to read and write the pending 'event' state that is not 1204visible to the guest. To save, restore or migrate a VCPU the struct representing 1205the state can be read then written using this GET/SET API, along with the other 1206guest-visible registers. It is not possible to 'cancel' an SError that has been 1207made pending. 1208 1209A device being emulated in user-space may also wish to generate an SError. To do 1210this the events structure can be populated by user-space. The current state 1211should be read first, to ensure no existing SError is pending. If an existing 1212SError is pending, the architecture's 'Multiple SError interrupts' rules should 1213be followed. (2.5.3 of DDI0587.a "ARM Reliability, Availability, and 1214Serviceability (RAS) Specification"). 1215 1216SError exceptions always have an ESR value. Some CPUs have the ability to 1217specify what the virtual SError's ESR value should be. These systems will 1218advertise KVM_CAP_ARM_INJECT_SERROR_ESR. In this case exception.has_esr will 1219always have a non-zero value when read, and the agent making an SError pending 1220should specify the ISS field in the lower 24 bits of exception.serror_esr. If 1221the system supports KVM_CAP_ARM_INJECT_SERROR_ESR, but user-space sets the events 1222with exception.has_esr as zero, KVM will choose an ESR. 1223 1224Specifying exception.has_esr on a system that does not support it will return 1225-EINVAL. Setting anything other than the lower 24bits of exception.serror_esr 1226will return -EINVAL. 1227 1228It is not possible to read back a pending external abort (injected via 1229KVM_SET_VCPU_EVENTS or otherwise) because such an exception is always delivered 1230directly to the virtual CPU). 1231 1232:: 1233 1234 struct kvm_vcpu_events { 1235 struct { 1236 __u8 serror_pending; 1237 __u8 serror_has_esr; 1238 __u8 ext_dabt_pending; 1239 /* Align it to 8 bytes */ 1240 __u8 pad[5]; 1241 __u64 serror_esr; 1242 } exception; 1243 __u32 reserved[12]; 1244 }; 1245 12464.32 KVM_SET_VCPU_EVENTS 1247------------------------ 1248 1249:Capability: KVM_CAP_VCPU_EVENTS 1250:Extended by: KVM_CAP_INTR_SHADOW 1251:Architectures: x86, arm64 1252:Type: vcpu ioctl 1253:Parameters: struct kvm_vcpu_events (in) 1254:Returns: 0 on success, -1 on error 1255 1256X86: 1257^^^^ 1258 1259Set pending exceptions, interrupts, and NMIs as well as related states of the 1260vcpu. 1261 1262See KVM_GET_VCPU_EVENTS for the data structure. 1263 1264Fields that may be modified asynchronously by running VCPUs can be excluded 1265from the update. These fields are nmi.pending, sipi_vector, smi.smm, 1266smi.pending. Keep the corresponding bits in the flags field cleared to 1267suppress overwriting the current in-kernel state. The bits are: 1268 1269=============================== ================================== 1270KVM_VCPUEVENT_VALID_NMI_PENDING transfer nmi.pending to the kernel 1271KVM_VCPUEVENT_VALID_SIPI_VECTOR transfer sipi_vector 1272KVM_VCPUEVENT_VALID_SMM transfer the smi sub-struct. 1273=============================== ================================== 1274 1275If KVM_CAP_INTR_SHADOW is available, KVM_VCPUEVENT_VALID_SHADOW can be set in 1276the flags field to signal that interrupt.shadow contains a valid state and 1277shall be written into the VCPU. 1278 1279KVM_VCPUEVENT_VALID_SMM can only be set if KVM_CAP_X86_SMM is available. 1280 1281If KVM_CAP_EXCEPTION_PAYLOAD is enabled, KVM_VCPUEVENT_VALID_PAYLOAD 1282can be set in the flags field to signal that the 1283exception_has_payload, exception_payload, and exception.pending fields 1284contain a valid state and shall be written into the VCPU. 1285 1286If KVM_CAP_X86_TRIPLE_FAULT_EVENT is enabled, KVM_VCPUEVENT_VALID_TRIPLE_FAULT 1287can be set in flags field to signal that the triple_fault field contains 1288a valid state and shall be written into the VCPU. 1289 1290ARM64: 1291^^^^^^ 1292 1293User space may need to inject several types of events to the guest. 1294 1295Set the pending SError exception state for this VCPU. It is not possible to 1296'cancel' an Serror that has been made pending. 1297 1298If the guest performed an access to I/O memory which could not be handled by 1299userspace, for example because of missing instruction syndrome decode 1300information or because there is no device mapped at the accessed IPA, then 1301userspace can ask the kernel to inject an external abort using the address 1302from the exiting fault on the VCPU. It is a programming error to set 1303ext_dabt_pending after an exit which was not either KVM_EXIT_MMIO or 1304KVM_EXIT_ARM_NISV. This feature is only available if the system supports 1305KVM_CAP_ARM_INJECT_EXT_DABT. This is a helper which provides commonality in 1306how userspace reports accesses for the above cases to guests, across different 1307userspace implementations. Nevertheless, userspace can still emulate all Arm 1308exceptions by manipulating individual registers using the KVM_SET_ONE_REG API. 1309 1310See KVM_GET_VCPU_EVENTS for the data structure. 1311 1312 13134.33 KVM_GET_DEBUGREGS 1314---------------------- 1315 1316:Capability: KVM_CAP_DEBUGREGS 1317:Architectures: x86 1318:Type: vcpu ioctl 1319:Parameters: struct kvm_debugregs (out) 1320:Returns: 0 on success, -1 on error 1321 1322Reads debug registers from the vcpu. 1323 1324:: 1325 1326 struct kvm_debugregs { 1327 __u64 db[4]; 1328 __u64 dr6; 1329 __u64 dr7; 1330 __u64 flags; 1331 __u64 reserved[9]; 1332 }; 1333 1334 13354.34 KVM_SET_DEBUGREGS 1336---------------------- 1337 1338:Capability: KVM_CAP_DEBUGREGS 1339:Architectures: x86 1340:Type: vcpu ioctl 1341:Parameters: struct kvm_debugregs (in) 1342:Returns: 0 on success, -1 on error 1343 1344Writes debug registers into the vcpu. 1345 1346See KVM_GET_DEBUGREGS for the data structure. The flags field is unused 1347yet and must be cleared on entry. 1348 1349 13504.35 KVM_SET_USER_MEMORY_REGION 1351------------------------------- 1352 1353:Capability: KVM_CAP_USER_MEMORY 1354:Architectures: all 1355:Type: vm ioctl 1356:Parameters: struct kvm_userspace_memory_region (in) 1357:Returns: 0 on success, -1 on error 1358 1359:: 1360 1361 struct kvm_userspace_memory_region { 1362 __u32 slot; 1363 __u32 flags; 1364 __u64 guest_phys_addr; 1365 __u64 memory_size; /* bytes */ 1366 __u64 userspace_addr; /* start of the userspace allocated memory */ 1367 }; 1368 1369 /* for kvm_userspace_memory_region::flags */ 1370 #define KVM_MEM_LOG_DIRTY_PAGES (1UL << 0) 1371 #define KVM_MEM_READONLY (1UL << 1) 1372 1373This ioctl allows the user to create, modify or delete a guest physical 1374memory slot. Bits 0-15 of "slot" specify the slot id and this value 1375should be less than the maximum number of user memory slots supported per 1376VM. The maximum allowed slots can be queried using KVM_CAP_NR_MEMSLOTS. 1377Slots may not overlap in guest physical address space. 1378 1379If KVM_CAP_MULTI_ADDRESS_SPACE is available, bits 16-31 of "slot" 1380specifies the address space which is being modified. They must be 1381less than the value that KVM_CHECK_EXTENSION returns for the 1382KVM_CAP_MULTI_ADDRESS_SPACE capability. Slots in separate address spaces 1383are unrelated; the restriction on overlapping slots only applies within 1384each address space. 1385 1386Deleting a slot is done by passing zero for memory_size. When changing 1387an existing slot, it may be moved in the guest physical memory space, 1388or its flags may be modified, but it may not be resized. 1389 1390Memory for the region is taken starting at the address denoted by the 1391field userspace_addr, which must point at user addressable memory for 1392the entire memory slot size. Any object may back this memory, including 1393anonymous memory, ordinary files, and hugetlbfs. 1394 1395On architectures that support a form of address tagging, userspace_addr must 1396be an untagged address. 1397 1398It is recommended that the lower 21 bits of guest_phys_addr and userspace_addr 1399be identical. This allows large pages in the guest to be backed by large 1400pages in the host. 1401 1402The flags field supports two flags: KVM_MEM_LOG_DIRTY_PAGES and 1403KVM_MEM_READONLY. The former can be set to instruct KVM to keep track of 1404writes to memory within the slot. See KVM_GET_DIRTY_LOG ioctl to know how to 1405use it. The latter can be set, if KVM_CAP_READONLY_MEM capability allows it, 1406to make a new slot read-only. In this case, writes to this memory will be 1407posted to userspace as KVM_EXIT_MMIO exits. 1408 1409When the KVM_CAP_SYNC_MMU capability is available, changes in the backing of 1410the memory region are automatically reflected into the guest. For example, an 1411mmap() that affects the region will be made visible immediately. Another 1412example is madvise(MADV_DROP). 1413 1414For TDX guest, deleting/moving memory region loses guest memory contents. 1415Read only region isn't supported. Only as-id 0 is supported. 1416 1417Note: On arm64, a write generated by the page-table walker (to update 1418the Access and Dirty flags, for example) never results in a 1419KVM_EXIT_MMIO exit when the slot has the KVM_MEM_READONLY flag. This 1420is because KVM cannot provide the data that would be written by the 1421page-table walker, making it impossible to emulate the access. 1422Instead, an abort (data abort if the cause of the page-table update 1423was a load or a store, instruction abort if it was an instruction 1424fetch) is injected in the guest. 1425 1426S390: 1427^^^^^ 1428 1429Returns -EINVAL or -EEXIST if the VM has the KVM_VM_S390_UCONTROL flag set. 1430Returns -EINVAL if called on a protected VM. 1431 14324.36 KVM_SET_TSS_ADDR 1433--------------------- 1434 1435:Capability: KVM_CAP_SET_TSS_ADDR 1436:Architectures: x86 1437:Type: vm ioctl 1438:Parameters: unsigned long tss_address (in) 1439:Returns: 0 on success, -1 on error 1440 1441This ioctl defines the physical address of a three-page region in the guest 1442physical address space. The region must be within the first 4GB of the 1443guest physical address space and must not conflict with any memory slot 1444or any mmio address. The guest may malfunction if it accesses this memory 1445region. 1446 1447This ioctl is required on Intel-based hosts. This is needed on Intel hardware 1448because of a quirk in the virtualization implementation (see the internals 1449documentation when it pops into existence). 1450 1451 1452.. _KVM_ENABLE_CAP: 1453 14544.37 KVM_ENABLE_CAP 1455------------------- 1456 1457:Capability: KVM_CAP_ENABLE_CAP 1458:Architectures: mips, ppc, s390, x86, loongarch 1459:Type: vcpu ioctl 1460:Parameters: struct kvm_enable_cap (in) 1461:Returns: 0 on success; -1 on error 1462 1463:Capability: KVM_CAP_ENABLE_CAP_VM 1464:Architectures: all 1465:Type: vm ioctl 1466:Parameters: struct kvm_enable_cap (in) 1467:Returns: 0 on success; -1 on error 1468 1469.. note:: 1470 1471 Not all extensions are enabled by default. Using this ioctl the application 1472 can enable an extension, making it available to the guest. 1473 1474On systems that do not support this ioctl, it always fails. On systems that 1475do support it, it only works for extensions that are supported for enablement. 1476 1477To check if a capability can be enabled, the KVM_CHECK_EXTENSION ioctl should 1478be used. 1479 1480:: 1481 1482 struct kvm_enable_cap { 1483 /* in */ 1484 __u32 cap; 1485 1486The capability that is supposed to get enabled. 1487 1488:: 1489 1490 __u32 flags; 1491 1492A bitfield indicating future enhancements. Has to be 0 for now. 1493 1494:: 1495 1496 __u64 args[4]; 1497 1498Arguments for enabling a feature. If a feature needs initial values to 1499function properly, this is the place to put them. 1500 1501:: 1502 1503 __u8 pad[64]; 1504 }; 1505 1506The vcpu ioctl should be used for vcpu-specific capabilities, the vm ioctl 1507for vm-wide capabilities. 1508 15094.38 KVM_GET_MP_STATE 1510--------------------- 1511 1512:Capability: KVM_CAP_MP_STATE 1513:Architectures: x86, s390, arm64, riscv, loongarch 1514:Type: vcpu ioctl 1515:Parameters: struct kvm_mp_state (out) 1516:Returns: 0 on success; -1 on error 1517 1518:: 1519 1520 struct kvm_mp_state { 1521 __u32 mp_state; 1522 }; 1523 1524Returns the vcpu's current "multiprocessing state" (though also valid on 1525uniprocessor guests). 1526 1527Possible values are: 1528 1529 ========================== =============================================== 1530 KVM_MP_STATE_RUNNABLE the vcpu is currently running 1531 [x86,arm64,riscv,loongarch] 1532 KVM_MP_STATE_UNINITIALIZED the vcpu is an application processor (AP) 1533 which has not yet received an INIT signal [x86] 1534 KVM_MP_STATE_INIT_RECEIVED the vcpu has received an INIT signal, and is 1535 now ready for a SIPI [x86] 1536 KVM_MP_STATE_HALTED the vcpu has executed a HLT instruction and 1537 is waiting for an interrupt [x86] 1538 KVM_MP_STATE_SIPI_RECEIVED the vcpu has just received a SIPI (vector 1539 accessible via KVM_GET_VCPU_EVENTS) [x86] 1540 KVM_MP_STATE_STOPPED the vcpu is stopped [s390,arm64,riscv] 1541 KVM_MP_STATE_CHECK_STOP the vcpu is in a special error state [s390] 1542 KVM_MP_STATE_OPERATING the vcpu is operating (running or halted) 1543 [s390] 1544 KVM_MP_STATE_LOAD the vcpu is in a special load/startup state 1545 [s390] 1546 KVM_MP_STATE_SUSPENDED the vcpu is in a suspend state and is waiting 1547 for a wakeup event [arm64] 1548 ========================== =============================================== 1549 1550On x86, this ioctl is only useful after KVM_CREATE_IRQCHIP. Without an 1551in-kernel irqchip, the multiprocessing state must be maintained by userspace on 1552these architectures. 1553 1554For arm64: 1555^^^^^^^^^^ 1556 1557If a vCPU is in the KVM_MP_STATE_SUSPENDED state, KVM will emulate the 1558architectural execution of a WFI instruction. 1559 1560If a wakeup event is recognized, KVM will exit to userspace with a 1561KVM_SYSTEM_EVENT exit, where the event type is KVM_SYSTEM_EVENT_WAKEUP. If 1562userspace wants to honor the wakeup, it must set the vCPU's MP state to 1563KVM_MP_STATE_RUNNABLE. If it does not, KVM will continue to await a wakeup 1564event in subsequent calls to KVM_RUN. 1565 1566.. warning:: 1567 1568 If userspace intends to keep the vCPU in a SUSPENDED state, it is 1569 strongly recommended that userspace take action to suppress the 1570 wakeup event (such as masking an interrupt). Otherwise, subsequent 1571 calls to KVM_RUN will immediately exit with a KVM_SYSTEM_EVENT_WAKEUP 1572 event and inadvertently waste CPU cycles. 1573 1574 Additionally, if userspace takes action to suppress a wakeup event, 1575 it is strongly recommended that it also restores the vCPU to its 1576 original state when the vCPU is made RUNNABLE again. For example, 1577 if userspace masked a pending interrupt to suppress the wakeup, 1578 the interrupt should be unmasked before returning control to the 1579 guest. 1580 1581For riscv: 1582^^^^^^^^^^ 1583 1584The only states that are valid are KVM_MP_STATE_STOPPED and 1585KVM_MP_STATE_RUNNABLE which reflect if the vcpu is paused or not. 1586 1587On LoongArch, only the KVM_MP_STATE_RUNNABLE state is used to reflect 1588whether the vcpu is runnable. 1589 15904.39 KVM_SET_MP_STATE 1591--------------------- 1592 1593:Capability: KVM_CAP_MP_STATE 1594:Architectures: x86, s390, arm64, riscv, loongarch 1595:Type: vcpu ioctl 1596:Parameters: struct kvm_mp_state (in) 1597:Returns: 0 on success; -1 on error 1598 1599Sets the vcpu's current "multiprocessing state"; see KVM_GET_MP_STATE for 1600arguments. 1601 1602On x86, this ioctl is only useful after KVM_CREATE_IRQCHIP. Without an 1603in-kernel irqchip, the multiprocessing state must be maintained by userspace on 1604these architectures. 1605 1606For arm64/riscv: 1607^^^^^^^^^^^^^^^^ 1608 1609The only states that are valid are KVM_MP_STATE_STOPPED and 1610KVM_MP_STATE_RUNNABLE which reflect if the vcpu should be paused or not. 1611 1612On LoongArch, only the KVM_MP_STATE_RUNNABLE state is used to reflect 1613whether the vcpu is runnable. 1614 16154.40 KVM_SET_IDENTITY_MAP_ADDR 1616------------------------------ 1617 1618:Capability: KVM_CAP_SET_IDENTITY_MAP_ADDR 1619:Architectures: x86 1620:Type: vm ioctl 1621:Parameters: unsigned long identity (in) 1622:Returns: 0 on success, -1 on error 1623 1624This ioctl defines the physical address of a one-page region in the guest 1625physical address space. The region must be within the first 4GB of the 1626guest physical address space and must not conflict with any memory slot 1627or any mmio address. The guest may malfunction if it accesses this memory 1628region. 1629 1630Setting the address to 0 will result in resetting the address to its default 1631(0xfffbc000). 1632 1633This ioctl is required on Intel-based hosts. This is needed on Intel hardware 1634because of a quirk in the virtualization implementation (see the internals 1635documentation when it pops into existence). 1636 1637Fails if any VCPU has already been created. 1638 16394.41 KVM_SET_BOOT_CPU_ID 1640------------------------ 1641 1642:Capability: KVM_CAP_SET_BOOT_CPU_ID 1643:Architectures: x86 1644:Type: vm ioctl 1645:Parameters: unsigned long vcpu_id 1646:Returns: 0 on success, -1 on error 1647 1648Define which vcpu is the Bootstrap Processor (BSP). Values are the same 1649as the vcpu id in KVM_CREATE_VCPU. If this ioctl is not called, the default 1650is vcpu 0. This ioctl has to be called before vcpu creation, 1651otherwise it will return EBUSY error. 1652 1653 16544.42 KVM_GET_XSAVE 1655------------------ 1656 1657:Capability: KVM_CAP_XSAVE 1658:Architectures: x86 1659:Type: vcpu ioctl 1660:Parameters: struct kvm_xsave (out) 1661:Returns: 0 on success, -1 on error 1662 1663 1664:: 1665 1666 struct kvm_xsave { 1667 __u32 region[1024]; 1668 __u32 extra[0]; 1669 }; 1670 1671This ioctl would copy current vcpu's xsave struct to the userspace. 1672 1673 16744.43 KVM_SET_XSAVE 1675------------------ 1676 1677:Capability: KVM_CAP_XSAVE and KVM_CAP_XSAVE2 1678:Architectures: x86 1679:Type: vcpu ioctl 1680:Parameters: struct kvm_xsave (in) 1681:Returns: 0 on success, -1 on error 1682 1683:: 1684 1685 1686 struct kvm_xsave { 1687 __u32 region[1024]; 1688 __u32 extra[0]; 1689 }; 1690 1691This ioctl would copy userspace's xsave struct to the kernel. It copies 1692as many bytes as are returned by KVM_CHECK_EXTENSION(KVM_CAP_XSAVE2), 1693when invoked on the vm file descriptor. The size value returned by 1694KVM_CHECK_EXTENSION(KVM_CAP_XSAVE2) will always be at least 4096. 1695Currently, it is only greater than 4096 if a dynamic feature has been 1696enabled with ``arch_prctl()``, but this may change in the future. 1697 1698The offsets of the state save areas in struct kvm_xsave follow the 1699contents of CPUID leaf 0xD on the host. 1700 1701 17024.44 KVM_GET_XCRS 1703----------------- 1704 1705:Capability: KVM_CAP_XCRS 1706:Architectures: x86 1707:Type: vcpu ioctl 1708:Parameters: struct kvm_xcrs (out) 1709:Returns: 0 on success, -1 on error 1710 1711:: 1712 1713 struct kvm_xcr { 1714 __u32 xcr; 1715 __u32 reserved; 1716 __u64 value; 1717 }; 1718 1719 struct kvm_xcrs { 1720 __u32 nr_xcrs; 1721 __u32 flags; 1722 struct kvm_xcr xcrs[KVM_MAX_XCRS]; 1723 __u64 padding[16]; 1724 }; 1725 1726This ioctl would copy current vcpu's xcrs to the userspace. 1727 1728 17294.45 KVM_SET_XCRS 1730----------------- 1731 1732:Capability: KVM_CAP_XCRS 1733:Architectures: x86 1734:Type: vcpu ioctl 1735:Parameters: struct kvm_xcrs (in) 1736:Returns: 0 on success, -1 on error 1737 1738:: 1739 1740 struct kvm_xcr { 1741 __u32 xcr; 1742 __u32 reserved; 1743 __u64 value; 1744 }; 1745 1746 struct kvm_xcrs { 1747 __u32 nr_xcrs; 1748 __u32 flags; 1749 struct kvm_xcr xcrs[KVM_MAX_XCRS]; 1750 __u64 padding[16]; 1751 }; 1752 1753This ioctl would set vcpu's xcr to the value userspace specified. 1754 1755 17564.46 KVM_GET_SUPPORTED_CPUID 1757---------------------------- 1758 1759:Capability: KVM_CAP_EXT_CPUID 1760:Architectures: x86 1761:Type: system ioctl 1762:Parameters: struct kvm_cpuid2 (in/out) 1763:Returns: 0 on success, -1 on error 1764 1765:: 1766 1767 struct kvm_cpuid2 { 1768 __u32 nent; 1769 __u32 padding; 1770 struct kvm_cpuid_entry2 entries[0]; 1771 }; 1772 1773 #define KVM_CPUID_FLAG_SIGNIFCANT_INDEX BIT(0) 1774 #define KVM_CPUID_FLAG_STATEFUL_FUNC BIT(1) /* deprecated */ 1775 #define KVM_CPUID_FLAG_STATE_READ_NEXT BIT(2) /* deprecated */ 1776 1777 struct kvm_cpuid_entry2 { 1778 __u32 function; 1779 __u32 index; 1780 __u32 flags; 1781 __u32 eax; 1782 __u32 ebx; 1783 __u32 ecx; 1784 __u32 edx; 1785 __u32 padding[3]; 1786 }; 1787 1788This ioctl returns x86 cpuid features which are supported by both the 1789hardware and kvm in its default configuration. Userspace can use the 1790information returned by this ioctl to construct cpuid information (for 1791KVM_SET_CPUID2) that is consistent with hardware, kernel, and 1792userspace capabilities, and with user requirements (for example, the 1793user may wish to constrain cpuid to emulate older hardware, or for 1794feature consistency across a cluster). 1795 1796Dynamically-enabled feature bits need to be requested with 1797``arch_prctl()`` before calling this ioctl. Feature bits that have not 1798been requested are excluded from the result. 1799 1800Note that certain capabilities, such as KVM_CAP_X86_DISABLE_EXITS, may 1801expose cpuid features (e.g. MONITOR) which are not supported by kvm in 1802its default configuration. If userspace enables such capabilities, it 1803is responsible for modifying the results of this ioctl appropriately. 1804 1805Userspace invokes KVM_GET_SUPPORTED_CPUID by passing a kvm_cpuid2 structure 1806with the 'nent' field indicating the number of entries in the variable-size 1807array 'entries'. If the number of entries is too low to describe the cpu 1808capabilities, an error (E2BIG) is returned. If the number is too high, 1809the 'nent' field is adjusted and an error (ENOMEM) is returned. If the 1810number is just right, the 'nent' field is adjusted to the number of valid 1811entries in the 'entries' array, which is then filled. 1812 1813The entries returned are the host cpuid as returned by the cpuid instruction, 1814with unknown or unsupported features masked out. Some features (for example, 1815x2apic), may not be present in the host cpu, but are exposed by kvm if it can 1816emulate them efficiently. The fields in each entry are defined as follows: 1817 1818 function: 1819 the eax value used to obtain the entry 1820 1821 index: 1822 the ecx value used to obtain the entry (for entries that are 1823 affected by ecx) 1824 1825 flags: 1826 an OR of zero or more of the following: 1827 1828 KVM_CPUID_FLAG_SIGNIFCANT_INDEX: 1829 if the index field is valid 1830 1831 eax, ebx, ecx, edx: 1832 the values returned by the cpuid instruction for 1833 this function/index combination 1834 1835x2APIC (CPUID leaf 1, ecx[21) and TSC deadline timer (CPUID leaf 1, ecx[24]) 1836may be returned as true, but they depend on KVM_CREATE_IRQCHIP for in-kernel 1837emulation of the local APIC. TSC deadline timer support is also reported via:: 1838 1839 ioctl(KVM_CHECK_EXTENSION, KVM_CAP_TSC_DEADLINE_TIMER) 1840 1841if that returns true and you use KVM_CREATE_IRQCHIP, or if you emulate the 1842feature in userspace, then you can enable the feature for KVM_SET_CPUID2. 1843 1844Enabling x2APIC in KVM_SET_CPUID2 requires KVM_CREATE_IRQCHIP as KVM doesn't 1845support forwarding x2APIC MSR accesses to userspace, i.e. KVM does not support 1846emulating x2APIC in userspace. 1847 18484.47 KVM_PPC_GET_PVINFO 1849----------------------- 1850 1851:Capability: KVM_CAP_PPC_GET_PVINFO 1852:Architectures: ppc 1853:Type: vm ioctl 1854:Parameters: struct kvm_ppc_pvinfo (out) 1855:Returns: 0 on success, !0 on error 1856 1857:: 1858 1859 struct kvm_ppc_pvinfo { 1860 __u32 flags; 1861 __u32 hcall[4]; 1862 __u8 pad[108]; 1863 }; 1864 1865This ioctl fetches PV specific information that need to be passed to the guest 1866using the device tree or other means from vm context. 1867 1868The hcall array defines 4 instructions that make up a hypercall. 1869 1870If any additional field gets added to this structure later on, a bit for that 1871additional piece of information will be set in the flags bitmap. 1872 1873The flags bitmap is defined as:: 1874 1875 /* the host supports the ePAPR idle hcall 1876 #define KVM_PPC_PVINFO_FLAGS_EV_IDLE (1<<0) 1877 18784.52 KVM_SET_GSI_ROUTING 1879------------------------ 1880 1881:Capability: KVM_CAP_IRQ_ROUTING 1882:Architectures: x86 s390 arm64 1883:Type: vm ioctl 1884:Parameters: struct kvm_irq_routing (in) 1885:Returns: 0 on success, -1 on error 1886 1887Sets the GSI routing table entries, overwriting any previously set entries. 1888 1889On arm64, GSI routing has the following limitation: 1890 1891- GSI routing does not apply to KVM_IRQ_LINE but only to KVM_IRQFD. 1892 1893:: 1894 1895 struct kvm_irq_routing { 1896 __u32 nr; 1897 __u32 flags; 1898 struct kvm_irq_routing_entry entries[0]; 1899 }; 1900 1901No flags are specified so far, the corresponding field must be set to zero. 1902 1903:: 1904 1905 struct kvm_irq_routing_entry { 1906 __u32 gsi; 1907 __u32 type; 1908 __u32 flags; 1909 __u32 pad; 1910 union { 1911 struct kvm_irq_routing_irqchip irqchip; 1912 struct kvm_irq_routing_msi msi; 1913 struct kvm_irq_routing_s390_adapter adapter; 1914 struct kvm_irq_routing_hv_sint hv_sint; 1915 struct kvm_irq_routing_xen_evtchn xen_evtchn; 1916 __u32 pad[8]; 1917 } u; 1918 }; 1919 1920 /* gsi routing entry types */ 1921 #define KVM_IRQ_ROUTING_IRQCHIP 1 1922 #define KVM_IRQ_ROUTING_MSI 2 1923 #define KVM_IRQ_ROUTING_S390_ADAPTER 3 1924 #define KVM_IRQ_ROUTING_HV_SINT 4 1925 #define KVM_IRQ_ROUTING_XEN_EVTCHN 5 1926 1927On s390, adding a KVM_IRQ_ROUTING_S390_ADAPTER is rejected on ucontrol VMs with 1928error -EINVAL. 1929 1930flags: 1931 1932- KVM_MSI_VALID_DEVID: used along with KVM_IRQ_ROUTING_MSI routing entry 1933 type, specifies that the devid field contains a valid value. The per-VM 1934 KVM_CAP_MSI_DEVID capability advertises the requirement to provide 1935 the device ID. If this capability is not available, userspace should 1936 never set the KVM_MSI_VALID_DEVID flag as the ioctl might fail. 1937- zero otherwise 1938 1939:: 1940 1941 struct kvm_irq_routing_irqchip { 1942 __u32 irqchip; 1943 __u32 pin; 1944 }; 1945 1946 struct kvm_irq_routing_msi { 1947 __u32 address_lo; 1948 __u32 address_hi; 1949 __u32 data; 1950 union { 1951 __u32 pad; 1952 __u32 devid; 1953 }; 1954 }; 1955 1956If KVM_MSI_VALID_DEVID is set, devid contains a unique device identifier 1957for the device that wrote the MSI message. For PCI, this is usually a 1958BDF identifier in the lower 16 bits. 1959 1960On x86, address_hi is ignored unless the KVM_X2APIC_API_USE_32BIT_IDS 1961feature of KVM_CAP_X2APIC_API capability is enabled. If it is enabled, 1962address_hi bits 31-8 provide bits 31-8 of the destination id. Bits 7-0 of 1963address_hi must be zero. 1964 1965:: 1966 1967 struct kvm_irq_routing_s390_adapter { 1968 __u64 ind_addr; 1969 __u64 summary_addr; 1970 __u64 ind_offset; 1971 __u32 summary_offset; 1972 __u32 adapter_id; 1973 }; 1974 1975 struct kvm_irq_routing_hv_sint { 1976 __u32 vcpu; 1977 __u32 sint; 1978 }; 1979 1980 struct kvm_irq_routing_xen_evtchn { 1981 __u32 port; 1982 __u32 vcpu; 1983 __u32 priority; 1984 }; 1985 1986 1987When KVM_CAP_XEN_HVM includes the KVM_XEN_HVM_CONFIG_EVTCHN_2LEVEL bit 1988in its indication of supported features, routing to Xen event channels 1989is supported. Although the priority field is present, only the value 1990KVM_XEN_HVM_CONFIG_EVTCHN_2LEVEL is supported, which means delivery by 19912 level event channels. FIFO event channel support may be added in 1992the future. 1993 1994 19954.55 KVM_SET_TSC_KHZ 1996-------------------- 1997 1998:Capability: KVM_CAP_TSC_CONTROL / KVM_CAP_VM_TSC_CONTROL 1999:Architectures: x86 2000:Type: vcpu ioctl / vm ioctl 2001:Parameters: virtual tsc_khz 2002:Returns: 0 on success, -1 on error 2003 2004Specifies the tsc frequency for the virtual machine. The unit of the 2005frequency is KHz. 2006 2007If the KVM_CAP_VM_TSC_CONTROL capability is advertised, this can also 2008be used as a vm ioctl to set the initial tsc frequency of subsequently 2009created vCPUs. Note, the vm ioctl is only allowed prior to creating vCPUs. 2010 2011For TSC protected Confidential Computing (CoCo) VMs where TSC frequency 2012is configured once at VM scope and remains unchanged during VM's 2013lifetime, the vm ioctl should be used to configure the TSC frequency 2014and the vcpu ioctl is not supported. 2015 2016Example of such CoCo VMs: TDX guests. 2017 20184.56 KVM_GET_TSC_KHZ 2019-------------------- 2020 2021:Capability: KVM_CAP_GET_TSC_KHZ / KVM_CAP_VM_TSC_CONTROL 2022:Architectures: x86 2023:Type: vcpu ioctl / vm ioctl 2024:Parameters: none 2025:Returns: virtual tsc-khz on success, negative value on error 2026 2027Returns the tsc frequency of the guest. The unit of the return value is 2028KHz. If the host has unstable tsc this ioctl returns -EIO instead as an 2029error. 2030 2031 20324.57 KVM_GET_LAPIC 2033------------------ 2034 2035:Capability: KVM_CAP_IRQCHIP 2036:Architectures: x86 2037:Type: vcpu ioctl 2038:Parameters: struct kvm_lapic_state (out) 2039:Returns: 0 on success, -1 on error 2040 2041:: 2042 2043 #define KVM_APIC_REG_SIZE 0x400 2044 struct kvm_lapic_state { 2045 char regs[KVM_APIC_REG_SIZE]; 2046 }; 2047 2048Reads the Local APIC registers and copies them into the input argument. The 2049data format and layout are the same as documented in the architecture manual. 2050 2051If KVM_X2APIC_API_USE_32BIT_IDS feature of KVM_CAP_X2APIC_API is 2052enabled, then the format of APIC_ID register depends on the APIC mode 2053(reported by MSR_IA32_APICBASE) of its VCPU. x2APIC stores APIC ID in 2054the APIC_ID register (bytes 32-35). xAPIC only allows an 8-bit APIC ID 2055which is stored in bits 31-24 of the APIC register, or equivalently in 2056byte 35 of struct kvm_lapic_state's regs field. KVM_GET_LAPIC must then 2057be called after MSR_IA32_APICBASE has been set with KVM_SET_MSR. 2058 2059If KVM_X2APIC_API_USE_32BIT_IDS feature is disabled, struct kvm_lapic_state 2060always uses xAPIC format. 2061 2062 20634.58 KVM_SET_LAPIC 2064------------------ 2065 2066:Capability: KVM_CAP_IRQCHIP 2067:Architectures: x86 2068:Type: vcpu ioctl 2069:Parameters: struct kvm_lapic_state (in) 2070:Returns: 0 on success, -1 on error 2071 2072:: 2073 2074 #define KVM_APIC_REG_SIZE 0x400 2075 struct kvm_lapic_state { 2076 char regs[KVM_APIC_REG_SIZE]; 2077 }; 2078 2079Copies the input argument into the Local APIC registers. The data format 2080and layout are the same as documented in the architecture manual. 2081 2082The format of the APIC ID register (bytes 32-35 of struct kvm_lapic_state's 2083regs field) depends on the state of the KVM_CAP_X2APIC_API capability. 2084See the note in KVM_GET_LAPIC. 2085 2086 20874.59 KVM_IOEVENTFD 2088------------------ 2089 2090:Capability: KVM_CAP_IOEVENTFD 2091:Architectures: all 2092:Type: vm ioctl 2093:Parameters: struct kvm_ioeventfd (in) 2094:Returns: 0 on success, !0 on error 2095 2096This ioctl attaches or detaches an ioeventfd to a legal pio/mmio address 2097within the guest. A guest write in the registered address will signal the 2098provided event instead of triggering an exit. 2099 2100:: 2101 2102 struct kvm_ioeventfd { 2103 __u64 datamatch; 2104 __u64 addr; /* legal pio/mmio address */ 2105 __u32 len; /* 0, 1, 2, 4, or 8 bytes */ 2106 __s32 fd; 2107 __u32 flags; 2108 __u8 pad[36]; 2109 }; 2110 2111For the special case of virtio-ccw devices on s390, the ioevent is matched 2112to a subchannel/virtqueue tuple instead. 2113 2114The following flags are defined:: 2115 2116 #define KVM_IOEVENTFD_FLAG_DATAMATCH (1 << kvm_ioeventfd_flag_nr_datamatch) 2117 #define KVM_IOEVENTFD_FLAG_PIO (1 << kvm_ioeventfd_flag_nr_pio) 2118 #define KVM_IOEVENTFD_FLAG_DEASSIGN (1 << kvm_ioeventfd_flag_nr_deassign) 2119 #define KVM_IOEVENTFD_FLAG_VIRTIO_CCW_NOTIFY \ 2120 (1 << kvm_ioeventfd_flag_nr_virtio_ccw_notify) 2121 2122If datamatch flag is set, the event will be signaled only if the written value 2123to the registered address is equal to datamatch in struct kvm_ioeventfd. 2124 2125For virtio-ccw devices, addr contains the subchannel id and datamatch the 2126virtqueue index. 2127 2128With KVM_CAP_IOEVENTFD_ANY_LENGTH, a zero length ioeventfd is allowed, and 2129the kernel will ignore the length of guest write and may get a faster vmexit. 2130The speedup may only apply to specific architectures, but the ioeventfd will 2131work anyway. 2132 21334.60 KVM_DIRTY_TLB 2134------------------ 2135 2136:Capability: KVM_CAP_SW_TLB 2137:Architectures: ppc 2138:Type: vcpu ioctl 2139:Parameters: struct kvm_dirty_tlb (in) 2140:Returns: 0 on success, -1 on error 2141 2142:: 2143 2144 struct kvm_dirty_tlb { 2145 __u64 bitmap; 2146 __u32 num_dirty; 2147 }; 2148 2149This must be called whenever userspace has changed an entry in the shared 2150TLB, prior to calling KVM_RUN on the associated vcpu. 2151 2152The "bitmap" field is the userspace address of an array. This array 2153consists of a number of bits, equal to the total number of TLB entries as 2154determined by the last successful call to ``KVM_ENABLE_CAP(KVM_CAP_SW_TLB)``, 2155rounded up to the nearest multiple of 64. 2156 2157Each bit corresponds to one TLB entry, ordered the same as in the shared TLB 2158array. 2159 2160The array is little-endian: the bit 0 is the least significant bit of the 2161first byte, bit 8 is the least significant bit of the second byte, etc. 2162This avoids any complications with differing word sizes. 2163 2164The "num_dirty" field is a performance hint for KVM to determine whether it 2165should skip processing the bitmap and just invalidate everything. It must 2166be set to the number of set bits in the bitmap. 2167 2168 21694.62 KVM_CREATE_SPAPR_TCE 2170------------------------- 2171 2172:Capability: KVM_CAP_SPAPR_TCE 2173:Architectures: powerpc 2174:Type: vm ioctl 2175:Parameters: struct kvm_create_spapr_tce (in) 2176:Returns: file descriptor for manipulating the created TCE table 2177 2178This creates a virtual TCE (translation control entry) table, which 2179is an IOMMU for PAPR-style virtual I/O. It is used to translate 2180logical addresses used in virtual I/O into guest physical addresses, 2181and provides a scatter/gather capability for PAPR virtual I/O. 2182 2183:: 2184 2185 /* for KVM_CAP_SPAPR_TCE */ 2186 struct kvm_create_spapr_tce { 2187 __u64 liobn; 2188 __u32 window_size; 2189 }; 2190 2191The liobn field gives the logical IO bus number for which to create a 2192TCE table. The window_size field specifies the size of the DMA window 2193which this TCE table will translate - the table will contain one 64 2194bit TCE entry for every 4kiB of the DMA window. 2195 2196When the guest issues an H_PUT_TCE hcall on a liobn for which a TCE 2197table has been created using this ioctl(), the kernel will handle it 2198in real mode, updating the TCE table. H_PUT_TCE calls for other 2199liobns will cause a vm exit and must be handled by userspace. 2200 2201The return value is a file descriptor which can be passed to mmap(2) 2202to map the created TCE table into userspace. This lets userspace read 2203the entries written by kernel-handled H_PUT_TCE calls, and also lets 2204userspace update the TCE table directly which is useful in some 2205circumstances. 2206 2207 22084.64 KVM_NMI 2209------------ 2210 2211:Capability: KVM_CAP_USER_NMI 2212:Architectures: x86 2213:Type: vcpu ioctl 2214:Parameters: none 2215:Returns: 0 on success, -1 on error 2216 2217Queues an NMI on the thread's vcpu. Note this is well defined only 2218when KVM_CREATE_IRQCHIP has not been called, since this is an interface 2219between the virtual cpu core and virtual local APIC. After KVM_CREATE_IRQCHIP 2220has been called, this interface is completely emulated within the kernel. 2221 2222To use this to emulate the LINT1 input with KVM_CREATE_IRQCHIP, use the 2223following algorithm: 2224 2225 - pause the vcpu 2226 - read the local APIC's state (KVM_GET_LAPIC) 2227 - check whether changing LINT1 will queue an NMI (see the LVT entry for LINT1) 2228 - if so, issue KVM_NMI 2229 - resume the vcpu 2230 2231Some guests configure the LINT1 NMI input to cause a panic, aiding in 2232debugging. 2233 2234 22354.65 KVM_S390_UCAS_MAP 2236---------------------- 2237 2238:Capability: KVM_CAP_S390_UCONTROL 2239:Architectures: s390 2240:Type: vcpu ioctl 2241:Parameters: struct kvm_s390_ucas_mapping (in) 2242:Returns: 0 in case of success 2243 2244The parameter is defined like this:: 2245 2246 struct kvm_s390_ucas_mapping { 2247 __u64 user_addr; 2248 __u64 vcpu_addr; 2249 __u64 length; 2250 }; 2251 2252This ioctl maps the memory at "user_addr" with the length "length" to 2253the vcpu's address space starting at "vcpu_addr". All parameters need to 2254be aligned by 1 megabyte. 2255 2256 22574.66 KVM_S390_UCAS_UNMAP 2258------------------------ 2259 2260:Capability: KVM_CAP_S390_UCONTROL 2261:Architectures: s390 2262:Type: vcpu ioctl 2263:Parameters: struct kvm_s390_ucas_mapping (in) 2264:Returns: 0 in case of success 2265 2266The parameter is defined like this:: 2267 2268 struct kvm_s390_ucas_mapping { 2269 __u64 user_addr; 2270 __u64 vcpu_addr; 2271 __u64 length; 2272 }; 2273 2274This ioctl unmaps the memory in the vcpu's address space starting at 2275"vcpu_addr" with the length "length". The field "user_addr" is ignored. 2276All parameters need to be aligned by 1 megabyte. 2277 2278 22794.67 KVM_S390_VCPU_FAULT 2280------------------------ 2281 2282:Capability: KVM_CAP_S390_UCONTROL 2283:Architectures: s390 2284:Type: vcpu ioctl 2285:Parameters: vcpu absolute address (in) 2286:Returns: 0 in case of success 2287 2288This call creates a page table entry on the virtual cpu's address space 2289(for user controlled virtual machines) or the virtual machine's address 2290space (for regular virtual machines). This only works for minor faults, 2291thus it's recommended to access subject memory page via the user page 2292table upfront. This is useful to handle validity intercepts for user 2293controlled virtual machines to fault in the virtual cpu's lowcore pages 2294prior to calling the KVM_RUN ioctl. 2295 2296 22974.68 KVM_SET_ONE_REG 2298-------------------- 2299 2300:Capability: KVM_CAP_ONE_REG 2301:Architectures: all 2302:Type: vcpu ioctl 2303:Parameters: struct kvm_one_reg (in) 2304:Returns: 0 on success, negative value on failure 2305 2306Errors: 2307 2308 ====== ============================================================ 2309 ENOENT no such register 2310 EINVAL invalid register ID, or no such register or used with VMs in 2311 protected virtualization mode on s390 2312 EPERM (arm64) register access not allowed before vcpu finalization 2313 EBUSY (riscv) changing register value not allowed after the vcpu 2314 has run at least once 2315 ====== ============================================================ 2316 2317(These error codes are indicative only: do not rely on a specific error 2318code being returned in a specific situation.) 