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 1003:: 1004 1005 struct kvm_xen_hvm_config { 1006 __u32 flags; 1007 __u32 msr; 1008 __u64 blob_addr_32; 1009 __u64 blob_addr_64; 1010 __u8 blob_size_32; 1011 __u8 blob_size_64; 1012 __u8 pad2[30]; 1013 }; 1014 1015If certain flags are returned from the KVM_CAP_XEN_HVM check, they may 1016be set in the flags field of this ioctl: 1017 1018The KVM_XEN_HVM_CONFIG_INTERCEPT_HCALL flag requests KVM to generate 1019the contents of the hypercall page automatically; hypercalls will be 1020intercepted and passed to userspace through KVM_EXIT_XEN. In this 1021case, all of the blob size and address fields must be zero. 1022 1023The KVM_XEN_HVM_CONFIG_EVTCHN_SEND flag indicates to KVM that userspace 1024will always use the KVM_XEN_HVM_EVTCHN_SEND ioctl to deliver event 1025channel interrupts rather than manipulating the guest's shared_info 1026structures directly. This, in turn, may allow KVM to enable features 1027such as intercepting the SCHEDOP_poll hypercall to accelerate PV 1028spinlock operation for the guest. Userspace may still use the ioctl 1029to deliver events if it was advertised, even if userspace does not 1030send this indication that it will always do so 1031 1032No other flags are currently valid in the struct kvm_xen_hvm_config. 1033 10344.29 KVM_GET_CLOCK 1035------------------ 1036 1037:Capability: KVM_CAP_ADJUST_CLOCK 1038:Architectures: x86 1039:Type: vm ioctl 1040:Parameters: struct kvm_clock_data (out) 1041:Returns: 0 on success, -1 on error 1042 1043Gets the current timestamp of kvmclock as seen by the current guest. In 1044conjunction with KVM_SET_CLOCK, it is used to ensure monotonicity on scenarios 1045such as migration. 1046 1047When KVM_CAP_ADJUST_CLOCK is passed to KVM_CHECK_EXTENSION, it returns the 1048set of bits that KVM can return in struct kvm_clock_data's flag member. 1049 1050The following flags are defined: 1051 1052KVM_CLOCK_TSC_STABLE 1053 If set, the returned value is the exact kvmclock 1054 value seen by all VCPUs at the instant when KVM_GET_CLOCK was called. 1055 If clear, the returned value is simply CLOCK_MONOTONIC plus a constant 1056 offset; the offset can be modified with KVM_SET_CLOCK. KVM will try 1057 to make all VCPUs follow this clock, but the exact value read by each 1058 VCPU could differ, because the host TSC is not stable. 1059 1060KVM_CLOCK_REALTIME 1061 If set, the `realtime` field in the kvm_clock_data 1062 structure is populated with the value of the host's real time 1063 clocksource at the instant when KVM_GET_CLOCK was called. If clear, 1064 the `realtime` field does not contain a value. 1065 1066KVM_CLOCK_HOST_TSC 1067 If set, the `host_tsc` field in the kvm_clock_data 1068 structure is populated with the value of the host's timestamp counter (TSC) 1069 at the instant when KVM_GET_CLOCK was called. If clear, the `host_tsc` field 1070 does not contain a value. 1071 1072:: 1073 1074 struct kvm_clock_data { 1075 __u64 clock; /* kvmclock current value */ 1076 __u32 flags; 1077 __u32 pad0; 1078 __u64 realtime; 1079 __u64 host_tsc; 1080 __u32 pad[4]; 1081 }; 1082 1083 10844.30 KVM_SET_CLOCK 1085------------------ 1086 1087:Capability: KVM_CAP_ADJUST_CLOCK 1088:Architectures: x86 1089:Type: vm ioctl 1090:Parameters: struct kvm_clock_data (in) 1091:Returns: 0 on success, -1 on error 1092 1093Sets the current timestamp of kvmclock to the value specified in its parameter. 1094In conjunction with KVM_GET_CLOCK, it is used to ensure monotonicity on scenarios 1095such as migration. 1096 1097The following flags can be passed: 1098 1099KVM_CLOCK_REALTIME 1100 If set, KVM will compare the value of the `realtime` field 1101 with the value of the host's real time clocksource at the instant when 1102 KVM_SET_CLOCK was called. The difference in elapsed time is added to the final 1103 kvmclock value that will be provided to guests. 1104 1105Other flags returned by ``KVM_GET_CLOCK`` are accepted but ignored. 1106 1107:: 1108 1109 struct kvm_clock_data { 1110 __u64 clock; /* kvmclock current value */ 1111 __u32 flags; 1112 __u32 pad0; 1113 __u64 realtime; 1114 __u64 host_tsc; 1115 __u32 pad[4]; 1116 }; 1117 1118 11194.31 KVM_GET_VCPU_EVENTS 1120------------------------ 1121 1122:Capability: KVM_CAP_VCPU_EVENTS 1123:Extended by: KVM_CAP_INTR_SHADOW 1124:Architectures: x86, arm64 1125:Type: vcpu ioctl 1126:Parameters: struct kvm_vcpu_events (out) 1127:Returns: 0 on success, -1 on error 1128 1129X86: 1130^^^^ 1131 1132Gets currently pending exceptions, interrupts, and NMIs as well as related 1133states of the vcpu. 1134 1135:: 1136 1137 struct kvm_vcpu_events { 1138 struct { 1139 __u8 injected; 1140 __u8 nr; 1141 __u8 has_error_code; 1142 __u8 pending; 1143 __u32 error_code; 1144 } exception; 1145 struct { 1146 __u8 injected; 1147 __u8 nr; 1148 __u8 soft; 1149 __u8 shadow; 1150 } interrupt; 1151 struct { 1152 __u8 injected; 1153 __u8 pending; 1154 __u8 masked; 1155 __u8 pad; 1156 } nmi; 1157 __u32 sipi_vector; 1158 __u32 flags; 1159 struct { 1160 __u8 smm; 1161 __u8 pending; 1162 __u8 smm_inside_nmi; 1163 __u8 latched_init; 1164 } smi; 1165 __u8 reserved[27]; 1166 __u8 exception_has_payload; 1167 __u64 exception_payload; 1168 }; 1169 1170The following bits are defined in the flags field: 1171 1172- KVM_VCPUEVENT_VALID_SHADOW may be set to signal that 1173 interrupt.shadow contains a valid state. 1174 1175- KVM_VCPUEVENT_VALID_SMM may be set to signal that smi contains a 1176 valid state. 1177 1178- KVM_VCPUEVENT_VALID_PAYLOAD may be set to signal that the 1179 exception_has_payload, exception_payload, and exception.pending 1180 fields contain a valid state. This bit will be set whenever 1181 KVM_CAP_EXCEPTION_PAYLOAD is enabled. 1182 1183- KVM_VCPUEVENT_VALID_TRIPLE_FAULT may be set to signal that the 1184 triple_fault_pending field contains a valid state. This bit will 1185 be set whenever KVM_CAP_X86_TRIPLE_FAULT_EVENT is enabled. 1186 1187ARM64: 1188^^^^^^ 1189 1190If the guest accesses a device that is being emulated by the host kernel in 1191such a way that a real device would generate a physical SError, KVM may make 1192a virtual SError pending for that VCPU. This system error interrupt remains 1193pending until the guest takes the exception by unmasking PSTATE.A. 1194 1195Running the VCPU may cause it to take a pending SError, or make an access that 1196causes an SError to become pending. The event's description is only valid while 1197the VPCU is not running. 1198 1199This API provides a way to read and write the pending 'event' state that is not 1200visible to the guest. To save, restore or migrate a VCPU the struct representing 1201the state can be read then written using this GET/SET API, along with the other 1202guest-visible registers. It is not possible to 'cancel' an SError that has been 1203made pending. 1204 1205A device being emulated in user-space may also wish to generate an SError. To do 1206this the events structure can be populated by user-space. The current state 1207should be read first, to ensure no existing SError is pending. If an existing 1208SError is pending, the architecture's 'Multiple SError interrupts' rules should 1209be followed. (2.5.3 of DDI0587.a "ARM Reliability, Availability, and 1210Serviceability (RAS) Specification"). 1211 1212SError exceptions always have an ESR value. Some CPUs have the ability to 1213specify what the virtual SError's ESR value should be. These systems will 1214advertise KVM_CAP_ARM_INJECT_SERROR_ESR. In this case exception.has_esr will 1215always have a non-zero value when read, and the agent making an SError pending 1216should specify the ISS field in the lower 24 bits of exception.serror_esr. If 1217the system supports KVM_CAP_ARM_INJECT_SERROR_ESR, but user-space sets the events 1218with exception.has_esr as zero, KVM will choose an ESR. 1219 1220Specifying exception.has_esr on a system that does not support it will return 1221-EINVAL. Setting anything other than the lower 24bits of exception.serror_esr 1222will return -EINVAL. 1223 1224It is not possible to read back a pending external abort (injected via 1225KVM_SET_VCPU_EVENTS or otherwise) because such an exception is always delivered 1226directly to the virtual CPU). 