2319 2320:: 2321 2322 struct kvm_one_reg { 2323 __u64 id; 2324 __u64 addr; 2325 }; 2326 2327Using this ioctl, a single vcpu register can be set to a specific value 2328defined by user space with the passed in struct kvm_one_reg, where id 2329refers to the register identifier as described below and addr is a pointer 2330to a variable with the respective size. There can be architecture agnostic 2331and architecture specific registers. Each have their own range of operation 2332and their own constants and width. To keep track of the implemented 2333registers, find a list below: 2334 2335 ======= =============================== ============ 2336 Arch Register Width (bits) 2337 ======= =============================== ============ 2338 PPC KVM_REG_PPC_HIOR 64 2339 PPC KVM_REG_PPC_IAC1 64 2340 PPC KVM_REG_PPC_IAC2 64 2341 PPC KVM_REG_PPC_IAC3 64 2342 PPC KVM_REG_PPC_IAC4 64 2343 PPC KVM_REG_PPC_DAC1 64 2344 PPC KVM_REG_PPC_DAC2 64 2345 PPC KVM_REG_PPC_DABR 64 2346 PPC KVM_REG_PPC_DSCR 64 2347 PPC KVM_REG_PPC_PURR 64 2348 PPC KVM_REG_PPC_SPURR 64 2349 PPC KVM_REG_PPC_DAR 64 2350 PPC KVM_REG_PPC_DSISR 32 2351 PPC KVM_REG_PPC_AMR 64 2352 PPC KVM_REG_PPC_UAMOR 64 2353 PPC KVM_REG_PPC_MMCR0 64 2354 PPC KVM_REG_PPC_MMCR1 64 2355 PPC KVM_REG_PPC_MMCRA 64 2356 PPC KVM_REG_PPC_MMCR2 64 2357 PPC KVM_REG_PPC_MMCRS 64 2358 PPC KVM_REG_PPC_MMCR3 64 2359 PPC KVM_REG_PPC_SIAR 64 2360 PPC KVM_REG_PPC_SDAR 64 2361 PPC KVM_REG_PPC_SIER 64 2362 PPC KVM_REG_PPC_SIER2 64 2363 PPC KVM_REG_PPC_SIER3 64 2364 PPC KVM_REG_PPC_PMC1 32 2365 PPC KVM_REG_PPC_PMC2 32 2366 PPC KVM_REG_PPC_PMC3 32 2367 PPC KVM_REG_PPC_PMC4 32 2368 PPC KVM_REG_PPC_PMC5 32 2369 PPC KVM_REG_PPC_PMC6 32 2370 PPC KVM_REG_PPC_PMC7 32 2371 PPC KVM_REG_PPC_PMC8 32 2372 PPC KVM_REG_PPC_FPR0 64 2373 ... 2374 PPC KVM_REG_PPC_FPR31 64 2375 PPC KVM_REG_PPC_VR0 128 2376 ... 2377 PPC KVM_REG_PPC_VR31 128 2378 PPC KVM_REG_PPC_VSR0 128 2379 ... 2380 PPC KVM_REG_PPC_VSR31 128 2381 PPC KVM_REG_PPC_FPSCR 64 2382 PPC KVM_REG_PPC_VSCR 32 2383 PPC KVM_REG_PPC_VPA_ADDR 64 2384 PPC KVM_REG_PPC_VPA_SLB 128 2385 PPC KVM_REG_PPC_VPA_DTL 128 2386 PPC KVM_REG_PPC_EPCR 32 2387 PPC KVM_REG_PPC_EPR 32 2388 PPC KVM_REG_PPC_TCR 32 2389 PPC KVM_REG_PPC_TSR 32 2390 PPC KVM_REG_PPC_OR_TSR 32 2391 PPC KVM_REG_PPC_CLEAR_TSR 32 2392 PPC KVM_REG_PPC_MAS0 32 2393 PPC KVM_REG_PPC_MAS1 32 2394 PPC KVM_REG_PPC_MAS2 64 2395 PPC KVM_REG_PPC_MAS7_3 64 2396 PPC KVM_REG_PPC_MAS4 32 2397 PPC KVM_REG_PPC_MAS6 32 2398 PPC KVM_REG_PPC_MMUCFG 32 2399 PPC KVM_REG_PPC_TLB0CFG 32 2400 PPC KVM_REG_PPC_TLB1CFG 32 2401 PPC KVM_REG_PPC_TLB2CFG 32 2402 PPC KVM_REG_PPC_TLB3CFG 32 2403 PPC KVM_REG_PPC_TLB0PS 32 2404 PPC KVM_REG_PPC_TLB1PS 32 2405 PPC KVM_REG_PPC_TLB2PS 32 2406 PPC KVM_REG_PPC_TLB3PS 32 2407 PPC KVM_REG_PPC_EPTCFG 32 2408 PPC KVM_REG_PPC_ICP_STATE 64 2409 PPC KVM_REG_PPC_VP_STATE 128 2410 PPC KVM_REG_PPC_TB_OFFSET 64 2411 PPC KVM_REG_PPC_SPMC1 32 2412 PPC KVM_REG_PPC_SPMC2 32 2413 PPC KVM_REG_PPC_IAMR 64 2414 PPC KVM_REG_PPC_TFHAR 64 2415 PPC KVM_REG_PPC_TFIAR 64 2416 PPC KVM_REG_PPC_TEXASR 64 2417 PPC KVM_REG_PPC_FSCR 64 2418 PPC KVM_REG_PPC_PSPB 32 2419 PPC KVM_REG_PPC_EBBHR 64 2420 PPC KVM_REG_PPC_EBBRR 64 2421 PPC KVM_REG_PPC_BESCR 64 2422 PPC KVM_REG_PPC_TAR 64 2423 PPC KVM_REG_PPC_DPDES 64 2424 PPC KVM_REG_PPC_DAWR 64 2425 PPC KVM_REG_PPC_DAWRX 64 2426 PPC KVM_REG_PPC_CIABR 64 2427 PPC KVM_REG_PPC_IC 64 2428 PPC KVM_REG_PPC_VTB 64 2429 PPC KVM_REG_PPC_CSIGR 64 2430 PPC KVM_REG_PPC_TACR 64 2431 PPC KVM_REG_PPC_TCSCR 64 2432 PPC KVM_REG_PPC_PID 64 2433 PPC KVM_REG_PPC_ACOP 64 2434 PPC KVM_REG_PPC_VRSAVE 32 2435 PPC KVM_REG_PPC_LPCR 32 2436 PPC KVM_REG_PPC_LPCR_64 64 2437 PPC KVM_REG_PPC_PPR 64 2438 PPC KVM_REG_PPC_ARCH_COMPAT 32 2439 PPC KVM_REG_PPC_DABRX 32 2440 PPC KVM_REG_PPC_WORT 64 2441 PPC KVM_REG_PPC_SPRG9 64 2442 PPC KVM_REG_PPC_DBSR 32 2443 PPC KVM_REG_PPC_TIDR 64 2444 PPC KVM_REG_PPC_PSSCR 64 2445 PPC KVM_REG_PPC_DEC_EXPIRY 64 2446 PPC KVM_REG_PPC_PTCR 64 2447 PPC KVM_REG_PPC_HASHKEYR 64 2448 PPC KVM_REG_PPC_HASHPKEYR 64 2449 PPC KVM_REG_PPC_DAWR1 64 2450 PPC KVM_REG_PPC_DAWRX1 64 2451 PPC KVM_REG_PPC_DEXCR 64 2452 PPC KVM_REG_PPC_TM_GPR0 64 2453 ... 2454 PPC KVM_REG_PPC_TM_GPR31 64 2455 PPC KVM_REG_PPC_TM_VSR0 128 2456 ... 2457 PPC KVM_REG_PPC_TM_VSR63 128 2458 PPC KVM_REG_PPC_TM_CR 64 2459 PPC KVM_REG_PPC_TM_LR 64 2460 PPC KVM_REG_PPC_TM_CTR 64 2461 PPC KVM_REG_PPC_TM_FPSCR 64 2462 PPC KVM_REG_PPC_TM_AMR 64 2463 PPC KVM_REG_PPC_TM_PPR 64 2464 PPC KVM_REG_PPC_TM_VRSAVE 64 2465 PPC KVM_REG_PPC_TM_VSCR 32 2466 PPC KVM_REG_PPC_TM_DSCR 64 2467 PPC KVM_REG_PPC_TM_TAR 64 2468 PPC KVM_REG_PPC_TM_XER 64 2469 2470 MIPS KVM_REG_MIPS_R0 64 2471 ... 2472 MIPS KVM_REG_MIPS_R31 64 2473 MIPS KVM_REG_MIPS_HI 64 2474 MIPS KVM_REG_MIPS_LO 64 2475 MIPS KVM_REG_MIPS_PC 64 2476 MIPS KVM_REG_MIPS_CP0_INDEX 32 2477 MIPS KVM_REG_MIPS_CP0_ENTRYLO0 64 2478 MIPS KVM_REG_MIPS_CP0_ENTRYLO1 64 2479 MIPS KVM_REG_MIPS_CP0_CONTEXT 64 2480 MIPS KVM_REG_MIPS_CP0_CONTEXTCONFIG 32 2481 MIPS KVM_REG_MIPS_CP0_USERLOCAL 64 2482 MIPS KVM_REG_MIPS_CP0_XCONTEXTCONFIG 64 2483 MIPS KVM_REG_MIPS_CP0_PAGEMASK 32 2484 MIPS KVM_REG_MIPS_CP0_PAGEGRAIN 32 2485 MIPS KVM_REG_MIPS_CP0_SEGCTL0 64 2486 MIPS KVM_REG_MIPS_CP0_SEGCTL1 64 2487 MIPS KVM_REG_MIPS_CP0_SEGCTL2 64 2488 MIPS KVM_REG_MIPS_CP0_PWBASE 64 2489 MIPS KVM_REG_MIPS_CP0_PWFIELD 64 2490 MIPS KVM_REG_MIPS_CP0_PWSIZE 64 2491 MIPS KVM_REG_MIPS_CP0_WIRED 32 2492 MIPS KVM_REG_MIPS_CP0_PWCTL 32 2493 MIPS KVM_REG_MIPS_CP0_HWRENA 32 2494 MIPS KVM_REG_MIPS_CP0_BADVADDR 64 2495 MIPS KVM_REG_MIPS_CP0_BADINSTR 32 2496 MIPS KVM_REG_MIPS_CP0_BADINSTRP 32 2497 MIPS KVM_REG_MIPS_CP0_COUNT 32 2498 MIPS KVM_REG_MIPS_CP0_ENTRYHI 64 2499 MIPS KVM_REG_MIPS_CP0_COMPARE 32 2500 MIPS KVM_REG_MIPS_CP0_STATUS 32 2501 MIPS KVM_REG_MIPS_CP0_INTCTL 32 2502 MIPS KVM_REG_MIPS_CP0_CAUSE 32 2503 MIPS KVM_REG_MIPS_CP0_EPC 64 2504 MIPS KVM_REG_MIPS_CP0_PRID 32 2505 MIPS KVM_REG_MIPS_CP0_EBASE 64 2506 MIPS KVM_REG_MIPS_CP0_CONFIG 32 2507 MIPS KVM_REG_MIPS_CP0_CONFIG1 32 2508 MIPS KVM_REG_MIPS_CP0_CONFIG2 32 2509 MIPS KVM_REG_MIPS_CP0_CONFIG3 32 2510 MIPS KVM_REG_MIPS_CP0_CONFIG4 32 2511 MIPS KVM_REG_MIPS_CP0_CONFIG5 32 2512 MIPS KVM_REG_MIPS_CP0_CONFIG7 32 2513 MIPS KVM_REG_MIPS_CP0_XCONTEXT 64 2514 MIPS KVM_REG_MIPS_CP0_ERROREPC 64 2515 MIPS KVM_REG_MIPS_CP0_KSCRATCH1 64 2516 MIPS KVM_REG_MIPS_CP0_KSCRATCH2 64 2517 MIPS KVM_REG_MIPS_CP0_KSCRATCH3 64 2518 MIPS KVM_REG_MIPS_CP0_KSCRATCH4 64 2519 MIPS KVM_REG_MIPS_CP0_KSCRATCH5 64 2520 MIPS KVM_REG_MIPS_CP0_KSCRATCH6 64 2521 MIPS KVM_REG_MIPS_CP0_MAAR(0..63) 64 2522 MIPS KVM_REG_MIPS_COUNT_CTL 64 2523 MIPS KVM_REG_MIPS_COUNT_RESUME 64 2524 MIPS KVM_REG_MIPS_COUNT_HZ 64 2525 MIPS KVM_REG_MIPS_FPR_32(0..31) 32 2526 MIPS KVM_REG_MIPS_FPR_64(0..31) 64 2527 MIPS KVM_REG_MIPS_VEC_128(0..31) 128 2528 MIPS KVM_REG_MIPS_FCR_IR 32 2529 MIPS KVM_REG_MIPS_FCR_CSR 32 2530 MIPS KVM_REG_MIPS_MSA_IR 32 2531 MIPS KVM_REG_MIPS_MSA_CSR 32 2532 ======= =============================== ============ 2533 2534ARM registers are mapped using the lower 32 bits. The upper 16 of that 2535is the register group type, or coprocessor number: 2536 2537ARM core registers have the following id bit patterns:: 2538 2539 0x4020 0000 0010 <index into the kvm_regs struct:16> 2540 2541ARM 32-bit CP15 registers have the following id bit patterns:: 2542 2543 0x4020 0000 000F <zero:1> <crn:4> <crm:4> <opc1:4> <opc2:3> 2544 2545ARM 64-bit CP15 registers have the following id bit patterns:: 2546 2547 0x4030 0000 000F <zero:1> <zero:4> <crm:4> <opc1:4> <zero:3> 2548 2549ARM CCSIDR registers are demultiplexed by CSSELR value:: 2550 2551 0x4020 0000 0011 00 <csselr:8> 2552 2553ARM 32-bit VFP control registers have the following id bit patterns:: 2554 2555 0x4020 0000 0012 1 <regno:12> 2556 2557ARM 64-bit FP registers have the following id bit patterns:: 2558 2559 0x4030 0000 0012 0 <regno:12> 2560 2561ARM firmware pseudo-registers have the following bit pattern:: 2562 2563 0x4030 0000 0014 <regno:16> 2564 2565 2566arm64 registers are mapped using the lower 32 bits. The upper 16 of 2567that is the register group type, or coprocessor number: 2568 2569arm64 core/FP-SIMD registers have the following id bit patterns. Note 2570that the size of the access is variable, as the kvm_regs structure 2571contains elements ranging from 32 to 128 bits. The index is a 32bit 2572value in the kvm_regs structure seen as a 32bit array:: 2573 2574 0x60x0 0000 0010 <index into the kvm_regs struct:16> 2575 2576Specifically: 2577 2578======================= ========= ===== ======================================= 2579 Encoding Register Bits kvm_regs member 2580======================= ========= ===== ======================================= 2581 0x6030 0000 0010 0000 X0 64 regs.regs[0] 2582 0x6030 0000 0010 0002 X1 64 regs.regs[1] 2583 ... 2584 0x6030 0000 0010 003c X30 64 regs.regs[30] 2585 0x6030 0000 0010 003e SP 64 regs.sp 2586 0x6030 0000 0010 0040 PC 64 regs.pc 2587 0x6030 0000 0010 0042 PSTATE 64 regs.pstate 2588 0x6030 0000 0010 0044 SP_EL1 64 sp_el1 2589 0x6030 0000 0010 0046 ELR_EL1 64 elr_el1 2590 0x6030 0000 0010 0048 SPSR_EL1 64 spsr[KVM_SPSR_EL1] (alias SPSR_SVC) 2591 0x6030 0000 0010 004a SPSR_ABT 64 spsr[KVM_SPSR_ABT] 2592 0x6030 0000 0010 004c SPSR_UND 64 spsr[KVM_SPSR_UND] 2593 0x6030 0000 0010 004e SPSR_IRQ 64 spsr[KVM_SPSR_IRQ] 2594 0x6030 0000 0010 0050 SPSR_FIQ 64 spsr[KVM_SPSR_FIQ] 2595 0x6040 0000 0010 0054 V0 128 fp_regs.vregs[0] [1]_ 2596 0x6040 0000 0010 0058 V1 128 fp_regs.vregs[1] [1]_ 2597 ... 2598 0x6040 0000 0010 00d0 V31 128 fp_regs.vregs[31] [1]_ 2599 0x6020 0000 0010 00d4 FPSR 32 fp_regs.fpsr 2600 0x6020 0000 0010 00d5 FPCR 32 fp_regs.fpcr 2601======================= ========= ===== ======================================= 2602 2603.. [1] These encodings are not accepted for SVE-enabled vcpus. See 2604 :ref:`KVM_ARM_VCPU_INIT`. 2605 2606 The equivalent register content can be accessed via bits [127:0] of 2607 the corresponding SVE Zn registers instead for vcpus that have SVE 2608 enabled (see below). 2609 2610arm64 CCSIDR registers are demultiplexed by CSSELR value:: 2611 2612 0x6020 0000 0011 00 <csselr:8> 2613 2614arm64 system registers have the following id bit patterns:: 2615 2616 0x6030 0000 0013 <op0:2> <op1:3> <crn:4> <crm:4> <op2:3> 2617 2618.. warning:: 2619 2620 Two system register IDs do not follow the specified pattern. These 2621 are KVM_REG_ARM_TIMER_CVAL and KVM_REG_ARM_TIMER_CNT, which map to 2622 system registers CNTV_CVAL_EL0 and CNTVCT_EL0 respectively. These 2623 two had their values accidentally swapped, which means TIMER_CVAL is 2624 derived from the register encoding for CNTVCT_EL0 and TIMER_CNT is 2625 derived from the register encoding for CNTV_CVAL_EL0. As this is 2626 API, it must remain this way. 2627 2628arm64 firmware pseudo-registers have the following bit pattern:: 2629 2630 0x6030 0000 0014 <regno:16> 2631 2632arm64 SVE registers have the following bit patterns:: 2633 2634 0x6080 0000 0015 00 <n:5> <slice:5> Zn bits[2048*slice + 2047 : 2048*slice] 2635 0x6050 0000 0015 04 <n:4> <slice:5> Pn bits[256*slice + 255 : 256*slice] 2636 0x6050 0000 0015 060 <slice:5> FFR bits[256*slice + 255 : 256*slice] 2637 0x6060 0000 0015 ffff KVM_REG_ARM64_SVE_VLS pseudo-register 2638 2639Access to register IDs where 2048 * slice >= 128 * max_vq will fail with 2640ENOENT. max_vq is the vcpu's maximum supported vector length in 128-bit 2641quadwords: see [2]_ below. 2642 2643These registers are only accessible on vcpus for which SVE is enabled. 2644See KVM_ARM_VCPU_INIT for details. 2645 2646In addition, except for KVM_REG_ARM64_SVE_VLS, these registers are not 2647accessible until the vcpu's SVE configuration has been finalized 2648using KVM_ARM_VCPU_FINALIZE(KVM_ARM_VCPU_SVE). See KVM_ARM_VCPU_INIT 2649and KVM_ARM_VCPU_FINALIZE for more information about this procedure. 2650 2651KVM_REG_ARM64_SVE_VLS is a pseudo-register that allows the set of vector 2652lengths supported by the vcpu to be discovered and configured by 2653userspace. When transferred to or from user memory via KVM_GET_ONE_REG 2654or KVM_SET_ONE_REG, the value of this register is of type 2655__u64[KVM_ARM64_SVE_VLS_WORDS], and encodes the set of vector lengths as 2656follows:: 2657 2658 __u64 vector_lengths[KVM_ARM64_SVE_VLS_WORDS]; 2659 2660 if (vq >= SVE_VQ_MIN && vq <= SVE_VQ_MAX && 2661 ((vector_lengths[(vq - KVM_ARM64_SVE_VQ_MIN) / 64] >> 2662 ((vq - KVM_ARM64_SVE_VQ_MIN) % 64)) & 1)) 2663 /* Vector length vq * 16 bytes supported */ 2664 else 2665 /* Vector length vq * 16 bytes not supported */ 2666 2667.. [2] The maximum value vq for which the above condition is true is 2668 max_vq. This is the maximum vector length available to the guest on 2669 this vcpu, and determines which register slices are visible through 2670 this ioctl interface. 2671 2672(See Documentation/arch/arm64/sve.rst for an explanation of the "vq" 2673nomenclature.) 2674 2675KVM_REG_ARM64_SVE_VLS is only accessible after KVM_ARM_VCPU_INIT. 2676KVM_ARM_VCPU_INIT initialises it to the best set of vector lengths that 2677the host supports. 2678 2679Userspace may subsequently modify it if desired until the vcpu's SVE 2680configuration is finalized using KVM_ARM_VCPU_FINALIZE(KVM_ARM_VCPU_SVE). 2681 2682Apart from simply removing all vector lengths from the host set that 2683exceed some value, support for arbitrarily chosen sets of vector lengths 2684is hardware-dependent and may not be available. Attempting to configure 2685an invalid set of vector lengths via KVM_SET_ONE_REG will fail with 2686EINVAL. 2687 2688After the vcpu's SVE configuration is finalized, further attempts to 2689write this register will fail with EPERM. 2690 2691arm64 bitmap feature firmware pseudo-registers have the following bit pattern:: 2692 2693 0x6030 0000 0016 <regno:16> 2694 2695The bitmap feature firmware registers exposes the hypercall services that 2696are available for userspace to configure. The set bits corresponds to the 2697services that are available for the guests to access. By default, KVM 2698sets all the supported bits during VM initialization. The userspace can 2699discover the available services via KVM_GET_ONE_REG, and write back the 2700bitmap corresponding to the features that it wishes guests to see via 2701KVM_SET_ONE_REG. 2702 2703Note: These registers are immutable once any of the vCPUs of the VM has 2704run at least once. A KVM_SET_ONE_REG in such a scenario will return 2705a -EBUSY to userspace. 2706 2707(See Documentation/virt/kvm/arm/hypercalls.rst for more details.) 2708 2709 2710MIPS registers are mapped using the lower 32 bits. The upper 16 of that is 2711the register group type: 2712 2713MIPS core registers (see above) have the following id bit patterns:: 2714 2715 0x7030 0000 0000 <reg:16> 2716 2717MIPS CP0 registers (see KVM_REG_MIPS_CP0_* above) have the following id bit 2718patterns depending on whether they're 32-bit or 64-bit registers:: 2719 2720 0x7020 0000 0001 00 <reg:5> <sel:3> (32-bit) 2721 0x7030 0000 0001 00 <reg:5> <sel:3> (64-bit) 2722 2723Note: KVM_REG_MIPS_CP0_ENTRYLO0 and KVM_REG_MIPS_CP0_ENTRYLO1 are the MIPS64 2724versions of the EntryLo registers regardless of the word size of the host 2725hardware, host kernel, guest, and whether XPA is present in the guest, i.e. 2726with the RI and XI bits (if they exist) in bits 63 and 62 respectively, and 2727the PFNX field starting at bit 30. 2728 2729MIPS MAARs (see KVM_REG_MIPS_CP0_MAAR(*) above) have the following id bit 2730patterns:: 2731 2732 0x7030 0000 0001 01 <reg:8> 2733 2734MIPS KVM control registers (see above) have the following id bit patterns:: 2735 2736 0x7030 0000 0002 <reg:16> 2737 2738MIPS FPU registers (see KVM_REG_MIPS_FPR_{32,64}() above) have the following 2739id bit patterns depending on the size of the register being accessed. They are 2740always accessed according to the current guest FPU mode (Status.FR and 2741Config5.FRE), i.e. as the guest would see them, and they become unpredictable 2742if the guest FPU mode is changed. MIPS SIMD Architecture (MSA) vector 2743registers (see KVM_REG_MIPS_VEC_128() above) have similar patterns as they 2744overlap the FPU registers:: 2745 2746 0x7020 0000 0003 00 <0:3> <reg:5> (32-bit FPU registers) 2747 0x7030 0000 0003 00 <0:3> <reg:5> (64-bit FPU registers) 2748 0x7040 0000 0003 00 <0:3> <reg:5> (128-bit MSA vector registers) 2749 2750MIPS FPU control registers (see KVM_REG_MIPS_FCR_{IR,CSR} above) have the 2751following id bit patterns:: 2752 2753 0x7020 0000 0003 01 <0:3> <reg:5> 2754 2755MIPS MSA control registers (see KVM_REG_MIPS_MSA_{IR,CSR} above) have the 2756following id bit patterns:: 2757 2758 0x7020 0000 0003 02 <0:3> <reg:5> 2759 2760RISC-V registers are mapped using the lower 32 bits. The upper 8 bits of 2761that is the register group type. 2762 2763RISC-V config registers are meant for configuring a Guest VCPU and it has 2764the following id bit patterns:: 2765 2766 0x8020 0000 01 <index into the kvm_riscv_config struct:24> (32bit Host) 2767 0x8030 0000 01 <index into the kvm_riscv_config struct:24> (64bit Host) 2768 2769Following are the RISC-V config registers: 2770 2771======================= ========= ============================================= 2772 Encoding Register Description 2773======================= ========= ============================================= 2774 0x80x0 0000 0100 0000 isa ISA feature bitmap of Guest VCPU 2775======================= ========= ============================================= 2776 2777The isa config register can be read anytime but can only be written before 2778a Guest VCPU runs. It will have ISA feature bits matching underlying host 2779set by default. 2780 2781RISC-V core registers represent the general execution state of a Guest VCPU 2782and it has the following id bit patterns:: 2783 2784 0x8020 0000 02 <index into the kvm_riscv_core struct:24> (32bit Host) 2785 0x8030 0000 02 <index into the kvm_riscv_core struct:24> (64bit Host) 2786 2787Following are the RISC-V core registers: 2788 2789======================= ========= ============================================= 2790 Encoding Register Description 2791======================= ========= ============================================= 2792 0x80x0 0000 0200 0000 regs.pc Program counter 2793 0x80x0 0000 0200 0001 regs.ra Return address 2794 0x80x0 0000 0200 0002 regs.sp Stack pointer 2795 0x80x0 0000 0200 0003 regs.gp Global pointer 2796 0x80x0 0000 0200 0004 regs.tp Task pointer 2797 0x80x0 0000 0200 0005 regs.t0 Caller saved register 0 2798 0x80x0 0000 0200 0006 regs.t1 Caller saved register 1 2799 0x80x0 0000 0200 0007 regs.t2 Caller saved register 2 2800 0x80x0 0000 0200 0008 regs.s0 Callee saved register 0 2801 0x80x0 0000 0200 0009 regs.s1 Callee saved register 1 2802 0x80x0 0000 0200 000a regs.a0 Function argument (or return value) 0 2803 0x80x0 0000 0200 000b regs.a1 Function argument (or return value) 1 2804 0x80x0 0000 0200 000c regs.a2 Function argument 2 2805 0x80x0 0000 0200 000d regs.a3 Function argument 3 2806 0x80x0 0000 0200 000e regs.a4 Function argument 4 2807 0x80x0 0000 0200 000f regs.a5 Function argument 5 2808 0x80x0 0000 0200 0010 regs.a6 Function argument 6 2809 0x80x0 0000 0200 0011 regs.a7 Function argument 7 2810 0x80x0 0000 0200 0012 regs.s2 Callee saved register 2 2811 0x80x0 0000 0200 0013 regs.s3 Callee saved register 3 2812 0x80x0 0000 0200 0014 regs.s4 Callee saved register 4 2813 0x80x0 0000 0200 0015 regs.s5 Callee saved register 5 2814 0x80x0 0000 0200 0016 regs.s6 Callee saved register 6 2815 0x80x0 0000 0200 0017 regs.s7 Callee saved register 7 2816 0x80x0 0000 0200 0018 regs.s8 Callee saved register 8 2817 0x80x0 0000 0200 0019 regs.s9 Callee saved register 9 2818 0x80x0 0000 0200 001a regs.s10 Callee saved register 10 2819 0x80x0 0000 0200 001b regs.s11 Callee saved register 11 2820 0x80x0 0000 0200 001c regs.t3 Caller saved register 3 2821 0x80x0 0000 0200 001d regs.t4 Caller saved register 4 2822 0x80x0 0000 0200 001e regs.t5 Caller saved register 5 2823 0x80x0 0000 0200 001f regs.t6 Caller saved register 6 2824 0x80x0 0000 0200 0020 mode Privilege mode (1 = S-mode or 0 = U-mode) 2825======================= ========= ============================================= 2826 2827RISC-V csr registers represent the supervisor mode control/status registers 2828of a Guest VCPU and it has the following id bit patterns:: 2829 2830 0x8020 0000 03 <index into the kvm_riscv_csr struct:24> (32bit Host) 2831 0x8030 0000 03 <index into the kvm_riscv_csr struct:24> (64bit Host) 2832 2833Following are the RISC-V csr registers: 2834 2835======================= ========= ============================================= 2836 Encoding Register Description 2837======================= ========= ============================================= 2838 0x80x0 0000 0300 0000 sstatus Supervisor status 2839 0x80x0 0000 0300 0001 sie Supervisor interrupt enable 2840 0x80x0 0000 0300 0002 stvec Supervisor trap vector base 2841 0x80x0 0000 0300 0003 sscratch Supervisor scratch register 2842 0x80x0 0000 0300 0004 sepc Supervisor exception program counter 2843 0x80x0 0000 0300 0005 scause Supervisor trap cause 2844 0x80x0 0000 0300 0006 stval Supervisor bad address or instruction 2845 0x80x0 0000 0300 0007 sip Supervisor interrupt pending 2846 0x80x0 0000 0300 0008 satp Supervisor address translation and protection 2847======================= ========= ============================================= 2848 2849RISC-V timer registers represent the timer state of a Guest VCPU and it has 2850the following id bit patterns:: 2851 2852 0x8030 0000 04 <index into the kvm_riscv_timer struct:24> 2853 2854Following are the RISC-V timer registers: 2855 2856======================= ========= ============================================= 2857 Encoding Register Description 2858======================= ========= ============================================= 2859 0x8030 0000 0400 0000 frequency Time base frequency (read-only) 2860 0x8030 0000 0400 0001 time Time value visible to Guest 2861 0x8030 0000 0400 0002 compare Time compare programmed by Guest 2862 0x8030 0000 0400 0003 state Time compare state (1 = ON or 0 = OFF) 2863======================= ========= ============================================= 2864 2865RISC-V F-extension registers represent the single precision floating point 2866state of a Guest VCPU and it has the following id bit patterns:: 2867 2868 0x8020 0000 05 <index into the __riscv_f_ext_state struct:24> 2869 2870Following are the RISC-V F-extension registers: 2871 2872======================= ========= ============================================= 2873 Encoding Register Description 2874======================= ========= ============================================= 2875 0x8020 0000 0500 0000 f[0] Floating point register 0 2876 ... 2877 0x8020 0000 0500 001f f[31] Floating point register 31 2878 0x8020 0000 0500 0020 fcsr Floating point control and status register 2879======================= ========= ============================================= 2880 2881RISC-V D-extension registers represent the double precision floating point 2882state of a Guest VCPU and it has the following id bit patterns:: 2883 2884 0x8020 0000 06 <index into the __riscv_d_ext_state struct:24> (fcsr) 2885 0x8030 0000 06 <index into the __riscv_d_ext_state struct:24> (non-fcsr) 2886 2887Following are the RISC-V D-extension registers: 2888 2889======================= ========= ============================================= 2890 Encoding Register Description 2891======================= ========= ============================================= 2892 0x8030 0000 0600 0000 f[0] Floating point register 0 2893 ... 2894 0x8030 0000 0600 001f f[31] Floating point register 31 2895 0x8020 0000 0600 0020 fcsr Floating point control and status register 2896======================= ========= ============================================= 2897 2898LoongArch registers are mapped using the lower 32 bits. The upper 16 bits of 2899that is the register group type. 2900 2901LoongArch csr registers are used to control guest cpu or get status of guest 2902cpu, and they have the following id bit patterns:: 2903 2904 0x9030 0000 0001 00 <reg:5> <sel:3> (64-bit) 2905 2906LoongArch KVM control registers are used to implement some new defined functions 2907such as set vcpu counter or reset vcpu, and they have the following id bit patterns:: 2908 2909 0x9030 0000 0002 <reg:16> 2910 2911 29124.69 KVM_GET_ONE_REG 2913-------------------- 2914 2915:Capability: KVM_CAP_ONE_REG 2916:Architectures: all 2917:Type: vcpu ioctl 2918:Parameters: struct kvm_one_reg (in and out) 2919:Returns: 0 on success, negative value on failure 2920 2921Errors include: 2922 2923 ======== ============================================================ 2924 ENOENT no such register 2925 EINVAL invalid register ID, or no such register or used with VMs in 2926 protected virtualization mode on s390 2927 EPERM (arm64) register access not allowed before vcpu finalization 2928 ======== ============================================================ 2929 2930(These error codes are indicative only: do not rely on a specific error 2931code being returned in a specific situation.) 2932 2933This ioctl allows to receive the value of a single register implemented 2934in a vcpu. The register to read is indicated by the "id" field of the 2935kvm_one_reg struct passed in. On success, the register value can be found 2936at the memory location pointed to by "addr". 2937 2938The list of registers accessible using this interface is identical to the 2939list in 4.68. 2940 2941 29424.70 KVM_KVMCLOCK_CTRL 2943---------------------- 2944 2945:Capability: KVM_CAP_KVMCLOCK_CTRL 2946:Architectures: Any that implement pvclocks (currently x86 only) 2947:Type: vcpu ioctl 2948:Parameters: None 2949:Returns: 0 on success, -1 on error 2950 2951This ioctl sets a flag accessible to the guest indicating that the specified 2952vCPU has been paused by the host userspace. 2953 2954The host will set a flag in the pvclock structure that is checked from the 2955soft lockup watchdog. The flag is part of the pvclock structure that is 2956shared between guest and host, specifically the second bit of the flags 2957field of the pvclock_vcpu_time_info structure. It will be set exclusively by 2958the host and read/cleared exclusively by the guest. The guest operation of 2959checking and clearing the flag must be an atomic operation so 2960load-link/store-conditional, or equivalent must be used. There are two cases 2961where the guest will clear the flag: when the soft lockup watchdog timer resets 2962itself or when a soft lockup is detected. This ioctl can be called any time 2963after pausing the vcpu, but before it is resumed. 2964 2965 29664.71 KVM_SIGNAL_MSI 2967------------------- 2968 2969:Capability: KVM_CAP_SIGNAL_MSI 2970:Architectures: x86 arm64 2971:Type: vm ioctl 2972:Parameters: struct kvm_msi (in) 2973:Returns: >0 on delivery, 0 if guest blocked the MSI, and -1 on error 2974 2975Directly inject a MSI message. Only valid with in-kernel irqchip that handles 2976MSI messages. 2977 2978:: 2979 2980 struct kvm_msi { 2981 __u32 address_lo; 2982 __u32 address_hi; 2983 __u32 data; 2984 __u32 flags; 2985 __u32 devid; 2986 __u8 pad[12]; 2987 }; 2988 2989flags: 2990 KVM_MSI_VALID_DEVID: devid contains a valid value. The per-VM 2991 KVM_CAP_MSI_DEVID capability advertises the requirement to provide 2992 the device ID. If this capability is not available, userspace 2993 should never set the KVM_MSI_VALID_DEVID flag as the ioctl might fail. 2994 2995If KVM_MSI_VALID_DEVID is set, devid contains a unique device identifier 2996for the device that wrote the MSI message. For PCI, this is usually a 2997BDF identifier in the lower 16 bits. 2998 2999On x86, address_hi is ignored unless the KVM_X2APIC_API_USE_32BIT_IDS 3000feature of KVM_CAP_X2APIC_API capability is enabled. If it is enabled, 3001address_hi bits 31-8 provide bits 31-8 of the destination id. Bits 7-0 of 3002address_hi must be zero. 3003 3004 30054.71 KVM_CREATE_PIT2 3006-------------------- 3007 3008:Capability: KVM_CAP_PIT2 3009:Architectures: x86 3010:Type: vm ioctl 3011:Parameters: struct kvm_pit_config (in) 3012:Returns: 0 on success, -1 on error 3013 3014Creates an in-kernel device model for the i8254 PIT. This call is only valid 3015after enabling in-kernel irqchip support via KVM_CREATE_IRQCHIP. The following 3016parameters have to be passed:: 3017 3018 struct kvm_pit_config { 3019 __u32 flags; 3020 __u32 pad[15]; 3021 }; 3022 3023Valid flags are:: 3024 3025 #define KVM_PIT_SPEAKER_DUMMY 1 /* emulate speaker port stub */ 3026 3027PIT timer interrupts may use a per-VM kernel thread for injection. If it 3028exists, this thread will have a name of the following pattern:: 3029 3030 kvm-pit/<owner-process-pid> 3031 3032When running a guest with elevated priorities, the scheduling parameters of 3033this thread may have to be adjusted accordingly. 3034 3035This IOCTL replaces the obsolete KVM_CREATE_PIT. 3036 3037 30384.72 KVM_GET_PIT2 3039----------------- 3040 3041:Capability: KVM_CAP_PIT_STATE2 3042:Architectures: x86 3043:Type: vm ioctl 3044:Parameters: struct kvm_pit_state2 (out) 3045:Returns: 0 on success, -1 on error 3046 3047Retrieves the state of the in-kernel PIT model. Only valid after 3048KVM_CREATE_PIT2. The state is returned in the following structure:: 3049 3050 struct kvm_pit_state2 { 3051 struct kvm_pit_channel_state channels[3]; 3052 __u32 flags; 3053 __u32 reserved[9]; 3054 }; 3055 3056Valid flags are:: 3057 3058 /* disable PIT in HPET legacy mode */ 3059 #define KVM_PIT_FLAGS_HPET_LEGACY 0x00000001 3060 /* speaker port data bit enabled */ 3061 #define KVM_PIT_FLAGS_SPEAKER_DATA_ON 0x00000002 3062 3063This IOCTL replaces the obsolete KVM_GET_PIT. 3064 3065 30664.73 KVM_SET_PIT2 3067----------------- 3068 3069:Capability: KVM_CAP_PIT_STATE2 3070:Architectures: x86 3071:Type: vm ioctl 3072:Parameters: struct kvm_pit_state2 (in) 3073:Returns: 0 on success, -1 on error 3074 3075Sets the state of the in-kernel PIT model. Only valid after KVM_CREATE_PIT2. 3076See KVM_GET_PIT2 for details on struct kvm_pit_state2. 3077 3078This IOCTL replaces the obsolete KVM_SET_PIT. 3079 3080 30814.74 KVM_PPC_GET_SMMU_INFO 3082-------------------------- 3083 3084:Capability: KVM_CAP_PPC_GET_SMMU_INFO 3085:Architectures: powerpc 3086:Type: vm ioctl 3087:Parameters: None 3088:Returns: 0 on success, -1 on error 3089 3090This populates and returns a structure describing the features of 3091the "Server" class MMU emulation supported by KVM. 3092This can in turn be used by userspace to generate the appropriate 3093device-tree properties for the guest operating system. 3094 3095The structure contains some global information, followed by an 3096array of supported segment page sizes:: 3097 3098 struct kvm_ppc_smmu_info { 3099 __u64 flags; 3100 __u32 slb_size; 3101 __u32 pad; 3102 struct kvm_ppc_one_seg_page_size sps[KVM_PPC_PAGE_SIZES_MAX_SZ]; 3103 }; 3104 3105The supported flags are: 3106 3107 - KVM_PPC_PAGE_SIZES_REAL: 3108 When that flag is set, guest page sizes must "fit" the backing 3109 store page sizes. When not set, any page size in the list can 3110 be used regardless of how they are backed by userspace. 3111 3112 - KVM_PPC_1T_SEGMENTS 3113 The emulated MMU supports 1T segments in addition to the 3114 standard 256M ones. 3115 3116 - KVM_PPC_NO_HASH 3117 This flag indicates that HPT guests are not supported by KVM, 3118 thus all guests must use radix MMU mode. 3119 3120The "slb_size" field indicates how many SLB entries are supported 3121 3122The "sps" array contains 8 entries indicating the supported base 3123page sizes for a segment in increasing order. Each entry is defined 3124as follow:: 3125 3126 struct kvm_ppc_one_seg_page_size { 3127 __u32 page_shift; /* Base page shift of segment (or 0) */ 3128 __u32 slb_enc; /* SLB encoding for BookS */ 3129 struct kvm_ppc_one_page_size enc[KVM_PPC_PAGE_SIZES_MAX_SZ]; 3130 }; 3131 3132An entry with a "page_shift" of 0 is unused. Because the array is 3133organized in increasing order, a lookup can stop when encountering 3134such an entry. 3135 3136The "slb_enc" field provides the encoding to use in the SLB for the 3137page size. The bits are in positions such as the value can directly 3138be OR'ed into the "vsid" argument of the slbmte instruction. 3139 3140The "enc" array is a list which for each of those segment base page 3141size provides the list of supported actual page sizes (which can be 3142only larger or equal to the base page size), along with the 3143corresponding encoding in the hash PTE. Similarly, the array is 31448 entries sorted by increasing sizes and an entry with a "0" shift 3145is an empty entry and a terminator:: 3146 3147 struct kvm_ppc_one_page_size { 3148 __u32 page_shift; /* Page shift (or 0) */ 3149 __u32 pte_enc; /* Encoding in the HPTE (>>12) */ 3150 }; 3151 3152The "pte_enc" field provides a value that can OR'ed into the hash 3153PTE's RPN field (ie, it needs to be shifted left by 12 to OR it 3154into the hash PTE second double word). 3155 31564.75 KVM_IRQFD 3157-------------- 3158 3159:Capability: KVM_CAP_IRQFD 3160:Architectures: x86 s390 arm64 3161:Type: vm ioctl 3162:Parameters: struct kvm_irqfd (in) 3163:Returns: 0 on success, -1 on error 3164 3165Allows setting an eventfd to directly trigger a guest interrupt. 3166kvm_irqfd.fd specifies the file descriptor to use as the eventfd and 3167kvm_irqfd.gsi specifies the irqchip pin toggled by this event. When 3168an event is triggered on the eventfd, an interrupt is injected into 3169the guest using the specified gsi pin. The irqfd is removed using 3170the KVM_IRQFD_FLAG_DEASSIGN flag, specifying both kvm_irqfd.fd 3171and kvm_irqfd.gsi. 3172 3173With KVM_CAP_IRQFD_RESAMPLE, KVM_IRQFD supports a de-assert and notify 3174mechanism allowing emulation of level-triggered, irqfd-based 3175interrupts. When KVM_IRQFD_FLAG_RESAMPLE is set the user must pass an 3176additional eventfd in the kvm_irqfd.resamplefd field. When operating 3177in resample mode, posting of an interrupt through kvm_irq.fd asserts 3178the specified gsi in the irqchip. When the irqchip is resampled, such 3179as from an EOI, the gsi is de-asserted and the user is notified via 3180kvm_irqfd.resamplefd. It is the user's responsibility to re-queue 3181the interrupt if the device making use of it still requires service. 3182Note that closing the resamplefd is not sufficient to disable the 3183irqfd. The KVM_IRQFD_FLAG_RESAMPLE is only necessary on assignment 3184and need not be specified with KVM_IRQFD_FLAG_DEASSIGN. 3185 3186On arm64, gsi routing being supported, the following can happen: 3187 3188- in case no routing entry is associated to this gsi, injection fails 3189- in case the gsi is associated to an irqchip routing entry, 3190 irqchip.pin + 32 corresponds to the injected SPI ID. 3191- in case the gsi is associated to an MSI routing entry, the MSI 3192 message and device ID are translated into an LPI (support restricted 3193 to GICv3 ITS in-kernel emulation). 3194 31954.76 KVM_PPC_ALLOCATE_HTAB 3196-------------------------- 3197 3198:Capability: KVM_CAP_PPC_ALLOC_HTAB 3199:Architectures: powerpc 3200:Type: vm ioctl 3201:Parameters: Pointer to u32 containing hash table order (in/out) 3202:Returns: 0 on success, -1 on error 3203 3204This requests the host kernel to allocate an MMU hash table for a 3205guest using the PAPR paravirtualization interface. This only does 3206anything if the kernel is configured to use the Book 3S HV style of 3207virtualization. Otherwise the capability doesn't exist and the ioctl 3208returns an ENOTTY error. The rest of this description assumes Book 3S 3209HV. 3210 3211There must be no vcpus running when this ioctl is called; if there 3212are, it will do nothing and return an EBUSY error. 3213 3214The parameter is a pointer to a 32-bit unsigned integer variable 3215containing the order (log base 2) of the desired size of the hash 3216table, which must be between 18 and 46. On successful return from the 3217ioctl, the value will not be changed by the kernel. 3218 3219If no hash table has been allocated when any vcpu is asked to run 3220(with the KVM_RUN ioctl), the host kernel will allocate a 3221default-sized hash table (16 MB). 3222 3223If this ioctl is called when a hash table has already been allocated, 3224with a different order from the existing hash table, the existing hash 3225table will be freed and a new one allocated. If this is ioctl is 3226called when a hash table has already been allocated of the same order 3227as specified, the kernel will clear out the existing hash table (zero 3228all HPTEs). In either case, if the guest is using the virtualized 3229real-mode area (VRMA) facility, the kernel will re-create the VMRA 3230HPTEs on the next KVM_RUN of any vcpu. 3231 32324.77 KVM_S390_INTERRUPT 3233----------------------- 3234 3235:Capability: basic 3236:Architectures: s390 3237:Type: vm ioctl, vcpu ioctl 3238:Parameters: struct kvm_s390_interrupt (in) 3239:Returns: 0 on success, -1 on error 3240 3241Allows to inject an interrupt to the guest. Interrupts can be floating 3242(vm ioctl) or per cpu (vcpu ioctl), depending on the interrupt type. 3243 3244Interrupt parameters are passed via kvm_s390_interrupt:: 3245 3246 struct kvm_s390_interrupt { 3247 __u32 type; 3248 __u32 parm; 3249 __u64 parm64; 3250 }; 3251 3252type can be one of the following: 3253 3254KVM_S390_SIGP_STOP (vcpu) 3255 - sigp stop; optional flags in parm 3256KVM_S390_PROGRAM_INT (vcpu) 3257 - program check; code in parm 3258KVM_S390_SIGP_SET_PREFIX (vcpu) 3259 - sigp set prefix; prefix address in parm 3260KVM_S390_RESTART (vcpu) 3261 - restart 3262KVM_S390_INT_CLOCK_COMP (vcpu) 3263 - clock comparator interrupt 3264KVM_S390_INT_CPU_TIMER (vcpu) 3265 - CPU timer interrupt 3266KVM_S390_INT_VIRTIO (vm) 3267 - virtio external interrupt; external interrupt 3268 parameters in parm and parm64 3269KVM_S390_INT_SERVICE (vm) 3270 - sclp external interrupt; sclp parameter in parm 3271KVM_S390_INT_EMERGENCY (vcpu) 3272 - sigp emergency; source cpu in parm 3273KVM_S390_INT_EXTERNAL_CALL (vcpu) 3274 - sigp external call; source cpu in parm 3275KVM_S390_INT_IO(ai,cssid,ssid,schid) (vm) 3276 - compound value to indicate an 3277 I/O interrupt (ai - adapter interrupt; cssid,ssid,schid - subchannel); 3278 I/O interruption parameters in parm (subchannel) and parm64 (intparm, 3279 interruption subclass) 3280KVM_S390_MCHK (vm, vcpu) 3281 - machine check interrupt; cr 14 bits in parm, machine check interrupt 3282 code in parm64 (note that machine checks needing further payload are not 3283 supported by this ioctl) 3284 3285This is an asynchronous vcpu ioctl and can be invoked from any thread. 3286 32874.78 KVM_PPC_GET_HTAB_FD 3288------------------------ 3289 3290:Capability: KVM_CAP_PPC_HTAB_FD 3291:Architectures: powerpc 3292:Type: vm ioctl 3293:Parameters: Pointer to struct kvm_get_htab_fd (in) 3294:Returns: file descriptor number (>= 0) on success, -1 on error 3295 3296This returns a file descriptor that can be used either to read out the 3297entries in the guest's hashed page table (HPT), or to write entries to 3298initialize the HPT. The returned fd can only be written to if the 3299KVM_GET_HTAB_WRITE bit is set in the flags field of the argument, and 3300can only be read if that bit is clear. The argument struct looks like 3301this:: 3302 3303 /* For KVM_PPC_GET_HTAB_FD */ 3304 struct kvm_get_htab_fd { 3305 __u64 flags; 3306 __u64 start_index; 3307 __u64 reserved[2]; 3308 }; 3309 3310 /* Values for kvm_get_htab_fd.flags */ 3311 #define KVM_GET_HTAB_BOLTED_ONLY ((__u64)0x1) 3312 #define KVM_GET_HTAB_WRITE ((__u64)0x2) 3313 3314The 'start_index' field gives the index in the HPT of the entry at 3315which to start reading. It is ignored when writing. 