1227 1228:: 1229 1230 struct kvm_vcpu_events { 1231 struct { 1232 __u8 serror_pending; 1233 __u8 serror_has_esr; 1234 __u8 ext_dabt_pending; 1235 /* Align it to 8 bytes */ 1236 __u8 pad[5]; 1237 __u64 serror_esr; 1238 } exception; 1239 __u32 reserved[12]; 1240 }; 1241 12424.32 KVM_SET_VCPU_EVENTS 1243------------------------ 1244 1245:Capability: KVM_CAP_VCPU_EVENTS 1246:Extended by: KVM_CAP_INTR_SHADOW 1247:Architectures: x86, arm64 1248:Type: vcpu ioctl 1249:Parameters: struct kvm_vcpu_events (in) 1250:Returns: 0 on success, -1 on error 1251 1252X86: 1253^^^^ 1254 1255Set pending exceptions, interrupts, and NMIs as well as related states of the 1256vcpu. 1257 1258See KVM_GET_VCPU_EVENTS for the data structure. 1259 1260Fields that may be modified asynchronously by running VCPUs can be excluded 1261from the update. These fields are nmi.pending, sipi_vector, smi.smm, 1262smi.pending. Keep the corresponding bits in the flags field cleared to 1263suppress overwriting the current in-kernel state. The bits are: 1264 1265=============================== ================================== 1266KVM_VCPUEVENT_VALID_NMI_PENDING transfer nmi.pending to the kernel 1267KVM_VCPUEVENT_VALID_SIPI_VECTOR transfer sipi_vector 1268KVM_VCPUEVENT_VALID_SMM transfer the smi sub-struct. 1269=============================== ================================== 1270 1271If KVM_CAP_INTR_SHADOW is available, KVM_VCPUEVENT_VALID_SHADOW can be set in 1272the flags field to signal that interrupt.shadow contains a valid state and 1273shall be written into the VCPU. 1274 1275KVM_VCPUEVENT_VALID_SMM can only be set if KVM_CAP_X86_SMM is available. 1276 1277If KVM_CAP_EXCEPTION_PAYLOAD is enabled, KVM_VCPUEVENT_VALID_PAYLOAD 1278can be set in the flags field to signal that the 1279exception_has_payload, exception_payload, and exception.pending fields 1280contain a valid state and shall be written into the VCPU. 1281 1282If KVM_CAP_X86_TRIPLE_FAULT_EVENT is enabled, KVM_VCPUEVENT_VALID_TRIPLE_FAULT 1283can be set in flags field to signal that the triple_fault field contains 1284a valid state and shall be written into the VCPU. 1285 1286ARM64: 1287^^^^^^ 1288 1289User space may need to inject several types of events to the guest. 1290 1291Set the pending SError exception state for this VCPU. It is not possible to 1292'cancel' an Serror that has been made pending. 1293 1294If the guest performed an access to I/O memory which could not be handled by 1295userspace, for example because of missing instruction syndrome decode 1296information or because there is no device mapped at the accessed IPA, then 1297userspace can ask the kernel to inject an external abort using the address 1298from the exiting fault on the VCPU. It is a programming error to set 1299ext_dabt_pending after an exit which was not either KVM_EXIT_MMIO or 1300KVM_EXIT_ARM_NISV. This feature is only available if the system supports 1301KVM_CAP_ARM_INJECT_EXT_DABT. This is a helper which provides commonality in 1302how userspace reports accesses for the above cases to guests, across different 1303userspace implementations. Nevertheless, userspace can still emulate all Arm 1304exceptions by manipulating individual registers using the KVM_SET_ONE_REG API. 1305 1306See KVM_GET_VCPU_EVENTS for the data structure. 1307 1308 13094.33 KVM_GET_DEBUGREGS 1310---------------------- 1311 1312:Capability: KVM_CAP_DEBUGREGS 1313:Architectures: x86 1314:Type: vcpu ioctl 1315:Parameters: struct kvm_debugregs (out) 1316:Returns: 0 on success, -1 on error 1317 1318Reads debug registers from the vcpu. 1319 1320:: 1321 1322 struct kvm_debugregs { 1323 __u64 db[4]; 1324 __u64 dr6; 1325 __u64 dr7; 1326 __u64 flags; 1327 __u64 reserved[9]; 1328 }; 1329 1330 13314.34 KVM_SET_DEBUGREGS 1332---------------------- 1333 1334:Capability: KVM_CAP_DEBUGREGS 1335:Architectures: x86 1336:Type: vcpu ioctl 1337:Parameters: struct kvm_debugregs (in) 1338:Returns: 0 on success, -1 on error 1339 1340Writes debug registers into the vcpu. 1341 1342See KVM_GET_DEBUGREGS for the data structure. The flags field is unused 1343yet and must be cleared on entry. 1344 1345 13464.35 KVM_SET_USER_MEMORY_REGION 1347------------------------------- 1348 1349:Capability: KVM_CAP_USER_MEMORY 1350:Architectures: all 1351:Type: vm ioctl 1352:Parameters: struct kvm_userspace_memory_region (in) 1353:Returns: 0 on success, -1 on error 1354 1355:: 1356 1357 struct kvm_userspace_memory_region { 1358 __u32 slot; 1359 __u32 flags; 1360 __u64 guest_phys_addr; 1361 __u64 memory_size; /* bytes */ 1362 __u64 userspace_addr; /* start of the userspace allocated memory */ 1363 }; 1364 1365 /* for kvm_userspace_memory_region::flags */ 1366 #define KVM_MEM_LOG_DIRTY_PAGES (1UL << 0) 1367 #define KVM_MEM_READONLY (1UL << 1) 1368 1369This ioctl allows the user to create, modify or delete a guest physical 1370memory slot. Bits 0-15 of "slot" specify the slot id and this value 1371should be less than the maximum number of user memory slots supported per 1372VM. The maximum allowed slots can be queried using KVM_CAP_NR_MEMSLOTS. 1373Slots may not overlap in guest physical address space. 1374 1375If KVM_CAP_MULTI_ADDRESS_SPACE is available, bits 16-31 of "slot" 1376specifies the address space which is being modified. They must be 1377less than the value that KVM_CHECK_EXTENSION returns for the 1378KVM_CAP_MULTI_ADDRESS_SPACE capability. Slots in separate address spaces 1379are unrelated; the restriction on overlapping slots only applies within 1380each address space. 1381 1382Deleting a slot is done by passing zero for memory_size. When changing 1383an existing slot, it may be moved in the guest physical memory space, 1384or its flags may be modified, but it may not be resized. 1385 1386Memory for the region is taken starting at the address denoted by the 1387field userspace_addr, which must point at user addressable memory for 1388the entire memory slot size. Any object may back this memory, including 1389anonymous memory, ordinary files, and hugetlbfs. 1390 1391On architectures that support a form of address tagging, userspace_addr must 1392be an untagged address. 1393 1394It is recommended that the lower 21 bits of guest_phys_addr and userspace_addr 1395be identical. This allows large pages in the guest to be backed by large 1396pages in the host. 1397 1398The flags field supports two flags: KVM_MEM_LOG_DIRTY_PAGES and 1399KVM_MEM_READONLY. The former can be set to instruct KVM to keep track of 1400writes to memory within the slot. See KVM_GET_DIRTY_LOG ioctl to know how to 1401use it. The latter can be set, if KVM_CAP_READONLY_MEM capability allows it, 1402to make a new slot read-only. In this case, writes to this memory will be 1403posted to userspace as KVM_EXIT_MMIO exits. 1404 1405When the KVM_CAP_SYNC_MMU capability is available, changes in the backing of 1406the memory region are automatically reflected into the guest. For example, an 1407mmap() that affects the region will be made visible immediately. Another 1408example is madvise(MADV_DROP). 1409 1410Note: On arm64, a write generated by the page-table walker (to update 1411the Access and Dirty flags, for example) never results in a 1412KVM_EXIT_MMIO exit when the slot has the KVM_MEM_READONLY flag. This 1413is because KVM cannot provide the data that would be written by the 1414page-table walker, making it impossible to emulate the access. 1415Instead, an abort (data abort if the cause of the page-table update 1416was a load or a store, instruction abort if it was an instruction 1417fetch) is injected in the guest. 1418 1419S390: 1420^^^^^ 1421 1422Returns -EINVAL if the VM has the KVM_VM_S390_UCONTROL flag set. 1423Returns -EINVAL if called on a protected VM. 1424 14254.36 KVM_SET_TSS_ADDR 1426--------------------- 1427 1428:Capability: KVM_CAP_SET_TSS_ADDR 1429:Architectures: x86 1430:Type: vm ioctl 1431:Parameters: unsigned long tss_address (in) 1432:Returns: 0 on success, -1 on error 1433 1434This ioctl defines the physical address of a three-page region in the guest 1435physical address space. The region must be within the first 4GB of the 1436guest physical address space and must not conflict with any memory slot 1437or any mmio address. The guest may malfunction if it accesses this memory 1438region. 1439 1440This ioctl is required on Intel-based hosts. This is needed on Intel hardware 1441because of a quirk in the virtualization implementation (see the internals 1442documentation when it pops into existence). 1443 1444 1445.. _KVM_ENABLE_CAP: 1446 14474.37 KVM_ENABLE_CAP 1448------------------- 1449 1450:Capability: KVM_CAP_ENABLE_CAP 1451:Architectures: mips, ppc, s390, x86, loongarch 1452:Type: vcpu ioctl 1453:Parameters: struct kvm_enable_cap (in) 1454:Returns: 0 on success; -1 on error 1455 1456:Capability: KVM_CAP_ENABLE_CAP_VM 1457:Architectures: all 1458:Type: vm ioctl 1459:Parameters: struct kvm_enable_cap (in) 1460:Returns: 0 on success; -1 on error 1461 1462.. note:: 1463 1464 Not all extensions are enabled by default. Using this ioctl the application 1465 can enable an extension, making it available to the guest. 1466 1467On systems that do not support this ioctl, it always fails. On systems that 1468do support it, it only works for extensions that are supported for enablement. 1469 1470To check if a capability can be enabled, the KVM_CHECK_EXTENSION ioctl should 1471be used. 1472 1473:: 1474 1475 struct kvm_enable_cap { 1476 /* in */ 1477 __u32 cap; 1478 1479The capability that is supposed to get enabled. 1480 1481:: 1482 1483 __u32 flags; 1484 1485A bitfield indicating future enhancements. Has to be 0 for now. 1486 1487:: 1488 1489 __u64 args[4]; 1490 1491Arguments for enabling a feature. If a feature needs initial values to 1492function properly, this is the place to put them. 1493 1494:: 1495 1496 __u8 pad[64]; 1497 }; 1498 1499The vcpu ioctl should be used for vcpu-specific capabilities, the vm ioctl 1500for vm-wide capabilities. 1501 15024.38 KVM_GET_MP_STATE 1503--------------------- 1504 1505:Capability: KVM_CAP_MP_STATE 1506:Architectures: x86, s390, arm64, riscv, loongarch 1507:Type: vcpu ioctl 1508:Parameters: struct kvm_mp_state (out) 1509:Returns: 0 on success; -1 on error 1510 1511:: 1512 1513 struct kvm_mp_state { 1514 __u32 mp_state; 1515 }; 1516 1517Returns the vcpu's current "multiprocessing state" (though also valid on 1518uniprocessor guests). 1519 1520Possible values are: 1521 1522 ========================== =============================================== 1523 KVM_MP_STATE_RUNNABLE the vcpu is currently running 1524 [x86,arm64,riscv,loongarch] 1525 KVM_MP_STATE_UNINITIALIZED the vcpu is an application processor (AP) 1526 which has not yet received an INIT signal [x86] 1527 KVM_MP_STATE_INIT_RECEIVED the vcpu has received an INIT signal, and is 1528 now ready for a SIPI [x86] 1529 KVM_MP_STATE_HALTED the vcpu has executed a HLT instruction and 1530 is waiting for an interrupt [x86] 1531 KVM_MP_STATE_SIPI_RECEIVED the vcpu has just received a SIPI (vector 1532 accessible via KVM_GET_VCPU_EVENTS) [x86] 1533 KVM_MP_STATE_STOPPED the vcpu is stopped [s390,arm64,riscv] 1534 KVM_MP_STATE_CHECK_STOP the vcpu is in a special error state [s390] 1535 KVM_MP_STATE_OPERATING the vcpu is operating (running or halted) 1536 [s390] 1537 KVM_MP_STATE_LOAD the vcpu is in a special load/startup state 1538 [s390] 1539 KVM_MP_STATE_SUSPENDED the vcpu is in a suspend state and is waiting 1540 for a wakeup event [arm64] 1541 ========================== =============================================== 1542 1543On x86, this ioctl is only useful after KVM_CREATE_IRQCHIP. Without an 1544in-kernel irqchip, the multiprocessing state must be maintained by userspace on 1545these architectures. 1546 1547For arm64: 1548^^^^^^^^^^ 1549 1550If a vCPU is in the KVM_MP_STATE_SUSPENDED state, KVM will emulate the 1551architectural execution of a WFI instruction. 1552 1553If a wakeup event is recognized, KVM will exit to userspace with a 1554KVM_SYSTEM_EVENT exit, where the event type is KVM_SYSTEM_EVENT_WAKEUP. If 1555userspace wants to honor the wakeup, it must set the vCPU's MP state to 1556KVM_MP_STATE_RUNNABLE. If it does not, KVM will continue to await a wakeup 1557event in subsequent calls to KVM_RUN. 1558 1559.. warning:: 1560 1561 If userspace intends to keep the vCPU in a SUSPENDED state, it is 1562 strongly recommended that userspace take action to suppress the 1563 wakeup event (such as masking an interrupt). Otherwise, subsequent 1564 calls to KVM_RUN will immediately exit with a KVM_SYSTEM_EVENT_WAKEUP 1565 event and inadvertently waste CPU cycles. 1566 1567 Additionally, if userspace takes action to suppress a wakeup event, 1568 it is strongly recommended that it also restores the vCPU to its 1569 original state when the vCPU is made RUNNABLE again. For example, 1570 if userspace masked a pending interrupt to suppress the wakeup, 1571 the interrupt should be unmasked before returning control to the 1572 guest. 1573 1574For riscv: 1575^^^^^^^^^^ 1576 1577The only states that are valid are KVM_MP_STATE_STOPPED and 1578KVM_MP_STATE_RUNNABLE which reflect if the vcpu is paused or not. 1579 1580On LoongArch, only the KVM_MP_STATE_RUNNABLE state is used to reflect 1581whether the vcpu is runnable. 1582 15834.39 KVM_SET_MP_STATE 1584--------------------- 1585 1586:Capability: KVM_CAP_MP_STATE 1587:Architectures: x86, s390, arm64, riscv, loongarch 1588:Type: vcpu ioctl 1589:Parameters: struct kvm_mp_state (in) 1590:Returns: 0 on success; -1 on error 1591 1592Sets the vcpu's current "multiprocessing state"; see KVM_GET_MP_STATE for 1593arguments. 1594 1595On x86, this ioctl is only useful after KVM_CREATE_IRQCHIP. Without an 1596in-kernel irqchip, the multiprocessing state must be maintained by userspace on 1597these architectures. 1598 1599For arm64/riscv: 1600^^^^^^^^^^^^^^^^ 1601 1602The only states that are valid are KVM_MP_STATE_STOPPED and 1603KVM_MP_STATE_RUNNABLE which reflect if the vcpu should be paused or not. 1604 1605On LoongArch, only the KVM_MP_STATE_RUNNABLE state is used to reflect 1606whether the vcpu is runnable. 1607 16084.40 KVM_SET_IDENTITY_MAP_ADDR 1609------------------------------ 1610 1611:Capability: KVM_CAP_SET_IDENTITY_MAP_ADDR 1612:Architectures: x86 1613:Type: vm ioctl 1614:Parameters: unsigned long identity (in) 1615:Returns: 0 on success, -1 on error 1616 1617This ioctl defines the physical address of a one-page region in the guest 1618physical address space. The region must be within the first 4GB of the 1619guest physical address space and must not conflict with any memory slot 1620or any mmio address. The guest may malfunction if it accesses this memory 1621region. 1622 1623Setting the address to 0 will result in resetting the address to its default 1624(0xfffbc000). 1625 1626This ioctl is required on Intel-based hosts. This is needed on Intel hardware 1627because of a quirk in the virtualization implementation (see the internals 1628documentation when it pops into existence). 1629 1630Fails if any VCPU has already been created. 1631 16324.41 KVM_SET_BOOT_CPU_ID 1633------------------------ 1634 1635:Capability: KVM_CAP_SET_BOOT_CPU_ID 1636:Architectures: x86 1637:Type: vm ioctl 1638:Parameters: unsigned long vcpu_id 1639:Returns: 0 on success, -1 on error 1640 1641Define which vcpu is the Bootstrap Processor (BSP). Values are the same 1642as the vcpu id in KVM_CREATE_VCPU. If this ioctl is not called, the default 1643is vcpu 0. This ioctl has to be called before vcpu creation, 1644otherwise it will return EBUSY error. 1645 1646 16474.42 KVM_GET_XSAVE 1648------------------ 1649 1650:Capability: KVM_CAP_XSAVE 1651:Architectures: x86 1652:Type: vcpu ioctl 1653:Parameters: struct kvm_xsave (out) 1654:Returns: 0 on success, -1 on error 1655 1656 1657:: 1658 1659 struct kvm_xsave { 1660 __u32 region[1024]; 1661 __u32 extra[0]; 1662 }; 1663 1664This ioctl would copy current vcpu's xsave struct to the userspace. 