3316 3317Reads on the fd will initially supply information about all 3318"interesting" HPT entries. Interesting entries are those with the 3319bolted bit set, if the KVM_GET_HTAB_BOLTED_ONLY bit is set, otherwise 3320all entries. When the end of the HPT is reached, the read() will 3321return. If read() is called again on the fd, it will start again from 3322the beginning of the HPT, but will only return HPT entries that have 3323changed since they were last read. 3324 3325Data read or written is structured as a header (8 bytes) followed by a 3326series of valid HPT entries (16 bytes) each. The header indicates how 3327many valid HPT entries there are and how many invalid entries follow 3328the valid entries. The invalid entries are not represented explicitly 3329in the stream. The header format is:: 3330 3331 struct kvm_get_htab_header { 3332 __u32 index; 3333 __u16 n_valid; 3334 __u16 n_invalid; 3335 }; 3336 3337Writes to the fd create HPT entries starting at the index given in the 3338header; first 'n_valid' valid entries with contents from the data 3339written, then 'n_invalid' invalid entries, invalidating any previously 3340valid entries found. 3341 33424.79 KVM_CREATE_DEVICE 3343---------------------- 3344 3345:Capability: KVM_CAP_DEVICE_CTRL 3346:Architectures: all 3347:Type: vm ioctl 3348:Parameters: struct kvm_create_device (in/out) 3349:Returns: 0 on success, -1 on error 3350 3351Errors: 3352 3353 ====== ======================================================= 3354 ENODEV The device type is unknown or unsupported 3355 EEXIST Device already created, and this type of device may not 3356 be instantiated multiple times 3357 ====== ======================================================= 3358 3359 Other error conditions may be defined by individual device types or 3360 have their standard meanings. 3361 3362Creates an emulated device in the kernel. The file descriptor returned 3363in fd can be used with KVM_SET/GET/HAS_DEVICE_ATTR. 3364 3365If the KVM_CREATE_DEVICE_TEST flag is set, only test whether the 3366device type is supported (not necessarily whether it can be created 3367in the current vm). 3368 3369Individual devices should not define flags. Attributes should be used 3370for specifying any behavior that is not implied by the device type 3371number. 3372 3373:: 3374 3375 struct kvm_create_device { 3376 __u32 type; /* in: KVM_DEV_TYPE_xxx */ 3377 __u32 fd; /* out: device handle */ 3378 __u32 flags; /* in: KVM_CREATE_DEVICE_xxx */ 3379 }; 3380 33814.80 KVM_SET_DEVICE_ATTR/KVM_GET_DEVICE_ATTR 3382-------------------------------------------- 3383 3384:Capability: KVM_CAP_DEVICE_CTRL, KVM_CAP_VM_ATTRIBUTES for vm device, 3385 KVM_CAP_VCPU_ATTRIBUTES for vcpu device 3386 KVM_CAP_SYS_ATTRIBUTES for system (/dev/kvm) device (no set) 3387:Architectures: x86, arm64, s390 3388:Type: device ioctl, vm ioctl, vcpu ioctl 3389:Parameters: struct kvm_device_attr 3390:Returns: 0 on success, -1 on error 3391 3392Errors: 3393 3394 ===== ============================================================= 3395 ENXIO The group or attribute is unknown/unsupported for this device 3396 or hardware support is missing. 3397 EPERM The attribute cannot (currently) be accessed this way 3398 (e.g. read-only attribute, or attribute that only makes 3399 sense when the device is in a different state) 3400 ===== ============================================================= 3401 3402 Other error conditions may be defined by individual device types. 3403 3404Gets/sets a specified piece of device configuration and/or state. The 3405semantics are device-specific. See individual device documentation in 3406the "devices" directory. As with ONE_REG, the size of the data 3407transferred is defined by the particular attribute. 3408 3409:: 3410 3411 struct kvm_device_attr { 3412 __u32 flags; /* no flags currently defined */ 3413 __u32 group; /* device-defined */ 3414 __u64 attr; /* group-defined */ 3415 __u64 addr; /* userspace address of attr data */ 3416 }; 3417 34184.81 KVM_HAS_DEVICE_ATTR 3419------------------------ 3420 3421:Capability: KVM_CAP_DEVICE_CTRL, KVM_CAP_VM_ATTRIBUTES for vm device, 3422 KVM_CAP_VCPU_ATTRIBUTES for vcpu device 3423 KVM_CAP_SYS_ATTRIBUTES for system (/dev/kvm) device 3424:Type: device ioctl, vm ioctl, vcpu ioctl 3425:Parameters: struct kvm_device_attr 3426:Returns: 0 on success, -1 on error 3427 3428Errors: 3429 3430 ===== ============================================================= 3431 ENXIO The group or attribute is unknown/unsupported for this device 3432 or hardware support is missing. 3433 ===== ============================================================= 3434 3435Tests whether a device supports a particular attribute. A successful 3436return indicates the attribute is implemented. It does not necessarily 3437indicate that the attribute can be read or written in the device's 3438current state. "addr" is ignored. 3439 3440.. _KVM_ARM_VCPU_INIT: 3441 34424.82 KVM_ARM_VCPU_INIT 3443---------------------- 3444 3445:Capability: basic 3446:Architectures: arm64 3447:Type: vcpu ioctl 3448:Parameters: struct kvm_vcpu_init (in) 3449:Returns: 0 on success; -1 on error 3450 3451Errors: 3452 3453 ====== ================================================================= 3454 EINVAL the target is unknown, or the combination of features is invalid. 3455 ENOENT a features bit specified is unknown. 3456 ====== ================================================================= 3457 3458This tells KVM what type of CPU to present to the guest, and what 3459optional features it should have. This will cause a reset of the cpu 3460registers to their initial values. If this is not called, KVM_RUN will 3461return ENOEXEC for that vcpu. 3462 3463The initial values are defined as: 3464 - Processor state: 3465 * AArch64: EL1h, D, A, I and F bits set. All other bits 3466 are cleared. 3467 * AArch32: SVC, A, I and F bits set. All other bits are 3468 cleared. 3469 - General Purpose registers, including PC and SP: set to 0 3470 - FPSIMD/NEON registers: set to 0 3471 - SVE registers: set to 0 3472 - System registers: Reset to their architecturally defined 3473 values as for a warm reset to EL1 (resp. SVC) or EL2 (in the 3474 case of EL2 being enabled). 3475 3476Note that because some registers reflect machine topology, all vcpus 3477should be created before this ioctl is invoked. 3478 3479Userspace can call this function multiple times for a given vcpu, including 3480after the vcpu has been run. This will reset the vcpu to its initial 3481state. All calls to this function after the initial call must use the same 3482target and same set of feature flags, otherwise EINVAL will be returned. 3483 3484Possible features: 3485 3486 - KVM_ARM_VCPU_POWER_OFF: Starts the CPU in a power-off state. 3487 Depends on KVM_CAP_ARM_PSCI. If not set, the CPU will be powered on 3488 and execute guest code when KVM_RUN is called. 3489 - KVM_ARM_VCPU_EL1_32BIT: Starts the CPU in a 32bit mode. 3490 Depends on KVM_CAP_ARM_EL1_32BIT (arm64 only). 3491 - KVM_ARM_VCPU_PSCI_0_2: Emulate PSCI v0.2 (or a future revision 3492 backward compatible with v0.2) for the CPU. 3493 Depends on KVM_CAP_ARM_PSCI_0_2. 3494 - KVM_ARM_VCPU_PMU_V3: Emulate PMUv3 for the CPU. 3495 Depends on KVM_CAP_ARM_PMU_V3. 3496 3497 - KVM_ARM_VCPU_PTRAUTH_ADDRESS: Enables Address Pointer authentication 3498 for arm64 only. 3499 Depends on KVM_CAP_ARM_PTRAUTH_ADDRESS. 3500 If KVM_CAP_ARM_PTRAUTH_ADDRESS and KVM_CAP_ARM_PTRAUTH_GENERIC are 3501 both present, then both KVM_ARM_VCPU_PTRAUTH_ADDRESS and 3502 KVM_ARM_VCPU_PTRAUTH_GENERIC must be requested or neither must be 3503 requested. 3504 3505 - KVM_ARM_VCPU_PTRAUTH_GENERIC: Enables Generic Pointer authentication 3506 for arm64 only. 3507 Depends on KVM_CAP_ARM_PTRAUTH_GENERIC. 3508 If KVM_CAP_ARM_PTRAUTH_ADDRESS and KVM_CAP_ARM_PTRAUTH_GENERIC are 3509 both present, then both KVM_ARM_VCPU_PTRAUTH_ADDRESS and 3510 KVM_ARM_VCPU_PTRAUTH_GENERIC must be requested or neither must be 3511 requested. 3512 3513 - KVM_ARM_VCPU_SVE: Enables SVE for the CPU (arm64 only). 3514 Depends on KVM_CAP_ARM_SVE. 3515 Requires KVM_ARM_VCPU_FINALIZE(KVM_ARM_VCPU_SVE): 3516 3517 * After KVM_ARM_VCPU_INIT: 3518 3519 - KVM_REG_ARM64_SVE_VLS may be read using KVM_GET_ONE_REG: the 3520 initial value of this pseudo-register indicates the best set of 3521 vector lengths possible for a vcpu on this host. 3522 3523 * Before KVM_ARM_VCPU_FINALIZE(KVM_ARM_VCPU_SVE): 3524 3525 - KVM_RUN and KVM_GET_REG_LIST are not available; 3526 3527 - KVM_GET_ONE_REG and KVM_SET_ONE_REG cannot be used to access 3528 the scalable architectural SVE registers 3529 KVM_REG_ARM64_SVE_ZREG(), KVM_REG_ARM64_SVE_PREG() or 3530 KVM_REG_ARM64_SVE_FFR; 3531 3532 - KVM_REG_ARM64_SVE_VLS may optionally be written using 3533 KVM_SET_ONE_REG, to modify the set of vector lengths available 3534 for the vcpu. 3535 3536 * After KVM_ARM_VCPU_FINALIZE(KVM_ARM_VCPU_SVE): 3537 3538 - the KVM_REG_ARM64_SVE_VLS pseudo-register is immutable, and can 3539 no longer be written using KVM_SET_ONE_REG. 3540 3541 - KVM_ARM_VCPU_HAS_EL2: Enable Nested Virtualisation support, 3542 booting the guest from EL2 instead of EL1. 3543 Depends on KVM_CAP_ARM_EL2. 3544 The VM is running with HCR_EL2.E2H being RES1 (VHE) unless 3545 KVM_ARM_VCPU_HAS_EL2_E2H0 is also set. 3546 3547 - KVM_ARM_VCPU_HAS_EL2_E2H0: Restrict Nested Virtualisation 3548 support to HCR_EL2.E2H being RES0 (non-VHE). 3549 Depends on KVM_CAP_ARM_EL2_E2H0. 3550 KVM_ARM_VCPU_HAS_EL2 must also be set. 3551 35524.83 KVM_ARM_PREFERRED_TARGET 3553----------------------------- 3554 3555:Capability: basic 3556:Architectures: arm64 3557:Type: vm ioctl 3558:Parameters: struct kvm_vcpu_init (out) 3559:Returns: 0 on success; -1 on error 3560 3561Errors: 3562 3563 ====== ========================================== 3564 ENODEV no preferred target available for the host 3565 ====== ========================================== 3566 3567This queries KVM for preferred CPU target type which can be emulated 3568by KVM on underlying host. 3569 3570The ioctl returns struct kvm_vcpu_init instance containing information 3571about preferred CPU target type and recommended features for it. The 3572kvm_vcpu_init->features bitmap returned will have feature bits set if 3573the preferred target recommends setting these features, but this is 3574not mandatory. 3575 3576The information returned by this ioctl can be used to prepare an instance 3577of struct kvm_vcpu_init for KVM_ARM_VCPU_INIT ioctl which will result in 3578VCPU matching underlying host. 3579 3580 35814.84 KVM_GET_REG_LIST 3582--------------------- 3583 3584:Capability: basic 3585:Architectures: arm64, mips, riscv 3586:Type: vcpu ioctl 3587:Parameters: struct kvm_reg_list (in/out) 3588:Returns: 0 on success; -1 on error 3589 3590Errors: 3591 3592 ===== ============================================================== 3593 E2BIG the reg index list is too big to fit in the array specified by 3594 the user (the number required will be written into n). 3595 ===== ============================================================== 3596 3597:: 3598 3599 struct kvm_reg_list { 3600 __u64 n; /* number of registers in reg[] */ 3601 __u64 reg[0]; 3602 }; 3603 3604This ioctl returns the guest registers that are supported for the 3605KVM_GET_ONE_REG/KVM_SET_ONE_REG calls. 3606 3607Note that s390 does not support KVM_GET_REG_LIST for historical reasons 3608(read: nobody cared). The set of registers in kernels 4.x and newer is: 3609 3610- KVM_REG_S390_TODPR 3611 3612- KVM_REG_S390_EPOCHDIFF 3613 3614- KVM_REG_S390_CPU_TIMER 3615 3616- KVM_REG_S390_CLOCK_COMP 3617 3618- KVM_REG_S390_PFTOKEN 3619 3620- KVM_REG_S390_PFCOMPARE 3621 3622- KVM_REG_S390_PFSELECT 3623 3624- KVM_REG_S390_PP 3625 3626- KVM_REG_S390_GBEA 3627 3628 36294.85 KVM_ARM_SET_DEVICE_ADDR (deprecated) 3630----------------------------------------- 3631 3632:Capability: KVM_CAP_ARM_SET_DEVICE_ADDR 3633:Architectures: arm64 3634:Type: vm ioctl 3635:Parameters: struct kvm_arm_device_address (in) 3636:Returns: 0 on success, -1 on error 3637 3638Errors: 3639 3640 ====== ============================================ 3641 ENODEV The device id is unknown 3642 ENXIO Device not supported on current system 3643 EEXIST Address already set 3644 E2BIG Address outside guest physical address space 3645 EBUSY Address overlaps with other device range 3646 ====== ============================================ 3647 3648:: 3649 3650 struct kvm_arm_device_addr { 3651 __u64 id; 3652 __u64 addr; 3653 }; 3654 3655Specify a device address in the guest's physical address space where guests 3656can access emulated or directly exposed devices, which the host kernel needs 3657to know about. The id field is an architecture specific identifier for a 3658specific device. 3659 3660arm64 divides the id field into two parts, a device id and an 3661address type id specific to the individual device:: 3662 3663 bits: | 63 ... 32 | 31 ... 16 | 15 ... 0 | 3664 field: | 0x00000000 | device id | addr type id | 3665 3666arm64 currently only require this when using the in-kernel GIC 3667support for the hardware VGIC features, using KVM_ARM_DEVICE_VGIC_V2 3668as the device id. When setting the base address for the guest's 3669mapping of the VGIC virtual CPU and distributor interface, the ioctl 3670must be called after calling KVM_CREATE_IRQCHIP, but before calling 3671KVM_RUN on any of the VCPUs. Calling this ioctl twice for any of the 3672base addresses will return -EEXIST. 3673 3674Note, this IOCTL is deprecated and the more flexible SET/GET_DEVICE_ATTR API 3675should be used instead. 3676 3677 36784.86 KVM_PPC_RTAS_DEFINE_TOKEN 3679------------------------------ 3680 3681:Capability: KVM_CAP_PPC_RTAS 3682:Architectures: ppc 3683:Type: vm ioctl 3684:Parameters: struct kvm_rtas_token_args 3685:Returns: 0 on success, -1 on error 3686 3687Defines a token value for a RTAS (Run Time Abstraction Services) 3688service in order to allow it to be handled in the kernel. The 3689argument struct gives the name of the service, which must be the name 3690of a service that has a kernel-side implementation. If the token 3691value is non-zero, it will be associated with that service, and 3692subsequent RTAS calls by the guest specifying that token will be 3693handled by the kernel. If the token value is 0, then any token 3694associated with the service will be forgotten, and subsequent RTAS 3695calls by the guest for that service will be passed to userspace to be 3696handled. 3697 36984.87 KVM_SET_GUEST_DEBUG 3699------------------------ 3700 3701:Capability: KVM_CAP_SET_GUEST_DEBUG 3702:Architectures: x86, s390, ppc, arm64 3703:Type: vcpu ioctl 3704:Parameters: struct kvm_guest_debug (in) 3705:Returns: 0 on success; -1 on error 3706 3707:: 3708 3709 struct kvm_guest_debug { 3710 __u32 control; 3711 __u32 pad; 3712 struct kvm_guest_debug_arch arch; 3713 }; 3714 3715Set up the processor specific debug registers and configure vcpu for 3716handling guest debug events. There are two parts to the structure, the 3717first a control bitfield indicates the type of debug events to handle 3718when running. Common control bits are: 3719 3720 - KVM_GUESTDBG_ENABLE: guest debugging is enabled 3721 - KVM_GUESTDBG_SINGLESTEP: the next run should single-step 3722 3723The top 16 bits of the control field are architecture specific control 3724flags which can include the following: 3725 3726 - KVM_GUESTDBG_USE_SW_BP: using software breakpoints [x86, arm64] 3727 - KVM_GUESTDBG_USE_HW_BP: using hardware breakpoints [x86, s390] 3728 - KVM_GUESTDBG_USE_HW: using hardware debug events [arm64] 3729 - KVM_GUESTDBG_INJECT_DB: inject DB type exception [x86] 3730 - KVM_GUESTDBG_INJECT_BP: inject BP type exception [x86] 3731 - KVM_GUESTDBG_EXIT_PENDING: trigger an immediate guest exit [s390] 3732 - KVM_GUESTDBG_BLOCKIRQ: avoid injecting interrupts/NMI/SMI [x86] 3733 3734For example KVM_GUESTDBG_USE_SW_BP indicates that software breakpoints 3735are enabled in memory so we need to ensure breakpoint exceptions are 3736correctly trapped and the KVM run loop exits at the breakpoint and not 3737running off into the normal guest vector. For KVM_GUESTDBG_USE_HW_BP 3738we need to ensure the guest vCPUs architecture specific registers are 3739updated to the correct (supplied) values. 3740 3741The second part of the structure is architecture specific and 3742typically contains a set of debug registers. 3743 3744For arm64 the number of debug registers is implementation defined and 3745can be determined by querying the KVM_CAP_GUEST_DEBUG_HW_BPS and 3746KVM_CAP_GUEST_DEBUG_HW_WPS capabilities which return a positive number 3747indicating the number of supported registers. 3748 3749For ppc, the KVM_CAP_PPC_GUEST_DEBUG_SSTEP capability indicates whether 3750the single-step debug event (KVM_GUESTDBG_SINGLESTEP) is supported. 3751 3752Also when supported, KVM_CAP_SET_GUEST_DEBUG2 capability indicates the 3753supported KVM_GUESTDBG_* bits in the control field. 3754 3755When debug events exit the main run loop with the reason 3756KVM_EXIT_DEBUG with the kvm_debug_exit_arch part of the kvm_run 3757structure containing architecture specific debug information. 3758 37594.88 KVM_GET_EMULATED_CPUID 3760--------------------------- 3761 3762:Capability: KVM_CAP_EXT_EMUL_CPUID 3763:Architectures: x86 3764:Type: system ioctl 3765:Parameters: struct kvm_cpuid2 (in/out) 3766:Returns: 0 on success, -1 on error 3767 3768:: 3769 3770 struct kvm_cpuid2 { 3771 __u32 nent; 3772 __u32 flags; 3773 struct kvm_cpuid_entry2 entries[0]; 3774 }; 3775 3776The member 'flags' is used for passing flags from userspace. 3777 3778:: 3779 3780 #define KVM_CPUID_FLAG_SIGNIFCANT_INDEX BIT(0) 3781 #define KVM_CPUID_FLAG_STATEFUL_FUNC BIT(1) /* deprecated */ 3782 #define KVM_CPUID_FLAG_STATE_READ_NEXT BIT(2) /* deprecated */ 3783 3784 struct kvm_cpuid_entry2 { 3785 __u32 function; 3786 __u32 index; 3787 __u32 flags; 3788 __u32 eax; 3789 __u32 ebx; 3790 __u32 ecx; 3791 __u32 edx; 3792 __u32 padding[3]; 3793 }; 3794 3795This ioctl returns x86 cpuid features which are emulated by 3796kvm.Userspace can use the information returned by this ioctl to query 3797which features are emulated by kvm instead of being present natively. 3798 3799Userspace invokes KVM_GET_EMULATED_CPUID by passing a kvm_cpuid2 3800structure with the 'nent' field indicating the number of entries in 3801the variable-size array 'entries'. If the number of entries is too low 3802to describe the cpu capabilities, an error (E2BIG) is returned. If the 3803number is too high, the 'nent' field is adjusted and an error (ENOMEM) 3804is returned. If the number is just right, the 'nent' field is adjusted 3805to the number of valid entries in the 'entries' array, which is then 3806filled. 3807 3808The entries returned are the set CPUID bits of the respective features 3809which kvm emulates, as returned by the CPUID instruction, with unknown 3810or unsupported feature bits cleared. 3811 3812Features like x2apic, for example, may not be present in the host cpu 3813but are exposed by kvm in KVM_GET_SUPPORTED_CPUID because they can be 3814emulated efficiently and thus not included here. 3815 3816The fields in each entry are defined as follows: 3817 3818 function: 3819 the eax value used to obtain the entry 3820 index: 3821 the ecx value used to obtain the entry (for entries that are 3822 affected by ecx) 3823 flags: 3824 an OR of zero or more of the following: 3825 3826 KVM_CPUID_FLAG_SIGNIFCANT_INDEX: 3827 if the index field is valid 3828 3829 eax, ebx, ecx, edx: 3830 3831 the values returned by the cpuid instruction for 3832 this function/index combination 3833 38344.89 KVM_S390_MEM_OP 3835-------------------- 3836 3837:Capability: KVM_CAP_S390_MEM_OP, KVM_CAP_S390_PROTECTED, KVM_CAP_S390_MEM_OP_EXTENSION 3838:Architectures: s390 3839:Type: vm ioctl, vcpu ioctl 3840:Parameters: struct kvm_s390_mem_op (in) 3841:Returns: = 0 on success, 3842 < 0 on generic error (e.g. -EFAULT or -ENOMEM), 3843 16 bit program exception code if the access causes such an exception 3844 3845Read or write data from/to the VM's memory. 3846The KVM_CAP_S390_MEM_OP_EXTENSION capability specifies what functionality is 3847supported. 3848 3849Parameters are specified via the following structure:: 3850 3851 struct kvm_s390_mem_op { 3852 __u64 gaddr; /* the guest address */ 3853 __u64 flags; /* flags */ 3854 __u32 size; /* amount of bytes */ 3855 __u32 op; /* type of operation */ 3856 __u64 buf; /* buffer in userspace */ 3857 union { 3858 struct { 3859 __u8 ar; /* the access register number */ 3860 __u8 key; /* access key, ignored if flag unset */ 3861 __u8 pad1[6]; /* ignored */ 3862 __u64 old_addr; /* ignored if flag unset */ 3863 }; 3864 __u32 sida_offset; /* offset into the sida */ 3865 __u8 reserved[32]; /* ignored */ 3866 }; 3867 }; 3868 3869The start address of the memory region has to be specified in the "gaddr" 3870field, and the length of the region in the "size" field (which must not 3871be 0). The maximum value for "size" can be obtained by checking the 3872KVM_CAP_S390_MEM_OP capability. "buf" is the buffer supplied by the 3873userspace application where the read data should be written to for 3874a read access, or where the data that should be written is stored for 3875a write access. The "reserved" field is meant for future extensions. 3876Reserved and unused values are ignored. Future extension that add members must 3877introduce new flags. 3878 3879The type of operation is specified in the "op" field. Flags modifying 3880their behavior can be set in the "flags" field. Undefined flag bits must 3881be set to 0. 3882 3883Possible operations are: 3884 * ``KVM_S390_MEMOP_LOGICAL_READ`` 3885 * ``KVM_S390_MEMOP_LOGICAL_WRITE`` 3886 * ``KVM_S390_MEMOP_ABSOLUTE_READ`` 3887 * ``KVM_S390_MEMOP_ABSOLUTE_WRITE`` 3888 * ``KVM_S390_MEMOP_SIDA_READ`` 3889 * ``KVM_S390_MEMOP_SIDA_WRITE`` 3890 * ``KVM_S390_MEMOP_ABSOLUTE_CMPXCHG`` 3891 3892Logical read/write: 3893^^^^^^^^^^^^^^^^^^^ 3894 3895Access logical memory, i.e. translate the given guest address to an absolute 3896address given the state of the VCPU and use the absolute address as target of 3897the access. "ar" designates the access register number to be used; the valid 3898range is 0..15. 3899Logical accesses are permitted for the VCPU ioctl only. 3900Logical accesses are permitted for non-protected guests only. 3901 3902Supported flags: 3903 * ``KVM_S390_MEMOP_F_CHECK_ONLY`` 3904 * ``KVM_S390_MEMOP_F_INJECT_EXCEPTION`` 3905 * ``KVM_S390_MEMOP_F_SKEY_PROTECTION`` 3906 3907The KVM_S390_MEMOP_F_CHECK_ONLY flag can be set to check whether the 3908corresponding memory access would cause an access exception; however, 3909no actual access to the data in memory at the destination is performed. 3910In this case, "buf" is unused and can be NULL. 3911 3912In case an access exception occurred during the access (or would occur 3913in case of KVM_S390_MEMOP_F_CHECK_ONLY), the ioctl returns a positive 3914error number indicating the type of exception. This exception is also 3915raised directly at the corresponding VCPU if the flag 3916KVM_S390_MEMOP_F_INJECT_EXCEPTION is set. 3917On protection exceptions, unless specified otherwise, the injected 3918translation-exception identifier (TEID) indicates suppression. 3919 3920If the KVM_S390_MEMOP_F_SKEY_PROTECTION flag is set, storage key 3921protection is also in effect and may cause exceptions if accesses are 3922prohibited given the access key designated by "key"; the valid range is 0..15. 3923KVM_S390_MEMOP_F_SKEY_PROTECTION is available if KVM_CAP_S390_MEM_OP_EXTENSION 3924is > 0. 3925Since the accessed memory may span multiple pages and those pages might have 3926different storage keys, it is possible that a protection exception occurs 3927after memory has been modified. In this case, if the exception is injected, 3928the TEID does not indicate suppression. 3929 3930Absolute read/write: 3931^^^^^^^^^^^^^^^^^^^^ 3932 3933Access absolute memory. This operation is intended to be used with the 3934KVM_S390_MEMOP_F_SKEY_PROTECTION flag, to allow accessing memory and performing 3935the checks required for storage key protection as one operation (as opposed to 3936user space getting the storage keys, performing the checks, and accessing 3937memory thereafter, which could lead to a delay between check and access). 3938Absolute accesses are permitted for the VM ioctl if KVM_CAP_S390_MEM_OP_EXTENSION 3939has the KVM_S390_MEMOP_EXTENSION_CAP_BASE bit set. 3940Currently absolute accesses are not permitted for VCPU ioctls. 3941Absolute accesses are permitted for non-protected guests only. 3942 3943Supported flags: 3944 * ``KVM_S390_MEMOP_F_CHECK_ONLY`` 3945 * ``KVM_S390_MEMOP_F_SKEY_PROTECTION`` 3946 3947The semantics of the flags common with logical accesses are as for logical 3948accesses. 3949 3950Absolute cmpxchg: 3951^^^^^^^^^^^^^^^^^ 3952 3953Perform cmpxchg on absolute guest memory. Intended for use with the 3954KVM_S390_MEMOP_F_SKEY_PROTECTION flag. 3955Instead of doing an unconditional write, the access occurs only if the target 3956location contains the value pointed to by "old_addr". 3957This is performed as an atomic cmpxchg with the length specified by the "size" 3958parameter. "size" must be a power of two up to and including 16. 3959If the exchange did not take place because the target value doesn't match the 3960old value, the value "old_addr" points to is replaced by the target value. 3961User space can tell if an exchange took place by checking if this replacement 3962occurred. The cmpxchg op is permitted for the VM ioctl if 3963KVM_CAP_S390_MEM_OP_EXTENSION has flag KVM_S390_MEMOP_EXTENSION_CAP_CMPXCHG set. 3964 3965Supported flags: 3966 * ``KVM_S390_MEMOP_F_SKEY_PROTECTION`` 3967 3968SIDA read/write: 3969^^^^^^^^^^^^^^^^ 3970 3971Access the secure instruction data area which contains memory operands necessary 3972for instruction emulation for protected guests. 3973SIDA accesses are available if the KVM_CAP_S390_PROTECTED capability is available. 3974SIDA accesses are permitted for the VCPU ioctl only. 3975SIDA accesses are permitted for protected guests only. 3976 3977No flags are supported. 3978 39794.90 KVM_S390_GET_SKEYS 3980----------------------- 3981 3982:Capability: KVM_CAP_S390_SKEYS 3983:Architectures: s390 3984:Type: vm ioctl 3985:Parameters: struct kvm_s390_skeys 3986:Returns: 0 on success, KVM_S390_GET_SKEYS_NONE if guest is not using storage 3987 keys, negative value on error 3988 3989This ioctl is used to get guest storage key values on the s390 3990architecture. The ioctl takes parameters via the kvm_s390_skeys struct:: 3991 3992 struct kvm_s390_skeys { 3993 __u64 start_gfn; 3994 __u64 count; 3995 __u64 skeydata_addr; 3996 __u32 flags; 3997 __u32 reserved[9]; 3998 }; 3999 4000The start_gfn field is the number of the first guest frame whose storage keys 4001you want to get. 4002 4003The count field is the number of consecutive frames (starting from start_gfn) 4004whose storage keys to get. The count field must be at least 1 and the maximum 4005allowed value is defined as KVM_S390_SKEYS_MAX. Values outside this range 4006will cause the ioctl to return -EINVAL. 4007 4008The skeydata_addr field is the address to a buffer large enough to hold count 4009bytes. This buffer will be filled with storage key data by the ioctl. 4010 40114.91 KVM_S390_SET_SKEYS 4012----------------------- 4013 4014:Capability: KVM_CAP_S390_SKEYS 4015:Architectures: s390 4016:Type: vm ioctl 4017:Parameters: struct kvm_s390_skeys 4018:Returns: 0 on success, negative value on error 4019 4020This ioctl is used to set guest storage key values on the s390 4021architecture. The ioctl takes parameters via the kvm_s390_skeys struct. 4022See section on KVM_S390_GET_SKEYS for struct definition. 4023 4024The start_gfn field is the number of the first guest frame whose storage keys 4025you want to set. 4026 4027The count field is the number of consecutive frames (starting from start_gfn) 4028whose storage keys to get. The count field must be at least 1 and the maximum 4029allowed value is defined as KVM_S390_SKEYS_MAX. Values outside this range 4030will cause the ioctl to return -EINVAL. 4031 4032The skeydata_addr field is the address to a buffer containing count bytes of 4033storage keys. Each byte in the buffer will be set as the storage key for a 4034single frame starting at start_gfn for count frames. 4035 4036Note: If any architecturally invalid key value is found in the given data then 4037the ioctl will return -EINVAL. 4038 40394.92 KVM_S390_IRQ 4040----------------- 4041 4042:Capability: KVM_CAP_S390_INJECT_IRQ 4043:Architectures: s390 4044:Type: vcpu ioctl 4045:Parameters: struct kvm_s390_irq (in) 4046:Returns: 0 on success, -1 on error 4047 4048Errors: 4049 4050 4051 ====== ================================================================= 4052 EINVAL interrupt type is invalid 4053 type is KVM_S390_SIGP_STOP and flag parameter is invalid value, 4054 type is KVM_S390_INT_EXTERNAL_CALL and code is bigger 4055 than the maximum of VCPUs 4056 EBUSY type is KVM_S390_SIGP_SET_PREFIX and vcpu is not stopped, 4057 type is KVM_S390_SIGP_STOP and a stop irq is already pending, 4058 type is KVM_S390_INT_EXTERNAL_CALL and an external call interrupt 4059 is already pending 4060 ====== ================================================================= 4061 4062Allows to inject an interrupt to the guest. 4063 4064Using struct kvm_s390_irq as a parameter allows 4065to inject additional payload which is not 4066possible via KVM_S390_INTERRUPT. 4067 4068Interrupt parameters are passed via kvm_s390_irq:: 4069 4070 struct kvm_s390_irq { 4071 __u64 type; 4072 union { 4073 struct kvm_s390_io_info io; 4074 struct kvm_s390_ext_info ext; 4075 struct kvm_s390_pgm_info pgm; 4076 struct kvm_s390_emerg_info emerg; 4077 struct kvm_s390_extcall_info extcall; 4078 struct kvm_s390_prefix_info prefix; 4079 struct kvm_s390_stop_info stop; 4080 struct kvm_s390_mchk_info mchk; 4081 char reserved[64]; 4082 } u; 4083 }; 4084 4085type can be one of the following: 4086 4087- KVM_S390_SIGP_STOP - sigp stop; parameter in .stop 4088- KVM_S390_PROGRAM_INT - program check; parameters in .pgm 4089- KVM_S390_SIGP_SET_PREFIX - sigp set prefix; parameters in .prefix 4090- KVM_S390_RESTART - restart; no parameters 4091- KVM_S390_INT_CLOCK_COMP - clock comparator interrupt; no parameters 4092- KVM_S390_INT_CPU_TIMER - CPU timer interrupt; no parameters 4093- KVM_S390_INT_EMERGENCY - sigp emergency; parameters in .emerg 4094- KVM_S390_INT_EXTERNAL_CALL - sigp external call; parameters in .extcall 4095- KVM_S390_MCHK - machine check interrupt; parameters in .mchk 4096 4097This is an asynchronous vcpu ioctl and can be invoked from any thread. 4098 40994.94 KVM_S390_GET_IRQ_STATE 4100--------------------------- 4101 4102:Capability: KVM_CAP_S390_IRQ_STATE 4103:Architectures: s390 4104:Type: vcpu ioctl 4105:Parameters: struct kvm_s390_irq_state (out) 4106:Returns: >= number of bytes copied into buffer, 4107 -EINVAL if buffer size is 0, 4108 -ENOBUFS if buffer size is too small to fit all pending interrupts, 4109 -EFAULT if the buffer address was invalid 4110 4111This ioctl allows userspace to retrieve the complete state of all currently 4112pending interrupts in a single buffer. Use cases include migration 4113and introspection. The parameter structure contains the address of a 4114userspace buffer and its length:: 4115 4116 struct kvm_s390_irq_state { 4117 __u64 buf; 4118 __u32 flags; /* will stay unused for compatibility reasons */ 4119 __u32 len; 4120 __u32 reserved[4]; /* will stay unused for compatibility reasons */ 4121 }; 4122 4123Userspace passes in the above struct and for each pending interrupt a 4124struct kvm_s390_irq is copied to the provided buffer. 4125 4126The structure contains a flags and a reserved field for future extensions. As 4127the kernel never checked for flags == 0 and QEMU never pre-zeroed flags and 4128reserved, these fields can not be used in the future without breaking 4129compatibility. 4130 4131If -ENOBUFS is returned the buffer provided was too small and userspace 4132may retry with a bigger buffer. 4133 41344.95 KVM_S390_SET_IRQ_STATE 4135--------------------------- 4136 4137:Capability: KVM_CAP_S390_IRQ_STATE 4138:Architectures: s390 4139:Type: vcpu ioctl 4140:Parameters: struct kvm_s390_irq_state (in) 4141:Returns: 0 on success, 4142 -EFAULT if the buffer address was invalid, 4143 -EINVAL for an invalid buffer length (see below), 4144 -EBUSY if there were already interrupts pending, 4145 errors occurring when actually injecting the 4146 interrupt. See KVM_S390_IRQ. 4147 4148This ioctl allows userspace to set the complete state of all cpu-local 4149interrupts currently pending for the vcpu. It is intended for restoring 4150interrupt state after a migration. The input parameter is a userspace buffer 4151containing a struct kvm_s390_irq_state:: 4152 4153 struct kvm_s390_irq_state { 4154 __u64 buf; 4155 __u32 flags; /* will stay unused for compatibility reasons */ 4156 __u32 len; 4157 __u32 reserved[4]; /* will stay unused for compatibility reasons */ 4158 }; 4159 4160The restrictions for flags and reserved apply as well. 4161(see KVM_S390_GET_IRQ_STATE) 4162 4163The userspace memory referenced by buf contains a struct kvm_s390_irq 4164for each interrupt to be injected into the guest. 4165If one of the interrupts could not be injected for some reason the 4166ioctl aborts. 4167 4168len must be a multiple of sizeof(struct kvm_s390_irq). It must be > 0 4169and it must not exceed (max_vcpus + 32) * sizeof(struct kvm_s390_irq), 4170which is the maximum number of possibly pending cpu-local interrupts. 4171 41724.96 KVM_SMI 4173------------ 4174 4175:Capability: KVM_CAP_X86_SMM 4176:Architectures: x86 4177:Type: vcpu ioctl 4178:Parameters: none 4179:Returns: 0 on success, -1 on error 4180 4181Queues an SMI on the thread's vcpu. 4182 41834.97 KVM_X86_SET_MSR_FILTER 4184---------------------------- 4185 4186:Capability: KVM_CAP_X86_MSR_FILTER 4187:Architectures: x86 4188:Type: vm ioctl 4189:Parameters: struct kvm_msr_filter 4190:Returns: 0 on success, < 0 on error 4191 4192:: 4193 4194 struct kvm_msr_filter_range { 4195 #define KVM_MSR_FILTER_READ (1 << 0) 4196 #define KVM_MSR_FILTER_WRITE (1 << 1) 4197 __u32 flags; 4198 __u32 nmsrs; /* number of msrs in bitmap */ 4199 __u32 base; /* MSR index the bitmap starts at */ 4200 __u8 *bitmap; /* a 1 bit allows the operations in flags, 0 denies */ 4201 }; 4202 4203 #define KVM_MSR_FILTER_MAX_RANGES 16 4204 struct kvm_msr_filter { 4205 #define KVM_MSR_FILTER_DEFAULT_ALLOW (0 << 0) 4206 #define KVM_MSR_FILTER_DEFAULT_DENY (1 << 0) 4207 __u32 flags; 4208 struct kvm_msr_filter_range ranges[KVM_MSR_FILTER_MAX_RANGES]; 4209 }; 4210 4211flags values for ``struct kvm_msr_filter_range``: 4212 4213``KVM_MSR_FILTER_READ`` 4214 4215 Filter read accesses to MSRs using the given bitmap. A 0 in the bitmap 4216 indicates that read accesses should be denied, while a 1 indicates that 4217 a read for a particular MSR should be allowed regardless of the default 4218 filter action. 4219 4220``KVM_MSR_FILTER_WRITE`` 4221 4222 Filter write accesses to MSRs using the given bitmap. A 0 in the bitmap 4223 indicates that write accesses should be denied, while a 1 indicates that 4224 a write for a particular MSR should be allowed regardless of the default 4225 filter action. 4226 4227flags values for ``struct kvm_msr_filter``: 4228 4229``KVM_MSR_FILTER_DEFAULT_ALLOW`` 4230 4231 If no filter range matches an MSR index that is getting accessed, KVM will 4232 allow accesses to all MSRs by default. 4233 4234``KVM_MSR_FILTER_DEFAULT_DENY`` 4235 4236 If no filter range matches an MSR index that is getting accessed, KVM will 4237 deny accesses to all MSRs by default. 4238 4239This ioctl allows userspace to define up to 16 bitmaps of MSR ranges to deny 4240guest MSR accesses that would normally be allowed by KVM. If an MSR is not 4241covered by a specific range, the "default" filtering behavior applies. Each 4242bitmap range covers MSRs from [base .. base+nmsrs). 4243 4244If an MSR access is denied by userspace, the resulting KVM behavior depends on 4245whether or not KVM_CAP_X86_USER_SPACE_MSR's KVM_MSR_EXIT_REASON_FILTER is 4246enabled. If KVM_MSR_EXIT_REASON_FILTER is enabled, KVM will exit to userspace 4247on denied accesses, i.e. userspace effectively intercepts the MSR access. If 4248KVM_MSR_EXIT_REASON_FILTER is not enabled, KVM will inject a #GP into the guest 4249on denied accesses. Note, if an MSR access is denied during emulation of MSR 4250load/stores during VMX transitions, KVM ignores KVM_MSR_EXIT_REASON_FILTER. 4251See the below warning for full details. 4252 4253If an MSR access is allowed by userspace, KVM will emulate and/or virtualize 4254the access in accordance with the vCPU model. Note, KVM may still ultimately 4255inject a #GP if an access is allowed by userspace, e.g. if KVM doesn't support 4256the MSR, or to follow architectural behavior for the MSR. 4257 4258By default, KVM operates in KVM_MSR_FILTER_DEFAULT_ALLOW mode with no MSR range 4259filters. 4260 4261Calling this ioctl with an empty set of ranges (all nmsrs == 0) disables MSR 4262filtering. In that mode, ``KVM_MSR_FILTER_DEFAULT_DENY`` is invalid and causes 4263an error. 4264 4265.. warning:: 4266 MSR accesses that are side effects of instruction execution (emulated or 4267 native) are not filtered as hardware does not honor MSR bitmaps outside of 4268 RDMSR and WRMSR, and KVM mimics that behavior when emulating instructions 4269 to avoid pointless divergence from hardware. E.g. RDPID reads MSR_TSC_AUX, 4270 SYSENTER reads the SYSENTER MSRs, etc. 4271 4272 MSRs that are loaded/stored via dedicated VMCS fields are not filtered as 4273 part of VM-Enter/VM-Exit emulation. 4274 4275 MSRs that are loaded/store via VMX's load/store lists _are_ filtered as part 4276 of VM-Enter/VM-Exit emulation. If an MSR access is denied on VM-Enter, KVM 4277 synthesizes a consistency check VM-Exit(EXIT_REASON_MSR_LOAD_FAIL). If an 4278 MSR access is denied on VM-Exit, KVM synthesizes a VM-Abort. In short, KVM 4279 extends Intel's architectural list of MSRs that cannot be loaded/saved via 4280 the VM-Enter/VM-Exit MSR list. It is platform owner's responsibility to 4281 to communicate any such restrictions to their end users. 4282 4283 x2APIC MSR accesses cannot be filtered (KVM silently ignores filters that 4284 cover any x2APIC MSRs). 4285 4286Note, invoking this ioctl while a vCPU is running is inherently racy. However, 4287KVM does guarantee that vCPUs will see either the previous filter or the new 4288filter, e.g. MSRs with identical settings in both the old and new filter will 4289have deterministic behavior. 4290 4291Similarly, if userspace wishes to intercept on denied accesses, 4292KVM_MSR_EXIT_REASON_FILTER must be enabled before activating any filters, and 4293left enabled until after all filters are deactivated. Failure to do so may 4294result in KVM injecting a #GP instead of exiting to userspace. 4295 42964.98 KVM_CREATE_SPAPR_TCE_64 4297---------------------------- 4298 4299:Capability: KVM_CAP_SPAPR_TCE_64 4300:Architectures: powerpc 4301:Type: vm ioctl 4302:Parameters: struct kvm_create_spapr_tce_64 (in) 4303:Returns: file descriptor for manipulating the created TCE table 4304 4305This is an extension for KVM_CAP_SPAPR_TCE which only supports 32bit 4306windows, described in 4.62 KVM_CREATE_SPAPR_TCE 4307 4308This capability uses extended struct in ioctl interface:: 4309 4310 /* for KVM_CAP_SPAPR_TCE_64 */ 4311 struct kvm_create_spapr_tce_64 { 4312 __u64 liobn; 4313 __u32 page_shift; 4314 __u32 flags; 4315 __u64 offset; /* in pages */ 4316 __u64 size; /* in pages */ 4317 }; 4318 4319The aim of extension is to support an additional bigger DMA window with 4320a variable page size. 4321KVM_CREATE_SPAPR_TCE_64 receives a 64bit window size, an IOMMU page shift and 4322a bus offset of the corresponding DMA window, @size and @offset are numbers 4323of IOMMU pages. 4324 4325@flags are not used at the moment. 4326 4327The rest of functionality is identical to KVM_CREATE_SPAPR_TCE. 4328 43294.99 KVM_REINJECT_CONTROL 4330------------------------- 4331 4332:Capability: KVM_CAP_REINJECT_CONTROL 4333:Architectures: x86 4334:Type: vm ioctl 4335:Parameters: struct kvm_reinject_control (in) 4336:Returns: 0 on success, 4337 -EFAULT if struct kvm_reinject_control cannot be read, 4338 -ENXIO if KVM_CREATE_PIT or KVM_CREATE_PIT2 didn't succeed earlier. 4339 4340i8254 (PIT) has two modes, reinject and !reinject. The default is reinject, 4341where KVM queues elapsed i8254 ticks and monitors completion of interrupt from 4342vector(s) that i8254 injects. Reinject mode dequeues a tick and injects its 4343interrupt whenever there isn't a pending interrupt from i8254. 4344!reinject mode injects an interrupt as soon as a tick arrives. 4345 4346:: 4347 4348 struct kvm_reinject_control { 4349 __u8 pit_reinject; 4350 __u8 reserved[31]; 4351 }; 4352 4353pit_reinject = 0 (!reinject mode) is recommended, unless running an old 4354operating system that uses the PIT for timing (e.g. Linux 2.4.x). 4355 43564.100 KVM_PPC_CONFIGURE_V3_MMU 4357------------------------------ 4358 4359:Capability: KVM_CAP_PPC_MMU_RADIX or KVM_CAP_PPC_MMU_HASH_V3 4360:Architectures: ppc 4361:Type: vm ioctl 4362:Parameters: struct kvm_ppc_mmuv3_cfg (in) 4363:Returns: 0 on success, 4364 -EFAULT if struct kvm_ppc_mmuv3_cfg cannot be read, 4365 -EINVAL if the configuration is invalid 4366 4367This ioctl controls whether the guest will use radix or HPT (hashed 4368page table) translation, and sets the pointer to the process table for 4369the guest. 4370 4371:: 4372 4373 struct kvm_ppc_mmuv3_cfg { 4374 __u64 flags; 4375 __u64 process_table; 4376 }; 4377 4378There are two bits that can be set in flags; KVM_PPC_MMUV3_RADIX and 4379KVM_PPC_MMUV3_GTSE. KVM_PPC_MMUV3_RADIX, if set, configures the guest 4380to use radix tree translation, and if clear, to use HPT translation. 4381KVM_PPC_MMUV3_GTSE, if set and if KVM permits it, configures the guest 4382to be able to use the global TLB and SLB invalidation instructions; 4383if clear, the guest may not use these instructions. 4384 4385The process_table field specifies the address and size of the guest 4386process table, which is in the guest's space. This field is formatted 4387as the second doubleword of the partition table entry, as defined in 4388the Power ISA V3.00, Book III section 5.7.6.1. 4389 43904.101 KVM_PPC_GET_RMMU_INFO 4391--------------------------- 4392 4393:Capability: KVM_CAP_PPC_MMU_RADIX 4394:Architectures: ppc 4395:Type: vm ioctl 4396:Parameters: struct kvm_ppc_rmmu_info (out) 4397:Returns: 0 on success, 4398 -EFAULT if struct kvm_ppc_rmmu_info cannot be written, 4399 -EINVAL if no useful information can be returned 4400 4401This ioctl returns a structure containing two things: (a) a list 4402containing supported radix tree geometries, and (b) a list that maps 4403page sizes to put in the "AP" (actual page size) field for the tlbie 4404(TLB invalidate entry) instruction. 4405 4406:: 4407 4408 struct kvm_ppc_rmmu_info { 4409 struct kvm_ppc_radix_geom { 4410 __u8 page_shift; 4411 __u8 level_bits[4]; 4412 __u8 pad[3]; 4413 } geometries[8]; 4414 __u32 ap_encodings[8]; 4415 }; 4416 4417The geometries[] field gives up to 8 supported geometries for the 4418radix page table, in terms of the log base 2 of the smallest page 4419size, and the number of bits indexed at each level of the tree, from 4420the PTE level up to the PGD level in that order. Any unused entries 4421will have 0 in the page_shift field. 4422 4423The ap_encodings gives the supported page sizes and their AP field 4424encodings, encoded with the AP value in the top 3 bits and the log 4425base 2 of the page size in the bottom 6 bits. 4426 44274.102 KVM_PPC_RESIZE_HPT_PREPARE 4428-------------------------------- 4429 4430:Capability: KVM_CAP_SPAPR_RESIZE_HPT 4431:Architectures: powerpc 4432:Type: vm ioctl 4433:Parameters: struct kvm_ppc_resize_hpt (in) 4434:Returns: 0 on successful completion, 4435 >0 if a new HPT is being prepared, the value is an estimated 4436 number of milliseconds until preparation is complete, 4437 -EFAULT if struct kvm_reinject_control cannot be read, 4438 -EINVAL if the supplied shift or flags are invalid, 4439 -ENOMEM if unable to allocate the new HPT, 4440 4441Used to implement the PAPR extension for runtime resizing of a guest's 4442Hashed Page Table (HPT). Specifically this starts, stops or monitors 4443the preparation of a new potential HPT for the guest, essentially 4444implementing the H_RESIZE_HPT_PREPARE hypercall. 4445 4446:: 4447 4448 struct kvm_ppc_resize_hpt { 4449 __u64 flags; 4450 __u32 shift; 4451 __u32 pad; 4452 }; 4453 4454If called with shift > 0 when there is no pending HPT for the guest, 4455this begins preparation of a new pending HPT of size 2^(shift) bytes. 4456It then returns a positive integer with the estimated number of 4457milliseconds until preparation is complete. 4458 4459If called when there is a pending HPT whose size does not match that 4460requested in the parameters, discards the existing pending HPT and 4461creates a new one as above. 4462 4463If called when there is a pending HPT of the size requested, will: 4464 4465 * If preparation of the pending HPT is already complete, return 0 4466 * If preparation of the pending HPT has failed, return an error 4467 code, then discard the pending HPT. 4468 * If preparation of the pending HPT is still in progress, return an 4469 estimated number of milliseconds until preparation is complete. 4470 4471If called with shift == 0, discards any currently pending HPT and 4472returns 0 (i.e. cancels any in-progress preparation). 4473 4474flags is reserved for future expansion, currently setting any bits in 4475flags will result in an -EINVAL. 4476 4477Normally this will be called repeatedly with the same parameters until 4478it returns <= 0. The first call will initiate preparation, subsequent 4479ones will monitor preparation until it completes or fails. 4480 44814.103 KVM_PPC_RESIZE_HPT_COMMIT 4482------------------------------- 4483 4484:Capability: KVM_CAP_SPAPR_RESIZE_HPT 4485:Architectures: powerpc 4486:Type: vm ioctl 4487:Parameters: struct kvm_ppc_resize_hpt (in) 4488:Returns: 0 on successful completion, 4489 -EFAULT if struct kvm_reinject_control cannot be read, 4490 -EINVAL if the supplied shift or flags are invalid, 4491 -ENXIO is there is no pending HPT, or the pending HPT doesn't 4492 have the requested size, 4493 -EBUSY if the pending HPT is not fully prepared, 4494 -ENOSPC if there was a hash collision when moving existing 4495 HPT entries to the new HPT, 4496 -EIO on other error conditions 4497 4498Used to implement the PAPR extension for runtime resizing of a guest's 4499Hashed Page Table (HPT). Specifically this requests that the guest be 4500transferred to working with the new HPT, essentially implementing the 4501H_RESIZE_HPT_COMMIT hypercall. 4502 4503:: 4504 4505 struct kvm_ppc_resize_hpt { 4506 __u64 flags; 4507 __u32 shift; 4508 __u32 pad; 4509 }; 4510 4511This should only be called after KVM_PPC_RESIZE_HPT_PREPARE has 4512returned 0 with the same parameters. In other cases 4513KVM_PPC_RESIZE_HPT_COMMIT will return an error (usually -ENXIO or 4514-EBUSY, though others may be possible if the preparation was started, 4515but failed). 4516 4517This will have undefined effects on the guest if it has not already 4518placed itself in a quiescent state where no vcpu will make MMU enabled 4519memory accesses. 4520 4521On successful completion, the pending HPT will become the guest's active 4522HPT and the previous HPT will be discarded. 4523 4524On failure, the guest will still be operating on its previous HPT. 4525 45264.104 KVM_X86_GET_MCE_CAP_SUPPORTED 4527----------------------------------- 4528 4529:Capability: KVM_CAP_MCE 4530:Architectures: x86 4531:Type: system ioctl 4532:Parameters: u64 mce_cap (out) 4533:Returns: 0 on success, -1 on error 4534 4535Returns supported MCE capabilities. The u64 mce_cap parameter 4536has the same format as the MSR_IA32_MCG_CAP register. Supported 4537capabilities will have the corresponding bits set. 4538 45394.105 KVM_X86_SETUP_MCE 4540----------------------- 4541 4542:Capability: KVM_CAP_MCE 4543:Architectures: x86 4544:Type: vcpu ioctl 4545:Parameters: u64 mcg_cap (in) 4546:Returns: 0 on success, 4547 -EFAULT if u64 mcg_cap cannot be read, 4548 -EINVAL if the requested number of banks is invalid, 4549 -EINVAL if requested MCE capability is not supported. 4550 4551Initializes MCE support for use. The u64 mcg_cap parameter 4552has the same format as the MSR_IA32_MCG_CAP register and 4553specifies which capabilities should be enabled. The maximum 4554supported number of error-reporting banks can be retrieved when 4555checking for KVM_CAP_MCE. The supported capabilities can be 4556retrieved with KVM_X86_GET_MCE_CAP_SUPPORTED. 4557 45584.106 KVM_X86_SET_MCE 4559--------------------- 4560 4561:Capability: KVM_CAP_MCE 4562:Architectures: x86 4563:Type: vcpu ioctl 4564:Parameters: struct kvm_x86_mce (in) 4565:Returns: 0 on success, 4566 -EFAULT if struct kvm_x86_mce cannot be read, 4567 -EINVAL if the bank number is invalid, 4568 -EINVAL if VAL bit is not set in status field. 4569 4570Inject a machine check error (MCE) into the guest. The input 4571parameter is:: 4572 4573 struct kvm_x86_mce { 4574 __u64 status; 4575 __u64 addr; 4576 __u64 misc; 4577 __u64 mcg_status; 4578 __u8 bank; 4579 __u8 pad1[7]; 4580 __u64 pad2[3]; 4581 }; 4582 4583If the MCE being reported is an uncorrected error, KVM will 4584inject it as an MCE exception into the guest. If the guest 4585MCG_STATUS register reports that an MCE is in progress, KVM 4586causes an KVM_EXIT_SHUTDOWN vmexit. 4587 4588Otherwise, if the MCE is a corrected error, KVM will just 4589store it in the corresponding bank (provided this bank is 4590not holding a previously reported uncorrected error). 4591 45924.107 KVM_S390_GET_CMMA_BITS 4593---------------------------- 4594 4595:Capability: KVM_CAP_S390_CMMA_MIGRATION 4596:Architectures: s390 4597:Type: vm ioctl 4598:Parameters: struct kvm_s390_cmma_log (in, out) 4599:Returns: 0 on success, a negative value on error 4600 4601Errors: 4602 4603 ====== ============================================================= 4604 ENOMEM not enough memory can be allocated to complete the task 4605 ENXIO if CMMA is not enabled 4606 EINVAL if KVM_S390_CMMA_PEEK is not set but migration mode was not enabled 4607 EINVAL if KVM_S390_CMMA_PEEK is not set but dirty tracking has been 4608 disabled (and thus migration mode was automatically disabled) 4609 EFAULT if the userspace address is invalid or if no page table is 4610 present for the addresses (e.g. when using hugepages). 4611 ====== ============================================================= 4612 4613This ioctl is used to get the values of the CMMA bits on the s390 4614architecture. It is meant to be used in two scenarios: 4615 4616- During live migration to save the CMMA values. Live migration needs 4617 to be enabled via the KVM_REQ_START_MIGRATION VM property. 4618- To non-destructively peek at the CMMA values, with the flag 4619 KVM_S390_CMMA_PEEK set. 4620 4621The ioctl takes parameters via the kvm_s390_cmma_log struct. The desired 4622values are written to a buffer whose location is indicated via the "values" 4623member in the kvm_s390_cmma_log struct. The values in the input struct are 4624also updated as needed. 4625 4626Each CMMA value takes up one byte. 4627 4628:: 4629 4630 struct kvm_s390_cmma_log { 4631 __u64 start_gfn; 4632 __u32 count; 4633 __u32 flags; 4634 union { 4635 __u64 remaining; 4636 __u64 mask; 4637 }; 4638 __u64 values; 4639 }; 4640 4641start_gfn is the number of the first guest frame whose CMMA values are 4642to be retrieved, 4643 4644count is the length of the buffer in bytes, 4645 4646values points to the buffer where the result will be written to. 4647 4648If count is greater than KVM_S390_SKEYS_MAX, then it is considered to be 4649KVM_S390_SKEYS_MAX. KVM_S390_SKEYS_MAX is re-used for consistency with 4650other ioctls. 4651 4652The result is written in the buffer pointed to by the field values, and 4653the values of the input parameter are updated as follows. 4654 4655Depending on the flags, different actions are performed. The only 4656supported flag so far is KVM_S390_CMMA_PEEK. 4657 4658The default behaviour if KVM_S390_CMMA_PEEK is not set is: 4659start_gfn will indicate the first page frame whose CMMA bits were dirty. 4660It is not necessarily the same as the one passed as input, as clean pages 4661are skipped. 4662 4663count will indicate the number of bytes actually written in the buffer. 4664It can (and very often will) be smaller than the input value, since the 4665buffer is only filled until 16 bytes of clean values are found (which 4666are then not copied in the buffer). Since a CMMA migration block needs 4667the base address and the length, for a total of 16 bytes, we will send 4668back some clean data if there is some dirty data afterwards, as long as 4669the size of the clean data does not exceed the size of the header. This 4670allows to minimize the amount of data to be saved or transferred over 4671the network at the expense of more roundtrips to userspace. The next 4672invocation of the ioctl will skip over all the clean values, saving 4673potentially more than just the 16 bytes we found. 4674 4675If KVM_S390_CMMA_PEEK is set: 4676the existing storage attributes are read even when not in migration 4677mode, and no other action is performed; 4678 4679the output start_gfn will be equal to the input start_gfn, 4680 4681the output count will be equal to the input count, except if the end of 4682memory has been reached. 4683 4684In both cases: 4685the field "remaining" will indicate the total number of dirty CMMA values 4686still remaining, or 0 if KVM_S390_CMMA_PEEK is set and migration mode is 4687not enabled. 4688 4689mask is unused. 4690 4691values points to the userspace buffer where the result will be stored. 4692 46934.108 KVM_S390_SET_CMMA_BITS 4694---------------------------- 4695 4696:Capability: KVM_CAP_S390_CMMA_MIGRATION 4697:Architectures: s390 4698:Type: vm ioctl 4699:Parameters: struct kvm_s390_cmma_log (in) 4700:Returns: 0 on success, a negative value on error 4701 4702This ioctl is used to set the values of the CMMA bits on the s390 4703architecture. It is meant to be used during live migration to restore 4704the CMMA values, but there are no restrictions on its use. 4705The ioctl takes parameters via the kvm_s390_cmma_values struct. 4706Each CMMA value takes up one byte. 4707 4708:: 4709 4710 struct kvm_s390_cmma_log { 4711 __u64 start_gfn; 4712 __u32 count; 4713 __u32 flags; 4714 union { 4715 __u64 remaining; 4716 __u64 mask; 4717 }; 4718 __u64 values; 4719 }; 4720 4721start_gfn indicates the starting guest frame number, 4722 4723count indicates how many values are to be considered in the buffer, 4724 4725flags is not used and must be 0. 4726 4727mask indicates which PGSTE bits are to be considered. 4728 4729remaining is not used. 4730 4731values points to the buffer in userspace where to store the values. 4732 4733This ioctl can fail with -ENOMEM if not enough memory can be allocated to 4734complete the task, with -ENXIO if CMMA is not enabled, with -EINVAL if 4735the count field is too large (e.g. more than KVM_S390_CMMA_SIZE_MAX) or 4736if the flags field was not 0, with -EFAULT if the userspace address is 4737invalid, if invalid pages are written to (e.g. after the end of memory) 4738or if no page table is present for the addresses (e.g. when using 4739hugepages). 4740 47414.109 KVM_PPC_GET_CPU_CHAR 4742-------------------------- 4743 4744:Capability: KVM_CAP_PPC_GET_CPU_CHAR 4745:Architectures: powerpc 4746:Type: vm ioctl 4747:Parameters: struct kvm_ppc_cpu_char (out) 4748:Returns: 0 on successful completion, 4749 -EFAULT if struct kvm_ppc_cpu_char cannot be written 4750 4751This ioctl gives userspace information about certain characteristics 4752of the CPU relating to speculative execution of instructions and 4753possible information leakage resulting from speculative execution (see 4754CVE-2017-5715, CVE-2017-5753 and CVE-2017-5754). The information is 4755returned in struct kvm_ppc_cpu_char, which looks like this:: 4756 4757 struct kvm_ppc_cpu_char { 4758 __u64 character; /* characteristics of the CPU */ 4759 __u64 behaviour; /* recommended software behaviour */ 4760 __u64 character_mask; /* valid bits in character */ 4761 __u64 behaviour_mask; /* valid bits in behaviour */ 4762 }; 4763 4764For extensibility, the character_mask and behaviour_mask fields 4765indicate which bits of character and behaviour have been filled in by 4766the kernel. If the set of defined bits is extended in future then 4767userspace will be able to tell whether it is running on a kernel that 4768knows about the new bits. 4769 4770The character field describes attributes of the CPU which can help 4771with preventing inadvertent information disclosure - specifically, 4772whether there is an instruction to flash-invalidate the L1 data cache 4773(ori 30,30,0 or mtspr SPRN_TRIG2,rN), whether the L1 data cache is set 4774to a mode where entries can only be used by the thread that created 4775them, whether the bcctr[l] instruction prevents speculation, and 4776whether a speculation barrier instruction (ori 31,31,0) is provided. 4777 4778The behaviour field describes actions that software should take to 4779prevent inadvertent information disclosure, and thus describes which 4780vulnerabilities the hardware is subject to; specifically whether the 4781L1 data cache should be flushed when returning to user mode from the 4782kernel, and whether a speculation barrier should be placed between an 4783array bounds check and the array access. 4784 4785These fields use the same bit definitions as the new 4786H_GET_CPU_CHARACTERISTICS hypercall. 4787 47884.110 KVM_MEMORY_ENCRYPT_OP 4789--------------------------- 4790 4791:Capability: basic 4792:Architectures: x86 4793:Type: vm ioctl, vcpu ioctl 4794:Parameters: an opaque platform specific structure (in/out) 4795:Returns: 0 on success; -1 on error 4796 4797If the platform supports creating encrypted VMs then this ioctl can be used 4798for issuing platform-specific memory encryption commands to manage those 4799encrypted VMs. 4800 4801Currently, this ioctl is used for issuing both Secure Encrypted Virtualization 4802(SEV) commands on AMD Processors and Trusted Domain Extensions (TDX) commands 4803on Intel Processors. The detailed commands are defined in 4804Documentation/virt/kvm/x86/amd-memory-encryption.rst and 4805Documentation/virt/kvm/x86/intel-tdx.rst. 4806 48074.111 KVM_MEMORY_ENCRYPT_REG_REGION 4808----------------------------------- 4809 4810:Capability: basic 4811:Architectures: x86 4812:Type: system 4813:Parameters: struct kvm_enc_region (in) 4814:Returns: 0 on success; -1 on error 4815 4816This ioctl can be used to register a guest memory region which may 4817contain encrypted data (e.g. guest RAM, SMRAM etc). 4818 4819It is used in the SEV-enabled guest. When encryption is enabled, a guest 4820memory region may contain encrypted data. The SEV memory encryption 4821engine uses a tweak such that two identical plaintext pages, each at 4822different locations will have differing ciphertexts. So swapping or 4823moving ciphertext of those pages will not result in plaintext being 4824swapped. So relocating (or migrating) physical backing pages for the SEV 4825guest will require some additional steps. 4826 4827Note: The current SEV key management spec does not provide commands to 4828swap or migrate (move) ciphertext pages. Hence, for now we pin the guest 4829memory region registered with the ioctl. 4830 48314.112 KVM_MEMORY_ENCRYPT_UNREG_REGION 4832------------------------------------- 4833 4834:Capability: basic 4835:Architectures: x86 4836:Type: system 4837:Parameters: struct kvm_enc_region (in) 4838:Returns: 0 on success; -1 on error 4839 4840This ioctl can be used to unregister the guest memory region registered 4841with KVM_MEMORY_ENCRYPT_REG_REGION ioctl above. 4842 48434.113 KVM_HYPERV_EVENTFD 4844------------------------ 4845 4846:Capability: KVM_CAP_HYPERV_EVENTFD 4847:Architectures: x86 4848:Type: vm ioctl 4849:Parameters: struct kvm_hyperv_eventfd (in) 4850 4851This ioctl (un)registers an eventfd to receive notifications from the guest on 4852the specified Hyper-V connection id through the SIGNAL_EVENT hypercall, without 4853causing a user exit. SIGNAL_EVENT hypercall with non-zero event flag number 4854(bits 24-31) still triggers a KVM_EXIT_HYPERV_HCALL user exit. 4855 4856:: 4857 4858 struct kvm_hyperv_eventfd { 4859 __u32 conn_id; 4860 __s32 fd; 4861 __u32 flags; 4862 __u32 padding[3]; 4863 }; 4864 4865The conn_id field should fit within 24 bits:: 4866 4867 #define KVM_HYPERV_CONN_ID_MASK 0x00ffffff 4868 4869The acceptable values for the flags field are:: 4870 4871 #define KVM_HYPERV_EVENTFD_DEASSIGN (1 << 0) 4872 4873:Returns: 0 on success, 4874 -EINVAL if conn_id or flags is outside the allowed range, 4875 -ENOENT on deassign if the conn_id isn't registered, 4876 -EEXIST on assign if the conn_id is already registered 4877 48784.114 KVM_GET_NESTED_STATE 4879-------------------------- 4880 4881:Capability: KVM_CAP_NESTED_STATE 4882:Architectures: x86 4883:Type: vcpu ioctl 4884:Parameters: struct kvm_nested_state (in/out) 4885:Returns: 0 on success, -1 on error 4886 4887Errors: 4888 4889 ===== ============================================================= 4890 E2BIG the total state size exceeds the value of 'size' specified by 4891 the user; the size required will be written into size. 4892 ===== ============================================================= 4893 4894:: 4895 4896 struct kvm_nested_state { 4897 __u16 flags; 4898 __u16 format; 4899 __u32 size; 4900 4901 union { 4902 struct kvm_vmx_nested_state_hdr vmx; 4903 struct kvm_svm_nested_state_hdr svm; 4904 4905 /* Pad the header to 128 bytes. */ 4906 __u8 pad[120]; 4907 } hdr; 4908 4909 union { 4910 struct kvm_vmx_nested_state_data vmx[0]; 4911 struct kvm_svm_nested_state_data svm[0]; 4912 } data; 4913 }; 4914 4915 #define KVM_STATE_NESTED_GUEST_MODE 0x00000001 4916 #define KVM_STATE_NESTED_RUN_PENDING 0x00000002 4917 #define KVM_STATE_NESTED_EVMCS 0x00000004 4918 4919 #define KVM_STATE_NESTED_FORMAT_VMX 0 4920 #define KVM_STATE_NESTED_FORMAT_SVM 1 4921 4922 #define KVM_STATE_NESTED_VMX_VMCS_SIZE 0x1000 4923 4924 #define KVM_STATE_NESTED_VMX_SMM_GUEST_MODE 0x00000001 4925 #define KVM_STATE_NESTED_VMX_SMM_VMXON 0x00000002 4926 4927 #define KVM_STATE_VMX_PREEMPTION_TIMER_DEADLINE 0x00000001 4928 4929 struct kvm_vmx_nested_state_hdr { 4930 __u64 vmxon_pa; 4931 __u64 vmcs12_pa; 4932 4933 struct { 4934 __u16 flags; 4935 } smm; 4936 4937 __u32 flags; 4938 __u64 preemption_timer_deadline; 4939 }; 4940 4941 struct kvm_vmx_nested_state_data { 4942 __u8 vmcs12[KVM_STATE_NESTED_VMX_VMCS_SIZE]; 4943 __u8 shadow_vmcs12[KVM_STATE_NESTED_VMX_VMCS_SIZE]; 4944 }; 4945 4946This ioctl copies the vcpu's nested virtualization state from the kernel to 4947userspace. 4948 4949The maximum size of the state can be retrieved by passing KVM_CAP_NESTED_STATE 4950to the KVM_CHECK_EXTENSION ioctl(). 4951 49524.115 KVM_SET_NESTED_STATE 4953-------------------------- 4954 4955:Capability: KVM_CAP_NESTED_STATE 4956:Architectures: x86 4957:Type: vcpu ioctl 4958:Parameters: struct kvm_nested_state (in) 4959:Returns: 0 on success, -1 on error 4960 4961This copies the vcpu's kvm_nested_state struct from userspace to the kernel. 4962For the definition of struct kvm_nested_state, see KVM_GET_NESTED_STATE. 4963 49644.116 KVM_(UN)REGISTER_COALESCED_MMIO 4965------------------------------------- 4966 4967:Capability: KVM_CAP_COALESCED_MMIO (for coalesced mmio) 4968 KVM_CAP_COALESCED_PIO (for coalesced pio) 4969:Architectures: all 4970:Type: vm ioctl 4971:Parameters: struct kvm_coalesced_mmio_zone 4972:Returns: 0 on success, < 0 on error 4973 4974Coalesced I/O is a performance optimization that defers hardware 4975register write emulation so that userspace exits are avoided. It is 4976typically used to reduce the overhead of emulating frequently accessed 4977hardware registers. 4978 4979When a hardware register is configured for coalesced I/O, write accesses 4980do not exit to userspace and their value is recorded in a ring buffer 4981that is shared between kernel and userspace. 4982 4983Coalesced I/O is used if one or more write accesses to a hardware 4984register can be deferred until a read or a write to another hardware 4985register on the same device. This last access will cause a vmexit and 4986userspace will process accesses from the ring buffer before emulating 4987it. That will avoid exiting to userspace on repeated writes. 4988 4989Coalesced pio is based on coalesced mmio. There is little difference 4990between coalesced mmio and pio except that coalesced pio records accesses 4991to I/O ports. 4992 49934.117 KVM_CLEAR_DIRTY_LOG 4994------------------------- 4995 4996:Capability: KVM_CAP_MANUAL_DIRTY_LOG_PROTECT2 4997:Architectures: x86, arm64, mips 4998:Type: vm ioctl 4999:Parameters: struct kvm_clear_dirty_log (in) 5000:Returns: 0 on success, -1 on error 5001 5002:: 5003 5004 /* for KVM_CLEAR_DIRTY_LOG */ 5005 struct kvm_clear_dirty_log { 5006 __u32 slot; 5007 __u32 num_pages; 5008 __u64 first_page; 5009 union { 5010 void __user *dirty_bitmap; /* one bit per page */ 5011 __u64 padding; 5012 }; 5013 }; 5014 5015The ioctl clears the dirty status of pages in a memory slot, according to 5016the bitmap that is passed in struct kvm_clear_dirty_log's dirty_bitmap 5017field. Bit 0 of the bitmap corresponds to page "first_page" in the 5018memory slot, and num_pages is the size in bits of the input bitmap. 5019first_page must be a multiple of 64; num_pages must also be a multiple of 502064 unless first_page + num_pages is the size of the memory slot. For each 5021bit that is set in the input bitmap, the corresponding page is marked "clean" 5022in KVM's dirty bitmap, and dirty tracking is re-enabled for that page 5023(for example via write-protection, or by clearing the dirty bit in 5024a page table entry). 5025 5026If KVM_CAP_MULTI_ADDRESS_SPACE is available, bits 16-31 of slot field specifies 5027the address space for which you want to clear the dirty status. See 5028KVM_SET_USER_MEMORY_REGION for details on the usage of slot field. 5029 5030This ioctl is mostly useful when KVM_CAP_MANUAL_DIRTY_LOG_PROTECT2 5031is enabled; for more information, see the description of the capability. 5032However, it can always be used as long as KVM_CHECK_EXTENSION confirms 5033that KVM_CAP_MANUAL_DIRTY_LOG_PROTECT2 is present. 5034 50354.118 KVM_GET_SUPPORTED_HV_CPUID 5036-------------------------------- 5037 5038:Capability: KVM_CAP_HYPERV_CPUID (vcpu), KVM_CAP_SYS_HYPERV_CPUID (system) 5039:Architectures: x86 5040:Type: system ioctl, vcpu ioctl 5041:Parameters: struct kvm_cpuid2 (in/out) 5042:Returns: 0 on success, -1 on error 5043 5044:: 5045 5046 struct kvm_cpuid2 { 5047 __u32 nent; 5048 __u32 padding; 5049 struct kvm_cpuid_entry2 entries[0]; 5050 }; 5051 5052 struct kvm_cpuid_entry2 { 5053 __u32 function; 5054 __u32 index; 5055 __u32 flags; 5056 __u32 eax; 5057 __u32 ebx; 5058 __u32 ecx; 5059 __u32 edx; 5060 __u32 padding[3]; 5061 }; 5062 5063This ioctl returns x86 cpuid features leaves related to Hyper-V emulation in 5064KVM. Userspace can use the information returned by this ioctl to construct 5065cpuid information presented to guests consuming Hyper-V enlightenments (e.g. 5066Windows or Hyper-V guests). 5067 5068CPUID feature leaves returned by this ioctl are defined by Hyper-V Top Level 5069Functional Specification (TLFS). These leaves can't be obtained with 5070KVM_GET_SUPPORTED_CPUID ioctl because some of them intersect with KVM feature 5071leaves (0x40000000, 0x40000001). 5072 5073Currently, the following list of CPUID leaves are returned: 5074 5075 - HYPERV_CPUID_VENDOR_AND_MAX_FUNCTIONS 5076 - HYPERV_CPUID_INTERFACE 5077 - HYPERV_CPUID_VERSION 5078 - HYPERV_CPUID_FEATURES 5079 - HYPERV_CPUID_ENLIGHTMENT_INFO 5080 - HYPERV_CPUID_IMPLEMENT_LIMITS 5081 - HYPERV_CPUID_NESTED_FEATURES 5082 - HYPERV_CPUID_SYNDBG_VENDOR_AND_MAX_FUNCTIONS 5083 - HYPERV_CPUID_SYNDBG_INTERFACE 5084 - HYPERV_CPUID_SYNDBG_PLATFORM_CAPABILITIES 5085 5086Userspace invokes KVM_GET_SUPPORTED_HV_CPUID by passing a kvm_cpuid2 structure 5087with the 'nent' field indicating the number of entries in the variable-size 5088array 'entries'. If the number of entries is too low to describe all Hyper-V 5089feature leaves, an error (E2BIG) is returned. If the number is more or equal 5090to the number of Hyper-V feature leaves, the 'nent' field is adjusted to the 5091number of valid entries in the 'entries' array, which is then filled. 5092 5093'index' and 'flags' fields in 'struct kvm_cpuid_entry2' are currently reserved, 5094userspace should not expect to get any particular value there. 5095 5096Note, vcpu version of KVM_GET_SUPPORTED_HV_CPUID is currently deprecated. Unlike 5097system ioctl which exposes all supported feature bits unconditionally, vcpu 5098version has the following quirks: 5099 5100- HYPERV_CPUID_NESTED_FEATURES leaf and HV_X64_ENLIGHTENED_VMCS_RECOMMENDED 5101 feature bit are only exposed when Enlightened VMCS was previously enabled 5102 on the corresponding vCPU (KVM_CAP_HYPERV_ENLIGHTENED_VMCS). 5103- HV_STIMER_DIRECT_MODE_AVAILABLE bit is only exposed with in-kernel LAPIC. 5104 (presumes KVM_CREATE_IRQCHIP has already been called). 5105 51064.119 KVM_ARM_VCPU_FINALIZE 5107--------------------------- 5108 5109:Architectures: arm64 5110:Type: vcpu ioctl 5111:Parameters: int feature (in) 5112:Returns: 0 on success, -1 on error 5113 5114Errors: 5115 5116 ====== ============================================================== 5117 EPERM feature not enabled, needs configuration, or already finalized 5118 EINVAL feature unknown or not present 5119 ====== ============================================================== 5120 5121Recognised values for feature: 5122 5123 ===== =========================================== 5124 arm64 KVM_ARM_VCPU_SVE (requires KVM_CAP_ARM_SVE) 5125 ===== =========================================== 5126 5127Finalizes the configuration of the specified vcpu feature. 5128 5129The vcpu must already have been initialised, enabling the affected feature, by 5130means of a successful :ref:`KVM_ARM_VCPU_INIT <KVM_ARM_VCPU_INIT>` call with the 5131appropriate flag set in features[]. 5132 5133For affected vcpu features, this is a mandatory step that must be performed 5134before the vcpu is fully usable. 5135 5136Between KVM_ARM_VCPU_INIT and KVM_ARM_VCPU_FINALIZE, the feature may be 5137configured by use of ioctls such as KVM_SET_ONE_REG. The exact configuration 5138that should be performed and how to do it are feature-dependent. 5139 5140Other calls that depend on a particular feature being finalized, such as 5141KVM_RUN, KVM_GET_REG_LIST, KVM_GET_ONE_REG and KVM_SET_ONE_REG, will fail with 5142-EPERM unless the feature has already been finalized by means of a 5143KVM_ARM_VCPU_FINALIZE call. 5144 5145See KVM_ARM_VCPU_INIT for details of vcpu features that require finalization 5146using this ioctl. 5147 51484.120 KVM_SET_PMU_EVENT_FILTER 5149------------------------------ 5150 5151:Capability: KVM_CAP_PMU_EVENT_FILTER 5152:Architectures: x86 5153:Type: vm ioctl 5154:Parameters: struct kvm_pmu_event_filter (in) 5155:Returns: 0 on success, -1 on error 5156 5157Errors: 5158 5159 ====== ============================================================ 5160 EFAULT args[0] cannot be accessed 5161 EINVAL args[0] contains invalid data in the filter or filter events 5162 E2BIG nevents is too large 5163 EBUSY not enough memory to allocate the filter 5164 ====== ============================================================ 5165 5166:: 5167 5168 struct kvm_pmu_event_filter { 5169 __u32 action; 5170 __u32 nevents; 5171 __u32 fixed_counter_bitmap; 5172 __u32 flags; 5173 __u32 pad[4]; 5174 __u64 events[0]; 5175 }; 5176 5177This ioctl restricts the set of PMU events the guest can program by limiting 5178which event select and unit mask combinations are permitted. 5179 5180The argument holds a list of filter events which will be allowed or denied. 5181 5182Filter events only control general purpose counters; fixed purpose counters 5183are controlled by the fixed_counter_bitmap. 5184 5185Valid values for 'flags':: 5186 5187``0`` 5188 5189To use this mode, clear the 'flags' field. 5190 5191In this mode each event will contain an event select + unit mask. 5192 5193When the guest attempts to program the PMU the guest's event select + 5194unit mask is compared against the filter events to determine whether the 5195guest should have access. 5196 5197``KVM_PMU_EVENT_FLAG_MASKED_EVENTS`` 5198:Capability: KVM_CAP_PMU_EVENT_MASKED_EVENTS 5199 5200In this mode each filter event will contain an event select, mask, match, and 5201exclude value. To encode a masked event use:: 5202 5203 KVM_PMU_ENCODE_MASKED_ENTRY() 5204 5205An encoded event will follow this layout:: 5206 5207 Bits Description 5208 ---- ----------- 5209 7:0 event select (low bits) 5210 15:8 umask match 5211 31:16 unused 5212 35:32 event select (high bits) 5213 36:54 unused 5214 55 exclude bit 5215 63:56 umask mask 5216 5217When the guest attempts to program the PMU, these steps are followed in 5218determining if the guest should have access: 5219 5220 1. Match the event select from the guest against the filter events. 5221 2. If a match is found, match the guest's unit mask to the mask and match 5222 values of the included filter events. 5223 I.e. (unit mask & mask) == match && !exclude. 5224 3. If a match is found, match the guest's unit mask to the mask and match 5225 values of the excluded filter events. 5226 I.e. (unit mask & mask) == match && exclude. 5227 4. 5228 a. If an included match is found and an excluded match is not found, filter 5229 the event. 5230 b. For everything else, do not filter the event. 5231 5. 5232 a. If the event is filtered and it's an allow list, allow the guest to 5233 program the event. 5234 b. If the event is filtered and it's a deny list, do not allow the guest to 5235 program the event. 5236 5237When setting a new pmu event filter, -EINVAL will be returned if any of the 5238unused fields are set or if any of the high bits (35:32) in the event 5239select are set when called on Intel. 5240 5241Valid values for 'action':: 5242 5243 #define KVM_PMU_EVENT_ALLOW 0 5244 #define KVM_PMU_EVENT_DENY 1 5245 5246Via this API, KVM userspace can also control the behavior of the VM's fixed 5247counters (if any) by configuring the "action" and "fixed_counter_bitmap" fields. 5248 5249Specifically, KVM follows the following pseudo-code when determining whether to 5250allow the guest FixCtr[i] to count its pre-defined fixed event:: 5251 5252 FixCtr[i]_is_allowed = (action == ALLOW) && (bitmap & BIT(i)) || 5253 (action == DENY) && !(bitmap & BIT(i)); 5254 FixCtr[i]_is_denied = !FixCtr[i]_is_allowed; 5255 5256KVM always consumes fixed_counter_bitmap, it's userspace's responsibility to 5257ensure fixed_counter_bitmap is set correctly, e.g. if userspace wants to define 5258a filter that only affects general purpose counters. 5259 5260Note, the "events" field also applies to fixed counters' hardcoded event_select 5261and unit_mask values. "fixed_counter_bitmap" has higher priority than "events" 5262if there is a contradiction between the two. 5263 52644.121 KVM_PPC_SVM_OFF 5265--------------------- 5266 5267:Capability: basic 5268:Architectures: powerpc 5269:Type: vm ioctl 5270:Parameters: none 5271:Returns: 0 on successful completion, 5272 5273Errors: 5274 5275 ====== ================================================================ 5276 EINVAL if ultravisor failed to terminate the secure guest 5277 ENOMEM if hypervisor failed to allocate new radix page tables for guest 5278 ====== ================================================================ 5279 5280This ioctl is used to turn off the secure mode of the guest or transition 5281the guest from secure mode to normal mode. This is invoked when the guest 5282is reset. This has no effect if called for a normal guest. 5283 5284This ioctl issues an ultravisor call to terminate the secure guest, 5285unpins the VPA pages and releases all the device pages that are used to 5286track the secure pages by hypervisor. 5287 52884.122 KVM_S390_NORMAL_RESET 5289--------------------------- 5290 5291:Capability: KVM_CAP_S390_VCPU_RESETS 5292:Architectures: s390 5293:Type: vcpu ioctl 5294:Parameters: none 5295:Returns: 0 5296 5297This ioctl resets VCPU registers and control structures according to 5298the cpu reset definition in the POP (Principles Of Operation). 5299 53004.123 KVM_S390_INITIAL_RESET 5301---------------------------- 5302 5303:Capability: basic 5304:Architectures: s390 5305:Type: vcpu ioctl 5306:Parameters: none 5307:Returns: 0 5308 5309This ioctl resets VCPU registers and control structures according to 5310the initial cpu reset definition in the POP. However, the cpu is not 5311put into ESA mode. This reset is a superset of the normal reset. 5312 53134.124 KVM_S390_CLEAR_RESET 5314-------------------------- 5315 5316:Capability: KVM_CAP_S390_VCPU_RESETS 5317:Architectures: s390 5318:Type: vcpu ioctl 5319:Parameters: none 5320:Returns: 0 5321 5322This ioctl resets VCPU registers and control structures according to 5323the clear cpu reset definition in the POP. However, the cpu is not put 5324into ESA mode. This reset is a superset of the initial reset. 5325 5326 53274.125 KVM_S390_PV_COMMAND 5328------------------------- 5329 5330:Capability: KVM_CAP_S390_PROTECTED 5331:Architectures: s390 5332:Type: vm ioctl 5333:Parameters: struct kvm_pv_cmd 5334:Returns: 0 on success, < 0 on error 5335 5336:: 5337 5338 struct kvm_pv_cmd { 5339 __u32 cmd; /* Command to be executed */ 5340 __u16 rc; /* Ultravisor return code */ 5341 __u16 rrc; /* Ultravisor return reason code */ 5342 __u64 data; /* Data or address */ 5343 __u32 flags; /* flags for future extensions. Must be 0 for now */ 5344 __u32 reserved[3]; 5345 }; 5346 5347**Ultravisor return codes** 5348The Ultravisor return (reason) codes are provided by the kernel if a 5349Ultravisor call has been executed to achieve the results expected by 5350the command. Therefore they are independent of the IOCTL return 5351code. If KVM changes `rc`, its value will always be greater than 0 5352hence setting it to 0 before issuing a PV command is advised to be 5353able to detect a change of `rc`. 