1665 1666 16674.43 KVM_SET_XSAVE 1668------------------ 1669 1670:Capability: KVM_CAP_XSAVE and KVM_CAP_XSAVE2 1671:Architectures: x86 1672:Type: vcpu ioctl 1673:Parameters: struct kvm_xsave (in) 1674:Returns: 0 on success, -1 on error 1675 1676:: 1677 1678 1679 struct kvm_xsave { 1680 __u32 region[1024]; 1681 __u32 extra[0]; 1682 }; 1683 1684This ioctl would copy userspace's xsave struct to the kernel. It copies 1685as many bytes as are returned by KVM_CHECK_EXTENSION(KVM_CAP_XSAVE2), 1686when invoked on the vm file descriptor. The size value returned by 1687KVM_CHECK_EXTENSION(KVM_CAP_XSAVE2) will always be at least 4096. 1688Currently, it is only greater than 4096 if a dynamic feature has been 1689enabled with ``arch_prctl()``, but this may change in the future. 1690 1691The offsets of the state save areas in struct kvm_xsave follow the 1692contents of CPUID leaf 0xD on the host. 1693 1694 16954.44 KVM_GET_XCRS 1696----------------- 1697 1698:Capability: KVM_CAP_XCRS 1699:Architectures: x86 1700:Type: vcpu ioctl 1701:Parameters: struct kvm_xcrs (out) 1702:Returns: 0 on success, -1 on error 1703 1704:: 1705 1706 struct kvm_xcr { 1707 __u32 xcr; 1708 __u32 reserved; 1709 __u64 value; 1710 }; 1711 1712 struct kvm_xcrs { 1713 __u32 nr_xcrs; 1714 __u32 flags; 1715 struct kvm_xcr xcrs[KVM_MAX_XCRS]; 1716 __u64 padding[16]; 1717 }; 1718 1719This ioctl would copy current vcpu's xcrs to the userspace. 1720 1721 17224.45 KVM_SET_XCRS 1723----------------- 1724 1725:Capability: KVM_CAP_XCRS 1726:Architectures: x86 1727:Type: vcpu ioctl 1728:Parameters: struct kvm_xcrs (in) 1729:Returns: 0 on success, -1 on error 1730 1731:: 1732 1733 struct kvm_xcr { 1734 __u32 xcr; 1735 __u32 reserved; 1736 __u64 value; 1737 }; 1738 1739 struct kvm_xcrs { 1740 __u32 nr_xcrs; 1741 __u32 flags; 1742 struct kvm_xcr xcrs[KVM_MAX_XCRS]; 1743 __u64 padding[16]; 1744 }; 1745 1746This ioctl would set vcpu's xcr to the value userspace specified. 1747 1748 17494.46 KVM_GET_SUPPORTED_CPUID 1750---------------------------- 1751 1752:Capability: KVM_CAP_EXT_CPUID 1753:Architectures: x86 1754:Type: system ioctl 1755:Parameters: struct kvm_cpuid2 (in/out) 1756:Returns: 0 on success, -1 on error 1757 1758:: 1759 1760 struct kvm_cpuid2 { 1761 __u32 nent; 1762 __u32 padding; 1763 struct kvm_cpuid_entry2 entries[0]; 1764 }; 1765 1766 #define KVM_CPUID_FLAG_SIGNIFCANT_INDEX BIT(0) 1767 #define KVM_CPUID_FLAG_STATEFUL_FUNC BIT(1) /* deprecated */ 1768 #define KVM_CPUID_FLAG_STATE_READ_NEXT BIT(2) /* deprecated */ 1769 1770 struct kvm_cpuid_entry2 { 1771 __u32 function; 1772 __u32 index; 1773 __u32 flags; 1774 __u32 eax; 1775 __u32 ebx; 1776 __u32 ecx; 1777 __u32 edx; 1778 __u32 padding[3]; 1779 }; 1780 1781This ioctl returns x86 cpuid features which are supported by both the 1782hardware and kvm in its default configuration. Userspace can use the 1783information returned by this ioctl to construct cpuid information (for 1784KVM_SET_CPUID2) that is consistent with hardware, kernel, and 1785userspace capabilities, and with user requirements (for example, the 1786user may wish to constrain cpuid to emulate older hardware, or for 1787feature consistency across a cluster). 1788 1789Dynamically-enabled feature bits need to be requested with 1790``arch_prctl()`` before calling this ioctl. Feature bits that have not 1791been requested are excluded from the result. 1792 1793Note that certain capabilities, such as KVM_CAP_X86_DISABLE_EXITS, may 1794expose cpuid features (e.g. MONITOR) which are not supported by kvm in 1795its default configuration. If userspace enables such capabilities, it 1796is responsible for modifying the results of this ioctl appropriately. 1797 1798Userspace invokes KVM_GET_SUPPORTED_CPUID by passing a kvm_cpuid2 structure 1799with the 'nent' field indicating the number of entries in the variable-size 1800array 'entries'. If the number of entries is too low to describe the cpu 1801capabilities, an error (E2BIG) is returned. If the number is too high, 1802the 'nent' field is adjusted and an error (ENOMEM) is returned. If the 1803number is just right, the 'nent' field is adjusted to the number of valid 1804entries in the 'entries' array, which is then filled. 1805 1806The entries returned are the host cpuid as returned by the cpuid instruction, 1807with unknown or unsupported features masked out. Some features (for example, 1808x2apic), may not be present in the host cpu, but are exposed by kvm if it can 1809emulate them efficiently. The fields in each entry are defined as follows: 1810 1811 function: 1812 the eax value used to obtain the entry 1813 1814 index: 1815 the ecx value used to obtain the entry (for entries that are 1816 affected by ecx) 1817 1818 flags: 1819 an OR of zero or more of the following: 1820 1821 KVM_CPUID_FLAG_SIGNIFCANT_INDEX: 1822 if the index field is valid 1823 1824 eax, ebx, ecx, edx: 1825 the values returned by the cpuid instruction for 1826 this function/index combination 1827 1828The TSC deadline timer feature (CPUID leaf 1, ecx[24]) is always returned 1829as false, since the feature depends on KVM_CREATE_IRQCHIP for local APIC 1830support. Instead it is reported via:: 1831 1832 ioctl(KVM_CHECK_EXTENSION, KVM_CAP_TSC_DEADLINE_TIMER) 1833 1834if that returns true and you use KVM_CREATE_IRQCHIP, or if you emulate the 1835feature in userspace, then you can enable the feature for KVM_SET_CPUID2. 1836 1837 18384.47 KVM_PPC_GET_PVINFO 1839----------------------- 1840 1841:Capability: KVM_CAP_PPC_GET_PVINFO 1842:Architectures: ppc 1843:Type: vm ioctl 1844:Parameters: struct kvm_ppc_pvinfo (out) 1845:Returns: 0 on success, !0 on error 1846 1847:: 1848 1849 struct kvm_ppc_pvinfo { 1850 __u32 flags; 1851 __u32 hcall[4]; 1852 __u8 pad[108]; 1853 }; 1854 1855This ioctl fetches PV specific information that need to be passed to the guest 1856using the device tree or other means from vm context. 1857 1858The hcall array defines 4 instructions that make up a hypercall. 1859 1860If any additional field gets added to this structure later on, a bit for that 1861additional piece of information will be set in the flags bitmap. 1862 1863The flags bitmap is defined as:: 1864 1865 /* the host supports the ePAPR idle hcall 1866 #define KVM_PPC_PVINFO_FLAGS_EV_IDLE (1<<0) 1867 18684.52 KVM_SET_GSI_ROUTING 1869------------------------ 1870 1871:Capability: KVM_CAP_IRQ_ROUTING 1872:Architectures: x86 s390 arm64 1873:Type: vm ioctl 1874:Parameters: struct kvm_irq_routing (in) 1875:Returns: 0 on success, -1 on error 1876 1877Sets the GSI routing table entries, overwriting any previously set entries. 1878 1879On arm64, GSI routing has the following limitation: 1880 1881- GSI routing does not apply to KVM_IRQ_LINE but only to KVM_IRQFD. 1882 1883:: 1884 1885 struct kvm_irq_routing { 1886 __u32 nr; 1887 __u32 flags; 1888 struct kvm_irq_routing_entry entries[0]; 1889 }; 1890 1891No flags are specified so far, the corresponding field must be set to zero. 1892 1893:: 1894 1895 struct kvm_irq_routing_entry { 1896 __u32 gsi; 1897 __u32 type; 1898 __u32 flags; 1899 __u32 pad; 1900 union { 1901 struct kvm_irq_routing_irqchip irqchip; 1902 struct kvm_irq_routing_msi msi; 1903 struct kvm_irq_routing_s390_adapter adapter; 1904 struct kvm_irq_routing_hv_sint hv_sint; 1905 struct kvm_irq_routing_xen_evtchn xen_evtchn; 1906 __u32 pad[8]; 1907 } u; 1908 }; 1909 1910 /* gsi routing entry types */ 1911 #define KVM_IRQ_ROUTING_IRQCHIP 1 1912 #define KVM_IRQ_ROUTING_MSI 2 1913 #define KVM_IRQ_ROUTING_S390_ADAPTER 3 1914 #define KVM_IRQ_ROUTING_HV_SINT 4 1915 #define KVM_IRQ_ROUTING_XEN_EVTCHN 5 1916 1917On s390, adding a KVM_IRQ_ROUTING_S390_ADAPTER is rejected on ucontrol VMs with 1918error -EINVAL. 