5354 5355**cmd values:** 5356 5357KVM_PV_ENABLE 5358 Allocate memory and register the VM with the Ultravisor, thereby 5359 donating memory to the Ultravisor that will become inaccessible to 5360 KVM. All existing CPUs are converted to protected ones. After this 5361 command has succeeded, any CPU added via hotplug will become 5362 protected during its creation as well. 5363 5364 Errors: 5365 5366 ===== ============================= 5367 EINTR an unmasked signal is pending 5368 ===== ============================= 5369 5370KVM_PV_DISABLE 5371 Deregister the VM from the Ultravisor and reclaim the memory that had 5372 been donated to the Ultravisor, making it usable by the kernel again. 5373 All registered VCPUs are converted back to non-protected ones. If a 5374 previous protected VM had been prepared for asynchronous teardown with 5375 KVM_PV_ASYNC_CLEANUP_PREPARE and not subsequently torn down with 5376 KVM_PV_ASYNC_CLEANUP_PERFORM, it will be torn down in this call 5377 together with the current protected VM. 5378 5379KVM_PV_VM_SET_SEC_PARMS 5380 Pass the image header from VM memory to the Ultravisor in 5381 preparation of image unpacking and verification. 5382 5383KVM_PV_VM_UNPACK 5384 Unpack (protect and decrypt) a page of the encrypted boot image. 5385 5386KVM_PV_VM_VERIFY 5387 Verify the integrity of the unpacked image. Only if this succeeds, 5388 KVM is allowed to start protected VCPUs. 5389 5390KVM_PV_INFO 5391 :Capability: KVM_CAP_S390_PROTECTED_DUMP 5392 5393 Presents an API that provides Ultravisor related data to userspace 5394 via subcommands. len_max is the size of the user space buffer, 5395 len_written is KVM's indication of how much bytes of that buffer 5396 were actually written to. len_written can be used to determine the 5397 valid fields if more response fields are added in the future. 5398 5399 :: 5400 5401 enum pv_cmd_info_id { 5402 KVM_PV_INFO_VM, 5403 KVM_PV_INFO_DUMP, 5404 }; 5405 5406 struct kvm_s390_pv_info_header { 5407 __u32 id; 5408 __u32 len_max; 5409 __u32 len_written; 5410 __u32 reserved; 5411 }; 5412 5413 struct kvm_s390_pv_info { 5414 struct kvm_s390_pv_info_header header; 5415 struct kvm_s390_pv_info_dump dump; 5416 struct kvm_s390_pv_info_vm vm; 5417 }; 5418 5419**subcommands:** 5420 5421 KVM_PV_INFO_VM 5422 This subcommand provides basic Ultravisor information for PV 5423 hosts. These values are likely also exported as files in the sysfs 5424 firmware UV query interface but they are more easily available to 5425 programs in this API. 5426 5427 The installed calls and feature_indication members provide the 5428 installed UV calls and the UV's other feature indications. 5429 5430 The max_* members provide information about the maximum number of PV 5431 vcpus, PV guests and PV guest memory size. 5432 5433 :: 5434 5435 struct kvm_s390_pv_info_vm { 5436 __u64 inst_calls_list[4]; 5437 __u64 max_cpus; 5438 __u64 max_guests; 5439 __u64 max_guest_addr; 5440 __u64 feature_indication; 5441 }; 5442 5443 5444 KVM_PV_INFO_DUMP 5445 This subcommand provides information related to dumping PV guests. 5446 5447 :: 5448 5449 struct kvm_s390_pv_info_dump { 5450 __u64 dump_cpu_buffer_len; 5451 __u64 dump_config_mem_buffer_per_1m; 5452 __u64 dump_config_finalize_len; 5453 }; 5454 5455KVM_PV_DUMP 5456 :Capability: KVM_CAP_S390_PROTECTED_DUMP 5457 5458 Presents an API that provides calls which facilitate dumping a 5459 protected VM. 5460 5461 :: 5462 5463 struct kvm_s390_pv_dmp { 5464 __u64 subcmd; 5465 __u64 buff_addr; 5466 __u64 buff_len; 5467 __u64 gaddr; /* For dump storage state */ 5468 }; 5469 5470 **subcommands:** 5471 5472 KVM_PV_DUMP_INIT 5473 Initializes the dump process of a protected VM. If this call does 5474 not succeed all other subcommands will fail with -EINVAL. This 5475 subcommand will return -EINVAL if a dump process has not yet been 5476 completed. 5477 5478 Not all PV vms can be dumped, the owner needs to set `dump 5479 allowed` PCF bit 34 in the SE header to allow dumping. 5480 5481 KVM_PV_DUMP_CONFIG_STOR_STATE 5482 Stores `buff_len` bytes of tweak component values starting with 5483 the 1MB block specified by the absolute guest address 5484 (`gaddr`). `buff_len` needs to be `conf_dump_storage_state_len` 5485 aligned and at least >= the `conf_dump_storage_state_len` value 5486 provided by the dump uv_info data. buff_user might be written to 5487 even if an error rc is returned. For instance if we encounter a 5488 fault after writing the first page of data. 5489 5490 KVM_PV_DUMP_COMPLETE 5491 If the subcommand succeeds it completes the dump process and lets 5492 KVM_PV_DUMP_INIT be called again. 5493 5494 On success `conf_dump_finalize_len` bytes of completion data will be 5495 stored to the `buff_addr`. The completion data contains a key 5496 derivation seed, IV, tweak nonce and encryption keys as well as an 5497 authentication tag all of which are needed to decrypt the dump at a 5498 later time. 5499 5500KVM_PV_ASYNC_CLEANUP_PREPARE 5501 :Capability: KVM_CAP_S390_PROTECTED_ASYNC_DISABLE 5502 5503 Prepare the current protected VM for asynchronous teardown. Most 5504 resources used by the current protected VM will be set aside for a 5505 subsequent asynchronous teardown. The current protected VM will then 5506 resume execution immediately as non-protected. There can be at most 5507 one protected VM prepared for asynchronous teardown at any time. If 5508 a protected VM had already been prepared for teardown without 5509 subsequently calling KVM_PV_ASYNC_CLEANUP_PERFORM, this call will 5510 fail. In that case, the userspace process should issue a normal 5511 KVM_PV_DISABLE. The resources set aside with this call will need to 5512 be cleaned up with a subsequent call to KVM_PV_ASYNC_CLEANUP_PERFORM 5513 or KVM_PV_DISABLE, otherwise they will be cleaned up when KVM 5514 terminates. KVM_PV_ASYNC_CLEANUP_PREPARE can be called again as soon 5515 as cleanup starts, i.e. before KVM_PV_ASYNC_CLEANUP_PERFORM finishes. 5516 5517KVM_PV_ASYNC_CLEANUP_PERFORM 5518 :Capability: KVM_CAP_S390_PROTECTED_ASYNC_DISABLE 5519 5520 Tear down the protected VM previously prepared for teardown with 5521 KVM_PV_ASYNC_CLEANUP_PREPARE. The resources that had been set aside 5522 will be freed during the execution of this command. This PV command 5523 should ideally be issued by userspace from a separate thread. If a 5524 fatal signal is received (or the process terminates naturally), the 5525 command will terminate immediately without completing, and the normal 5526 KVM shutdown procedure will take care of cleaning up all remaining 5527 protected VMs, including the ones whose teardown was interrupted by 5528 process termination. 5529 55304.126 KVM_XEN_HVM_SET_ATTR 5531-------------------------- 5532 5533:Capability: KVM_CAP_XEN_HVM / KVM_XEN_HVM_CONFIG_SHARED_INFO 5534:Architectures: x86 5535:Type: vm ioctl 5536:Parameters: struct kvm_xen_hvm_attr 5537:Returns: 0 on success, < 0 on error 5538 5539:: 5540 5541 struct kvm_xen_hvm_attr { 5542 __u16 type; 5543 __u16 pad[3]; 5544 union { 5545 __u8 long_mode; 5546 __u8 vector; 5547 __u8 runstate_update_flag; 5548 union { 5549 __u64 gfn; 5550 __u64 hva; 5551 } shared_info; 5552 struct { 5553 __u32 send_port; 5554 __u32 type; /* EVTCHNSTAT_ipi / EVTCHNSTAT_interdomain */ 5555 __u32 flags; 5556 union { 5557 struct { 5558 __u32 port; 5559 __u32 vcpu; 5560 __u32 priority; 5561 } port; 5562 struct { 5563 __u32 port; /* Zero for eventfd */ 5564 __s32 fd; 5565 } eventfd; 5566 __u32 padding[4]; 5567 } deliver; 5568 } evtchn; 5569 __u32 xen_version; 5570 __u64 pad[8]; 5571 } u; 5572 }; 5573 5574type values: 5575 5576KVM_XEN_ATTR_TYPE_LONG_MODE 5577 Sets the ABI mode of the VM to 32-bit or 64-bit (long mode). This 5578 determines the layout of the shared_info page exposed to the VM. 5579 5580KVM_XEN_ATTR_TYPE_SHARED_INFO 5581 Sets the guest physical frame number at which the Xen shared_info 5582 page resides. Note that although Xen places vcpu_info for the first 5583 32 vCPUs in the shared_info page, KVM does not automatically do so 5584 and instead requires that KVM_XEN_VCPU_ATTR_TYPE_VCPU_INFO or 5585 KVM_XEN_VCPU_ATTR_TYPE_VCPU_INFO_HVA be used explicitly even when 5586 the vcpu_info for a given vCPU resides at the "default" location 5587 in the shared_info page. This is because KVM may not be aware of 5588 the Xen CPU id which is used as the index into the vcpu_info[] 5589 array, so may know the correct default location. 5590 5591 Note that the shared_info page may be constantly written to by KVM; 5592 it contains the event channel bitmap used to deliver interrupts to 5593 a Xen guest, amongst other things. It is exempt from dirty tracking 5594 mechanisms — KVM will not explicitly mark the page as dirty each 5595 time an event channel interrupt is delivered to the guest! Thus, 5596 userspace should always assume that the designated GFN is dirty if 5597 any vCPU has been running or any event channel interrupts can be 5598 routed to the guest. 5599 5600 Setting the gfn to KVM_XEN_INVALID_GFN will disable the shared_info 5601 page. 5602 5603KVM_XEN_ATTR_TYPE_SHARED_INFO_HVA 5604 If the KVM_XEN_HVM_CONFIG_SHARED_INFO_HVA flag is also set in the 5605 Xen capabilities, then this attribute may be used to set the 5606 userspace address at which the shared_info page resides, which 5607 will always be fixed in the VMM regardless of where it is mapped 5608 in guest physical address space. This attribute should be used in 5609 preference to KVM_XEN_ATTR_TYPE_SHARED_INFO as it avoids 5610 unnecessary invalidation of an internal cache when the page is 5611 re-mapped in guest physical address space. 5612 5613 Setting the hva to zero will disable the shared_info page. 5614 5615KVM_XEN_ATTR_TYPE_UPCALL_VECTOR 5616 Sets the exception vector used to deliver Xen event channel upcalls. 5617 This is the HVM-wide vector injected directly by the hypervisor 5618 (not through the local APIC), typically configured by a guest via 5619 HVM_PARAM_CALLBACK_IRQ. This can be disabled again (e.g. for guest 5620 SHUTDOWN_soft_reset) by setting it to zero. 5621 5622KVM_XEN_ATTR_TYPE_EVTCHN 5623 This attribute is available when the KVM_CAP_XEN_HVM ioctl indicates 5624 support for KVM_XEN_HVM_CONFIG_EVTCHN_SEND features. It configures 5625 an outbound port number for interception of EVTCHNOP_send requests 5626 from the guest. A given sending port number may be directed back to 5627 a specified vCPU (by APIC ID) / port / priority on the guest, or to 5628 trigger events on an eventfd. The vCPU and priority can be changed 5629 by setting KVM_XEN_EVTCHN_UPDATE in a subsequent call, but other 5630 fields cannot change for a given sending port. A port mapping is 5631 removed by using KVM_XEN_EVTCHN_DEASSIGN in the flags field. Passing 5632 KVM_XEN_EVTCHN_RESET in the flags field removes all interception of 5633 outbound event channels. The values of the flags field are mutually 5634 exclusive and cannot be combined as a bitmask. 5635 5636KVM_XEN_ATTR_TYPE_XEN_VERSION 5637 This attribute is available when the KVM_CAP_XEN_HVM ioctl indicates 5638 support for KVM_XEN_HVM_CONFIG_EVTCHN_SEND features. It configures 5639 the 32-bit version code returned to the guest when it invokes the 5640 XENVER_version call; typically (XEN_MAJOR << 16 | XEN_MINOR). PV 5641 Xen guests will often use this to as a dummy hypercall to trigger 5642 event channel delivery, so responding within the kernel without 5643 exiting to userspace is beneficial. 5644 5645KVM_XEN_ATTR_TYPE_RUNSTATE_UPDATE_FLAG 5646 This attribute is available when the KVM_CAP_XEN_HVM ioctl indicates 5647 support for KVM_XEN_HVM_CONFIG_RUNSTATE_UPDATE_FLAG. It enables the 5648 XEN_RUNSTATE_UPDATE flag which allows guest vCPUs to safely read 5649 other vCPUs' vcpu_runstate_info. Xen guests enable this feature via 5650 the VMASST_TYPE_runstate_update_flag of the HYPERVISOR_vm_assist 5651 hypercall. 5652 56534.127 KVM_XEN_HVM_GET_ATTR 5654-------------------------- 5655 5656:Capability: KVM_CAP_XEN_HVM / KVM_XEN_HVM_CONFIG_SHARED_INFO 5657:Architectures: x86 5658:Type: vm ioctl 5659:Parameters: struct kvm_xen_hvm_attr 5660:Returns: 0 on success, < 0 on error 5661 5662Allows Xen VM attributes to be read. For the structure and types, 5663see KVM_XEN_HVM_SET_ATTR above. The KVM_XEN_ATTR_TYPE_EVTCHN 5664attribute cannot be read. 5665 56664.128 KVM_XEN_VCPU_SET_ATTR 5667--------------------------- 5668 5669:Capability: KVM_CAP_XEN_HVM / KVM_XEN_HVM_CONFIG_SHARED_INFO 5670:Architectures: x86 5671:Type: vcpu ioctl 5672:Parameters: struct kvm_xen_vcpu_attr 5673:Returns: 0 on success, < 0 on error 5674 5675:: 5676 5677 struct kvm_xen_vcpu_attr { 5678 __u16 type; 5679 __u16 pad[3]; 5680 union { 5681 __u64 gpa; 5682 __u64 pad[4]; 5683 struct { 5684 __u64 state; 5685 __u64 state_entry_time; 5686 __u64 time_running; 5687 __u64 time_runnable; 5688 __u64 time_blocked; 5689 __u64 time_offline; 5690 } runstate; 5691 __u32 vcpu_id; 5692 struct { 5693 __u32 port; 5694 __u32 priority; 5695 __u64 expires_ns; 5696 } timer; 5697 __u8 vector; 5698 } u; 5699 }; 5700 5701type values: 5702 5703KVM_XEN_VCPU_ATTR_TYPE_VCPU_INFO 5704 Sets the guest physical address of the vcpu_info for a given vCPU. 5705 As with the shared_info page for the VM, the corresponding page may be 5706 dirtied at any time if event channel interrupt delivery is enabled, so 5707 userspace should always assume that the page is dirty without relying 5708 on dirty logging. Setting the gpa to KVM_XEN_INVALID_GPA will disable 5709 the vcpu_info. 5710 5711KVM_XEN_VCPU_ATTR_TYPE_VCPU_INFO_HVA 5712 If the KVM_XEN_HVM_CONFIG_SHARED_INFO_HVA flag is also set in the 5713 Xen capabilities, then this attribute may be used to set the 5714 userspace address of the vcpu_info for a given vCPU. It should 5715 only be used when the vcpu_info resides at the "default" location 5716 in the shared_info page. In this case it is safe to assume the 5717 userspace address will not change, because the shared_info page is 5718 an overlay on guest memory and remains at a fixed host address 5719 regardless of where it is mapped in guest physical address space 5720 and hence unnecessary invalidation of an internal cache may be 5721 avoided if the guest memory layout is modified. 5722 If the vcpu_info does not reside at the "default" location then 5723 it is not guaranteed to remain at the same host address and 5724 hence the aforementioned cache invalidation is required. 5725 5726KVM_XEN_VCPU_ATTR_TYPE_VCPU_TIME_INFO 5727 Sets the guest physical address of an additional pvclock structure 5728 for a given vCPU. This is typically used for guest vsyscall support. 5729 Setting the gpa to KVM_XEN_INVALID_GPA will disable the structure. 5730 5731KVM_XEN_VCPU_ATTR_TYPE_RUNSTATE_ADDR 5732 Sets the guest physical address of the vcpu_runstate_info for a given 5733 vCPU. This is how a Xen guest tracks CPU state such as steal time. 5734 Setting the gpa to KVM_XEN_INVALID_GPA will disable the runstate area. 5735 5736KVM_XEN_VCPU_ATTR_TYPE_RUNSTATE_CURRENT 5737 Sets the runstate (RUNSTATE_running/_runnable/_blocked/_offline) of 5738 the given vCPU from the .u.runstate.state member of the structure. 5739 KVM automatically accounts running and runnable time but blocked 5740 and offline states are only entered explicitly. 5741 5742KVM_XEN_VCPU_ATTR_TYPE_RUNSTATE_DATA 5743 Sets all fields of the vCPU runstate data from the .u.runstate member 5744 of the structure, including the current runstate. The state_entry_time 5745 must equal the sum of the other four times. 5746 5747KVM_XEN_VCPU_ATTR_TYPE_RUNSTATE_ADJUST 5748 This *adds* the contents of the .u.runstate members of the structure 5749 to the corresponding members of the given vCPU's runstate data, thus 5750 permitting atomic adjustments to the runstate times. The adjustment 5751 to the state_entry_time must equal the sum of the adjustments to the 5752 other four times. The state field must be set to -1, or to a valid 5753 runstate value (RUNSTATE_running, RUNSTATE_runnable, RUNSTATE_blocked 5754 or RUNSTATE_offline) to set the current accounted state as of the 5755 adjusted state_entry_time. 5756 5757KVM_XEN_VCPU_ATTR_TYPE_VCPU_ID 5758 This attribute is available when the KVM_CAP_XEN_HVM ioctl indicates 5759 support for KVM_XEN_HVM_CONFIG_EVTCHN_SEND features. It sets the Xen 5760 vCPU ID of the given vCPU, to allow timer-related VCPU operations to 5761 be intercepted by KVM. 5762 5763KVM_XEN_VCPU_ATTR_TYPE_TIMER 5764 This attribute is available when the KVM_CAP_XEN_HVM ioctl indicates 5765 support for KVM_XEN_HVM_CONFIG_EVTCHN_SEND features. It sets the 5766 event channel port/priority for the VIRQ_TIMER of the vCPU, as well 5767 as allowing a pending timer to be saved/restored. Setting the timer 5768 port to zero disables kernel handling of the singleshot timer. 5769 5770KVM_XEN_VCPU_ATTR_TYPE_UPCALL_VECTOR 5771 This attribute is available when the KVM_CAP_XEN_HVM ioctl indicates 5772 support for KVM_XEN_HVM_CONFIG_EVTCHN_SEND features. It sets the 5773 per-vCPU local APIC upcall vector, configured by a Xen guest with 5774 the HVMOP_set_evtchn_upcall_vector hypercall. This is typically 5775 used by Windows guests, and is distinct from the HVM-wide upcall 5776 vector configured with HVM_PARAM_CALLBACK_IRQ. It is disabled by 5777 setting the vector to zero. 5778 5779 57804.129 KVM_XEN_VCPU_GET_ATTR 5781--------------------------- 5782 5783:Capability: KVM_CAP_XEN_HVM / KVM_XEN_HVM_CONFIG_SHARED_INFO 5784:Architectures: x86 5785:Type: vcpu ioctl 5786:Parameters: struct kvm_xen_vcpu_attr 5787:Returns: 0 on success, < 0 on error 5788 5789Allows Xen vCPU attributes to be read. For the structure and types, 5790see KVM_XEN_VCPU_SET_ATTR above. 5791 5792The KVM_XEN_VCPU_ATTR_TYPE_RUNSTATE_ADJUST type may not be used 5793with the KVM_XEN_VCPU_GET_ATTR ioctl. 5794 57954.130 KVM_ARM_MTE_COPY_TAGS 5796--------------------------- 5797 5798:Capability: KVM_CAP_ARM_MTE 5799:Architectures: arm64 5800:Type: vm ioctl 5801:Parameters: struct kvm_arm_copy_mte_tags 5802:Returns: number of bytes copied, < 0 on error (-EINVAL for incorrect 5803 arguments, -EFAULT if memory cannot be accessed). 5804 5805:: 5806 5807 struct kvm_arm_copy_mte_tags { 5808 __u64 guest_ipa; 5809 __u64 length; 5810 void __user *addr; 5811 __u64 flags; 5812 __u64 reserved[2]; 5813 }; 5814 5815Copies Memory Tagging Extension (MTE) tags to/from guest tag memory. The 5816``guest_ipa`` and ``length`` fields must be ``PAGE_SIZE`` aligned. 5817``length`` must not be bigger than 2^31 - PAGE_SIZE bytes. The ``addr`` 5818field must point to a buffer which the tags will be copied to or from. 5819 5820``flags`` specifies the direction of copy, either ``KVM_ARM_TAGS_TO_GUEST`` or 5821``KVM_ARM_TAGS_FROM_GUEST``. 5822 5823The size of the buffer to store the tags is ``(length / 16)`` bytes 5824(granules in MTE are 16 bytes long). Each byte contains a single tag 5825value. This matches the format of ``PTRACE_PEEKMTETAGS`` and 5826``PTRACE_POKEMTETAGS``. 5827 5828If an error occurs before any data is copied then a negative error code is 5829returned. If some tags have been copied before an error occurs then the number 5830of bytes successfully copied is returned. If the call completes successfully 5831then ``length`` is returned. 5832 58334.131 KVM_GET_SREGS2 5834-------------------- 5835 5836:Capability: KVM_CAP_SREGS2 5837:Architectures: x86 5838:Type: vcpu ioctl 5839:Parameters: struct kvm_sregs2 (out) 5840:Returns: 0 on success, -1 on error 5841 5842Reads special registers from the vcpu. 5843This ioctl (when supported) replaces the KVM_GET_SREGS. 5844 5845:: 5846 5847 struct kvm_sregs2 { 5848 /* out (KVM_GET_SREGS2) / in (KVM_SET_SREGS2) */ 5849 struct kvm_segment cs, ds, es, fs, gs, ss; 5850 struct kvm_segment tr, ldt; 5851 struct kvm_dtable gdt, idt; 5852 __u64 cr0, cr2, cr3, cr4, cr8; 5853 __u64 efer; 5854 __u64 apic_base; 5855 __u64 flags; 5856 __u64 pdptrs[4]; 5857 }; 5858 5859flags values for ``kvm_sregs2``: 5860 5861``KVM_SREGS2_FLAGS_PDPTRS_VALID`` 5862 5863 Indicates that the struct contains valid PDPTR values. 5864 5865 58664.132 KVM_SET_SREGS2 5867-------------------- 5868 5869:Capability: KVM_CAP_SREGS2 5870:Architectures: x86 5871:Type: vcpu ioctl 5872:Parameters: struct kvm_sregs2 (in) 5873:Returns: 0 on success, -1 on error 5874 5875Writes special registers into the vcpu. 5876See KVM_GET_SREGS2 for the data structures. 5877This ioctl (when supported) replaces the KVM_SET_SREGS. 5878 58794.133 KVM_GET_STATS_FD 5880---------------------- 5881 5882:Capability: KVM_CAP_STATS_BINARY_FD 5883:Architectures: all 5884:Type: vm ioctl, vcpu ioctl 5885:Parameters: none 5886:Returns: statistics file descriptor on success, < 0 on error 5887 5888Errors: 5889 5890 ====== ====================================================== 5891 ENOMEM if the fd could not be created due to lack of memory 5892 EMFILE if the number of opened files exceeds the limit 5893 ====== ====================================================== 5894 5895The returned file descriptor can be used to read VM/vCPU statistics data in 5896binary format. The data in the file descriptor consists of four blocks 5897organized as follows: 5898 5899+-------------+ 5900| Header | 5901+-------------+ 5902| id string | 5903+-------------+ 5904| Descriptors | 5905+-------------+ 5906| Stats Data | 5907+-------------+ 5908 5909Apart from the header starting at offset 0, please be aware that it is 5910not guaranteed that the four blocks are adjacent or in the above order; 5911the offsets of the id, descriptors and data blocks are found in the 5912header. However, all four blocks are aligned to 64 bit offsets in the 5913file and they do not overlap. 5914 5915All blocks except the data block are immutable. Userspace can read them 5916only one time after retrieving the file descriptor, and then use ``pread`` or 5917``lseek`` to read the statistics repeatedly. 5918 5919All data is in system endianness. 5920 5921The format of the header is as follows:: 5922 5923 struct kvm_stats_header { 5924 __u32 flags; 5925 __u32 name_size; 5926 __u32 num_desc; 5927 __u32 id_offset; 5928 __u32 desc_offset; 5929 __u32 data_offset; 5930 }; 5931 5932The ``flags`` field is not used at the moment. It is always read as 0. 5933 5934The ``name_size`` field is the size (in byte) of the statistics name string 5935(including trailing '\0') which is contained in the "id string" block and 5936appended at the end of every descriptor. 5937 5938The ``num_desc`` field is the number of descriptors that are included in the 5939descriptor block. (The actual number of values in the data block may be 5940larger, since each descriptor may comprise more than one value). 5941 5942The ``id_offset`` field is the offset of the id string from the start of the 5943file indicated by the file descriptor. It is a multiple of 8. 5944 5945The ``desc_offset`` field is the offset of the Descriptors block from the start 5946of the file indicated by the file descriptor. It is a multiple of 8. 5947 5948The ``data_offset`` field is the offset of the Stats Data block from the start 5949of the file indicated by the file descriptor. It is a multiple of 8. 5950 5951The id string block contains a string which identifies the file descriptor on 5952which KVM_GET_STATS_FD was invoked. The size of the block, including the 5953trailing ``'\0'``, is indicated by the ``name_size`` field in the header. 5954 5955The descriptors block is only needed to be read once for the lifetime of the 5956file descriptor contains a sequence of ``struct kvm_stats_desc``, each followed 5957by a string of size ``name_size``. 5958:: 5959 5960 #define KVM_STATS_TYPE_SHIFT 0 5961 #define KVM_STATS_TYPE_MASK (0xF << KVM_STATS_TYPE_SHIFT) 5962 #define KVM_STATS_TYPE_CUMULATIVE (0x0 << KVM_STATS_TYPE_SHIFT) 5963 #define KVM_STATS_TYPE_INSTANT (0x1 << KVM_STATS_TYPE_SHIFT) 5964 #define KVM_STATS_TYPE_PEAK (0x2 << KVM_STATS_TYPE_SHIFT) 5965 #define KVM_STATS_TYPE_LINEAR_HIST (0x3 << KVM_STATS_TYPE_SHIFT) 5966 #define KVM_STATS_TYPE_LOG_HIST (0x4 << KVM_STATS_TYPE_SHIFT) 5967 #define KVM_STATS_TYPE_MAX KVM_STATS_TYPE_LOG_HIST 5968 5969 #define KVM_STATS_UNIT_SHIFT 4 5970 #define KVM_STATS_UNIT_MASK (0xF << KVM_STATS_UNIT_SHIFT) 5971 #define KVM_STATS_UNIT_NONE (0x0 << KVM_STATS_UNIT_SHIFT) 5972 #define KVM_STATS_UNIT_BYTES (0x1 << KVM_STATS_UNIT_SHIFT) 5973 #define KVM_STATS_UNIT_SECONDS (0x2 << KVM_STATS_UNIT_SHIFT) 5974 #define KVM_STATS_UNIT_CYCLES (0x3 << KVM_STATS_UNIT_SHIFT) 5975 #define KVM_STATS_UNIT_BOOLEAN (0x4 << KVM_STATS_UNIT_SHIFT) 5976 #define KVM_STATS_UNIT_MAX KVM_STATS_UNIT_BOOLEAN 5977 5978 #define KVM_STATS_BASE_SHIFT 8 5979 #define KVM_STATS_BASE_MASK (0xF << KVM_STATS_BASE_SHIFT) 5980 #define KVM_STATS_BASE_POW10 (0x0 << KVM_STATS_BASE_SHIFT) 5981 #define KVM_STATS_BASE_POW2 (0x1 << KVM_STATS_BASE_SHIFT) 5982 #define KVM_STATS_BASE_MAX KVM_STATS_BASE_POW2 5983 5984 struct kvm_stats_desc { 5985 __u32 flags; 5986 __s16 exponent; 5987 __u16 size; 5988 __u32 offset; 5989 __u32 bucket_size; 5990 char name[]; 5991 }; 5992 5993The ``flags`` field contains the type and unit of the statistics data described 5994by this descriptor. Its endianness is CPU native. 5995The following flags are supported: 5996 5997Bits 0-3 of ``flags`` encode the type: 5998 5999 * ``KVM_STATS_TYPE_CUMULATIVE`` 6000 The statistics reports a cumulative count. The value of data can only be increased. 6001 Most of the counters used in KVM are of this type. 6002 The corresponding ``size`` field for this type is always 1. 6003 All cumulative statistics data are read/write. 6004 * ``KVM_STATS_TYPE_INSTANT`` 6005 The statistics reports an instantaneous value. Its value can be increased or 6006 decreased. This type is usually used as a measurement of some resources, 6007 like the number of dirty pages, the number of large pages, etc. 6008 All instant statistics are read only. 6009 The corresponding ``size`` field for this type is always 1. 6010 * ``KVM_STATS_TYPE_PEAK`` 6011 The statistics data reports a peak value, for example the maximum number 6012 of items in a hash table bucket, the longest time waited and so on. 6013 The value of data can only be increased. 6014 The corresponding ``size`` field for this type is always 1. 6015 * ``KVM_STATS_TYPE_LINEAR_HIST`` 6016 The statistic is reported as a linear histogram. The number of 6017 buckets is specified by the ``size`` field. The size of buckets is specified 6018 by the ``hist_param`` field. The range of the Nth bucket (1 <= N < ``size``) 6019 is [``hist_param``*(N-1), ``hist_param``*N), while the range of the last 6020 bucket is [``hist_param``*(``size``-1), +INF). (+INF means positive infinity 6021 value.) 6022 * ``KVM_STATS_TYPE_LOG_HIST`` 6023 The statistic is reported as a logarithmic histogram. The number of 6024 buckets is specified by the ``size`` field. The range of the first bucket is 6025 [0, 1), while the range of the last bucket is [pow(2, ``size``-2), +INF). 6026 Otherwise, The Nth bucket (1 < N < ``size``) covers 6027 [pow(2, N-2), pow(2, N-1)). 6028 6029Bits 4-7 of ``flags`` encode the unit: 6030 6031 * ``KVM_STATS_UNIT_NONE`` 6032 There is no unit for the value of statistics data. This usually means that 6033 the value is a simple counter of an event. 6034 * ``KVM_STATS_UNIT_BYTES`` 6035 It indicates that the statistics data is used to measure memory size, in the 6036 unit of Byte, KiByte, MiByte, GiByte, etc. The unit of the data is 6037 determined by the ``exponent`` field in the descriptor. 6038 * ``KVM_STATS_UNIT_SECONDS`` 6039 It indicates that the statistics data is used to measure time or latency. 6040 * ``KVM_STATS_UNIT_CYCLES`` 6041 It indicates that the statistics data is used to measure CPU clock cycles. 6042 * ``KVM_STATS_UNIT_BOOLEAN`` 6043 It indicates that the statistic will always be either 0 or 1. Boolean 6044 statistics of "peak" type will never go back from 1 to 0. Boolean 6045 statistics can be linear histograms (with two buckets) but not logarithmic 6046 histograms. 6047 6048Note that, in the case of histograms, the unit applies to the bucket 6049ranges, while the bucket value indicates how many samples fell in the 6050bucket's range. 6051 6052Bits 8-11 of ``flags``, together with ``exponent``, encode the scale of the 6053unit: 6054 6055 * ``KVM_STATS_BASE_POW10`` 6056 The scale is based on power of 10. It is used for measurement of time and 6057 CPU clock cycles. For example, an exponent of -9 can be used with 6058 ``KVM_STATS_UNIT_SECONDS`` to express that the unit is nanoseconds. 6059 * ``KVM_STATS_BASE_POW2`` 6060 The scale is based on power of 2. It is used for measurement of memory size. 6061 For example, an exponent of 20 can be used with ``KVM_STATS_UNIT_BYTES`` to 6062 express that the unit is MiB. 6063 6064The ``size`` field is the number of values of this statistics data. Its 6065value is usually 1 for most of simple statistics. 1 means it contains an 6066unsigned 64bit data. 6067 6068The ``offset`` field is the offset from the start of Data Block to the start of 6069the corresponding statistics data. 6070 6071The ``bucket_size`` field is used as a parameter for histogram statistics data. 6072It is only used by linear histogram statistics data, specifying the size of a 6073bucket in the unit expressed by bits 4-11 of ``flags`` together with ``exponent``. 6074 6075The ``name`` field is the name string of the statistics data. The name string 6076starts at the end of ``struct kvm_stats_desc``. The maximum length including 6077the trailing ``'\0'``, is indicated by ``name_size`` in the header. 6078 6079The Stats Data block contains an array of 64-bit values in the same order 6080as the descriptors in Descriptors block. 6081 60824.134 KVM_GET_XSAVE2 6083-------------------- 6084 6085:Capability: KVM_CAP_XSAVE2 6086:Architectures: x86 6087:Type: vcpu ioctl 6088:Parameters: struct kvm_xsave (out) 6089:Returns: 0 on success, -1 on error 6090 6091 6092:: 6093 6094 struct kvm_xsave { 6095 __u32 region[1024]; 6096 __u32 extra[0]; 6097 }; 6098 6099This ioctl would copy current vcpu's xsave struct to the userspace. It 6100copies as many bytes as are returned by KVM_CHECK_EXTENSION(KVM_CAP_XSAVE2) 6101when invoked on the vm file descriptor. The size value returned by 6102KVM_CHECK_EXTENSION(KVM_CAP_XSAVE2) will always be at least 4096. 6103Currently, it is only greater than 4096 if a dynamic feature has been 6104enabled with ``arch_prctl()``, but this may change in the future. 6105 6106The offsets of the state save areas in struct kvm_xsave follow the contents 6107of CPUID leaf 0xD on the host. 6108 61094.135 KVM_XEN_HVM_EVTCHN_SEND 6110----------------------------- 6111 6112:Capability: KVM_CAP_XEN_HVM / KVM_XEN_HVM_CONFIG_EVTCHN_SEND 6113:Architectures: x86 6114:Type: vm ioctl 6115:Parameters: struct kvm_irq_routing_xen_evtchn 6116:Returns: 0 on success, < 0 on error 6117 6118 6119:: 6120 6121 struct kvm_irq_routing_xen_evtchn { 6122 __u32 port; 6123 __u32 vcpu; 6124 __u32 priority; 6125 }; 6126 6127This ioctl injects an event channel interrupt directly to the guest vCPU. 6128 61294.136 KVM_S390_PV_CPU_COMMAND 6130----------------------------- 6131 6132:Capability: KVM_CAP_S390_PROTECTED_DUMP 6133:Architectures: s390 6134:Type: vcpu ioctl 6135:Parameters: none 6136:Returns: 0 on success, < 0 on error 6137 6138This ioctl closely mirrors `KVM_S390_PV_COMMAND` but handles requests 6139for vcpus. It re-uses the kvm_s390_pv_dmp struct and hence also shares 6140the command ids. 6141 6142**command:** 6143 6144KVM_PV_DUMP 6145 Presents an API that provides calls which facilitate dumping a vcpu 6146 of a protected VM. 6147 6148**subcommand:** 6149 6150KVM_PV_DUMP_CPU 6151 Provides encrypted dump data like register values. 6152 The length of the returned data is provided by uv_info.guest_cpu_stor_len. 6153 61544.137 KVM_S390_ZPCI_OP 6155---------------------- 6156 6157:Capability: KVM_CAP_S390_ZPCI_OP 6158:Architectures: s390 6159:Type: vm ioctl 6160:Parameters: struct kvm_s390_zpci_op (in) 6161:Returns: 0 on success, <0 on error 6162 6163Used to manage hardware-assisted virtualization features for zPCI devices. 6164 6165Parameters are specified via the following structure:: 6166 6167 struct kvm_s390_zpci_op { 6168 /* in */ 6169 __u32 fh; /* target device */ 6170 __u8 op; /* operation to perform */ 6171 __u8 pad[3]; 6172 union { 6173 /* for KVM_S390_ZPCIOP_REG_AEN */ 6174 struct { 6175 __u64 ibv; /* Guest addr of interrupt bit vector */ 6176 __u64 sb; /* Guest addr of summary bit */ 6177 __u32 flags; 6178 __u32 noi; /* Number of interrupts */ 6179 __u8 isc; /* Guest interrupt subclass */ 6180 __u8 sbo; /* Offset of guest summary bit vector */ 6181 __u16 pad; 6182 } reg_aen; 6183 __u64 reserved[8]; 6184 } u; 6185 }; 6186 6187The type of operation is specified in the "op" field. 6188KVM_S390_ZPCIOP_REG_AEN is used to register the VM for adapter event 6189notification interpretation, which will allow firmware delivery of adapter 6190events directly to the vm, with KVM providing a backup delivery mechanism; 6191KVM_S390_ZPCIOP_DEREG_AEN is used to subsequently disable interpretation of 6192adapter event notifications. 6193 6194The target zPCI function must also be specified via the "fh" field. For the 6195KVM_S390_ZPCIOP_REG_AEN operation, additional information to establish firmware 6196delivery must be provided via the "reg_aen" struct. 6197 6198The "pad" and "reserved" fields may be used for future extensions and should be 6199set to 0s by userspace. 6200 62014.138 KVM_ARM_SET_COUNTER_OFFSET 6202-------------------------------- 6203 6204:Capability: KVM_CAP_COUNTER_OFFSET 6205:Architectures: arm64 6206:Type: vm ioctl 6207:Parameters: struct kvm_arm_counter_offset (in) 6208:Returns: 0 on success, < 0 on error 6209 6210This capability indicates that userspace is able to apply a single VM-wide 6211offset to both the virtual and physical counters as viewed by the guest 6212using the KVM_ARM_SET_CNT_OFFSET ioctl and the following data structure: 6213 6214:: 6215 6216 struct kvm_arm_counter_offset { 6217 __u64 counter_offset; 6218 __u64 reserved; 6219 }; 6220 6221The offset describes a number of counter cycles that are subtracted from 6222both virtual and physical counter views (similar to the effects of the 6223CNTVOFF_EL2 and CNTPOFF_EL2 system registers, but only global). The offset 6224always applies to all vcpus (already created or created after this ioctl) 6225for this VM. 6226 6227It is userspace's responsibility to compute the offset based, for example, 6228on previous values of the guest counters. 6229 6230Any value other than 0 for the "reserved" field may result in an error 6231(-EINVAL) being returned. This ioctl can also return -EBUSY if any vcpu 6232ioctl is issued concurrently. 6233 6234Note that using this ioctl results in KVM ignoring subsequent userspace 6235writes to the CNTVCT_EL0 and CNTPCT_EL0 registers using the SET_ONE_REG 6236interface. No error will be returned, but the resulting offset will not be 6237applied. 6238 6239.. _KVM_ARM_GET_REG_WRITABLE_MASKS: 6240 62414.139 KVM_ARM_GET_REG_WRITABLE_MASKS 6242------------------------------------ 6243 6244:Capability: KVM_CAP_ARM_SUPPORTED_REG_MASK_RANGES 6245:Architectures: arm64 6246:Type: vm ioctl 6247:Parameters: struct reg_mask_range (in/out) 6248:Returns: 0 on success, < 0 on error 6249 6250 6251:: 6252 6253 #define KVM_ARM_FEATURE_ID_RANGE 0 6254 #define KVM_ARM_FEATURE_ID_RANGE_SIZE (3 * 8 * 8) 6255 6256 struct reg_mask_range { 6257 __u64 addr; /* Pointer to mask array */ 6258 __u32 range; /* Requested range */ 6259 __u32 reserved[13]; 6260 }; 6261 6262This ioctl copies the writable masks for a selected range of registers to 6263userspace. 6264 6265The ``addr`` field is a pointer to the destination array where KVM copies 6266the writable masks. 6267 6268The ``range`` field indicates the requested range of registers. 6269``KVM_CHECK_EXTENSION`` for the ``KVM_CAP_ARM_SUPPORTED_REG_MASK_RANGES`` 6270capability returns the supported ranges, expressed as a set of flags. Each 6271flag's bit index represents a possible value for the ``range`` field. 6272All other values are reserved for future use and KVM may return an error. 6273 6274The ``reserved[13]`` array is reserved for future use and should be 0, or 6275KVM may return an error. 6276 6277KVM_ARM_FEATURE_ID_RANGE (0) 6278^^^^^^^^^^^^^^^^^^^^^^^^^^^^ 6279 6280The Feature ID range is defined as the AArch64 System register space with 6281op0==3, op1=={0, 1, 3}, CRn==0, CRm=={0-7}, op2=={0-7}. 6282 6283The mask returned array pointed to by ``addr`` is indexed by the macro 6284``ARM64_FEATURE_ID_RANGE_IDX(op0, op1, crn, crm, op2)``, allowing userspace 6285to know what fields can be changed for the system register described by 6286``op0, op1, crn, crm, op2``. KVM rejects ID register values that describe a 6287superset of the features supported by the system. 6288 62894.140 KVM_SET_USER_MEMORY_REGION2 6290--------------------------------- 6291 6292:Capability: KVM_CAP_USER_MEMORY2 6293:Architectures: all 6294:Type: vm ioctl 6295:Parameters: struct kvm_userspace_memory_region2 (in) 6296:Returns: 0 on success, -1 on error 6297 6298KVM_SET_USER_MEMORY_REGION2 is an extension to KVM_SET_USER_MEMORY_REGION that 6299allows mapping guest_memfd memory into a guest. All fields shared with 6300KVM_SET_USER_MEMORY_REGION identically. Userspace can set KVM_MEM_GUEST_MEMFD 6301in flags to have KVM bind the memory region to a given guest_memfd range of 6302[guest_memfd_offset, guest_memfd_offset + memory_size]. The target guest_memfd 6303must point at a file created via KVM_CREATE_GUEST_MEMFD on the current VM, and 6304the target range must not be bound to any other memory region. All standard 6305bounds checks apply (use common sense). 6306 6307:: 6308 6309 struct kvm_userspace_memory_region2 { 6310 __u32 slot; 6311 __u32 flags; 6312 __u64 guest_phys_addr; 6313 __u64 memory_size; /* bytes */ 6314 __u64 userspace_addr; /* start of the userspace allocated memory */ 6315 __u64 guest_memfd_offset; 6316 __u32 guest_memfd; 6317 __u32 pad1; 6318 __u64 pad2[14]; 6319 }; 6320 6321A KVM_MEM_GUEST_MEMFD region _must_ have a valid guest_memfd (private memory) and 6322userspace_addr (shared memory). However, "valid" for userspace_addr simply 6323means that the address itself must be a legal userspace address. The backing 6324mapping for userspace_addr is not required to be valid/populated at the time of 6325KVM_SET_USER_MEMORY_REGION2, e.g. shared memory can be lazily mapped/allocated 6326on-demand. 6327 6328When mapping a gfn into the guest, KVM selects shared vs. private, i.e consumes 6329userspace_addr vs. guest_memfd, based on the gfn's KVM_MEMORY_ATTRIBUTE_PRIVATE 6330state. At VM creation time, all memory is shared, i.e. the PRIVATE attribute 6331is '0' for all gfns. Userspace can control whether memory is shared/private by 6332toggling KVM_MEMORY_ATTRIBUTE_PRIVATE via KVM_SET_MEMORY_ATTRIBUTES as needed. 6333 6334S390: 6335^^^^^ 6336 6337Returns -EINVAL if the VM has the KVM_VM_S390_UCONTROL flag set. 6338Returns -EINVAL if called on a protected VM. 6339 63404.141 KVM_SET_MEMORY_ATTRIBUTES 6341------------------------------- 6342 6343:Capability: KVM_CAP_MEMORY_ATTRIBUTES 6344:Architectures: x86 6345:Type: vm ioctl 6346:Parameters: struct kvm_memory_attributes (in) 6347:Returns: 0 on success, <0 on error 6348 6349KVM_SET_MEMORY_ATTRIBUTES allows userspace to set memory attributes for a range 6350of guest physical memory. 