1919 1920flags: 1921 1922- KVM_MSI_VALID_DEVID: used along with KVM_IRQ_ROUTING_MSI routing entry 1923 type, specifies that the devid field contains a valid value. The per-VM 1924 KVM_CAP_MSI_DEVID capability advertises the requirement to provide 1925 the device ID. If this capability is not available, userspace should 1926 never set the KVM_MSI_VALID_DEVID flag as the ioctl might fail. 1927- zero otherwise 1928 1929:: 1930 1931 struct kvm_irq_routing_irqchip { 1932 __u32 irqchip; 1933 __u32 pin; 1934 }; 1935 1936 struct kvm_irq_routing_msi { 1937 __u32 address_lo; 1938 __u32 address_hi; 1939 __u32 data; 1940 union { 1941 __u32 pad; 1942 __u32 devid; 1943 }; 1944 }; 1945 1946If KVM_MSI_VALID_DEVID is set, devid contains a unique device identifier 1947for the device that wrote the MSI message. For PCI, this is usually a 1948BDF identifier in the lower 16 bits. 1949 1950On x86, address_hi is ignored unless the KVM_X2APIC_API_USE_32BIT_IDS 1951feature of KVM_CAP_X2APIC_API capability is enabled. If it is enabled, 1952address_hi bits 31-8 provide bits 31-8 of the destination id. Bits 7-0 of 1953address_hi must be zero. 1954 1955:: 1956 1957 struct kvm_irq_routing_s390_adapter { 1958 __u64 ind_addr; 1959 __u64 summary_addr; 1960 __u64 ind_offset; 1961 __u32 summary_offset; 1962 __u32 adapter_id; 1963 }; 1964 1965 struct kvm_irq_routing_hv_sint { 1966 __u32 vcpu; 1967 __u32 sint; 1968 }; 1969 1970 struct kvm_irq_routing_xen_evtchn { 1971 __u32 port; 1972 __u32 vcpu; 1973 __u32 priority; 1974 }; 1975 1976 1977When KVM_CAP_XEN_HVM includes the KVM_XEN_HVM_CONFIG_EVTCHN_2LEVEL bit 1978in its indication of supported features, routing to Xen event channels 1979is supported. Although the priority field is present, only the value 1980KVM_XEN_HVM_CONFIG_EVTCHN_2LEVEL is supported, which means delivery by 19812 level event channels. FIFO event channel support may be added in 1982the future. 1983 1984 19854.55 KVM_SET_TSC_KHZ 1986-------------------- 1987 1988:Capability: KVM_CAP_TSC_CONTROL / KVM_CAP_VM_TSC_CONTROL 1989:Architectures: x86 1990:Type: vcpu ioctl / vm ioctl 1991:Parameters: virtual tsc_khz 1992:Returns: 0 on success, -1 on error 1993 1994Specifies the tsc frequency for the virtual machine. The unit of the 1995frequency is KHz. 1996 1997If the KVM_CAP_VM_TSC_CONTROL capability is advertised, this can also 1998be used as a vm ioctl to set the initial tsc frequency of subsequently 1999created vCPUs. 2000 20014.56 KVM_GET_TSC_KHZ 2002-------------------- 2003 2004:Capability: KVM_CAP_GET_TSC_KHZ / KVM_CAP_VM_TSC_CONTROL 2005:Architectures: x86 2006:Type: vcpu ioctl / vm ioctl 2007:Parameters: none 2008:Returns: virtual tsc-khz on success, negative value on error 2009 2010Returns the tsc frequency of the guest. The unit of the return value is 2011KHz. If the host has unstable tsc this ioctl returns -EIO instead as an 2012error. 2013 2014 20154.57 KVM_GET_LAPIC 2016------------------ 2017 2018:Capability: KVM_CAP_IRQCHIP 2019:Architectures: x86 2020:Type: vcpu ioctl 2021:Parameters: struct kvm_lapic_state (out) 2022:Returns: 0 on success, -1 on error 2023 2024:: 2025 2026 #define KVM_APIC_REG_SIZE 0x400 2027 struct kvm_lapic_state { 2028 char regs[KVM_APIC_REG_SIZE]; 2029 }; 2030 2031Reads the Local APIC registers and copies them into the input argument. The 2032data format and layout are the same as documented in the architecture manual. 2033 2034If KVM_X2APIC_API_USE_32BIT_IDS feature of KVM_CAP_X2APIC_API is 2035enabled, then the format of APIC_ID register depends on the APIC mode 2036(reported by MSR_IA32_APICBASE) of its VCPU. x2APIC stores APIC ID in 2037the APIC_ID register (bytes 32-35). xAPIC only allows an 8-bit APIC ID 2038which is stored in bits 31-24 of the APIC register, or equivalently in 2039byte 35 of struct kvm_lapic_state's regs field. KVM_GET_LAPIC must then 2040be called after MSR_IA32_APICBASE has been set with KVM_SET_MSR. 2041 2042If KVM_X2APIC_API_USE_32BIT_IDS feature is disabled, struct kvm_lapic_state 2043always uses xAPIC format. 2044 2045 20464.58 KVM_SET_LAPIC 2047------------------ 2048 2049:Capability: KVM_CAP_IRQCHIP 2050:Architectures: x86 2051:Type: vcpu ioctl 2052:Parameters: struct kvm_lapic_state (in) 2053:Returns: 0 on success, -1 on error 2054 2055:: 2056 2057 #define KVM_APIC_REG_SIZE 0x400 2058 struct kvm_lapic_state { 2059 char regs[KVM_APIC_REG_SIZE]; 2060 }; 2061 2062Copies the input argument into the Local APIC registers. The data format 2063and layout are the same as documented in the architecture manual. 2064 2065The format of the APIC ID register (bytes 32-35 of struct kvm_lapic_state's 2066regs field) depends on the state of the KVM_CAP_X2APIC_API capability. 2067See the note in KVM_GET_LAPIC. 2068 2069 20704.59 KVM_IOEVENTFD 2071------------------ 2072 2073:Capability: KVM_CAP_IOEVENTFD 2074:Architectures: all 2075:Type: vm ioctl 2076:Parameters: struct kvm_ioeventfd (in) 2077:Returns: 0 on success, !0 on error 2078 2079This ioctl attaches or detaches an ioeventfd to a legal pio/mmio address 2080within the guest. A guest write in the registered address will signal the 2081provided event instead of triggering an exit. 2082 2083:: 2084 2085 struct kvm_ioeventfd { 2086 __u64 datamatch; 2087 __u64 addr; /* legal pio/mmio address */ 2088 __u32 len; /* 0, 1, 2, 4, or 8 bytes */ 2089 __s32 fd; 2090 __u32 flags; 2091 __u8 pad[36]; 2092 }; 2093 2094For the special case of virtio-ccw devices on s390, the ioevent is matched 2095to a subchannel/virtqueue tuple instead. 2096 2097The following flags are defined:: 2098 2099 #define KVM_IOEVENTFD_FLAG_DATAMATCH (1 << kvm_ioeventfd_flag_nr_datamatch) 2100 #define KVM_IOEVENTFD_FLAG_PIO (1 << kvm_ioeventfd_flag_nr_pio) 2101 #define KVM_IOEVENTFD_FLAG_DEASSIGN (1 << kvm_ioeventfd_flag_nr_deassign) 2102 #define KVM_IOEVENTFD_FLAG_VIRTIO_CCW_NOTIFY \ 2103 (1 << kvm_ioeventfd_flag_nr_virtio_ccw_notify) 2104 2105If datamatch flag is set, the event will be signaled only if the written value 2106to the registered address is equal to datamatch in struct kvm_ioeventfd. 2107 2108For virtio-ccw devices, addr contains the subchannel id and datamatch the 2109virtqueue index. 2110 2111With KVM_CAP_IOEVENTFD_ANY_LENGTH, a zero length ioeventfd is allowed, and 2112the kernel will ignore the length of guest write and may get a faster vmexit. 2113The speedup may only apply to specific architectures, but the ioeventfd will 2114work anyway. 2115 21164.60 KVM_DIRTY_TLB 2117------------------ 2118 2119:Capability: KVM_CAP_SW_TLB 2120:Architectures: ppc 2121:Type: vcpu ioctl 2122:Parameters: struct kvm_dirty_tlb (in) 2123:Returns: 0 on success, -1 on error 2124 2125:: 2126 2127 struct kvm_dirty_tlb { 2128 __u64 bitmap; 2129 __u32 num_dirty; 2130 }; 2131 2132This must be called whenever userspace has changed an entry in the shared 2133TLB, prior to calling KVM_RUN on the associated vcpu. 2134 2135The "bitmap" field is the userspace address of an array. This array 2136consists of a number of bits, equal to the total number of TLB entries as 2137determined by the last successful call to ``KVM_ENABLE_CAP(KVM_CAP_SW_TLB)``, 2138rounded up to the nearest multiple of 64. 2139 2140Each bit corresponds to one TLB entry, ordered the same as in the shared TLB 2141array. 2142 2143The array is little-endian: the bit 0 is the least significant bit of the 2144first byte, bit 8 is the least significant bit of the second byte, etc. 2145This avoids any complications with differing word sizes. 2146 2147The "num_dirty" field is a performance hint for KVM to determine whether it 2148should skip processing the bitmap and just invalidate everything. It must 2149be set to the number of set bits in the bitmap. 2150 2151 21524.62 KVM_CREATE_SPAPR_TCE 2153------------------------- 2154 2155:Capability: KVM_CAP_SPAPR_TCE 2156:Architectures: powerpc 2157:Type: vm ioctl 2158:Parameters: struct kvm_create_spapr_tce (in) 2159:Returns: file descriptor for manipulating the created TCE table 2160 2161This creates a virtual TCE (translation control entry) table, which 2162is an IOMMU for PAPR-style virtual I/O. It is used to translate 2163logical addresses used in virtual I/O into guest physical addresses, 2164and provides a scatter/gather capability for PAPR virtual I/O. 2165 2166:: 2167 2168 /* for KVM_CAP_SPAPR_TCE */ 2169 struct kvm_create_spapr_tce { 2170 __u64 liobn; 2171 __u32 window_size; 2172 }; 2173 2174The liobn field gives the logical IO bus number for which to create a 2175TCE table. The window_size field specifies the size of the DMA window 2176which this TCE table will translate - the table will contain one 64 2177bit TCE entry for every 4kiB of the DMA window. 2178 2179When the guest issues an H_PUT_TCE hcall on a liobn for which a TCE 2180table has been created using this ioctl(), the kernel will handle it 2181in real mode, updating the TCE table. H_PUT_TCE calls for other 2182liobns will cause a vm exit and must be handled by userspace. 