6351 6352:: 6353 6354 struct kvm_memory_attributes { 6355 __u64 address; 6356 __u64 size; 6357 __u64 attributes; 6358 __u64 flags; 6359 }; 6360 6361 #define KVM_MEMORY_ATTRIBUTE_PRIVATE (1ULL << 3) 6362 6363The address and size must be page aligned. The supported attributes can be 6364retrieved via ioctl(KVM_CHECK_EXTENSION) on KVM_CAP_MEMORY_ATTRIBUTES. If 6365executed on a VM, KVM_CAP_MEMORY_ATTRIBUTES precisely returns the attributes 6366supported by that VM. If executed at system scope, KVM_CAP_MEMORY_ATTRIBUTES 6367returns all attributes supported by KVM. The only attribute defined at this 6368time is KVM_MEMORY_ATTRIBUTE_PRIVATE, which marks the associated gfn as being 6369guest private memory. 6370 6371Note, there is no "get" API. Userspace is responsible for explicitly tracking 6372the state of a gfn/page as needed. 6373 6374The "flags" field is reserved for future extensions and must be '0'. 6375 63764.142 KVM_CREATE_GUEST_MEMFD 6377---------------------------- 6378 6379:Capability: KVM_CAP_GUEST_MEMFD 6380:Architectures: none 6381:Type: vm ioctl 6382:Parameters: struct kvm_create_guest_memfd(in) 6383:Returns: A file descriptor on success, <0 on error 6384 6385KVM_CREATE_GUEST_MEMFD creates an anonymous file and returns a file descriptor 6386that refers to it. guest_memfd files are roughly analogous to files created 6387via memfd_create(), e.g. guest_memfd files live in RAM, have volatile storage, 6388and are automatically released when the last reference is dropped. Unlike 6389"regular" memfd_create() files, guest_memfd files are bound to their owning 6390virtual machine (see below), cannot be mapped, read, or written by userspace, 6391and cannot be resized (guest_memfd files do however support PUNCH_HOLE). 6392 6393:: 6394 6395 struct kvm_create_guest_memfd { 6396 __u64 size; 6397 __u64 flags; 6398 __u64 reserved[6]; 6399 }; 6400 6401Conceptually, the inode backing a guest_memfd file represents physical memory, 6402i.e. is coupled to the virtual machine as a thing, not to a "struct kvm". The 6403file itself, which is bound to a "struct kvm", is that instance's view of the 6404underlying memory, e.g. effectively provides the translation of guest addresses 6405to host memory. This allows for use cases where multiple KVM structures are 6406used to manage a single virtual machine, e.g. when performing intrahost 6407migration of a virtual machine. 6408 6409KVM currently only supports mapping guest_memfd via KVM_SET_USER_MEMORY_REGION2, 6410and more specifically via the guest_memfd and guest_memfd_offset fields in 6411"struct kvm_userspace_memory_region2", where guest_memfd_offset is the offset 6412into the guest_memfd instance. For a given guest_memfd file, there can be at 6413most one mapping per page, i.e. binding multiple memory regions to a single 6414guest_memfd range is not allowed (any number of memory regions can be bound to 6415a single guest_memfd file, but the bound ranges must not overlap). 6416 6417See KVM_SET_USER_MEMORY_REGION2 for additional details. 6418 64194.143 KVM_PRE_FAULT_MEMORY 6420--------------------------- 6421 6422:Capability: KVM_CAP_PRE_FAULT_MEMORY 6423:Architectures: none 6424:Type: vcpu ioctl 6425:Parameters: struct kvm_pre_fault_memory (in/out) 6426:Returns: 0 if at least one page is processed, < 0 on error 6427 6428Errors: 6429 6430 ========== =============================================================== 6431 EINVAL The specified `gpa` and `size` were invalid (e.g. not 6432 page aligned, causes an overflow, or size is zero). 6433 ENOENT The specified `gpa` is outside defined memslots. 6434 EINTR An unmasked signal is pending and no page was processed. 6435 EFAULT The parameter address was invalid. 6436 EOPNOTSUPP Mapping memory for a GPA is unsupported by the 6437 hypervisor, and/or for the current vCPU state/mode. 6438 EIO unexpected error conditions (also causes a WARN) 6439 ========== =============================================================== 6440 6441:: 6442 6443 struct kvm_pre_fault_memory { 6444 /* in/out */ 6445 __u64 gpa; 6446 __u64 size; 6447 /* in */ 6448 __u64 flags; 6449 __u64 padding[5]; 6450 }; 6451 6452KVM_PRE_FAULT_MEMORY populates KVM's stage-2 page tables used to map memory 6453for the current vCPU state. KVM maps memory as if the vCPU generated a 6454stage-2 read page fault, e.g. faults in memory as needed, but doesn't break 6455CoW. However, KVM does not mark any newly created stage-2 PTE as Accessed. 6456 6457In the case of confidential VM types where there is an initial set up of 6458private guest memory before the guest is 'finalized'/measured, this ioctl 6459should only be issued after completing all the necessary setup to put the 6460guest into a 'finalized' state so that the above semantics can be reliably 6461ensured. 6462 6463In some cases, multiple vCPUs might share the page tables. In this 6464case, the ioctl can be called in parallel. 6465 6466When the ioctl returns, the input values are updated to point to the 6467remaining range. If `size` > 0 on return, the caller can just issue 6468the ioctl again with the same `struct kvm_map_memory` argument. 6469 6470Shadow page tables cannot support this ioctl because they 6471are indexed by virtual address or nested guest physical address. 6472Calling this ioctl when the guest is using shadow page tables (for 6473example because it is running a nested guest with nested page tables) 6474will fail with `EOPNOTSUPP` even if `KVM_CHECK_EXTENSION` reports 6475the capability to be present. 6476 6477`flags` must currently be zero. 6478 6479 6480.. _kvm_run: 6481 64825. The kvm_run structure 6483======================== 6484 6485Application code obtains a pointer to the kvm_run structure by 6486mmap()ing a vcpu fd. From that point, application code can control 6487execution by changing fields in kvm_run prior to calling the KVM_RUN 6488ioctl, and obtain information about the reason KVM_RUN returned by 6489looking up structure members. 6490 6491:: 6492 6493 struct kvm_run { 6494 /* in */ 6495 __u8 request_interrupt_window; 6496 6497Request that KVM_RUN return when it becomes possible to inject external 6498interrupts into the guest. Useful in conjunction with KVM_INTERRUPT. 6499 6500:: 6501 6502 __u8 immediate_exit; 6503 6504This field is polled once when KVM_RUN starts; if non-zero, KVM_RUN 6505exits immediately, returning -EINTR. In the common scenario where a 6506signal is used to "kick" a VCPU out of KVM_RUN, this field can be used 6507to avoid usage of KVM_SET_SIGNAL_MASK, which has worse scalability. 6508Rather than blocking the signal outside KVM_RUN, userspace can set up 6509a signal handler that sets run->immediate_exit to a non-zero value. 6510 6511This field is ignored if KVM_CAP_IMMEDIATE_EXIT is not available. 6512 6513:: 6514 6515 __u8 padding1[6]; 6516 6517 /* out */ 6518 __u32 exit_reason; 6519 6520When KVM_RUN has returned successfully (return value 0), this informs 6521application code why KVM_RUN has returned. Allowable values for this 6522field are detailed below. 6523 6524:: 6525 6526 __u8 ready_for_interrupt_injection; 6527 6528If request_interrupt_window has been specified, this field indicates 6529an interrupt can be injected now with KVM_INTERRUPT. 6530 6531:: 6532 6533 __u8 if_flag; 6534 6535The value of the current interrupt flag. Only valid if in-kernel 6536local APIC is not used. 6537 6538:: 6539 6540 __u16 flags; 6541 6542More architecture-specific flags detailing state of the VCPU that may 6543affect the device's behavior. Current defined flags:: 6544 6545 /* x86, set if the VCPU is in system management mode */ 6546 #define KVM_RUN_X86_SMM (1 << 0) 6547 /* x86, set if bus lock detected in VM */ 6548 #define KVM_RUN_X86_BUS_LOCK (1 << 1) 6549 /* x86, set if the VCPU is executing a nested (L2) guest */ 6550 #define KVM_RUN_X86_GUEST_MODE (1 << 2) 6551 6552 /* arm64, set for KVM_EXIT_DEBUG */ 6553 #define KVM_DEBUG_ARCH_HSR_HIGH_VALID (1 << 0) 6554 6555:: 6556 6557 /* in (pre_kvm_run), out (post_kvm_run) */ 6558 __u64 cr8; 6559 6560The value of the cr8 register. Only valid if in-kernel local APIC is 6561not used. Both input and output. 6562 6563:: 6564 6565 __u64 apic_base; 6566 6567The value of the APIC BASE msr. Only valid if in-kernel local 6568APIC is not used. Both input and output. 6569 6570:: 6571 6572 union { 6573 /* KVM_EXIT_UNKNOWN */ 6574 struct { 6575 __u64 hardware_exit_reason; 6576 } hw; 6577 6578If exit_reason is KVM_EXIT_UNKNOWN, the vcpu has exited due to unknown 6579reasons. Further architecture-specific information is available in 6580hardware_exit_reason. 6581 6582:: 6583 6584 /* KVM_EXIT_FAIL_ENTRY */ 6585 struct { 6586 __u64 hardware_entry_failure_reason; 6587 __u32 cpu; /* if KVM_LAST_CPU */ 6588 } fail_entry; 6589 6590If exit_reason is KVM_EXIT_FAIL_ENTRY, the vcpu could not be run due 6591to unknown reasons. Further architecture-specific information is 6592available in hardware_entry_failure_reason. 6593 6594:: 6595 6596 /* KVM_EXIT_EXCEPTION */ 6597 struct { 6598 __u32 exception; 6599 __u32 error_code; 6600 } ex; 6601 6602Unused. 6603 6604:: 6605 6606 /* KVM_EXIT_IO */ 6607 struct { 6608 #define KVM_EXIT_IO_IN 0 6609 #define KVM_EXIT_IO_OUT 1 6610 __u8 direction; 6611 __u8 size; /* bytes */ 6612 __u16 port; 6613 __u32 count; 6614 __u64 data_offset; /* relative to kvm_run start */ 6615 } io; 6616 6617If exit_reason is KVM_EXIT_IO, then the vcpu has 6618executed a port I/O instruction which could not be satisfied by kvm. 6619data_offset describes where the data is located (KVM_EXIT_IO_OUT) or 6620where kvm expects application code to place the data for the next 6621KVM_RUN invocation (KVM_EXIT_IO_IN). Data format is a packed array. 6622 6623:: 6624 6625 /* KVM_EXIT_DEBUG */ 6626 struct { 6627 struct kvm_debug_exit_arch arch; 6628 } debug; 6629 6630If the exit_reason is KVM_EXIT_DEBUG, then a vcpu is processing a debug event 6631for which architecture specific information is returned. 6632 6633:: 6634 6635 /* KVM_EXIT_MMIO */ 6636 struct { 6637 __u64 phys_addr; 6638 __u8 data[8]; 6639 __u32 len; 6640 __u8 is_write; 6641 } mmio; 6642 6643If exit_reason is KVM_EXIT_MMIO, then the vcpu has 6644executed a memory-mapped I/O instruction which could not be satisfied 6645by kvm. The 'data' member contains the written data if 'is_write' is 6646true, and should be filled by application code otherwise. 6647 6648The 'data' member contains, in its first 'len' bytes, the value as it would 6649appear if the VCPU performed a load or store of the appropriate width directly 6650to the byte array. 6651 6652.. note:: 6653 6654 For KVM_EXIT_IO, KVM_EXIT_MMIO, KVM_EXIT_OSI, KVM_EXIT_PAPR, KVM_EXIT_XEN, 6655 KVM_EXIT_EPR, KVM_EXIT_HYPERCALL, KVM_EXIT_TDX, 6656 KVM_EXIT_X86_RDMSR and KVM_EXIT_X86_WRMSR the corresponding 6657 operations are complete (and guest state is consistent) only after userspace 6658 has re-entered the kernel with KVM_RUN. The kernel side will first finish 6659 incomplete operations and then check for pending signals. 6660 6661 The pending state of the operation is not preserved in state which is 6662 visible to userspace, thus userspace should ensure that the operation is 6663 completed before performing a live migration. Userspace can re-enter the 6664 guest with an unmasked signal pending or with the immediate_exit field set 6665 to complete pending operations without allowing any further instructions 6666 to be executed. 6667 6668:: 6669 6670 /* KVM_EXIT_HYPERCALL */ 6671 struct { 6672 __u64 nr; 6673 __u64 args[6]; 6674 __u64 ret; 6675 __u64 flags; 6676 } hypercall; 6677 6678 6679It is strongly recommended that userspace use ``KVM_EXIT_IO`` (x86) or 6680``KVM_EXIT_MMIO`` (all except s390) to implement functionality that 6681requires a guest to interact with host userspace. 6682 6683.. note:: KVM_EXIT_IO is significantly faster than KVM_EXIT_MMIO. 6684 6685For arm64: 6686---------- 6687 6688SMCCC exits can be enabled depending on the configuration of the SMCCC 6689filter. See the Documentation/virt/kvm/devices/vm.rst 6690``KVM_ARM_SMCCC_FILTER`` for more details. 6691 6692``nr`` contains the function ID of the guest's SMCCC call. Userspace is 6693expected to use the ``KVM_GET_ONE_REG`` ioctl to retrieve the call 6694parameters from the vCPU's GPRs. 6695 6696Definition of ``flags``: 6697 - ``KVM_HYPERCALL_EXIT_SMC``: Indicates that the guest used the SMC 6698 conduit to initiate the SMCCC call. If this bit is 0 then the guest 6699 used the HVC conduit for the SMCCC call. 6700 6701 - ``KVM_HYPERCALL_EXIT_16BIT``: Indicates that the guest used a 16bit 6702 instruction to initiate the SMCCC call. If this bit is 0 then the 6703 guest used a 32bit instruction. An AArch64 guest always has this 6704 bit set to 0. 6705 6706At the point of exit, PC points to the instruction immediately following 6707the trapping instruction. 6708 6709:: 6710 6711 /* KVM_EXIT_TPR_ACCESS */ 6712 struct { 6713 __u64 rip; 6714 __u32 is_write; 6715 __u32 pad; 6716 } tpr_access; 6717 6718To be documented (KVM_TPR_ACCESS_REPORTING). 6719 6720:: 6721 6722 /* KVM_EXIT_S390_SIEIC */ 6723 struct { 6724 __u8 icptcode; 6725 __u64 mask; /* psw upper half */ 6726 __u64 addr; /* psw lower half */ 6727 __u16 ipa; 6728 __u32 ipb; 6729 } s390_sieic; 6730 6731s390 specific. 6732 6733:: 6734 6735 /* KVM_EXIT_S390_RESET */ 6736 #define KVM_S390_RESET_POR 1 6737 #define KVM_S390_RESET_CLEAR 2 6738 #define KVM_S390_RESET_SUBSYSTEM 4 6739 #define KVM_S390_RESET_CPU_INIT 8 6740 #define KVM_S390_RESET_IPL 16 6741 __u64 s390_reset_flags; 6742 6743s390 specific. 6744 6745:: 6746 6747 /* KVM_EXIT_S390_UCONTROL */ 6748 struct { 6749 __u64 trans_exc_code; 6750 __u32 pgm_code; 6751 } s390_ucontrol; 6752 6753s390 specific. A page fault has occurred for a user controlled virtual 6754machine (KVM_VM_S390_UNCONTROL) on its host page table that cannot be 6755resolved by the kernel. 6756The program code and the translation exception code that were placed 6757in the cpu's lowcore are presented here as defined by the z Architecture 6758Principles of Operation Book in the Chapter for Dynamic Address Translation 6759(DAT) 6760 6761:: 6762 6763 /* KVM_EXIT_DCR */ 6764 struct { 6765 __u32 dcrn; 6766 __u32 data; 6767 __u8 is_write; 6768 } dcr; 6769 6770Deprecated - was used for 440 KVM. 6771 6772:: 6773 6774 /* KVM_EXIT_OSI */ 6775 struct { 6776 __u64 gprs[32]; 6777 } osi; 6778 6779MOL uses a special hypercall interface it calls 'OSI'. To enable it, we catch 6780hypercalls and exit with this exit struct that contains all the guest gprs. 6781 6782If exit_reason is KVM_EXIT_OSI, then the vcpu has triggered such a hypercall. 6783Userspace can now handle the hypercall and when it's done modify the gprs as 6784necessary. Upon guest entry all guest GPRs will then be replaced by the values 6785in this struct. 6786 6787:: 6788 6789 /* KVM_EXIT_PAPR_HCALL */ 6790 struct { 6791 __u64 nr; 6792 __u64 ret; 6793 __u64 args[9]; 6794 } papr_hcall; 6795 6796This is used on 64-bit PowerPC when emulating a pSeries partition, 6797e.g. with the 'pseries' machine type in qemu. It occurs when the 6798guest does a hypercall using the 'sc 1' instruction. The 'nr' field 6799contains the hypercall number (from the guest R3), and 'args' contains 6800the arguments (from the guest R4 - R12). Userspace should put the 6801return code in 'ret' and any extra returned values in args[]. 6802The possible hypercalls are defined in the Power Architecture Platform 6803Requirements (PAPR) document available from www.power.org (free 6804developer registration required to access it). 6805 6806:: 6807 6808 /* KVM_EXIT_S390_TSCH */ 6809 struct { 6810 __u16 subchannel_id; 6811 __u16 subchannel_nr; 6812 __u32 io_int_parm; 6813 __u32 io_int_word; 6814 __u32 ipb; 6815 __u8 dequeued; 6816 } s390_tsch; 6817 6818s390 specific. This exit occurs when KVM_CAP_S390_CSS_SUPPORT has been enabled 6819and TEST SUBCHANNEL was intercepted. If dequeued is set, a pending I/O 6820interrupt for the target subchannel has been dequeued and subchannel_id, 6821subchannel_nr, io_int_parm and io_int_word contain the parameters for that 6822interrupt. ipb is needed for instruction parameter decoding. 6823 6824:: 6825 6826 /* KVM_EXIT_EPR */ 6827 struct { 6828 __u32 epr; 6829 } epr; 6830 6831On FSL BookE PowerPC chips, the interrupt controller has a fast patch 6832interrupt acknowledge path to the core. When the core successfully 6833delivers an interrupt, it automatically populates the EPR register with 6834the interrupt vector number and acknowledges the interrupt inside 6835the interrupt controller. 6836 6837In case the interrupt controller lives in user space, we need to do 6838the interrupt acknowledge cycle through it to fetch the next to be 6839delivered interrupt vector using this exit. 6840 6841It gets triggered whenever both KVM_CAP_PPC_EPR are enabled and an 6842external interrupt has just been delivered into the guest. User space 6843should put the acknowledged interrupt vector into the 'epr' field. 6844 6845:: 6846 6847 /* KVM_EXIT_SYSTEM_EVENT */ 6848 struct { 6849 #define KVM_SYSTEM_EVENT_SHUTDOWN 1 6850 #define KVM_SYSTEM_EVENT_RESET 2 6851 #define KVM_SYSTEM_EVENT_CRASH 3 6852 #define KVM_SYSTEM_EVENT_WAKEUP 4 6853 #define KVM_SYSTEM_EVENT_SUSPEND 5 6854 #define KVM_SYSTEM_EVENT_SEV_TERM 6 6855 #define KVM_SYSTEM_EVENT_TDX_FATAL 7 6856 __u32 type; 6857 __u32 ndata; 6858 __u64 data[16]; 6859 } system_event; 6860 6861If exit_reason is KVM_EXIT_SYSTEM_EVENT then the vcpu has triggered 6862a system-level event using some architecture specific mechanism (hypercall 6863or some special instruction). In case of ARM64, this is triggered using 6864HVC instruction based PSCI call from the vcpu. 6865 6866The 'type' field describes the system-level event type. 6867Valid values for 'type' are: 6868 6869 - KVM_SYSTEM_EVENT_SHUTDOWN -- the guest has requested a shutdown of the 6870 VM. Userspace is not obliged to honour this, and if it does honour 6871 this does not need to destroy the VM synchronously (ie it may call 6872 KVM_RUN again before shutdown finally occurs). 6873 - KVM_SYSTEM_EVENT_RESET -- the guest has requested a reset of the VM. 6874 As with SHUTDOWN, userspace can choose to ignore the request, or 6875 to schedule the reset to occur in the future and may call KVM_RUN again. 6876 - KVM_SYSTEM_EVENT_CRASH -- the guest crash occurred and the guest 6877 has requested a crash condition maintenance. Userspace can choose 6878 to ignore the request, or to gather VM memory core dump and/or 6879 reset/shutdown of the VM. 6880 - KVM_SYSTEM_EVENT_SEV_TERM -- an AMD SEV guest requested termination. 6881 The guest physical address of the guest's GHCB is stored in `data[0]`. 6882 - KVM_SYSTEM_EVENT_TDX_FATAL -- a TDX guest reported a fatal error state. 6883 KVM doesn't do any parsing or conversion, it just dumps 16 general-purpose 6884 registers to userspace, in ascending order of the 4-bit indices for x86-64 6885 general-purpose registers in instruction encoding, as defined in the Intel 6886 SDM. 6887 - KVM_SYSTEM_EVENT_WAKEUP -- the exiting vCPU is in a suspended state and 6888 KVM has recognized a wakeup event. Userspace may honor this event by 6889 marking the exiting vCPU as runnable, or deny it and call KVM_RUN again. 6890 - KVM_SYSTEM_EVENT_SUSPEND -- the guest has requested a suspension of 6891 the VM. 6892 6893If KVM_CAP_SYSTEM_EVENT_DATA is present, the 'data' field can contain 6894architecture specific information for the system-level event. Only 6895the first `ndata` items (possibly zero) of the data array are valid. 6896 6897 - for arm64, data[0] is set to KVM_SYSTEM_EVENT_RESET_FLAG_PSCI_RESET2 if 6898 the guest issued a SYSTEM_RESET2 call according to v1.1 of the PSCI 6899 specification. 6900 6901 - for arm64, data[0] is set to KVM_SYSTEM_EVENT_SHUTDOWN_FLAG_PSCI_OFF2 6902 if the guest issued a SYSTEM_OFF2 call according to v1.3 of the PSCI 6903 specification. 6904 6905 - for RISC-V, data[0] is set to the value of the second argument of the 6906 ``sbi_system_reset`` call. 6907 6908Previous versions of Linux defined a `flags` member in this struct. The 6909field is now aliased to `data[0]`. Userspace can assume that it is only 6910written if ndata is greater than 0. 6911 6912For arm/arm64: 6913-------------- 6914 6915KVM_SYSTEM_EVENT_SUSPEND exits are enabled with the 6916KVM_CAP_ARM_SYSTEM_SUSPEND VM capability. If a guest invokes the PSCI 6917SYSTEM_SUSPEND function, KVM will exit to userspace with this event 6918type. 6919 6920It is the sole responsibility of userspace to implement the PSCI 6921SYSTEM_SUSPEND call according to ARM DEN0022D.b 5.19 "SYSTEM_SUSPEND". 6922KVM does not change the vCPU's state before exiting to userspace, so 6923the call parameters are left in-place in the vCPU registers. 6924 6925Userspace is _required_ to take action for such an exit. It must 6926either: 6927 6928 - Honor the guest request to suspend the VM. Userspace can request 6929 in-kernel emulation of suspension by setting the calling vCPU's 6930 state to KVM_MP_STATE_SUSPENDED. Userspace must configure the vCPU's 6931 state according to the parameters passed to the PSCI function when 6932 the calling vCPU is resumed. See ARM DEN0022D.b 5.19.1 "Intended use" 6933 for details on the function parameters. 6934 6935 - Deny the guest request to suspend the VM. See ARM DEN0022D.b 5.19.2 6936 "Caller responsibilities" for possible return values. 6937 6938Hibernation using the PSCI SYSTEM_OFF2 call is enabled when PSCI v1.3 6939is enabled. If a guest invokes the PSCI SYSTEM_OFF2 function, KVM will 6940exit to userspace with the KVM_SYSTEM_EVENT_SHUTDOWN event type and with 6941data[0] set to KVM_SYSTEM_EVENT_SHUTDOWN_FLAG_PSCI_OFF2. The only 6942supported hibernate type for the SYSTEM_OFF2 function is HIBERNATE_OFF. 6943 6944:: 6945 6946 /* KVM_EXIT_IOAPIC_EOI */ 6947 struct { 6948 __u8 vector; 6949 } eoi; 6950 6951Indicates that the VCPU's in-kernel local APIC received an EOI for a 6952level-triggered IOAPIC interrupt. This exit only triggers when the 6953IOAPIC is implemented in userspace (i.e. KVM_CAP_SPLIT_IRQCHIP is enabled); 6954the userspace IOAPIC should process the EOI and retrigger the interrupt if 6955it is still asserted. Vector is the LAPIC interrupt vector for which the 6956EOI was received. 6957 6958:: 6959 6960 struct kvm_hyperv_exit { 6961 #define KVM_EXIT_HYPERV_SYNIC 1 6962 #define KVM_EXIT_HYPERV_HCALL 2 6963 #define KVM_EXIT_HYPERV_SYNDBG 3 6964 __u32 type; 6965 __u32 pad1; 6966 union { 6967 struct { 6968 __u32 msr; 6969 __u32 pad2; 6970 __u64 control; 6971 __u64 evt_page; 6972 __u64 msg_page; 6973 } synic; 6974 struct { 6975 __u64 input; 6976 __u64 result; 6977 __u64 params[2]; 6978 } hcall; 6979 struct { 6980 __u32 msr; 6981 __u32 pad2; 6982 __u64 control; 6983 __u64 status; 6984 __u64 send_page; 6985 __u64 recv_page; 6986 __u64 pending_page; 6987 } syndbg; 6988 } u; 6989 }; 6990 /* KVM_EXIT_HYPERV */ 6991 struct kvm_hyperv_exit hyperv; 6992 6993Indicates that the VCPU exits into userspace to process some tasks 6994related to Hyper-V emulation. 6995 6996Valid values for 'type' are: 6997 6998 - KVM_EXIT_HYPERV_SYNIC -- synchronously notify user-space about 6999 7000Hyper-V SynIC state change. Notification is used to remap SynIC 7001event/message pages and to enable/disable SynIC messages/events processing 7002in userspace. 7003 7004 - KVM_EXIT_HYPERV_SYNDBG -- synchronously notify user-space about 7005 7006Hyper-V Synthetic debugger state change. Notification is used to either update 7007the pending_page location or to send a control command (send the buffer located 7008in send_page or recv a buffer to recv_page). 7009 7010:: 7011 7012 /* KVM_EXIT_ARM_NISV */ 7013 struct { 7014 __u64 esr_iss; 7015 __u64 fault_ipa; 7016 } arm_nisv; 7017 7018Used on arm64 systems. If a guest accesses memory not in a memslot, 7019KVM will typically return to userspace and ask it to do MMIO emulation on its 7020behalf. However, for certain classes of instructions, no instruction decode 7021(direction, length of memory access) is provided, and fetching and decoding 7022the instruction from the VM is overly complicated to live in the kernel. 7023 7024Historically, when this situation occurred, KVM would print a warning and kill 7025the VM. KVM assumed that if the guest accessed non-memslot memory, it was 7026trying to do I/O, which just couldn't be emulated, and the warning message was 7027phrased accordingly. However, what happened more often was that a guest bug 7028caused access outside the guest memory areas which should lead to a more 7029meaningful warning message and an external abort in the guest, if the access 7030did not fall within an I/O window. 7031 7032Userspace implementations can query for KVM_CAP_ARM_NISV_TO_USER, and enable 7033this capability at VM creation. Once this is done, these types of errors will 7034instead return to userspace with KVM_EXIT_ARM_NISV, with the valid bits from 7035the ESR_EL2 in the esr_iss field, and the faulting IPA in the fault_ipa field. 7036Userspace can either fix up the access if it's actually an I/O access by 7037decoding the instruction from guest memory (if it's very brave) and continue 7038executing the guest, or it can decide to suspend, dump, or restart the guest. 7039 7040Note that KVM does not skip the faulting instruction as it does for 7041KVM_EXIT_MMIO, but userspace has to emulate any change to the processing state 7042if it decides to decode and emulate the instruction. 7043 7044This feature isn't available to protected VMs, as userspace does not 7045have access to the state that is required to perform the emulation. 7046Instead, a data abort exception is directly injected in the guest. 7047Note that although KVM_CAP_ARM_NISV_TO_USER will be reported if 7048queried outside of a protected VM context, the feature will not be 7049exposed if queried on a protected VM file descriptor. 7050 7051:: 7052 7053 /* KVM_EXIT_X86_RDMSR / KVM_EXIT_X86_WRMSR */ 7054 struct { 7055 __u8 error; /* user -> kernel */ 7056 __u8 pad[7]; 7057 __u32 reason; /* kernel -> user */ 7058 __u32 index; /* kernel -> user */ 7059 __u64 data; /* kernel <-> user */ 7060 } msr; 7061 7062Used on x86 systems. When the VM capability KVM_CAP_X86_USER_SPACE_MSR is 7063enabled, MSR accesses to registers that would invoke a #GP by KVM kernel code 7064may instead trigger a KVM_EXIT_X86_RDMSR exit for reads and KVM_EXIT_X86_WRMSR 7065exit for writes. 7066 7067The "reason" field specifies why the MSR interception occurred. Userspace will 7068only receive MSR exits when a particular reason was requested during through 7069ENABLE_CAP. Currently valid exit reasons are: 7070 7071============================ ======================================== 7072 KVM_MSR_EXIT_REASON_UNKNOWN access to MSR that is unknown to KVM 7073 KVM_MSR_EXIT_REASON_INVAL access to invalid MSRs or reserved bits 7074 KVM_MSR_EXIT_REASON_FILTER access blocked by KVM_X86_SET_MSR_FILTER 7075============================ ======================================== 7076 7077For KVM_EXIT_X86_RDMSR, the "index" field tells userspace which MSR the guest 7078wants to read. To respond to this request with a successful read, userspace 7079writes the respective data into the "data" field and must continue guest 7080execution to ensure the read data is transferred into guest register state. 7081 7082If the RDMSR request was unsuccessful, userspace indicates that with a "1" in 7083the "error" field. This will inject a #GP into the guest when the VCPU is 7084executed again. 7085 7086For KVM_EXIT_X86_WRMSR, the "index" field tells userspace which MSR the guest 7087wants to write. Once finished processing the event, userspace must continue 7088vCPU execution. If the MSR write was unsuccessful, userspace also sets the 7089"error" field to "1". 7090 7091See KVM_X86_SET_MSR_FILTER for details on the interaction with MSR filtering. 7092 7093:: 7094 7095 7096 struct kvm_xen_exit { 7097 #define KVM_EXIT_XEN_HCALL 1 7098 __u32 type; 7099 union { 7100 struct { 7101 __u32 longmode; 7102 __u32 cpl; 7103 __u64 input; 7104 __u64 result; 7105 __u64 params[6]; 7106 } hcall; 7107 } u; 7108 }; 7109 /* KVM_EXIT_XEN */ 7110 struct kvm_hyperv_exit xen; 7111 7112Indicates that the VCPU exits into userspace to process some tasks 7113related to Xen emulation. 7114 7115Valid values for 'type' are: 7116 7117 - KVM_EXIT_XEN_HCALL -- synchronously notify user-space about Xen hypercall. 7118 Userspace is expected to place the hypercall result into the appropriate 7119 field before invoking KVM_RUN again. 7120 7121:: 7122 7123 /* KVM_EXIT_RISCV_SBI */ 7124 struct { 7125 unsigned long extension_id; 7126 unsigned long function_id; 7127 unsigned long args[6]; 7128 unsigned long ret[2]; 7129 } riscv_sbi; 7130 7131If exit reason is KVM_EXIT_RISCV_SBI then it indicates that the VCPU has 7132done a SBI call which is not handled by KVM RISC-V kernel module. The details 7133of the SBI call are available in 'riscv_sbi' member of kvm_run structure. The 7134'extension_id' field of 'riscv_sbi' represents SBI extension ID whereas the 7135'function_id' field represents function ID of given SBI extension. The 'args' 7136array field of 'riscv_sbi' represents parameters for the SBI call and 'ret' 7137array field represents return values. The userspace should update the return 7138values of SBI call before resuming the VCPU. For more details on RISC-V SBI 7139spec refer, https://github.com/riscv/riscv-sbi-doc. 7140 7141:: 7142 7143 /* KVM_EXIT_MEMORY_FAULT */ 7144 struct { 7145 #define KVM_MEMORY_EXIT_FLAG_PRIVATE (1ULL << 3) 7146 __u64 flags; 7147 __u64 gpa; 7148 __u64 size; 7149 } memory_fault; 7150 7151KVM_EXIT_MEMORY_FAULT indicates the vCPU has encountered a memory fault that 7152could not be resolved by KVM. The 'gpa' and 'size' (in bytes) describe the 7153guest physical address range [gpa, gpa + size) of the fault. The 'flags' field 7154describes properties of the faulting access that are likely pertinent: 7155 7156 - KVM_MEMORY_EXIT_FLAG_PRIVATE - When set, indicates the memory fault occurred 7157 on a private memory access. When clear, indicates the fault occurred on a 7158 shared access. 7159 7160Note! KVM_EXIT_MEMORY_FAULT is unique among all KVM exit reasons in that it 7161accompanies a return code of '-1', not '0'! errno will always be set to EFAULT 7162or EHWPOISON when KVM exits with KVM_EXIT_MEMORY_FAULT, userspace should assume 7163kvm_run.exit_reason is stale/undefined for all other error numbers. 7164 7165:: 7166 7167 /* KVM_EXIT_NOTIFY */ 7168 struct { 7169 #define KVM_NOTIFY_CONTEXT_INVALID (1 << 0) 7170 __u32 flags; 7171 } notify; 7172 7173Used on x86 systems. When the VM capability KVM_CAP_X86_NOTIFY_VMEXIT is 7174enabled, a VM exit generated if no event window occurs in VM non-root mode 7175for a specified amount of time. Once KVM_X86_NOTIFY_VMEXIT_USER is set when 7176enabling the cap, it would exit to userspace with the exit reason 7177KVM_EXIT_NOTIFY for further handling. The "flags" field contains more 7178detailed info. 7179 7180The valid value for 'flags' is: 7181 7182 - KVM_NOTIFY_CONTEXT_INVALID -- the VM context is corrupted and not valid 7183 in VMCS. It would run into unknown result if resume the target VM. 7184 7185:: 7186 7187 /* KVM_EXIT_TDX */ 7188 struct { 7189 __u64 flags; 7190 __u64 nr; 7191 union { 7192 struct { 7193 u64 ret; 7194 u64 data[5]; 7195 } unknown; 7196 struct { 7197 u64 ret; 7198 u64 gpa; 7199 u64 size; 7200 } get_quote; 7201 struct { 7202 u64 ret; 7203 u64 leaf; 7204 u64 r11, r12, r13, r14; 7205 } get_tdvmcall_info; 7206 struct { 7207 u64 ret; 7208 u64 vector; 7209 } setup_event_notify; 7210 }; 7211 } tdx; 7212 7213Process a TDVMCALL from the guest. KVM forwards select TDVMCALL based 7214on the Guest-Hypervisor Communication Interface (GHCI) specification; 7215KVM bridges these requests to the userspace VMM with minimal changes, 7216placing the inputs in the union and copying them back to the guest 7217on re-entry. 7218 7219Flags are currently always zero, whereas ``nr`` contains the TDVMCALL 7220number from register R11. The remaining field of the union provide the 7221inputs and outputs of the TDVMCALL. Currently the following values of 7222``nr`` are defined: 7223 7224 * ``TDVMCALL_GET_QUOTE``: the guest has requested to generate a TD-Quote 7225 signed by a service hosting TD-Quoting Enclave operating on the host. 7226 Parameters and return value are in the ``get_quote`` field of the union. 7227 The ``gpa`` field and ``size`` specify the guest physical address 7228 (without the shared bit set) and the size of a shared-memory buffer, in 7229 which the TDX guest passes a TD Report. The ``ret`` field represents 7230 the return value of the GetQuote request. When the request has been 7231 queued successfully, the TDX guest can poll the status field in the 7232 shared-memory area to check whether the Quote generation is completed or 7233 not. When completed, the generated Quote is returned via the same buffer. 7234 7235 * ``TDVMCALL_GET_TD_VM_CALL_INFO``: the guest has requested the support 7236 status of TDVMCALLs. The output values for the given leaf should be 7237 placed in fields from ``r11`` to ``r14`` of the ``get_tdvmcall_info`` 7238 field of the union. 7239 7240 * ``TDVMCALL_SETUP_EVENT_NOTIFY_INTERRUPT``: the guest has requested to 7241 set up a notification interrupt for vector ``vector``. 7242 7243KVM may add support for more values in the future that may cause a userspace 7244exit, even without calls to ``KVM_ENABLE_CAP`` or similar. In this case, 7245it will enter with output fields already valid; in the common case, the 7246``unknown.ret`` field of the union will be ``TDVMCALL_STATUS_SUBFUNC_UNSUPPORTED``. 7247Userspace need not do anything if it does not wish to support a TDVMCALL. 7248:: 7249 7250 /* Fix the size of the union. */ 7251 char padding[256]; 7252 }; 7253 7254 /* 7255 * shared registers between kvm and userspace. 7256 * kvm_valid_regs specifies the register classes set by the host 7257 * kvm_dirty_regs specified the register classes dirtied by userspace 7258 * struct kvm_sync_regs is architecture specific, as well as the 7259 * bits for kvm_valid_regs and kvm_dirty_regs 7260 */ 7261 __u64 kvm_valid_regs; 7262 __u64 kvm_dirty_regs; 7263 union { 7264 struct kvm_sync_regs regs; 7265 char padding[SYNC_REGS_SIZE_BYTES]; 7266 } s; 7267 7268If KVM_CAP_SYNC_REGS is defined, these fields allow userspace to access 7269certain guest registers without having to call SET/GET_*REGS. Thus we can 7270avoid some system call overhead if userspace has to handle the exit. 7271Userspace can query the validity of the structure by checking 7272kvm_valid_regs for specific bits. These bits are architecture specific 7273and usually define the validity of a groups of registers. (e.g. one bit 7274for general purpose registers) 7275 7276Please note that the kernel is allowed to use the kvm_run structure as the 7277primary storage for certain register types. Therefore, the kernel may use the 7278values in kvm_run even if the corresponding bit in kvm_dirty_regs is not set. 7279 7280 7281.. _cap_enable: 7282 72836. Capabilities that can be enabled on vCPUs 7284============================================ 7285 7286There are certain capabilities that change the behavior of the virtual CPU or 7287the virtual machine when enabled. To enable them, please see 7288:ref:`KVM_ENABLE_CAP`. 7289 7290Below you can find a list of capabilities and what their effect on the vCPU or 7291the virtual machine is when enabling them. 7292 7293The following information is provided along with the description: 7294 7295 Architectures: 7296 which instruction set architectures provide this ioctl. 7297 x86 includes both i386 and x86_64. 7298 7299 Target: 7300 whether this is a per-vcpu or per-vm capability. 7301 7302 Parameters: 7303 what parameters are accepted by the capability. 7304 7305 Returns: 7306 the return value. General error numbers (EBADF, ENOMEM, EINVAL) 7307 are not detailed, but errors with specific meanings are. 7308 7309 73106.1 KVM_CAP_PPC_OSI 7311------------------- 7312 7313:Architectures: ppc 7314:Target: vcpu 7315:Parameters: none 7316:Returns: 0 on success; -1 on error 7317 7318This capability enables interception of OSI hypercalls that otherwise would 7319be treated as normal system calls to be injected into the guest. OSI hypercalls 7320were invented by Mac-on-Linux to have a standardized communication mechanism 7321between the guest and the host. 7322 7323When this capability is enabled, KVM_EXIT_OSI can occur. 7324 7325 73266.2 KVM_CAP_PPC_PAPR 7327-------------------- 7328 7329:Architectures: ppc 7330:Target: vcpu 7331:Parameters: none 7332:Returns: 0 on success; -1 on error 7333 7334This capability enables interception of PAPR hypercalls. PAPR hypercalls are 7335done using the hypercall instruction "sc 1". 7336 7337It also sets the guest privilege level to "supervisor" mode. Usually the guest 7338runs in "hypervisor" privilege mode with a few missing features. 7339 7340In addition to the above, it changes the semantics of SDR1. In this mode, the 7341HTAB address part of SDR1 contains an HVA instead of a GPA, as PAPR keeps the 7342HTAB invisible to the guest. 7343 7344When this capability is enabled, KVM_EXIT_PAPR_HCALL can occur. 7345 7346 73476.3 KVM_CAP_SW_TLB 7348------------------ 7349 7350:Architectures: ppc 7351:Target: vcpu 7352:Parameters: args[0] is the address of a struct kvm_config_tlb 7353:Returns: 0 on success; -1 on error 7354 7355:: 7356 7357 struct kvm_config_tlb { 7358 __u64 params; 7359 __u64 array; 7360 __u32 mmu_type; 7361 __u32 array_len; 7362 }; 7363 7364Configures the virtual CPU's TLB array, establishing a shared memory area 7365between userspace and KVM. The "params" and "array" fields are userspace 7366addresses of mmu-type-specific data structures. The "array_len" field is an 7367safety mechanism, and should be set to the size in bytes of the memory that 7368userspace has reserved for the array. It must be at least the size dictated 7369by "mmu_type" and "params". 7370 7371While KVM_RUN is active, the shared region is under control of KVM. Its 7372contents are undefined, and any modification by userspace results in 7373boundedly undefined behavior. 7374 7375On return from KVM_RUN, the shared region will reflect the current state of 7376the guest's TLB. If userspace makes any changes, it must call KVM_DIRTY_TLB 7377to tell KVM which entries have been changed, prior to calling KVM_RUN again 7378on this vcpu. 7379 7380For mmu types KVM_MMU_FSL_BOOKE_NOHV and KVM_MMU_FSL_BOOKE_HV: 7381 7382 - The "params" field is of type "struct kvm_book3e_206_tlb_params". 7383 - The "array" field points to an array of type "struct 7384 kvm_book3e_206_tlb_entry". 7385 - The array consists of all entries in the first TLB, followed by all 7386 entries in the second TLB. 7387 - Within a TLB, entries are ordered first by increasing set number. Within a 7388 set, entries are ordered by way (increasing ESEL). 7389 - The hash for determining set number in TLB0 is: (MAS2 >> 12) & (num_sets - 1) 7390 where "num_sets" is the tlb_sizes[] value divided by the tlb_ways[] value. 7391 - The tsize field of mas1 shall be set to 4K on TLB0, even though the 7392 hardware ignores this value for TLB0. 