2183 2184The return value is a file descriptor which can be passed to mmap(2) 2185to map the created TCE table into userspace. This lets userspace read 2186the entries written by kernel-handled H_PUT_TCE calls, and also lets 2187userspace update the TCE table directly which is useful in some 2188circumstances. 2189 2190 21914.64 KVM_NMI 2192------------ 2193 2194:Capability: KVM_CAP_USER_NMI 2195:Architectures: x86 2196:Type: vcpu ioctl 2197:Parameters: none 2198:Returns: 0 on success, -1 on error 2199 2200Queues an NMI on the thread's vcpu. Note this is well defined only 2201when KVM_CREATE_IRQCHIP has not been called, since this is an interface 2202between the virtual cpu core and virtual local APIC. After KVM_CREATE_IRQCHIP 2203has been called, this interface is completely emulated within the kernel. 2204 2205To use this to emulate the LINT1 input with KVM_CREATE_IRQCHIP, use the 2206following algorithm: 2207 2208 - pause the vcpu 2209 - read the local APIC's state (KVM_GET_LAPIC) 2210 - check whether changing LINT1 will queue an NMI (see the LVT entry for LINT1) 2211 - if so, issue KVM_NMI 2212 - resume the vcpu 2213 2214Some guests configure the LINT1 NMI input to cause a panic, aiding in 2215debugging. 2216 2217 22184.65 KVM_S390_UCAS_MAP 2219---------------------- 2220 2221:Capability: KVM_CAP_S390_UCONTROL 2222:Architectures: s390 2223:Type: vcpu ioctl 2224:Parameters: struct kvm_s390_ucas_mapping (in) 2225:Returns: 0 in case of success 2226 2227The parameter is defined like this:: 2228 2229 struct kvm_s390_ucas_mapping { 2230 __u64 user_addr; 2231 __u64 vcpu_addr; 2232 __u64 length; 2233 }; 2234 2235This ioctl maps the memory at "user_addr" with the length "length" to 2236the vcpu's address space starting at "vcpu_addr". All parameters need to 2237be aligned by 1 megabyte. 2238 2239 22404.66 KVM_S390_UCAS_UNMAP 2241------------------------ 2242 2243:Capability: KVM_CAP_S390_UCONTROL 2244:Architectures: s390 2245:Type: vcpu ioctl 2246:Parameters: struct kvm_s390_ucas_mapping (in) 2247:Returns: 0 in case of success 2248 2249The parameter is defined like this:: 2250 2251 struct kvm_s390_ucas_mapping { 2252 __u64 user_addr; 2253 __u64 vcpu_addr; 2254 __u64 length; 2255 }; 2256 2257This ioctl unmaps the memory in the vcpu's address space starting at 2258"vcpu_addr" with the length "length". The field "user_addr" is ignored. 2259All parameters need to be aligned by 1 megabyte. 2260 2261 22624.67 KVM_S390_VCPU_FAULT 2263------------------------ 2264 2265:Capability: KVM_CAP_S390_UCONTROL 2266:Architectures: s390 2267:Type: vcpu ioctl 2268:Parameters: vcpu absolute address (in) 2269:Returns: 0 in case of success 2270 2271This call creates a page table entry on the virtual cpu's address space 2272(for user controlled virtual machines) or the virtual machine's address 2273space (for regular virtual machines). This only works for minor faults, 2274thus it's recommended to access subject memory page via the user page 2275table upfront. This is useful to handle validity intercepts for user 2276controlled virtual machines to fault in the virtual cpu's lowcore pages 2277prior to calling the KVM_RUN ioctl. 2278 2279 22804.68 KVM_SET_ONE_REG 2281-------------------- 2282 2283:Capability: KVM_CAP_ONE_REG 2284:Architectures: all 2285:Type: vcpu ioctl 2286:Parameters: struct kvm_one_reg (in) 2287:Returns: 0 on success, negative value on failure 2288 2289Errors: 2290 2291 ====== ============================================================ 2292 ENOENT no such register 2293 EINVAL invalid register ID, or no such register or used with VMs in 2294 protected virtualization mode on s390 2295 EPERM (arm64) register access not allowed before vcpu finalization 2296 EBUSY (riscv) changing register value not allowed after the vcpu 2297 has run at least once 2298 ====== ============================================================ 2299 2300(These error codes are indicative only: do not rely on a specific error 2301code being returned in a specific situation.) 2302 2303:: 2304 2305 struct kvm_one_reg { 2306 __u64 id; 2307 __u64 addr; 2308 }; 2309 2310Using this ioctl, a single vcpu register can be set to a specific value 2311defined by user space with the passed in struct kvm_one_reg, where id 2312refers to the register identifier as described below and addr is a pointer 2313to a variable with the respective size. There can be architecture agnostic 2314and architecture specific registers. Each have their own range of operation 2315and their own constants and width. To keep track of the implemented 2316registers, find a list below: 2317 2318 ======= =============================== ============ 2319 Arch Register Width (bits) 2320 ======= =============================== ============ 2321 PPC KVM_REG_PPC_HIOR 64 2322 PPC KVM_REG_PPC_IAC1 64 2323 PPC KVM_REG_PPC_IAC2 64 2324 PPC KVM_REG_PPC_IAC3 64 2325 PPC KVM_REG_PPC_IAC4 64 2326 PPC KVM_REG_PPC_DAC1 64 2327 PPC KVM_REG_PPC_DAC2 64 2328 PPC KVM_REG_PPC_DABR 64 2329 PPC KVM_REG_PPC_DSCR 64 2330 PPC KVM_REG_PPC_PURR 64 2331 PPC KVM_REG_PPC_SPURR 64 2332 PPC KVM_REG_PPC_DAR 64 2333 PPC KVM_REG_PPC_DSISR 32 2334 PPC KVM_REG_PPC_AMR 64 2335 PPC KVM_REG_PPC_UAMOR 64 2336 PPC KVM_REG_PPC_MMCR0 64 2337 PPC KVM_REG_PPC_MMCR1 64 2338 PPC KVM_REG_PPC_MMCRA 64 2339 PPC KVM_REG_PPC_MMCR2 64 2340 PPC KVM_REG_PPC_MMCRS 64 2341 PPC KVM_REG_PPC_MMCR3 64 2342 PPC KVM_REG_PPC_SIAR 64 2343 PPC KVM_REG_PPC_SDAR 64 2344 PPC KVM_REG_PPC_SIER 64 2345 PPC KVM_REG_PPC_SIER2 64 2346 PPC KVM_REG_PPC_SIER3 64 2347 PPC KVM_REG_PPC_PMC1 32 2348 PPC KVM_REG_PPC_PMC2 32 2349 PPC KVM_REG_PPC_PMC3 32 2350 PPC KVM_REG_PPC_PMC4 32 2351 PPC KVM_REG_PPC_PMC5 32 2352 PPC KVM_REG_PPC_PMC6 32 2353 PPC KVM_REG_PPC_PMC7 32 2354 PPC KVM_REG_PPC_PMC8 32 2355 PPC KVM_REG_PPC_FPR0 64 2356 ... 2357 PPC KVM_REG_PPC_FPR31 64 2358 PPC KVM_REG_PPC_VR0 128 2359 ... 2360 PPC KVM_REG_PPC_VR31 128 2361 PPC KVM_REG_PPC_VSR0 128 2362 ... 2363 PPC KVM_REG_PPC_VSR31 128 2364 PPC KVM_REG_PPC_FPSCR 64 2365 PPC KVM_REG_PPC_VSCR 32 2366 PPC KVM_REG_PPC_VPA_ADDR 64 2367 PPC KVM_REG_PPC_VPA_SLB 128 2368 PPC KVM_REG_PPC_VPA_DTL 128 2369 PPC KVM_REG_PPC_EPCR 32 2370 PPC KVM_REG_PPC_EPR 32 2371 PPC KVM_REG_PPC_TCR 32 2372 PPC KVM_REG_PPC_TSR 32 2373 PPC KVM_REG_PPC_OR_TSR 32 2374 PPC KVM_REG_PPC_CLEAR_TSR 32 2375 PPC KVM_REG_PPC_MAS0 32 2376 PPC KVM_REG_PPC_MAS1 32 2377 PPC KVM_REG_PPC_MAS2 64 2378 PPC KVM_REG_PPC_MAS7_3 64 2379 PPC KVM_REG_PPC_MAS4 32 2380 PPC KVM_REG_PPC_MAS6 32 2381 PPC KVM_REG_PPC_MMUCFG 32 2382 PPC KVM_REG_PPC_TLB0CFG 32 2383 PPC KVM_REG_PPC_TLB1CFG 32 2384 PPC KVM_REG_PPC_TLB2CFG 32 2385 PPC KVM_REG_PPC_TLB3CFG 32 2386 PPC KVM_REG_PPC_TLB0PS 32 2387 PPC KVM_REG_PPC_TLB1PS 32 2388 PPC KVM_REG_PPC_TLB2PS 32 2389 PPC KVM_REG_PPC_TLB3PS 32 2390 PPC KVM_REG_PPC_EPTCFG 32 2391 PPC KVM_REG_PPC_ICP_STATE 64 2392 PPC KVM_REG_PPC_VP_STATE 128 2393 PPC KVM_REG_PPC_TB_OFFSET 64 2394 PPC KVM_REG_PPC_SPMC1 32 2395 PPC KVM_REG_PPC_SPMC2 32 2396 PPC KVM_REG_PPC_IAMR 64 2397 PPC KVM_REG_PPC_TFHAR 64 2398 PPC KVM_REG_PPC_TFIAR 64 2399 PPC KVM_REG_PPC_TEXASR 64 2400 PPC KVM_REG_PPC_FSCR 64 2401 PPC KVM_REG_PPC_PSPB 32 2402 PPC KVM_REG_PPC_EBBHR 64 2403 PPC KVM_REG_PPC_EBBRR 64 2404 PPC KVM_REG_PPC_BESCR 64 2405 PPC KVM_REG_PPC_TAR 64 2406 PPC KVM_REG_PPC_DPDES 64 2407 PPC KVM_REG_PPC_DAWR 64 2408 PPC KVM_REG_PPC_DAWRX 64 2409 PPC KVM_REG_PPC_CIABR 64 2410 PPC KVM_REG_PPC_IC 64 2411 PPC KVM_REG_PPC_VTB 64 2412 PPC KVM_REG_PPC_CSIGR 64 2413 PPC KVM_REG_PPC_TACR 64 2414 PPC KVM_REG_PPC_TCSCR 64 2415 PPC KVM_REG_PPC_PID 64 2416 PPC KVM_REG_PPC_ACOP 64 2417 PPC KVM_REG_PPC_VRSAVE 32 2418 PPC KVM_REG_PPC_LPCR 32 2419 PPC KVM_REG_PPC_LPCR_64 64 2420 PPC KVM_REG_PPC_PPR 64 2421 PPC KVM_REG_PPC_ARCH_COMPAT 32 2422 PPC KVM_REG_PPC_DABRX 32 2423 PPC KVM_REG_PPC_WORT 64 2424 PPC KVM_REG_PPC_SPRG9 64 2425 PPC KVM_REG_PPC_DBSR 32 2426 PPC KVM_REG_PPC_TIDR 64 2427 PPC KVM_REG_PPC_PSSCR 64 2428 PPC KVM_REG_PPC_DEC_EXPIRY 64 2429 PPC KVM_REG_PPC_PTCR 64 2430 PPC KVM_REG_PPC_HASHKEYR 64 2431 PPC KVM_REG_PPC_HASHPKEYR 64 2432 PPC KVM_REG_PPC_DAWR1 64 2433 PPC KVM_REG_PPC_DAWRX1 64 2434 PPC KVM_REG_PPC_DEXCR 64 2435 PPC KVM_REG_PPC_TM_GPR0 64 2436 ... 2437 PPC KVM_REG_PPC_TM_GPR31 64 2438 PPC KVM_REG_PPC_TM_VSR0 128 2439 ... 2440 PPC KVM_REG_PPC_TM_VSR63 128 2441 PPC KVM_REG_PPC_TM_CR 64 2442 PPC KVM_REG_PPC_TM_LR 64 2443 PPC KVM_REG_PPC_TM_CTR 64 2444 PPC KVM_REG_PPC_TM_FPSCR 64 2445 PPC KVM_REG_PPC_TM_AMR 64 2446 PPC KVM_REG_PPC_TM_PPR 64 2447 PPC KVM_REG_PPC_TM_VRSAVE 64 2448 PPC KVM_REG_PPC_TM_VSCR 32 2449 PPC KVM_REG_PPC_TM_DSCR 64 2450 PPC KVM_REG_PPC_TM_TAR 64 2451 PPC KVM_REG_PPC_TM_XER 64 2452 2453 MIPS KVM_REG_MIPS_R0 64 2454 ... 2455 MIPS KVM_REG_MIPS_R31 64 2456 MIPS KVM_REG_MIPS_HI 64 2457 MIPS KVM_REG_MIPS_LO 64 2458 MIPS KVM_REG_MIPS_PC 64 2459 MIPS KVM_REG_MIPS_CP0_INDEX 32 2460 MIPS KVM_REG_MIPS_CP0_ENTRYLO0 64 2461 MIPS KVM_REG_MIPS_CP0_ENTRYLO1 64 2462 MIPS KVM_REG_MIPS_CP0_CONTEXT 64 2463 MIPS KVM_REG_MIPS_CP0_CONTEXTCONFIG 32 2464 MIPS KVM_REG_MIPS_CP0_USERLOCAL 64 2465 MIPS KVM_REG_MIPS_CP0_XCONTEXTCONFIG 64 2466 MIPS KVM_REG_MIPS_CP0_PAGEMASK 32 2467 MIPS KVM_REG_MIPS_CP0_PAGEGRAIN 32 2468 MIPS KVM_REG_MIPS_CP0_SEGCTL0 64 2469 MIPS KVM_REG_MIPS_CP0_SEGCTL1 64 2470 MIPS KVM_REG_MIPS_CP0_SEGCTL2 64 2471 MIPS KVM_REG_MIPS_CP0_PWBASE 64 2472 MIPS KVM_REG_MIPS_CP0_PWFIELD 64 2473 MIPS KVM_REG_MIPS_CP0_PWSIZE 64 2474 MIPS KVM_REG_MIPS_CP0_WIRED 32 2475 MIPS KVM_REG_MIPS_CP0_PWCTL 32 2476 MIPS KVM_REG_MIPS_CP0_HWRENA 32 2477 MIPS KVM_REG_MIPS_CP0_BADVADDR 64 2478 MIPS KVM_REG_MIPS_CP0_BADINSTR 32 2479 MIPS KVM_REG_MIPS_CP0_BADINSTRP 32 2480 MIPS KVM_REG_MIPS_CP0_COUNT 32 2481 MIPS KVM_REG_MIPS_CP0_ENTRYHI 64 2482 MIPS KVM_REG_MIPS_CP0_COMPARE 32 2483 MIPS KVM_REG_MIPS_CP0_STATUS 32 2484 MIPS KVM_REG_MIPS_CP0_INTCTL 32 2485 MIPS KVM_REG_MIPS_CP0_CAUSE 32 2486 MIPS KVM_REG_MIPS_CP0_EPC 64 2487 MIPS KVM_REG_MIPS_CP0_PRID 32 2488 MIPS KVM_REG_MIPS_CP0_EBASE 64 2489 MIPS KVM_REG_MIPS_CP0_CONFIG 32 2490 MIPS KVM_REG_MIPS_CP0_CONFIG1 32 2491 MIPS KVM_REG_MIPS_CP0_CONFIG2 32 2492 MIPS KVM_REG_MIPS_CP0_CONFIG3 32 2493 MIPS KVM_REG_MIPS_CP0_CONFIG4 32 2494 MIPS KVM_REG_MIPS_CP0_CONFIG5 32 2495 MIPS KVM_REG_MIPS_CP0_CONFIG7 32 2496 MIPS KVM_REG_MIPS_CP0_XCONTEXT 64 2497 MIPS KVM_REG_MIPS_CP0_ERROREPC 64 2498 MIPS KVM_REG_MIPS_CP0_KSCRATCH1 64 2499 MIPS KVM_REG_MIPS_CP0_KSCRATCH2 64 2500 MIPS KVM_REG_MIPS_CP0_KSCRATCH3 64 2501 MIPS KVM_REG_MIPS_CP0_KSCRATCH4 64 2502 MIPS KVM_REG_MIPS_CP0_KSCRATCH5 64 2503 MIPS KVM_REG_MIPS_CP0_KSCRATCH6 64 2504 MIPS KVM_REG_MIPS_CP0_MAAR(0..63) 64 2505 MIPS KVM_REG_MIPS_COUNT_CTL 64 2506 MIPS KVM_REG_MIPS_COUNT_RESUME 64 2507 MIPS KVM_REG_MIPS_COUNT_HZ 64 2508 MIPS KVM_REG_MIPS_FPR_32(0..31) 32 2509 MIPS KVM_REG_MIPS_FPR_64(0..31) 64 2510 MIPS KVM_REG_MIPS_VEC_128(0..31) 128 2511 MIPS KVM_REG_MIPS_FCR_IR 32 2512 MIPS KVM_REG_MIPS_FCR_CSR 32 2513 MIPS KVM_REG_MIPS_MSA_IR 32 2514 MIPS KVM_REG_MIPS_MSA_CSR 32 2515 ======= =============================== ============ 2516 2517ARM registers are mapped using the lower 32 bits. The upper 16 of that 2518is the register group type, or coprocessor number: 2519 2520ARM core registers have the following id bit patterns:: 2521 2522 0x4020 0000 0010 <index into the kvm_regs struct:16> 2523 2524ARM 32-bit CP15 registers have the following id bit patterns:: 2525 2526 0x4020 0000 000F <zero:1> <crn:4> <crm:4> <opc1:4> <opc2:3> 2527 2528ARM 64-bit CP15 registers have the following id bit patterns:: 2529 2530 0x4030 0000 000F <zero:1> <zero:4> <crm:4> <opc1:4> <zero:3> 2531 2532ARM CCSIDR registers are demultiplexed by CSSELR value:: 2533 2534 0x4020 0000 0011 00 <csselr:8> 2535 2536ARM 32-bit VFP control registers have the following id bit patterns:: 2537 2538 0x4020 0000 0012 1 <regno:12> 2539 2540ARM 64-bit FP registers have the following id bit patterns:: 2541 2542 0x4030 0000 0012 0 <regno:12> 2543 2544ARM firmware pseudo-registers have the following bit pattern:: 2545 2546 0x4030 0000 0014 <regno:16> 2547 2548 2549arm64 registers are mapped using the lower 32 bits. The upper 16 of 2550that is the register group type, or coprocessor number: 2551 2552arm64 core/FP-SIMD registers have the following id bit patterns. Note 2553that the size of the access is variable, as the kvm_regs structure 2554contains elements ranging from 32 to 128 bits. The index is a 32bit 2555value in the kvm_regs structure seen as a 32bit array:: 2556 2557 0x60x0 0000 0010 <index into the kvm_regs struct:16> 2558 2559Specifically: 2560 2561======================= ========= ===== ======================================= 2562 Encoding Register Bits kvm_regs member 2563======================= ========= ===== ======================================= 2564 0x6030 0000 0010 0000 X0 64 regs.regs[0] 2565 0x6030 0000 0010 0002 X1 64 regs.regs[1] 2566 ... 2567 0x6030 0000 0010 003c X30 64 regs.regs[30] 2568 0x6030 0000 0010 003e SP 64 regs.sp 2569 0x6030 0000 0010 0040 PC 64 regs.pc 2570 0x6030 0000 0010 0042 PSTATE 64 regs.pstate 2571 0x6030 0000 0010 0044 SP_EL1 64 sp_el1 2572 0x6030 0000 0010 0046 ELR_EL1 64 elr_el1 2573 0x6030 0000 0010 0048 SPSR_EL1 64 spsr[KVM_SPSR_EL1] (alias SPSR_SVC) 2574 0x6030 0000 0010 004a SPSR_ABT 64 spsr[KVM_SPSR_ABT] 2575 0x6030 0000 0010 004c SPSR_UND 64 spsr[KVM_SPSR_UND] 2576 0x6030 0000 0010 004e SPSR_IRQ 64 spsr[KVM_SPSR_IRQ] 2577 0x6030 0000 0010 0050 SPSR_FIQ 64 spsr[KVM_SPSR_FIQ] 2578 0x6040 0000 0010 0054 V0 128 fp_regs.vregs[0] [1]_ 2579 0x6040 0000 0010 0058 V1 128 fp_regs.vregs[1] [1]_ 2580 ... 2581 0x6040 0000 0010 00d0 V31 128 fp_regs.vregs[31] [1]_ 2582 0x6020 0000 0010 00d4 FPSR 32 fp_regs.fpsr 2583 0x6020 0000 0010 00d5 FPCR 32 fp_regs.fpcr 2584======================= ========= ===== ======================================= 2585 2586.. [1] These encodings are not accepted for SVE-enabled vcpus. See 2587 :ref:`KVM_ARM_VCPU_INIT`. 2588 2589 The equivalent register content can be accessed via bits [127:0] of 2590 the corresponding SVE Zn registers instead for vcpus that have SVE 2591 enabled (see below). 2592 2593arm64 CCSIDR registers are demultiplexed by CSSELR value:: 2594 2595 0x6020 0000 0011 00 <csselr:8> 2596 2597arm64 system registers have the following id bit patterns:: 2598 2599 0x6030 0000 0013 <op0:2> <op1:3> <crn:4> <crm:4> <op2:3> 2600 2601.. warning:: 2602 2603 Two system register IDs do not follow the specified pattern. These 2604 are KVM_REG_ARM_TIMER_CVAL and KVM_REG_ARM_TIMER_CNT, which map to 2605 system registers CNTV_CVAL_EL0 and CNTVCT_EL0 respectively. These 2606 two had their values accidentally swapped, which means TIMER_CVAL is 2607 derived from the register encoding for CNTVCT_EL0 and TIMER_CNT is 2608 derived from the register encoding for CNTV_CVAL_EL0. As this is 2609 API, it must remain this way. 2610 2611arm64 firmware pseudo-registers have the following bit pattern:: 2612 2613 0x6030 0000 0014 <regno:16> 2614 2615arm64 SVE registers have the following bit patterns:: 2616 2617 0x6080 0000 0015 00 <n:5> <slice:5> Zn bits[2048*slice + 2047 : 2048*slice] 2618 0x6050 0000 0015 04 <n:4> <slice:5> Pn bits[256*slice + 255 : 256*slice] 2619 0x6050 0000 0015 060 <slice:5> FFR bits[256*slice + 255 : 256*slice] 2620 0x6060 0000 0015 ffff KVM_REG_ARM64_SVE_VLS pseudo-register 2621 2622Access to register IDs where 2048 * slice >= 128 * max_vq will fail with 2623ENOENT. max_vq is the vcpu's maximum supported vector length in 128-bit 2624quadwords: see [2]_ below. 2625 2626These registers are only accessible on vcpus for which SVE is enabled. 2627See KVM_ARM_VCPU_INIT for details. 2628 2629In addition, except for KVM_REG_ARM64_SVE_VLS, these registers are not 2630accessible until the vcpu's SVE configuration has been finalized 2631using KVM_ARM_VCPU_FINALIZE(KVM_ARM_VCPU_SVE). See KVM_ARM_VCPU_INIT 2632and KVM_ARM_VCPU_FINALIZE for more information about this procedure. 