7393 73946.4 KVM_CAP_S390_CSS_SUPPORT 7395---------------------------- 7396 7397:Architectures: s390 7398:Target: vcpu 7399:Parameters: none 7400:Returns: 0 on success; -1 on error 7401 7402This capability enables support for handling of channel I/O instructions. 7403 7404TEST PENDING INTERRUPTION and the interrupt portion of TEST SUBCHANNEL are 7405handled in-kernel, while the other I/O instructions are passed to userspace. 7406 7407When this capability is enabled, KVM_EXIT_S390_TSCH will occur on TEST 7408SUBCHANNEL intercepts. 7409 7410Note that even though this capability is enabled per-vcpu, the complete 7411virtual machine is affected. 7412 74136.5 KVM_CAP_PPC_EPR 7414------------------- 7415 7416:Architectures: ppc 7417:Target: vcpu 7418:Parameters: args[0] defines whether the proxy facility is active 7419:Returns: 0 on success; -1 on error 7420 7421This capability enables or disables the delivery of interrupts through the 7422external proxy facility. 7423 7424When enabled (args[0] != 0), every time the guest gets an external interrupt 7425delivered, it automatically exits into user space with a KVM_EXIT_EPR exit 7426to receive the topmost interrupt vector. 7427 7428When disabled (args[0] == 0), behavior is as if this facility is unsupported. 7429 7430When this capability is enabled, KVM_EXIT_EPR can occur. 7431 74326.6 KVM_CAP_IRQ_MPIC 7433-------------------- 7434 7435:Architectures: ppc 7436:Parameters: args[0] is the MPIC device fd; 7437 args[1] is the MPIC CPU number for this vcpu 7438 7439This capability connects the vcpu to an in-kernel MPIC device. 7440 74416.7 KVM_CAP_IRQ_XICS 7442-------------------- 7443 7444:Architectures: ppc 7445:Target: vcpu 7446:Parameters: args[0] is the XICS device fd; 7447 args[1] is the XICS CPU number (server ID) for this vcpu 7448 7449This capability connects the vcpu to an in-kernel XICS device. 7450 74516.8 KVM_CAP_S390_IRQCHIP 7452------------------------ 7453 7454:Architectures: s390 7455:Target: vm 7456:Parameters: none 7457 7458This capability enables the in-kernel irqchip for s390. Please refer to 7459"4.24 KVM_CREATE_IRQCHIP" for details. 7460 74616.9 KVM_CAP_MIPS_FPU 7462-------------------- 7463 7464:Architectures: mips 7465:Target: vcpu 7466:Parameters: args[0] is reserved for future use (should be 0). 7467 7468This capability allows the use of the host Floating Point Unit by the guest. It 7469allows the Config1.FP bit to be set to enable the FPU in the guest. Once this is 7470done the ``KVM_REG_MIPS_FPR_*`` and ``KVM_REG_MIPS_FCR_*`` registers can be 7471accessed (depending on the current guest FPU register mode), and the Status.FR, 7472Config5.FRE bits are accessible via the KVM API and also from the guest, 7473depending on them being supported by the FPU. 7474 74756.10 KVM_CAP_MIPS_MSA 7476--------------------- 7477 7478:Architectures: mips 7479:Target: vcpu 7480:Parameters: args[0] is reserved for future use (should be 0). 7481 7482This capability allows the use of the MIPS SIMD Architecture (MSA) by the guest. 7483It allows the Config3.MSAP bit to be set to enable the use of MSA by the guest. 7484Once this is done the ``KVM_REG_MIPS_VEC_*`` and ``KVM_REG_MIPS_MSA_*`` 7485registers can be accessed, and the Config5.MSAEn bit is accessible via the 7486KVM API and also from the guest. 7487 74886.74 KVM_CAP_SYNC_REGS 7489---------------------- 7490 7491:Architectures: s390, x86 7492:Target: s390: always enabled, x86: vcpu 7493:Parameters: none 7494:Returns: x86: KVM_CHECK_EXTENSION returns a bit-array indicating which register 7495 sets are supported 7496 (bitfields defined in arch/x86/include/uapi/asm/kvm.h). 7497 7498As described above in the kvm_sync_regs struct info in section :ref:`kvm_run`, 7499KVM_CAP_SYNC_REGS "allow[s] userspace to access certain guest registers 7500without having to call SET/GET_*REGS". This reduces overhead by eliminating 7501repeated ioctl calls for setting and/or getting register values. This is 7502particularly important when userspace is making synchronous guest state 7503modifications, e.g. when emulating and/or intercepting instructions in 7504userspace. 7505 7506For s390 specifics, please refer to the source code. 7507 7508For x86: 7509 7510- the register sets to be copied out to kvm_run are selectable 7511 by userspace (rather that all sets being copied out for every exit). 7512- vcpu_events are available in addition to regs and sregs. 7513 7514For x86, the 'kvm_valid_regs' field of struct kvm_run is overloaded to 7515function as an input bit-array field set by userspace to indicate the 7516specific register sets to be copied out on the next exit. 7517 7518To indicate when userspace has modified values that should be copied into 7519the vCPU, the all architecture bitarray field, 'kvm_dirty_regs' must be set. 7520This is done using the same bitflags as for the 'kvm_valid_regs' field. 7521If the dirty bit is not set, then the register set values will not be copied 7522into the vCPU even if they've been modified. 7523 7524Unused bitfields in the bitarrays must be set to zero. 7525 7526:: 7527 7528 struct kvm_sync_regs { 7529 struct kvm_regs regs; 7530 struct kvm_sregs sregs; 7531 struct kvm_vcpu_events events; 7532 }; 7533 75346.75 KVM_CAP_PPC_IRQ_XIVE 7535------------------------- 7536 7537:Architectures: ppc 7538:Target: vcpu 7539:Parameters: args[0] is the XIVE device fd; 7540 args[1] is the XIVE CPU number (server ID) for this vcpu 7541 7542This capability connects the vcpu to an in-kernel XIVE device. 7543 75446.76 KVM_CAP_HYPERV_SYNIC 7545------------------------- 7546 7547:Architectures: x86 7548:Target: vcpu 7549 7550This capability, if KVM_CHECK_EXTENSION indicates that it is 7551available, means that the kernel has an implementation of the 7552Hyper-V Synthetic interrupt controller(SynIC). Hyper-V SynIC is 7553used to support Windows Hyper-V based guest paravirt drivers(VMBus). 7554 7555In order to use SynIC, it has to be activated by setting this 7556capability via KVM_ENABLE_CAP ioctl on the vcpu fd. Note that this 7557will disable the use of APIC hardware virtualization even if supported 7558by the CPU, as it's incompatible with SynIC auto-EOI behavior. 7559 75606.77 KVM_CAP_HYPERV_SYNIC2 7561-------------------------- 7562 7563:Architectures: x86 7564:Target: vcpu 7565 7566This capability enables a newer version of Hyper-V Synthetic interrupt 7567controller (SynIC). The only difference with KVM_CAP_HYPERV_SYNIC is that KVM 7568doesn't clear SynIC message and event flags pages when they are enabled by 7569writing to the respective MSRs. 7570 75716.78 KVM_CAP_HYPERV_DIRECT_TLBFLUSH 7572----------------------------------- 7573 7574:Architectures: x86 7575:Target: vcpu 7576 7577This capability indicates that KVM running on top of Hyper-V hypervisor 7578enables Direct TLB flush for its guests meaning that TLB flush 7579hypercalls are handled by Level 0 hypervisor (Hyper-V) bypassing KVM. 7580Due to the different ABI for hypercall parameters between Hyper-V and 7581KVM, enabling this capability effectively disables all hypercall 7582handling by KVM (as some KVM hypercall may be mistakenly treated as TLB 7583flush hypercalls by Hyper-V) so userspace should disable KVM identification 7584in CPUID and only exposes Hyper-V identification. In this case, guest 7585thinks it's running on Hyper-V and only use Hyper-V hypercalls. 7586 75876.79 KVM_CAP_HYPERV_ENFORCE_CPUID 7588--------------------------------- 7589 7590:Architectures: x86 7591:Target: vcpu 7592 7593When enabled, KVM will disable emulated Hyper-V features provided to the 7594guest according to the bits Hyper-V CPUID feature leaves. Otherwise, all 7595currently implemented Hyper-V features are provided unconditionally when 7596Hyper-V identification is set in the HYPERV_CPUID_INTERFACE (0x40000001) 7597leaf. 7598 75996.80 KVM_CAP_ENFORCE_PV_FEATURE_CPUID 7600------------------------------------- 7601 7602:Architectures: x86 7603:Target: vcpu 7604 7605When enabled, KVM will disable paravirtual features provided to the 7606guest according to the bits in the KVM_CPUID_FEATURES CPUID leaf 7607(0x40000001). Otherwise, a guest may use the paravirtual features 7608regardless of what has actually been exposed through the CPUID leaf. 7609 7610.. _KVM_CAP_DIRTY_LOG_RING: 7611 7612 7613.. _cap_enable_vm: 7614 76157. Capabilities that can be enabled on VMs 7616========================================== 7617 7618There are certain capabilities that change the behavior of the virtual 7619machine when enabled. To enable them, please see section 7620:ref:`KVM_ENABLE_CAP`. Below you can find a list of capabilities and 7621what their effect on the VM is when enabling them. 7622 7623The following information is provided along with the description: 7624 7625 Architectures: 7626 which instruction set architectures provide this ioctl. 7627 x86 includes both i386 and x86_64. 7628 7629 Parameters: 7630 what parameters are accepted by the capability. 7631 7632 Returns: 7633 the return value. General error numbers (EBADF, ENOMEM, EINVAL) 7634 are not detailed, but errors with specific meanings are. 7635 7636 76377.1 KVM_CAP_PPC_ENABLE_HCALL 7638---------------------------- 7639 7640:Architectures: ppc 7641:Parameters: args[0] is the sPAPR hcall number; 7642 args[1] is 0 to disable, 1 to enable in-kernel handling 7643 7644This capability controls whether individual sPAPR hypercalls (hcalls) 7645get handled by the kernel or not. Enabling or disabling in-kernel 7646handling of an hcall is effective across the VM. On creation, an 7647initial set of hcalls are enabled for in-kernel handling, which 7648consists of those hcalls for which in-kernel handlers were implemented 7649before this capability was implemented. If disabled, the kernel will 7650not to attempt to handle the hcall, but will always exit to userspace 7651to handle it. Note that it may not make sense to enable some and 7652disable others of a group of related hcalls, but KVM does not prevent 7653userspace from doing that. 7654 7655If the hcall number specified is not one that has an in-kernel 7656implementation, the KVM_ENABLE_CAP ioctl will fail with an EINVAL 7657error. 7658 76597.2 KVM_CAP_S390_USER_SIGP 7660-------------------------- 7661 7662:Architectures: s390 7663:Parameters: none 7664 7665This capability controls which SIGP orders will be handled completely in user 7666space. With this capability enabled, all fast orders will be handled completely 7667in the kernel: 7668 7669- SENSE 7670- SENSE RUNNING 7671- EXTERNAL CALL 7672- EMERGENCY SIGNAL 7673- CONDITIONAL EMERGENCY SIGNAL 7674 7675All other orders will be handled completely in user space. 7676 7677Only privileged operation exceptions will be checked for in the kernel (or even 7678in the hardware prior to interception). If this capability is not enabled, the 7679old way of handling SIGP orders is used (partially in kernel and user space). 7680 76817.3 KVM_CAP_S390_VECTOR_REGISTERS 7682--------------------------------- 7683 7684:Architectures: s390 7685:Parameters: none 7686:Returns: 0 on success, negative value on error 7687 7688Allows use of the vector registers introduced with z13 processor, and 7689provides for the synchronization between host and user space. Will 7690return -EINVAL if the machine does not support vectors. 7691 76927.4 KVM_CAP_S390_USER_STSI 7693-------------------------- 7694 7695:Architectures: s390 7696:Parameters: none 7697 7698This capability allows post-handlers for the STSI instruction. After 7699initial handling in the kernel, KVM exits to user space with 7700KVM_EXIT_S390_STSI to allow user space to insert further data. 7701 7702Before exiting to userspace, kvm handlers should fill in s390_stsi field of 7703vcpu->run:: 7704 7705 struct { 7706 __u64 addr; 7707 __u8 ar; 7708 __u8 reserved; 7709 __u8 fc; 7710 __u8 sel1; 7711 __u16 sel2; 7712 } s390_stsi; 7713 7714 @addr - guest address of STSI SYSIB 7715 @fc - function code 7716 @sel1 - selector 1 7717 @sel2 - selector 2 7718 @ar - access register number 7719 7720KVM handlers should exit to userspace with rc = -EREMOTE. 7721 77227.5 KVM_CAP_SPLIT_IRQCHIP 7723------------------------- 7724 7725:Architectures: x86 7726:Parameters: args[0] - number of routes reserved for userspace IOAPICs 7727:Returns: 0 on success, -1 on error 7728 7729Create a local apic for each processor in the kernel. This can be used 7730instead of KVM_CREATE_IRQCHIP if the userspace VMM wishes to emulate the 7731IOAPIC and PIC (and also the PIT, even though this has to be enabled 7732separately). 7733 7734This capability also enables in kernel routing of interrupt requests; 7735when KVM_CAP_SPLIT_IRQCHIP only routes of KVM_IRQ_ROUTING_MSI type are 7736used in the IRQ routing table. The first args[0] MSI routes are reserved 7737for the IOAPIC pins. Whenever the LAPIC receives an EOI for these routes, 7738a KVM_EXIT_IOAPIC_EOI vmexit will be reported to userspace. 7739 7740Fails if VCPU has already been created, or if the irqchip is already in the 7741kernel (i.e. KVM_CREATE_IRQCHIP has already been called). 7742 77437.6 KVM_CAP_S390_RI 7744------------------- 7745 7746:Architectures: s390 7747:Parameters: none 7748 7749Allows use of runtime-instrumentation introduced with zEC12 processor. 7750Will return -EINVAL if the machine does not support runtime-instrumentation. 7751Will return -EBUSY if a VCPU has already been created. 7752 77537.7 KVM_CAP_X2APIC_API 7754---------------------- 7755 7756:Architectures: x86 7757:Parameters: args[0] - features that should be enabled 7758:Returns: 0 on success, -EINVAL when args[0] contains invalid features 7759 7760Valid feature flags in args[0] are:: 7761 7762 #define KVM_X2APIC_API_USE_32BIT_IDS (1ULL << 0) 7763 #define KVM_X2APIC_API_DISABLE_BROADCAST_QUIRK (1ULL << 1) 7764 7765Enabling KVM_X2APIC_API_USE_32BIT_IDS changes the behavior of 7766KVM_SET_GSI_ROUTING, KVM_SIGNAL_MSI, KVM_SET_LAPIC, and KVM_GET_LAPIC, 7767allowing the use of 32-bit APIC IDs. See KVM_CAP_X2APIC_API in their 7768respective sections. 7769 7770KVM_X2APIC_API_DISABLE_BROADCAST_QUIRK must be enabled for x2APIC to work 7771in logical mode or with more than 255 VCPUs. Otherwise, KVM treats 0xff 7772as a broadcast even in x2APIC mode in order to support physical x2APIC 7773without interrupt remapping. This is undesirable in logical mode, 7774where 0xff represents CPUs 0-7 in cluster 0. 7775 77767.8 KVM_CAP_S390_USER_INSTR0 7777---------------------------- 7778 7779:Architectures: s390 7780:Parameters: none 7781 7782With this capability enabled, all illegal instructions 0x0000 (2 bytes) will 7783be intercepted and forwarded to user space. User space can use this 7784mechanism e.g. to realize 2-byte software breakpoints. The kernel will 7785not inject an operating exception for these instructions, user space has 7786to take care of that. 7787 7788This capability can be enabled dynamically even if VCPUs were already 7789created and are running. 7790 77917.9 KVM_CAP_S390_GS 7792------------------- 7793 7794:Architectures: s390 7795:Parameters: none 7796:Returns: 0 on success; -EINVAL if the machine does not support 7797 guarded storage; -EBUSY if a VCPU has already been created. 7798 7799Allows use of guarded storage for the KVM guest. 7800 78017.10 KVM_CAP_S390_AIS 7802--------------------- 7803 7804:Architectures: s390 7805:Parameters: none 7806 7807Allow use of adapter-interruption suppression. 7808:Returns: 0 on success; -EBUSY if a VCPU has already been created. 7809 78107.11 KVM_CAP_PPC_SMT 7811-------------------- 7812 7813:Architectures: ppc 7814:Parameters: vsmt_mode, flags 7815 7816Enabling this capability on a VM provides userspace with a way to set 7817the desired virtual SMT mode (i.e. the number of virtual CPUs per 7818virtual core). The virtual SMT mode, vsmt_mode, must be a power of 2 7819between 1 and 8. On POWER8, vsmt_mode must also be no greater than 7820the number of threads per subcore for the host. Currently flags must 7821be 0. A successful call to enable this capability will result in 7822vsmt_mode being returned when the KVM_CAP_PPC_SMT capability is 7823subsequently queried for the VM. This capability is only supported by 7824HV KVM, and can only be set before any VCPUs have been created. 7825The KVM_CAP_PPC_SMT_POSSIBLE capability indicates which virtual SMT 7826modes are available. 7827 78287.12 KVM_CAP_PPC_FWNMI 7829---------------------- 7830 7831:Architectures: ppc 7832:Parameters: none 7833 7834With this capability a machine check exception in the guest address 7835space will cause KVM to exit the guest with NMI exit reason. This 7836enables QEMU to build error log and branch to guest kernel registered 7837machine check handling routine. Without this capability KVM will 7838branch to guests' 0x200 interrupt vector. 7839 78407.13 KVM_CAP_X86_DISABLE_EXITS 7841------------------------------ 7842 7843:Architectures: x86 7844:Parameters: args[0] defines which exits are disabled 7845:Returns: 0 on success, -EINVAL when args[0] contains invalid exits 7846 or if any vCPUs have already been created 7847 7848Valid bits in args[0] are:: 7849 7850 #define KVM_X86_DISABLE_EXITS_MWAIT (1 << 0) 7851 #define KVM_X86_DISABLE_EXITS_HLT (1 << 1) 7852 #define KVM_X86_DISABLE_EXITS_PAUSE (1 << 2) 7853 #define KVM_X86_DISABLE_EXITS_CSTATE (1 << 3) 7854 #define KVM_X86_DISABLE_EXITS_APERFMPERF (1 << 4) 7855 7856Enabling this capability on a VM provides userspace with a way to no 7857longer intercept some instructions for improved latency in some 7858workloads, and is suggested when vCPUs are associated to dedicated 7859physical CPUs. More bits can be added in the future; userspace can 7860just pass the KVM_CHECK_EXTENSION result to KVM_ENABLE_CAP to disable 7861all such vmexits. 7862 7863Do not enable KVM_FEATURE_PV_UNHALT if you disable HLT exits. 7864 7865Virtualizing the ``IA32_APERF`` and ``IA32_MPERF`` MSRs requires more 7866than just disabling APERF/MPERF exits. While both Intel and AMD 7867document strict usage conditions for these MSRs--emphasizing that only 7868the ratio of their deltas over a time interval (T0 to T1) is 7869architecturally defined--simply passing through the MSRs can still 7870produce an incorrect ratio. 7871 7872This erroneous ratio can occur if, between T0 and T1: 7873 78741. The vCPU thread migrates between logical processors. 78752. Live migration or suspend/resume operations take place. 78763. Another task shares the vCPU's logical processor. 78774. C-states lower than C0 are emulated (e.g., via HLT interception). 78785. The guest TSC frequency doesn't match the host TSC frequency. 7879 7880Due to these complexities, KVM does not automatically associate this 7881passthrough capability with the guest CPUID bit, 7882``CPUID.6:ECX.APERFMPERF[bit 0]``. Userspace VMMs that deem this 7883mechanism adequate for virtualizing the ``IA32_APERF`` and 7884``IA32_MPERF`` MSRs must set the guest CPUID bit explicitly. 7885 7886 78877.14 KVM_CAP_S390_HPAGE_1M 7888-------------------------- 7889 7890:Architectures: s390 7891:Parameters: none 7892:Returns: 0 on success, -EINVAL if hpage module parameter was not set 7893 or cmma is enabled, or the VM has the KVM_VM_S390_UCONTROL 7894 flag set 7895 7896With this capability the KVM support for memory backing with 1m pages 7897through hugetlbfs can be enabled for a VM. After the capability is 7898enabled, cmma can't be enabled anymore and pfmfi and the storage key 7899interpretation are disabled. If cmma has already been enabled or the 7900hpage module parameter is not set to 1, -EINVAL is returned. 7901 7902While it is generally possible to create a huge page backed VM without 7903this capability, the VM will not be able to run. 7904 79057.15 KVM_CAP_MSR_PLATFORM_INFO 7906------------------------------ 7907 7908:Architectures: x86 7909:Parameters: args[0] whether feature should be enabled or not 7910 7911With this capability, a guest may read the MSR_PLATFORM_INFO MSR. Otherwise, 7912a #GP would be raised when the guest tries to access. Currently, this 7913capability does not enable write permissions of this MSR for the guest. 7914 79157.16 KVM_CAP_PPC_NESTED_HV 7916-------------------------- 7917 7918:Architectures: ppc 7919:Parameters: none 7920:Returns: 0 on success, -EINVAL when the implementation doesn't support 7921 nested-HV virtualization. 7922 7923HV-KVM on POWER9 and later systems allows for "nested-HV" 7924virtualization, which provides a way for a guest VM to run guests that 7925can run using the CPU's supervisor mode (privileged non-hypervisor 7926state). Enabling this capability on a VM depends on the CPU having 7927the necessary functionality and on the facility being enabled with a 7928kvm-hv module parameter. 7929 79307.17 KVM_CAP_EXCEPTION_PAYLOAD 7931------------------------------ 7932 7933:Architectures: x86 7934:Parameters: args[0] whether feature should be enabled or not 7935 7936With this capability enabled, CR2 will not be modified prior to the 7937emulated VM-exit when L1 intercepts a #PF exception that occurs in 7938L2. Similarly, for kvm-intel only, DR6 will not be modified prior to 7939the emulated VM-exit when L1 intercepts a #DB exception that occurs in 7940L2. As a result, when KVM_GET_VCPU_EVENTS reports a pending #PF (or 7941#DB) exception for L2, exception.has_payload will be set and the 7942faulting address (or the new DR6 bits*) will be reported in the 7943exception_payload field. Similarly, when userspace injects a #PF (or 7944#DB) into L2 using KVM_SET_VCPU_EVENTS, it is expected to set 7945exception.has_payload and to put the faulting address - or the new DR6 7946bits\ [#]_ - in the exception_payload field. 7947 7948This capability also enables exception.pending in struct 7949kvm_vcpu_events, which allows userspace to distinguish between pending 7950and injected exceptions. 7951 7952 7953.. [#] For the new DR6 bits, note that bit 16 is set iff the #DB exception 7954 will clear DR6.RTM. 7955 79567.18 KVM_CAP_MANUAL_DIRTY_LOG_PROTECT2 7957-------------------------------------- 7958 7959:Architectures: x86, arm64, mips 7960:Parameters: args[0] whether feature should be enabled or not 7961 7962Valid flags are:: 7963 7964 #define KVM_DIRTY_LOG_MANUAL_PROTECT_ENABLE (1 << 0) 7965 #define KVM_DIRTY_LOG_INITIALLY_SET (1 << 1) 7966 7967With KVM_DIRTY_LOG_MANUAL_PROTECT_ENABLE is set, KVM_GET_DIRTY_LOG will not 7968automatically clear and write-protect all pages that are returned as dirty. 7969Rather, userspace will have to do this operation separately using 7970KVM_CLEAR_DIRTY_LOG. 7971 7972At the cost of a slightly more complicated operation, this provides better 7973scalability and responsiveness for two reasons. First, 7974KVM_CLEAR_DIRTY_LOG ioctl can operate on a 64-page granularity rather 7975than requiring to sync a full memslot; this ensures that KVM does not 7976take spinlocks for an extended period of time. Second, in some cases a 7977large amount of time can pass between a call to KVM_GET_DIRTY_LOG and 7978userspace actually using the data in the page. Pages can be modified 7979during this time, which is inefficient for both the guest and userspace: 7980the guest will incur a higher penalty due to write protection faults, 7981while userspace can see false reports of dirty pages. Manual reprotection 7982helps reducing this time, improving guest performance and reducing the 7983number of dirty log false positives. 7984 7985With KVM_DIRTY_LOG_INITIALLY_SET set, all the bits of the dirty bitmap 7986will be initialized to 1 when created. This also improves performance because 7987dirty logging can be enabled gradually in small chunks on the first call 7988to KVM_CLEAR_DIRTY_LOG. KVM_DIRTY_LOG_INITIALLY_SET depends on 7989KVM_DIRTY_LOG_MANUAL_PROTECT_ENABLE (it is also only available on 7990x86 and arm64 for now). 7991 7992KVM_CAP_MANUAL_DIRTY_LOG_PROTECT2 was previously available under the name 7993KVM_CAP_MANUAL_DIRTY_LOG_PROTECT, but the implementation had bugs that make 7994it hard or impossible to use it correctly. The availability of 7995KVM_CAP_MANUAL_DIRTY_LOG_PROTECT2 signals that those bugs are fixed. 7996Userspace should not try to use KVM_CAP_MANUAL_DIRTY_LOG_PROTECT. 7997 79987.19 KVM_CAP_PPC_SECURE_GUEST 7999------------------------------ 8000 8001:Architectures: ppc 8002 8003This capability indicates that KVM is running on a host that has 8004ultravisor firmware and thus can support a secure guest. On such a 8005system, a guest can ask the ultravisor to make it a secure guest, 8006one whose memory is inaccessible to the host except for pages which 8007are explicitly requested to be shared with the host. The ultravisor 8008notifies KVM when a guest requests to become a secure guest, and KVM 8009has the opportunity to veto the transition. 8010 8011If present, this capability can be enabled for a VM, meaning that KVM 8012will allow the transition to secure guest mode. Otherwise KVM will 8013veto the transition. 8014 80157.20 KVM_CAP_HALT_POLL 8016---------------------- 8017 8018:Architectures: all 8019:Target: VM 8020:Parameters: args[0] is the maximum poll time in nanoseconds 8021:Returns: 0 on success; -1 on error 8022 8023KVM_CAP_HALT_POLL overrides the kvm.halt_poll_ns module parameter to set the 8024maximum halt-polling time for all vCPUs in the target VM. This capability can 8025be invoked at any time and any number of times to dynamically change the 8026maximum halt-polling time. 8027 8028See Documentation/virt/kvm/halt-polling.rst for more information on halt 8029polling. 8030 80317.21 KVM_CAP_X86_USER_SPACE_MSR 8032------------------------------- 8033 8034:Architectures: x86 8035:Target: VM 8036:Parameters: args[0] contains the mask of KVM_MSR_EXIT_REASON_* events to report 8037:Returns: 0 on success; -1 on error 8038 8039This capability allows userspace to intercept RDMSR and WRMSR instructions if 8040access to an MSR is denied. By default, KVM injects #GP on denied accesses. 8041 8042When a guest requests to read or write an MSR, KVM may not implement all MSRs 8043that are relevant to a respective system. It also does not differentiate by 8044CPU type. 8045 8046To allow more fine grained control over MSR handling, userspace may enable 8047this capability. With it enabled, MSR accesses that match the mask specified in 8048args[0] and would trigger a #GP inside the guest will instead trigger 8049KVM_EXIT_X86_RDMSR and KVM_EXIT_X86_WRMSR exit notifications. Userspace 8050can then implement model specific MSR handling and/or user notifications 8051to inform a user that an MSR was not emulated/virtualized by KVM. 8052 8053The valid mask flags are: 8054 8055============================ =============================================== 8056 KVM_MSR_EXIT_REASON_UNKNOWN intercept accesses to unknown (to KVM) MSRs 8057 KVM_MSR_EXIT_REASON_INVAL intercept accesses that are architecturally 8058 invalid according to the vCPU model and/or mode 8059 KVM_MSR_EXIT_REASON_FILTER intercept accesses that are denied by userspace 8060 via KVM_X86_SET_MSR_FILTER 8061============================ =============================================== 8062 80637.22 KVM_CAP_X86_BUS_LOCK_EXIT 8064------------------------------- 8065 8066:Architectures: x86 8067:Target: VM 8068:Parameters: args[0] defines the policy used when bus locks detected in guest 8069:Returns: 0 on success, -EINVAL when args[0] contains invalid bits 8070 8071Valid bits in args[0] are:: 8072 8073 #define KVM_BUS_LOCK_DETECTION_OFF (1 << 0) 8074 #define KVM_BUS_LOCK_DETECTION_EXIT (1 << 1) 8075 8076Enabling this capability on a VM provides userspace with a way to select a 8077policy to handle the bus locks detected in guest. Userspace can obtain the 8078supported modes from the result of KVM_CHECK_EXTENSION and define it through 8079the KVM_ENABLE_CAP. The supported modes are mutually-exclusive. 8080 8081This capability allows userspace to force VM exits on bus locks detected in the 8082guest, irrespective whether or not the host has enabled split-lock detection 8083(which triggers an #AC exception that KVM intercepts). This capability is 8084intended to mitigate attacks where a malicious/buggy guest can exploit bus 8085locks to degrade the performance of the whole system. 8086 8087If KVM_BUS_LOCK_DETECTION_OFF is set, KVM doesn't force guest bus locks to VM 8088exit, although the host kernel's split-lock #AC detection still applies, if 8089enabled. 8090 8091If KVM_BUS_LOCK_DETECTION_EXIT is set, KVM enables a CPU feature that ensures 8092bus locks in the guest trigger a VM exit, and KVM exits to userspace for all 8093such VM exits, e.g. to allow userspace to throttle the offending guest and/or 8094apply some other policy-based mitigation. When exiting to userspace, KVM sets 8095KVM_RUN_X86_BUS_LOCK in vcpu-run->flags, and conditionally sets the exit_reason 8096to KVM_EXIT_X86_BUS_LOCK. 8097 8098Due to differences in the underlying hardware implementation, the vCPU's RIP at 8099the time of exit diverges between Intel and AMD. On Intel hosts, RIP points at 8100the next instruction, i.e. the exit is trap-like. On AMD hosts, RIP points at 8101the offending instruction, i.e. the exit is fault-like. 8102 8103Note! Detected bus locks may be coincident with other exits to userspace, i.e. 8104KVM_RUN_X86_BUS_LOCK should be checked regardless of the primary exit reason if 8105userspace wants to take action on all detected bus locks. 8106 81077.23 KVM_CAP_PPC_DAWR1 8108---------------------- 8109 8110:Architectures: ppc 8111:Parameters: none 8112:Returns: 0 on success, -EINVAL when CPU doesn't support 2nd DAWR 8113 8114This capability can be used to check / enable 2nd DAWR feature provided 8115by POWER10 processor. 8116 8117 81187.24 KVM_CAP_VM_COPY_ENC_CONTEXT_FROM 8119------------------------------------- 8120 8121:Architectures: x86 SEV enabled 8122:Type: vm 8123:Parameters: args[0] is the fd of the source vm 8124:Returns: 0 on success; ENOTTY on error 8125 8126This capability enables userspace to copy encryption context from the vm 8127indicated by the fd to the vm this is called on. 8128 8129This is intended to support in-guest workloads scheduled by the host. This 8130allows the in-guest workload to maintain its own NPTs and keeps the two vms 8131from accidentally clobbering each other with interrupts and the like (separate 8132APIC/MSRs/etc). 8133 81347.25 KVM_CAP_SGX_ATTRIBUTE 8135-------------------------- 8136 8137:Architectures: x86 8138:Target: VM 8139:Parameters: args[0] is a file handle of a SGX attribute file in securityfs 8140:Returns: 0 on success, -EINVAL if the file handle is invalid or if a requested 8141 attribute is not supported by KVM. 8142 8143KVM_CAP_SGX_ATTRIBUTE enables a userspace VMM to grant a VM access to one or 8144more privileged enclave attributes. args[0] must hold a file handle to a valid 8145SGX attribute file corresponding to an attribute that is supported/restricted 8146by KVM (currently only PROVISIONKEY). 8147 8148The SGX subsystem restricts access to a subset of enclave attributes to provide 8149additional security for an uncompromised kernel, e.g. use of the PROVISIONKEY 8150is restricted to deter malware from using the PROVISIONKEY to obtain a stable 8151system fingerprint. To prevent userspace from circumventing such restrictions 8152by running an enclave in a VM, KVM prevents access to privileged attributes by 8153default. 8154 8155See Documentation/arch/x86/sgx.rst for more details. 8156 81577.27 KVM_CAP_EXIT_ON_EMULATION_FAILURE 8158-------------------------------------- 8159 8160:Architectures: x86 8161:Parameters: args[0] whether the feature should be enabled or not 8162 8163When this capability is enabled, an emulation failure will result in an exit 8164to userspace with KVM_INTERNAL_ERROR (except when the emulator was invoked 8165to handle a VMware backdoor instruction). Furthermore, KVM will now provide up 8166to 15 instruction bytes for any exit to userspace resulting from an emulation 8167failure. When these exits to userspace occur use the emulation_failure struct 8168instead of the internal struct. They both have the same layout, but the 8169emulation_failure struct matches the content better. It also explicitly 8170defines the 'flags' field which is used to describe the fields in the struct 8171that are valid (ie: if KVM_INTERNAL_ERROR_EMULATION_FLAG_INSTRUCTION_BYTES is 8172set in the 'flags' field then both 'insn_size' and 'insn_bytes' have valid data 8173in them.) 8174 81757.28 KVM_CAP_ARM_MTE 8176-------------------- 8177 8178:Architectures: arm64 8179:Parameters: none 8180 8181This capability indicates that KVM (and the hardware) supports exposing the 8182Memory Tagging Extensions (MTE) to the guest. It must also be enabled by the 8183VMM before creating any VCPUs to allow the guest access. Note that MTE is only 8184available to a guest running in AArch64 mode and enabling this capability will 8185cause attempts to create AArch32 VCPUs to fail. 8186 8187When enabled the guest is able to access tags associated with any memory given 8188to the guest. KVM will ensure that the tags are maintained during swap or 8189hibernation of the host; however the VMM needs to manually save/restore the 8190tags as appropriate if the VM is migrated. 8191 8192When this capability is enabled all memory in memslots must be mapped as 8193``MAP_ANONYMOUS`` or with a RAM-based file mapping (``tmpfs``, ``memfd``), 8194attempts to create a memslot with an invalid mmap will result in an 8195-EINVAL return. 8196 8197When enabled the VMM may make use of the ``KVM_ARM_MTE_COPY_TAGS`` ioctl to 8198perform a bulk copy of tags to/from the guest. 8199 82007.29 KVM_CAP_VM_MOVE_ENC_CONTEXT_FROM 8201------------------------------------- 8202 8203:Architectures: x86 SEV enabled 8204:Type: vm 8205:Parameters: args[0] is the fd of the source vm 8206:Returns: 0 on success 8207 8208This capability enables userspace to migrate the encryption context from the VM 8209indicated by the fd to the VM this is called on. 8210 8211This is intended to support intra-host migration of VMs between userspace VMMs, 8212upgrading the VMM process without interrupting the guest. 8213 82147.31 KVM_CAP_DISABLE_QUIRKS2 8215---------------------------- 8216 8217:Parameters: args[0] - set of KVM quirks to disable 8218:Architectures: x86 8219:Type: vm 8220 8221This capability, if enabled, will cause KVM to disable some behavior 8222quirks. 8223 8224Calling KVM_CHECK_EXTENSION for this capability returns a bitmask of 8225quirks that can be disabled in KVM. 8226 8227The argument to KVM_ENABLE_CAP for this capability is a bitmask of 8228quirks to disable, and must be a subset of the bitmask returned by 8229KVM_CHECK_EXTENSION. 8230 8231The valid bits in cap.args[0] are: 8232 8233=================================== ============================================ 8234 KVM_X86_QUIRK_LINT0_REENABLED By default, the reset value for the LVT 8235 LINT0 register is 0x700 (APIC_MODE_EXTINT). 8236 When this quirk is disabled, the reset value 8237 is 0x10000 (APIC_LVT_MASKED). 8238 8239 KVM_X86_QUIRK_CD_NW_CLEARED By default, KVM clears CR0.CD and CR0.NW on 8240 AMD CPUs to workaround buggy guest firmware 8241 that runs in perpetuity with CR0.CD, i.e. 8242 with caches in "no fill" mode. 8243 8244 When this quirk is disabled, KVM does not 8245 change the value of CR0.CD and CR0.NW. 8246 8247 KVM_X86_QUIRK_LAPIC_MMIO_HOLE By default, the MMIO LAPIC interface is 8248 available even when configured for x2APIC 8249 mode. When this quirk is disabled, KVM 8250 disables the MMIO LAPIC interface if the 8251 LAPIC is in x2APIC mode. 8252 8253 KVM_X86_QUIRK_OUT_7E_INC_RIP By default, KVM pre-increments %rip before 8254 exiting to userspace for an OUT instruction 8255 to port 0x7e. When this quirk is disabled, 8256 KVM does not pre-increment %rip before 8257 exiting to userspace. 8258 8259 KVM_X86_QUIRK_MISC_ENABLE_NO_MWAIT When this quirk is disabled, KVM sets 8260 CPUID.01H:ECX[bit 3] (MONITOR/MWAIT) if 8261 IA32_MISC_ENABLE[bit 18] (MWAIT) is set. 8262 Additionally, when this quirk is disabled, 8263 KVM clears CPUID.01H:ECX[bit 3] if 8264 IA32_MISC_ENABLE[bit 18] is cleared. 8265 8266 KVM_X86_QUIRK_FIX_HYPERCALL_INSN By default, KVM rewrites guest 8267 VMMCALL/VMCALL instructions to match the 8268 vendor's hypercall instruction for the 8269 system. When this quirk is disabled, KVM 8270 will no longer rewrite invalid guest 8271 hypercall instructions. Executing the 8272 incorrect hypercall instruction will 8273 generate a #UD within the guest. 8274 8275KVM_X86_QUIRK_MWAIT_NEVER_UD_FAULTS By default, KVM emulates MONITOR/MWAIT (if 8276 they are intercepted) as NOPs regardless of 8277 whether or not MONITOR/MWAIT are supported 8278 according to guest CPUID. When this quirk 8279 is disabled and KVM_X86_DISABLE_EXITS_MWAIT 8280 is not set (MONITOR/MWAIT are intercepted), 8281 KVM will inject a #UD on MONITOR/MWAIT if 8282 they're unsupported per guest CPUID. Note, 8283 KVM will modify MONITOR/MWAIT support in 8284 guest CPUID on writes to MISC_ENABLE if 8285 KVM_X86_QUIRK_MISC_ENABLE_NO_MWAIT is 8286 disabled. 8287 8288KVM_X86_QUIRK_SLOT_ZAP_ALL By default, for KVM_X86_DEFAULT_VM VMs, KVM 8289 invalidates all SPTEs in all memslots and 8290 address spaces when a memslot is deleted or 8291 moved. When this quirk is disabled (or the 8292 VM type isn't KVM_X86_DEFAULT_VM), KVM only 8293 ensures the backing memory of the deleted 8294 or moved memslot isn't reachable, i.e KVM 8295 _may_ invalidate only SPTEs related to the 8296 memslot. 8297 8298KVM_X86_QUIRK_STUFF_FEATURE_MSRS By default, at vCPU creation, KVM sets the 8299 vCPU's MSR_IA32_PERF_CAPABILITIES (0x345), 8300 MSR_IA32_ARCH_CAPABILITIES (0x10a), 8301 MSR_PLATFORM_INFO (0xce), and all VMX MSRs 8302 (0x480..0x492) to the maximal capabilities 8303 supported by KVM. KVM also sets 8304 MSR_IA32_UCODE_REV (0x8b) to an arbitrary 8305 value (which is different for Intel vs. 8306 AMD). Lastly, when guest CPUID is set (by 8307 userspace), KVM modifies select VMX MSR 8308 fields to force consistency between guest 8309 CPUID and L2's effective ISA. When this 8310 quirk is disabled, KVM zeroes the vCPU's MSR 8311 values (with two exceptions, see below), 8312 i.e. treats the feature MSRs like CPUID 8313 leaves and gives userspace full control of 8314 the vCPU model definition. This quirk does 8315 not affect VMX MSRs CR0/CR4_FIXED1 (0x487 8316 and 0x489), as KVM does now allow them to 8317 be set by userspace (KVM sets them based on 8318 guest CPUID, for safety purposes). 8319 8320KVM_X86_QUIRK_IGNORE_GUEST_PAT By default, on Intel platforms, KVM ignores 8321 guest PAT and forces the effective memory 8322 type to WB in EPT. The quirk is not available 8323 on Intel platforms which are incapable of 8324 safely honoring guest PAT (i.e., without CPU 8325 self-snoop, KVM always ignores guest PAT and 8326 forces effective memory type to WB). It is 8327 also ignored on AMD platforms or, on Intel, 8328 when a VM has non-coherent DMA devices 8329 assigned; KVM always honors guest PAT in 8330 such case. The quirk is needed to avoid 8331 slowdowns on certain Intel Xeon platforms 8332 (e.g. ICX, SPR) where self-snoop feature is 8333 supported but UC is slow enough to cause 8334 issues with some older guests that use 8335 UC instead of WC to map the video RAM. 8336 Userspace can disable the quirk to honor 8337 guest PAT if it knows that there is no such 8338 guest software, for example if it does not 8339 expose a bochs graphics device (which is 8340 known to have had a buggy driver). 8341=================================== ============================================ 8342 83437.32 KVM_CAP_MAX_VCPU_ID 8344------------------------ 8345 8346:Architectures: x86 8347:Target: VM 8348:Parameters: args[0] - maximum APIC ID value set for current VM 8349:Returns: 0 on success, -EINVAL if args[0] is beyond KVM_MAX_VCPU_IDS 8350 supported in KVM or if it has been set. 8351 8352This capability allows userspace to specify maximum possible APIC ID 8353assigned for current VM session prior to the creation of vCPUs, saving 8354memory for data structures indexed by the APIC ID. Userspace is able 8355to calculate the limit to APIC ID values from designated 8356CPU topology. 8357 8358The value can be changed only until KVM_ENABLE_CAP is set to a nonzero 8359value or until a vCPU is created. Upon creation of the first vCPU, 8360if the value was set to zero or KVM_ENABLE_CAP was not invoked, KVM 8361uses the return value of KVM_CHECK_EXTENSION(KVM_CAP_MAX_VCPU_ID) as 8362the maximum APIC ID. 8363 83647.33 KVM_CAP_X86_NOTIFY_VMEXIT 8365------------------------------ 8366 8367:Architectures: x86 8368:Target: VM 8369:Parameters: args[0] is the value of notify window as well as some flags 8370:Returns: 0 on success, -EINVAL if args[0] contains invalid flags or notify 8371 VM exit is unsupported. 8372 8373Bits 63:32 of args[0] are used for notify window. 8374Bits 31:0 of args[0] are for some flags. Valid bits are:: 8375 8376 #define KVM_X86_NOTIFY_VMEXIT_ENABLED (1 << 0) 8377 #define KVM_X86_NOTIFY_VMEXIT_USER (1 << 1) 8378 8379This capability allows userspace to configure the notify VM exit on/off 8380in per-VM scope during VM creation. Notify VM exit is disabled by default. 8381When userspace sets KVM_X86_NOTIFY_VMEXIT_ENABLED bit in args[0], VMM will 8382enable this feature with the notify window provided, which will generate 8383a VM exit if no event window occurs in VM non-root mode for a specified of 8384time (notify window). 8385 8386If KVM_X86_NOTIFY_VMEXIT_USER is set in args[0], upon notify VM exits happen, 8387KVM would exit to userspace for handling. 8388 8389This capability is aimed to mitigate the threat that malicious VMs can 8390cause CPU stuck (due to event windows don't open up) and make the CPU 8391unavailable to host or other VMs. 8392 83937.35 KVM_CAP_X86_APIC_BUS_CYCLES_NS 8394----------------------------------- 8395 8396:Architectures: x86 8397:Target: VM 8398:Parameters: args[0] is the desired APIC bus clock rate, in nanoseconds 8399:Returns: 0 on success, -EINVAL if args[0] contains an invalid value for the 8400 frequency or if any vCPUs have been created, -ENXIO if a virtual 8401 local APIC has not been created using KVM_CREATE_IRQCHIP. 8402 8403This capability sets the VM's APIC bus clock frequency, used by KVM's in-kernel 8404virtual APIC when emulating APIC timers. KVM's default value can be retrieved 8405by KVM_CHECK_EXTENSION. 8406 8407Note: Userspace is responsible for correctly configuring CPUID 0x15, a.k.a. the 8408core crystal clock frequency, if a non-zero CPUID 0x15 is exposed to the guest. 8409 84107.36 KVM_CAP_DIRTY_LOG_RING/KVM_CAP_DIRTY_LOG_RING_ACQ_REL 8411---------------------------------------------------------- 8412 8413:Architectures: x86, arm64, riscv 8414:Type: vm 8415:Parameters: args[0] - size of the dirty log ring 8416 8417KVM is capable of tracking dirty memory using ring buffers that are 8418mmapped into userspace; there is one dirty ring per vcpu. 8419 8420The dirty ring is available to userspace as an array of 8421``struct kvm_dirty_gfn``. Each dirty entry is defined as:: 8422 8423 struct kvm_dirty_gfn { 8424 __u32 flags; 8425 __u32 slot; /* as_id | slot_id */ 8426 __u64 offset; 8427 }; 8428 8429The following values are defined for the flags field to define the 8430current state of the entry:: 8431 8432 #define KVM_DIRTY_GFN_F_DIRTY BIT(0) 8433 #define KVM_DIRTY_GFN_F_RESET BIT(1) 8434 #define KVM_DIRTY_GFN_F_MASK 0x3 8435 8436Userspace should call KVM_ENABLE_CAP ioctl right after KVM_CREATE_VM 8437ioctl to enable this capability for the new guest and set the size of 8438the rings. Enabling the capability is only allowed before creating any 8439vCPU, and the size of the ring must be a power of two. The larger the 8440ring buffer, the less likely the ring is full and the VM is forced to 8441exit to userspace. The optimal size depends on the workload, but it is 8442recommended that it be at least 64 KiB (4096 entries). 8443 8444Just like for dirty page bitmaps, the buffer tracks writes to 8445all user memory regions for which the KVM_MEM_LOG_DIRTY_PAGES flag was 8446set in KVM_SET_USER_MEMORY_REGION. Once a memory region is registered 8447with the flag set, userspace can start harvesting dirty pages from the 8448ring buffer. 8449 8450An entry in the ring buffer can be unused (flag bits ``00``), 8451dirty (flag bits ``01``) or harvested (flag bits ``1X``). The 8452state machine for the entry is as follows:: 8453 8454 dirtied harvested reset 8455 00 -----------> 01 -------------> 1X -------+ 8456 ^ | 8457 | | 8458 +------------------------------------------+ 8459 8460To harvest the dirty pages, userspace accesses the mmapped ring buffer 8461to read the dirty GFNs. If the flags has the DIRTY bit set (at this stage 8462the RESET bit must be cleared), then it means this GFN is a dirty GFN. 8463The userspace should harvest this GFN and mark the flags from state 8464``01b`` to ``1Xb`` (bit 0 will be ignored by KVM, but bit 1 must be set 8465to show that this GFN is harvested and waiting for a reset), and move 8466on to the next GFN. The userspace should continue to do this until the 8467flags of a GFN have the DIRTY bit cleared, meaning that it has harvested 8468all the dirty GFNs that were available. 8469 8470Note that on weakly ordered architectures, userspace accesses to the 8471ring buffer (and more specifically the 'flags' field) must be ordered, 8472using load-acquire/store-release accessors when available, or any 8473other memory barrier that will ensure this ordering. 8474 8475It's not necessary for userspace to harvest the all dirty GFNs at once. 8476However it must collect the dirty GFNs in sequence, i.e., the userspace 8477program cannot skip one dirty GFN to collect the one next to it. 8478 8479After processing one or more entries in the ring buffer, userspace 8480calls the VM ioctl KVM_RESET_DIRTY_RINGS to notify the kernel about 8481it, so that the kernel will reprotect those collected GFNs. 8482Therefore, the ioctl must be called *before* reading the content of 8483the dirty pages. 8484 8485The dirty ring can get full. When it happens, the KVM_RUN of the 8486vcpu will return with exit reason KVM_EXIT_DIRTY_LOG_FULL. 8487 8488The dirty ring interface has a major difference comparing to the 8489KVM_GET_DIRTY_LOG interface in that, when reading the dirty ring from 8490userspace, it's still possible that the kernel has not yet flushed the 8491processor's dirty page buffers into the kernel buffer (with dirty bitmaps, the 8492flushing is done by the KVM_GET_DIRTY_LOG ioctl). To achieve that, one 8493needs to kick the vcpu out of KVM_RUN using a signal. The resulting 8494vmexit ensures that all dirty GFNs are flushed to the dirty rings. 8495 8496NOTE: KVM_CAP_DIRTY_LOG_RING_ACQ_REL is the only capability that 8497should be exposed by weakly ordered architecture, in order to indicate 8498the additional memory ordering requirements imposed on userspace when 8499reading the state of an entry and mutating it from DIRTY to HARVESTED. 8500Architecture with TSO-like ordering (such as x86) are allowed to 8501expose both KVM_CAP_DIRTY_LOG_RING and KVM_CAP_DIRTY_LOG_RING_ACQ_REL 8502to userspace. 8503 8504After enabling the dirty rings, the userspace needs to detect the 8505capability of KVM_CAP_DIRTY_LOG_RING_WITH_BITMAP to see whether the 8506ring structures can be backed by per-slot bitmaps. With this capability 8507advertised, it means the architecture can dirty guest pages without 8508vcpu/ring context, so that some of the dirty information will still be 8509maintained in the bitmap structure. KVM_CAP_DIRTY_LOG_RING_WITH_BITMAP 8510can't be enabled if the capability of KVM_CAP_DIRTY_LOG_RING_ACQ_REL 8511hasn't been enabled, or any memslot has been existing. 8512 8513Note that the bitmap here is only a backup of the ring structure. The 8514use of the ring and bitmap combination is only beneficial if there is 8515only a very small amount of memory that is dirtied out of vcpu/ring 8516context. Otherwise, the stand-alone per-slot bitmap mechanism needs to 8517be considered. 8518 8519To collect dirty bits in the backup bitmap, userspace can use the same 8520KVM_GET_DIRTY_LOG ioctl. KVM_CLEAR_DIRTY_LOG isn't needed as long as all 8521the generation of the dirty bits is done in a single pass. Collecting 8522the dirty bitmap should be the very last thing that the VMM does before 8523considering the state as complete. VMM needs to ensure that the dirty 8524state is final and avoid missing dirty pages from another ioctl ordered 8525after the bitmap collection. 8526 8527NOTE: Multiple examples of using the backup bitmap: (1) save vgic/its 8528tables through command KVM_DEV_ARM_{VGIC_GRP_CTRL, ITS_SAVE_TABLES} on 8529KVM device "kvm-arm-vgic-its". (2) restore vgic/its tables through 8530command KVM_DEV_ARM_{VGIC_GRP_CTRL, ITS_RESTORE_TABLES} on KVM device 8531"kvm-arm-vgic-its". VGICv3 LPI pending status is restored. (3) save 8532vgic3 pending table through KVM_DEV_ARM_VGIC_{GRP_CTRL, SAVE_PENDING_TABLES} 8533command on KVM device "kvm-arm-vgic-v3". 8534 85357.37 KVM_CAP_PMU_CAPABILITY 8536--------------------------- 8537 8538:Architectures: x86 8539:Type: vm 8540:Parameters: arg[0] is bitmask of PMU virtualization capabilities. 8541:Returns: 0 on success, -EINVAL when arg[0] contains invalid bits 8542 8543This capability alters PMU virtualization in KVM. 8544 8545Calling KVM_CHECK_EXTENSION for this capability returns a bitmask of 8546PMU virtualization capabilities that can be adjusted on a VM. 8547 8548The argument to KVM_ENABLE_CAP is also a bitmask and selects specific 8549PMU virtualization capabilities to be applied to the VM. This can 8550only be invoked on a VM prior to the creation of VCPUs. 8551 8552At this time, KVM_PMU_CAP_DISABLE is the only capability. Setting 8553this capability will disable PMU virtualization for that VM. Usermode 8554should adjust CPUID leaf 0xA to reflect that the PMU is disabled. 8555 85567.38 KVM_CAP_VM_DISABLE_NX_HUGE_PAGES 8557------------------------------------- 8558 8559:Architectures: x86 8560:Type: vm 8561:Parameters: arg[0] must be 0. 8562:Returns: 0 on success, -EPERM if the userspace process does not 8563 have CAP_SYS_BOOT, -EINVAL if args[0] is not 0 or any vCPUs have been 8564 created. 8565 8566This capability disables the NX huge pages mitigation for iTLB MULTIHIT. 8567 8568The capability has no effect if the nx_huge_pages module parameter is not set. 8569 8570This capability may only be set before any vCPUs are created. 8571 85727.39 KVM_CAP_ARM_EAGER_SPLIT_CHUNK_SIZE 8573--------------------------------------- 8574 8575:Architectures: arm64 8576:Type: vm 8577:Parameters: arg[0] is the new split chunk size. 8578:Returns: 0 on success, -EINVAL if any memslot was already created. 8579 8580This capability sets the chunk size used in Eager Page Splitting. 8581 8582Eager Page Splitting improves the performance of dirty-logging (used 8583in live migrations) when guest memory is backed by huge-pages. It 8584avoids splitting huge-pages (into PAGE_SIZE pages) on fault, by doing 8585it eagerly when enabling dirty logging (with the 8586KVM_MEM_LOG_DIRTY_PAGES flag for a memory region), or when using 8587KVM_CLEAR_DIRTY_LOG. 8588 8589The chunk size specifies how many pages to break at a time, using a 8590single allocation for each chunk. Bigger the chunk size, more pages 8591need to be allocated ahead of time. 8592 8593The chunk size needs to be a valid block size. The list of acceptable 8594block sizes is exposed in KVM_CAP_ARM_SUPPORTED_BLOCK_SIZES as a 859564-bit bitmap (each bit describing a block size). The default value is 85960, to disable the eager page splitting. 8597 85987.40 KVM_CAP_EXIT_HYPERCALL 8599--------------------------- 8600 8601:Architectures: x86 8602:Type: vm 8603 8604This capability, if enabled, will cause KVM to exit to userspace 8605with KVM_EXIT_HYPERCALL exit reason to process some hypercalls. 8606 8607Calling KVM_CHECK_EXTENSION for this capability will return a bitmask 8608of hypercalls that can be configured to exit to userspace. 8609Right now, the only such hypercall is KVM_HC_MAP_GPA_RANGE. 8610 8611The argument to KVM_ENABLE_CAP is also a bitmask, and must be a subset 8612of the result of KVM_CHECK_EXTENSION. KVM will forward to userspace 8613the hypercalls whose corresponding bit is in the argument, and return 8614ENOSYS for the others. 8615 86167.41 KVM_CAP_ARM_SYSTEM_SUSPEND 8617------------------------------- 8618 8619:Architectures: arm64 8620:Type: vm 8621 8622When enabled, KVM will exit to userspace with KVM_EXIT_SYSTEM_EVENT of 8623type KVM_SYSTEM_EVENT_SUSPEND to process the guest suspend request. 8624 86257.42 KVM_CAP_ARM_WRITABLE_IMP_ID_REGS 8626------------------------------------- 8627 8628:Architectures: arm64 8629:Target: VM 8630:Parameters: None 8631:Returns: 0 on success, -EINVAL if vCPUs have been created before enabling this 8632 capability. 8633 8634This capability changes the behavior of the registers that identify a PE 8635implementation of the Arm architecture: MIDR_EL1, REVIDR_EL1, and AIDR_EL1. 8636By default, these registers are visible to userspace but treated as invariant. 8637 8638When this capability is enabled, KVM allows userspace to change the 8639aforementioned registers before the first KVM_RUN. These registers are VM 8640scoped, meaning that the same set of values are presented on all vCPUs in a 8641given VM. 8642 86437.43 KVM_CAP_RISCV_MP_STATE_RESET 8644--------------------------------- 8645 8646:Architectures: riscv 8647:Type: VM 8648:Parameters: None 8649:Returns: 0 on success, -EINVAL if arg[0] is not zero 8650 8651When this capability is enabled, KVM resets the VCPU when setting 8652MP_STATE_INIT_RECEIVED through IOCTL. The original MP_STATE is preserved. 8653 86547.43 KVM_CAP_ARM_CACHEABLE_PFNMAP_SUPPORTED 8655------------------------------------------- 8656 8657:Architectures: arm64 8658:Target: VM 8659:Parameters: None 8660 8661This capability indicate to the userspace whether a PFNMAP memory region 8662can be safely mapped as cacheable. This relies on the presence of 8663force write back (FWB) feature support on the hardware. 8664 86658. Other capabilities. 8666====================== 8667 8668This section lists capabilities that give information about other 8669features of the KVM implementation. 8670 86718.1 KVM_CAP_PPC_HWRNG 8672--------------------- 8673 8674:Architectures: ppc 8675 8676This capability, if KVM_CHECK_EXTENSION indicates that it is 8677available, means that the kernel has an implementation of the 8678H_RANDOM hypercall backed by a hardware random-number generator. 8679If present, the kernel H_RANDOM handler can be enabled for guest use 8680with the KVM_CAP_PPC_ENABLE_HCALL capability. 8681 86828.3 KVM_CAP_PPC_MMU_RADIX 8683------------------------- 8684 8685:Architectures: ppc 8686 8687This capability, if KVM_CHECK_EXTENSION indicates that it is 8688available, means that the kernel can support guests using the 8689radix MMU defined in Power ISA V3.00 (as implemented in the POWER9 8690processor). 8691 86928.4 KVM_CAP_PPC_MMU_HASH_V3 8693--------------------------- 8694 8695:Architectures: ppc 8696 8697This capability, if KVM_CHECK_EXTENSION indicates that it is 8698available, means that the kernel can support guests using the 8699hashed page table MMU defined in Power ISA V3.00 (as implemented in 8700the POWER9 processor), including in-memory segment tables. 8701 87028.5 KVM_CAP_MIPS_VZ 8703------------------- 8704 8705:Architectures: mips 8706 8707This capability, if KVM_CHECK_EXTENSION on the main kvm handle indicates that 8708it is available, means that full hardware assisted virtualization capabilities 8709of the hardware are available for use through KVM. An appropriate 8710KVM_VM_MIPS_* type must be passed to KVM_CREATE_VM to create a VM which 8711utilises it. 8712 8713If KVM_CHECK_EXTENSION on a kvm VM handle indicates that this capability is 8714available, it means that the VM is using full hardware assisted virtualization 8715capabilities of the hardware. This is useful to check after creating a VM with 8716KVM_VM_MIPS_DEFAULT. 8717 8718The value returned by KVM_CHECK_EXTENSION should be compared against known 8719values (see below). All other values are reserved. This is to allow for the 8720possibility of other hardware assisted virtualization implementations which 8721may be incompatible with the MIPS VZ ASE. 8722 8723== ========================================================================== 8724 0 The trap & emulate implementation is in use to run guest code in user 8725 mode. Guest virtual memory segments are rearranged to fit the guest in the 8726 user mode address space. 8727 8728 1 The MIPS VZ ASE is in use, providing full hardware assisted 8729 virtualization, including standard guest virtual memory segments. 8730== ========================================================================== 8731 87328.7 KVM_CAP_MIPS_64BIT 8733---------------------- 8734 8735:Architectures: mips 8736 8737This capability indicates the supported architecture type of the guest, i.e. the 8738supported register and address width. 8739 8740The values returned when this capability is checked by KVM_CHECK_EXTENSION on a 8741kvm VM handle correspond roughly to the CP0_Config.AT register field, and should 8742be checked specifically against known values (see below). All other values are 8743reserved. 8744 8745== ======================================================================== 8746 0 MIPS32 or microMIPS32. 8747 Both registers and addresses are 32-bits wide. 8748 It will only be possible to run 32-bit guest code. 8749 8750 1 MIPS64 or microMIPS64 with access only to 32-bit compatibility segments. 8751 Registers are 64-bits wide, but addresses are 32-bits wide. 8752 64-bit guest code may run but cannot access MIPS64 memory segments. 8753 It will also be possible to run 32-bit guest code. 8754 8755 2 MIPS64 or microMIPS64 with access to all address segments. 8756 Both registers and addresses are 64-bits wide. 8757 It will be possible to run 64-bit or 32-bit guest code. 8758== ======================================================================== 8759 87608.9 KVM_CAP_ARM_USER_IRQ 8761------------------------ 8762 8763:Architectures: arm64 8764 8765This capability, if KVM_CHECK_EXTENSION indicates that it is available, means 8766that if userspace creates a VM without an in-kernel interrupt controller, it 8767will be notified of changes to the output level of in-kernel emulated devices, 8768which can generate virtual interrupts, presented to the VM. 8769For such VMs, on every return to userspace, the kernel 8770updates the vcpu's run->s.regs.device_irq_level field to represent the actual 8771output level of the device. 8772 8773Whenever kvm detects a change in the device output level, kvm guarantees at 8774least one return to userspace before running the VM. This exit could either 8775be a KVM_EXIT_INTR or any other exit event, like KVM_EXIT_MMIO. This way, 8776userspace can always sample the device output level and re-compute the state of 8777the userspace interrupt controller. Userspace should always check the state 8778of run->s.regs.device_irq_level on every kvm exit. 8779The value in run->s.regs.device_irq_level can represent both level and edge 8780triggered interrupt signals, depending on the device. Edge triggered interrupt 8781signals will exit to userspace with the bit in run->s.regs.device_irq_level 8782set exactly once per edge signal. 8783 8784The field run->s.regs.device_irq_level is available independent of 8785run->kvm_valid_regs or run->kvm_dirty_regs bits. 8786 8787If KVM_CAP_ARM_USER_IRQ is supported, the KVM_CHECK_EXTENSION ioctl returns a 8788number larger than 0 indicating the version of this capability is implemented 8789and thereby which bits in run->s.regs.device_irq_level can signal values. 8790 8791Currently the following bits are defined for the device_irq_level bitmap:: 8792 8793 KVM_CAP_ARM_USER_IRQ >= 1: 8794 8795 KVM_ARM_DEV_EL1_VTIMER - EL1 virtual timer 8796 KVM_ARM_DEV_EL1_PTIMER - EL1 physical timer 8797 KVM_ARM_DEV_PMU - ARM PMU overflow interrupt signal 8798 8799Future versions of kvm may implement additional events. These will get 8800indicated by returning a higher number from KVM_CHECK_EXTENSION and will be 8801listed above. 8802 88038.10 KVM_CAP_PPC_SMT_POSSIBLE 8804----------------------------- 8805 8806:Architectures: ppc 8807 8808Querying this capability returns a bitmap indicating the possible 8809virtual SMT modes that can be set using KVM_CAP_PPC_SMT. If bit N 8810(counting from the right) is set, then a virtual SMT mode of 2^N is 8811available. 8812 88138.12 KVM_CAP_HYPERV_VP_INDEX 8814---------------------------- 8815 8816:Architectures: x86 8817 8818This capability indicates that userspace can load HV_X64_MSR_VP_INDEX msr. Its 8819value is used to denote the target vcpu for a SynIC interrupt. For 8820compatibility, KVM initializes this msr to KVM's internal vcpu index. When this 8821capability is absent, userspace can still query this msr's value. 8822 88238.13 KVM_CAP_S390_AIS_MIGRATION 8824------------------------------- 8825 8826:Architectures: s390 8827 8828This capability indicates if the flic device will be able to get/set the 8829AIS states for migration via the KVM_DEV_FLIC_AISM_ALL attribute and allows 8830to discover this without having to create a flic device. 8831 88328.14 KVM_CAP_S390_PSW 8833--------------------- 8834 8835:Architectures: s390 8836 8837This capability indicates that the PSW is exposed via the kvm_run structure. 8838 88398.15 KVM_CAP_S390_GMAP 8840---------------------- 8841 8842:Architectures: s390 8843 8844This capability indicates that the user space memory used as guest mapping can 8845be anywhere in the user memory address space, as long as the memory slots are 8846aligned and sized to a segment (1MB) boundary. 8847 88488.16 KVM_CAP_S390_COW 8849--------------------- 8850 8851:Architectures: s390 8852 8853This capability indicates that the user space memory used as guest mapping can 8854use copy-on-write semantics as well as dirty pages tracking via read-only page 8855tables. 8856 88578.17 KVM_CAP_S390_BPB 8858--------------------- 8859 8860:Architectures: s390 8861 8862This capability indicates that kvm will implement the interfaces to handle 8863reset, migration and nested KVM for branch prediction blocking. The stfle 8864facility 82 should not be provided to the guest without this capability. 8865 88668.18 KVM_CAP_HYPERV_TLBFLUSH 8867---------------------------- 8868 8869:Architectures: x86 8870 8871This capability indicates that KVM supports paravirtualized Hyper-V TLB Flush 8872hypercalls: 8873HvFlushVirtualAddressSpace, HvFlushVirtualAddressSpaceEx, 8874HvFlushVirtualAddressList, HvFlushVirtualAddressListEx. 8875 88768.19 KVM_CAP_ARM_INJECT_SERROR_ESR 8877---------------------------------- 8878 8879:Architectures: arm64 8880 8881This capability indicates that userspace can specify (via the 8882KVM_SET_VCPU_EVENTS ioctl) the syndrome value reported to the guest when it 8883takes a virtual SError interrupt exception. 8884If KVM advertises this capability, userspace can only specify the ISS field for 8885the ESR syndrome. Other parts of the ESR, such as the EC are generated by the 8886CPU when the exception is taken. If this virtual SError is taken to EL1 using 8887AArch64, this value will be reported in the ISS field of ESR_ELx. 8888 8889See KVM_CAP_VCPU_EVENTS for more details. 8890 88918.20 KVM_CAP_HYPERV_SEND_IPI 8892---------------------------- 8893 8894:Architectures: x86 8895 8896This capability indicates that KVM supports paravirtualized Hyper-V IPI send 8897hypercalls: 8898HvCallSendSyntheticClusterIpi, HvCallSendSyntheticClusterIpiEx. 8899 89008.22 KVM_CAP_S390_VCPU_RESETS 8901----------------------------- 8902 8903:Architectures: s390 8904 8905This capability indicates that the KVM_S390_NORMAL_RESET and 8906KVM_S390_CLEAR_RESET ioctls are available. 8907 89088.23 KVM_CAP_S390_PROTECTED 8909--------------------------- 8910 8911:Architectures: s390 8912 8913This capability indicates that the Ultravisor has been initialized and 8914KVM can therefore start protected VMs. 8915This capability governs the KVM_S390_PV_COMMAND ioctl and the 8916KVM_MP_STATE_LOAD MP_STATE. KVM_SET_MP_STATE can fail for protected 8917guests when the state change is invalid. 8918 89198.24 KVM_CAP_STEAL_TIME 8920----------------------- 8921 8922:Architectures: arm64, x86 8923 8924This capability indicates that KVM supports steal time accounting. 8925When steal time accounting is supported it may be enabled with 8926architecture-specific interfaces. This capability and the architecture- 8927specific interfaces must be consistent, i.e. if one says the feature 8928is supported, than the other should as well and vice versa. For arm64 8929see Documentation/virt/kvm/devices/vcpu.rst "KVM_ARM_VCPU_PVTIME_CTRL". 8930For x86 see Documentation/virt/kvm/x86/msr.rst "MSR_KVM_STEAL_TIME". 8931 89328.25 KVM_CAP_S390_DIAG318 8933------------------------- 8934 8935:Architectures: s390 8936 8937This capability enables a guest to set information about its control program 8938(i.e. guest kernel type and version). The information is helpful during 8939system/firmware service events, providing additional data about the guest 8940environments running on the machine. 8941 8942The information is associated with the DIAGNOSE 0x318 instruction, which sets 8943an 8-byte value consisting of a one-byte Control Program Name Code (CPNC) and 8944a 7-byte Control Program Version Code (CPVC). The CPNC determines what 8945environment the control program is running in (e.g. Linux, z/VM...), and the 8946CPVC is used for information specific to OS (e.g. Linux version, Linux 8947distribution...) 8948 8949If this capability is available, then the CPNC and CPVC can be synchronized 8950between KVM and userspace via the sync regs mechanism (KVM_SYNC_DIAG318). 8951 89528.26 KVM_CAP_X86_USER_SPACE_MSR 8953------------------------------- 8954 8955:Architectures: x86 8956 8957This capability indicates that KVM supports deflection of MSR reads and 8958writes to user space. It can be enabled on a VM level. If enabled, MSR 8959accesses that would usually trigger a #GP by KVM into the guest will 8960instead get bounced to user space through the KVM_EXIT_X86_RDMSR and 8961KVM_EXIT_X86_WRMSR exit notifications. 8962 89638.27 KVM_CAP_X86_MSR_FILTER 8964--------------------------- 8965 8966:Architectures: x86 8967 8968This capability indicates that KVM supports that accesses to user defined MSRs 8969may be rejected. With this capability exposed, KVM exports new VM ioctl 8970KVM_X86_SET_MSR_FILTER which user space can call to specify bitmaps of MSR 8971ranges that KVM should deny access to. 8972 8973In combination with KVM_CAP_X86_USER_SPACE_MSR, this allows user space to 8974trap and emulate MSRs that are outside of the scope of KVM as well as 8975limit the attack surface on KVM's MSR emulation code. 8976 89778.30 KVM_CAP_XEN_HVM 8978-------------------- 8979 8980:Architectures: x86 8981 8982This capability indicates the features that Xen supports for hosting Xen 8983PVHVM guests. Valid flags are:: 8984 8985 #define KVM_XEN_HVM_CONFIG_HYPERCALL_MSR (1 << 0) 8986 #define KVM_XEN_HVM_CONFIG_INTERCEPT_HCALL (1 << 1) 8987 #define KVM_XEN_HVM_CONFIG_SHARED_INFO (1 << 2) 8988 #define KVM_XEN_HVM_CONFIG_RUNSTATE (1 << 3) 8989 #define KVM_XEN_HVM_CONFIG_EVTCHN_2LEVEL (1 << 4) 8990 #define KVM_XEN_HVM_CONFIG_EVTCHN_SEND (1 << 5) 8991 #define KVM_XEN_HVM_CONFIG_RUNSTATE_UPDATE_FLAG (1 << 6) 8992 #define KVM_XEN_HVM_CONFIG_PVCLOCK_TSC_UNSTABLE (1 << 7) 8993 8994The KVM_XEN_HVM_CONFIG_HYPERCALL_MSR flag indicates that the KVM_XEN_HVM_CONFIG 8995ioctl is available, for the guest to set its hypercall page. 8996 8997If KVM_XEN_HVM_CONFIG_INTERCEPT_HCALL is also set, the same flag may also be 8998provided in the flags to KVM_XEN_HVM_CONFIG, without providing hypercall page 8999contents, to request that KVM generate hypercall page content automatically 9000and also enable interception of guest hypercalls with KVM_EXIT_XEN. 9001 9002The KVM_XEN_HVM_CONFIG_SHARED_INFO flag indicates the availability of the 9003KVM_XEN_HVM_SET_ATTR, KVM_XEN_HVM_GET_ATTR, KVM_XEN_VCPU_SET_ATTR and 9004KVM_XEN_VCPU_GET_ATTR ioctls, as well as the delivery of exception vectors 9005for event channel upcalls when the evtchn_upcall_pending field of a vcpu's 9006vcpu_info is set. 9007 9008The KVM_XEN_HVM_CONFIG_RUNSTATE flag indicates that the runstate-related 9009features KVM_XEN_VCPU_ATTR_TYPE_RUNSTATE_ADDR/_CURRENT/_DATA/_ADJUST are 9010supported by the KVM_XEN_VCPU_SET_ATTR/KVM_XEN_VCPU_GET_ATTR ioctls. 9011 9012The KVM_XEN_HVM_CONFIG_EVTCHN_2LEVEL flag indicates that IRQ routing entries 9013of the type KVM_IRQ_ROUTING_XEN_EVTCHN are supported, with the priority 9014field set to indicate 2 level event channel delivery. 9015 9016The KVM_XEN_HVM_CONFIG_EVTCHN_SEND flag indicates that KVM supports 9017injecting event channel events directly into the guest with the 9018KVM_XEN_HVM_EVTCHN_SEND ioctl. It also indicates support for the 9019KVM_XEN_ATTR_TYPE_EVTCHN/XEN_VERSION HVM attributes and the 9020KVM_XEN_VCPU_ATTR_TYPE_VCPU_ID/TIMER/UPCALL_VECTOR vCPU attributes. 9021related to event channel delivery, timers, and the XENVER_version 9022interception. 9023 9024The KVM_XEN_HVM_CONFIG_RUNSTATE_UPDATE_FLAG flag indicates that KVM supports 9025the KVM_XEN_ATTR_TYPE_RUNSTATE_UPDATE_FLAG attribute in the KVM_XEN_SET_ATTR 9026and KVM_XEN_GET_ATTR ioctls. This controls whether KVM will set the 9027XEN_RUNSTATE_UPDATE flag in guest memory mapped vcpu_runstate_info during 9028updates of the runstate information. Note that versions of KVM which support 9029the RUNSTATE feature above, but not the RUNSTATE_UPDATE_FLAG feature, will 9030always set the XEN_RUNSTATE_UPDATE flag when updating the guest structure, 9031which is perhaps counterintuitive. When this flag is advertised, KVM will 9032behave more correctly, not using the XEN_RUNSTATE_UPDATE flag until/unless 9033specifically enabled (by the guest making the hypercall, causing the VMM 9034to enable the KVM_XEN_ATTR_TYPE_RUNSTATE_UPDATE_FLAG attribute). 9035 9036The KVM_XEN_HVM_CONFIG_PVCLOCK_TSC_UNSTABLE flag indicates that KVM supports 9037clearing the PVCLOCK_TSC_STABLE_BIT flag in Xen pvclock sources. This will be 9038done when the KVM_CAP_XEN_HVM ioctl sets the 9039KVM_XEN_HVM_CONFIG_PVCLOCK_TSC_UNSTABLE flag. 9040 90418.31 KVM_CAP_SPAPR_MULTITCE 9042--------------------------- 9043 9044:Architectures: ppc 9045:Type: vm 9046 9047This capability means the kernel is capable of handling hypercalls 9048H_PUT_TCE_INDIRECT and H_STUFF_TCE without passing those into the user 9049space. This significantly accelerates DMA operations for PPC KVM guests. 9050User space should expect that its handlers for these hypercalls 9051are not going to be called if user space previously registered LIOBN 9052in KVM (via KVM_CREATE_SPAPR_TCE or similar calls). 9053 9054In order to enable H_PUT_TCE_INDIRECT and H_STUFF_TCE use in the guest, 9055user space might have to advertise it for the guest. For example, 9056IBM pSeries (sPAPR) guest starts using them if "hcall-multi-tce" is 9057present in the "ibm,hypertas-functions" device-tree property. 9058 9059The hypercalls mentioned above may or may not be processed successfully 9060in the kernel based fast path. If they can not be handled by the kernel, 9061they will get passed on to user space. So user space still has to have 9062an implementation for these despite the in kernel acceleration. 9063 9064This capability is always enabled. 9065 90668.32 KVM_CAP_PTP_KVM 9067-------------------- 9068 9069:Architectures: arm64 9070 9071This capability indicates that the KVM virtual PTP service is 9072supported in the host. A VMM can check whether the service is 9073available to the guest on migration. 9074 90758.37 KVM_CAP_S390_PROTECTED_DUMP 9076-------------------------------- 9077 9078:Architectures: s390 9079:Type: vm 9080 9081This capability indicates that KVM and the Ultravisor support dumping 9082PV guests. The `KVM_PV_DUMP` command is available for the 9083`KVM_S390_PV_COMMAND` ioctl and the `KVM_PV_INFO` command provides 9084dump related UV data. Also the vcpu ioctl `KVM_S390_PV_CPU_COMMAND` is 9085available and supports the `KVM_PV_DUMP_CPU` subcommand. 9086 90878.39 KVM_CAP_S390_CPU_TOPOLOGY 9088------------------------------ 9089 9090:Architectures: s390 9091:Type: vm 9092 9093This capability indicates that KVM will provide the S390 CPU Topology 9094facility which consist of the interpretation of the PTF instruction for 9095the function code 2 along with interception and forwarding of both the 9096PTF instruction with function codes 0 or 1 and the STSI(15,1,x) 9097instruction to the userland hypervisor. 9098 9099The stfle facility 11, CPU Topology facility, should not be indicated 9100to the guest without this capability. 9101 9102When this capability is present, KVM provides a new attribute group 9103on vm fd, KVM_S390_VM_CPU_TOPOLOGY. 9104This new attribute allows to get, set or clear the Modified Change 9105Topology Report (MTCR) bit of the SCA through the kvm_device_attr 9106structure. 9107 9108When getting the Modified Change Topology Report value, the attr->addr 9109must point to a byte where the value will be stored or retrieved from. 9110 91118.41 KVM_CAP_VM_TYPES 9112--------------------- 9113 9114:Architectures: x86 9115:Type: system ioctl 9116 9117This capability returns a bitmap of support VM types. The 1-setting of bit @n 9118means the VM type with value @n is supported. Possible values of @n are:: 9119 9120 #define KVM_X86_DEFAULT_VM 0 9121 #define KVM_X86_SW_PROTECTED_VM 1 9122 #define KVM_X86_SEV_VM 2 9123 #define KVM_X86_SEV_ES_VM 3 9124 9125Note, KVM_X86_SW_PROTECTED_VM is currently only for development and testing. 9126Do not use KVM_X86_SW_PROTECTED_VM for "real" VMs, and especially not in 9127production. The behavior and effective ABI for software-protected VMs is 9128unstable. 9129 91308.42 KVM_CAP_PPC_RPT_INVALIDATE 9131------------------------------- 9132 9133:Architectures: ppc 9134 9135This capability indicates that the kernel is capable of handling 9136H_RPT_INVALIDATE hcall. 9137 9138In order to enable the use of H_RPT_INVALIDATE in the guest, 9139user space might have to advertise it for the guest. For example, 9140IBM pSeries (sPAPR) guest starts using it if "hcall-rpt-invalidate" is 9141present in the "ibm,hypertas-functions" device-tree property. 9142 9143This capability is enabled for hypervisors on platforms like POWER9 9144that support radix MMU. 9145 91468.43 KVM_CAP_PPC_AIL_MODE_3 9147--------------------------- 9148 9149:Architectures: ppc 9150 9151This capability indicates that the kernel supports the mode 3 setting for the 9152"Address Translation Mode on Interrupt" aka "Alternate Interrupt Location" 9153resource that is controlled with the H_SET_MODE hypercall. 9154 9155This capability allows a guest kernel to use a better-performance mode for 9156handling interrupts and system calls. 9157 91588.44 KVM_CAP_MEMORY_FAULT_INFO 9159------------------------------ 9160 9161:Architectures: x86 9162 9163The presence of this capability indicates that KVM_RUN will fill 9164kvm_run.memory_fault if KVM cannot resolve a guest page fault VM-Exit, e.g. if 9165there is a valid memslot but no backing VMA for the corresponding host virtual 9166address. 9167 9168The information in kvm_run.memory_fault is valid if and only if KVM_RUN returns 9169an error with errno=EFAULT or errno=EHWPOISON *and* kvm_run.exit_reason is set 9170to KVM_EXIT_MEMORY_FAULT. 9171 9172Note: Userspaces which attempt to resolve memory faults so that they can retry 9173KVM_RUN are encouraged to guard against repeatedly receiving the same 9174error/annotated fault. 9175 9176See KVM_EXIT_MEMORY_FAULT for more information. 9177 91788.45 KVM_CAP_X86_GUEST_MODE 9179--------------------------- 9180 9181:Architectures: x86 9182 9183The presence of this capability indicates that KVM_RUN will update the 9184KVM_RUN_X86_GUEST_MODE bit in kvm_run.flags to indicate whether the 9185vCPU was executing nested guest code when it exited. 9186 9187KVM exits with the register state of either the L1 or L2 guest 9188depending on which executed at the time of an exit. Userspace must 9189take care to differentiate between these cases. 9190 91919. Known KVM API problems 9192========================= 9193 9194In some cases, KVM's API has some inconsistencies or common pitfalls 9195that userspace need to be aware of. This section details some of 9196these issues. 9197 9198Most of them are architecture specific, so the section is split by 9199architecture. 9200 92019.1. x86 9202-------- 9203 9204``KVM_GET_SUPPORTED_CPUID`` issues 9205^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ 9206 9207In general, ``KVM_GET_SUPPORTED_CPUID`` is designed so that it is possible 9208to take its result and pass it directly to ``KVM_SET_CPUID2``. This section 9209documents some cases in which that requires some care. 9210 9211Local APIC features 9212~~~~~~~~~~~~~~~~~~~ 9213 9214CPU[EAX=1]:ECX[21] (X2APIC) is reported by ``KVM_GET_SUPPORTED_CPUID``, 9215but it can only be enabled if ``KVM_CREATE_IRQCHIP`` or 9216``KVM_ENABLE_CAP(KVM_CAP_IRQCHIP_SPLIT)`` are used to enable in-kernel emulation of 9217the local APIC. 9218 9219The same is true for the ``KVM_FEATURE_PV_UNHALT`` paravirtualized feature. 9220 9221On older versions of Linux, CPU[EAX=1]:ECX[24] (TSC_DEADLINE) is not reported by 9222``KVM_GET_SUPPORTED_CPUID``, but it can be enabled if ``KVM_CAP_TSC_DEADLINE_TIMER`` 9223is present and the kernel has enabled in-kernel emulation of the local APIC. 9224On newer versions, ``KVM_GET_SUPPORTED_CPUID`` does report the bit as available. 9225 9226CPU topology 9227~~~~~~~~~~~~ 9228 9229Several CPUID values include topology information for the host CPU: 92300x0b and 0x1f for Intel systems, 0x8000001e for AMD systems. Different 9231versions of KVM return different values for this information and userspace 9232should not rely on it. Currently they return all zeroes. 9233 9234If userspace wishes to set up a guest topology, it should be careful that 9235the values of these three leaves differ for each CPU. In particular, 9236the APIC ID is found in EDX for all subleaves of 0x0b and 0x1f, and in EAX 9237for 0x8000001e; the latter also encodes the core id and node id in bits 92387:0 of EBX and ECX respectively. 9239 9240Obsolete ioctls and capabilities 9241^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ 9242 9243KVM_CAP_DISABLE_QUIRKS does not let userspace know which quirks are actually 9244available. Use ``KVM_CHECK_EXTENSION(KVM_CAP_DISABLE_QUIRKS2)`` instead if 9245available. 9246 9247Ordering of KVM_GET_*/KVM_SET_* ioctls 9248^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ 9249 9250TBD 9251