2633 2634KVM_REG_ARM64_SVE_VLS is a pseudo-register that allows the set of vector 2635lengths supported by the vcpu to be discovered and configured by 2636userspace. When transferred to or from user memory via KVM_GET_ONE_REG 2637or KVM_SET_ONE_REG, the value of this register is of type 2638__u64[KVM_ARM64_SVE_VLS_WORDS], and encodes the set of vector lengths as 2639follows:: 2640 2641 __u64 vector_lengths[KVM_ARM64_SVE_VLS_WORDS]; 2642 2643 if (vq >= SVE_VQ_MIN && vq <= SVE_VQ_MAX && 2644 ((vector_lengths[(vq - KVM_ARM64_SVE_VQ_MIN) / 64] >> 2645 ((vq - KVM_ARM64_SVE_VQ_MIN) % 64)) & 1)) 2646 /* Vector length vq * 16 bytes supported */ 2647 else 2648 /* Vector length vq * 16 bytes not supported */ 2649 2650.. [2] The maximum value vq for which the above condition is true is 2651 max_vq. This is the maximum vector length available to the guest on 2652 this vcpu, and determines which register slices are visible through 2653 this ioctl interface. 2654 2655(See Documentation/arch/arm64/sve.rst for an explanation of the "vq" 2656nomenclature.) 2657 2658KVM_REG_ARM64_SVE_VLS is only accessible after KVM_ARM_VCPU_INIT. 2659KVM_ARM_VCPU_INIT initialises it to the best set of vector lengths that 2660the host supports. 2661 2662Userspace may subsequently modify it if desired until the vcpu's SVE 2663configuration is finalized using KVM_ARM_VCPU_FINALIZE(KVM_ARM_VCPU_SVE). 2664 2665Apart from simply removing all vector lengths from the host set that 2666exceed some value, support for arbitrarily chosen sets of vector lengths 2667is hardware-dependent and may not be available. Attempting to configure 2668an invalid set of vector lengths via KVM_SET_ONE_REG will fail with 2669EINVAL. 2670 2671After the vcpu's SVE configuration is finalized, further attempts to 2672write this register will fail with EPERM. 2673 2674arm64 bitmap feature firmware pseudo-registers have the following bit pattern:: 2675 2676 0x6030 0000 0016 <regno:16> 2677 2678The bitmap feature firmware registers exposes the hypercall services that 2679are available for userspace to configure. The set bits corresponds to the 2680services that are available for the guests to access. By default, KVM 2681sets all the supported bits during VM initialization. The userspace can 2682discover the available services via KVM_GET_ONE_REG, and write back the 2683bitmap corresponding to the features that it wishes guests to see via 2684KVM_SET_ONE_REG. 2685 2686Note: These registers are immutable once any of the vCPUs of the VM has 2687run at least once. A KVM_SET_ONE_REG in such a scenario will return 2688a -EBUSY to userspace. 2689 2690(See Documentation/virt/kvm/arm/hypercalls.rst for more details.) 2691 2692 2693MIPS registers are mapped using the lower 32 bits. The upper 16 of that is 2694the register group type: 2695 2696MIPS core registers (see above) have the following id bit patterns:: 2697 2698 0x7030 0000 0000 <reg:16> 2699 2700MIPS CP0 registers (see KVM_REG_MIPS_CP0_* above) have the following id bit 2701patterns depending on whether they're 32-bit or 64-bit registers:: 2702 2703 0x7020 0000 0001 00 <reg:5> <sel:3> (32-bit) 2704 0x7030 0000 0001 00 <reg:5> <sel:3> (64-bit) 2705 2706Note: KVM_REG_MIPS_CP0_ENTRYLO0 and KVM_REG_MIPS_CP0_ENTRYLO1 are the MIPS64 2707versions of the EntryLo registers regardless of the word size of the host 2708hardware, host kernel, guest, and whether XPA is present in the guest, i.e. 2709with the RI and XI bits (if they exist) in bits 63 and 62 respectively, and 2710the PFNX field starting at bit 30. 2711 2712MIPS MAARs (see KVM_REG_MIPS_CP0_MAAR(*) above) have the following id bit 2713patterns:: 2714 2715 0x7030 0000 0001 01 <reg:8> 2716 2717MIPS KVM control registers (see above) have the following id bit patterns:: 2718 2719 0x7030 0000 0002 <reg:16> 2720 2721MIPS FPU registers (see KVM_REG_MIPS_FPR_{32,64}() above) have the following 2722id bit patterns depending on the size of the register being accessed. They are 2723always accessed according to the current guest FPU mode (Status.FR and 2724Config5.FRE), i.e. as the guest would see them, and they become unpredictable 2725if the guest FPU mode is changed. MIPS SIMD Architecture (MSA) vector 2726registers (see KVM_REG_MIPS_VEC_128() above) have similar patterns as they 2727overlap the FPU registers:: 2728 2729 0x7020 0000 0003 00 <0:3> <reg:5> (32-bit FPU registers) 2730 0x7030 0000 0003 00 <0:3> <reg:5> (64-bit FPU registers) 2731 0x7040 0000 0003 00 <0:3> <reg:5> (128-bit MSA vector registers) 2732 2733MIPS FPU control registers (see KVM_REG_MIPS_FCR_{IR,CSR} above) have the 2734following id bit patterns:: 2735 2736 0x7020 0000 0003 01 <0:3> <reg:5> 2737 2738MIPS MSA control registers (see KVM_REG_MIPS_MSA_{IR,CSR} above) have the 2739following id bit patterns:: 2740 2741 0x7020 0000 0003 02 <0:3> <reg:5> 2742 2743RISC-V registers are mapped using the lower 32 bits. The upper 8 bits of 2744that is the register group type. 2745 2746RISC-V config registers are meant for configuring a Guest VCPU and it has 2747the following id bit patterns:: 2748 2749 0x8020 0000 01 <index into the kvm_riscv_config struct:24> (32bit Host) 2750 0x8030 0000 01 <index into the kvm_riscv_config struct:24> (64bit Host) 2751 2752Following are the RISC-V config registers: 2753 2754======================= ========= ============================================= 2755 Encoding Register Description 2756======================= ========= ============================================= 2757 0x80x0 0000 0100 0000 isa ISA feature bitmap of Guest VCPU 2758======================= ========= ============================================= 2759 2760The isa config register can be read anytime but can only be written before 2761a Guest VCPU runs. It will have ISA feature bits matching underlying host 2762set by default. 2763 2764RISC-V core registers represent the general execution state of a Guest VCPU 2765and it has the following id bit patterns:: 2766 2767 0x8020 0000 02 <index into the kvm_riscv_core struct:24> (32bit Host) 2768 0x8030 0000 02 <index into the kvm_riscv_core struct:24> (64bit Host) 2769 2770Following are the RISC-V core registers: 2771 2772======================= ========= ============================================= 2773 Encoding Register Description 2774======================= ========= ============================================= 2775 0x80x0 0000 0200 0000 regs.pc Program counter 2776 0x80x0 0000 0200 0001 regs.ra Return address 2777 0x80x0 0000 0200 0002 regs.sp Stack pointer 2778 0x80x0 0000 0200 0003 regs.gp Global pointer 2779 0x80x0 0000 0200 0004 regs.tp Task pointer 2780 0x80x0 0000 0200 0005 regs.t0 Caller saved register 0 2781 0x80x0 0000 0200 0006 regs.t1 Caller saved register 1 2782 0x80x0 0000 0200 0007 regs.t2 Caller saved register 2 2783 0x80x0 0000 0200 0008 regs.s0 Callee saved register 0 2784 0x80x0 0000 0200 0009 regs.s1 Callee saved register 1 2785 0x80x0 0000 0200 000a regs.a0 Function argument (or return value) 0 2786 0x80x0 0000 0200 000b regs.a1 Function argument (or return value) 1 2787 0x80x0 0000 0200 000c regs.a2 Function argument 2 2788 0x80x0 0000 0200 000d regs.a3 Function argument 3 2789 0x80x0 0000 0200 000e regs.a4 Function argument 4 2790 0x80x0 0000 0200 000f regs.a5 Function argument 5 2791 0x80x0 0000 0200 0010 regs.a6 Function argument 6 2792 0x80x0 0000 0200 0011 regs.a7 Function argument 7 2793 0x80x0 0000 0200 0012 regs.s2 Callee saved register 2 2794 0x80x0 0000 0200 0013 regs.s3 Callee saved register 3 2795 0x80x0 0000 0200 0014 regs.s4 Callee saved register 4 2796 0x80x0 0000 0200 0015 regs.s5 Callee saved register 5 2797 0x80x0 0000 0200 0016 regs.s6 Callee saved register 6 2798 0x80x0 0000 0200 0017 regs.s7 Callee saved register 7 2799 0x80x0 0000 0200 0018 regs.s8 Callee saved register 8 2800 0x80x0 0000 0200 0019 regs.s9 Callee saved register 9 2801 0x80x0 0000 0200 001a regs.s10 Callee saved register 10 2802 0x80x0 0000 0200 001b regs.s11 Callee saved register 11 2803 0x80x0 0000 0200 001c regs.t3 Caller saved register 3 2804 0x80x0 0000 0200 001d regs.t4 Caller saved register 4 2805 0x80x0 0000 0200 001e regs.t5 Caller saved register 5 2806 0x80x0 0000 0200 001f regs.t6 Caller saved register 6 2807 0x80x0 0000 0200 0020 mode Privilege mode (1 = S-mode or 0 = U-mode) 2808======================= ========= ============================================= 2809 2810RISC-V csr registers represent the supervisor mode control/status registers 2811of a Guest VCPU and it has the following id bit patterns:: 2812 2813 0x8020 0000 03 <index into the kvm_riscv_csr struct:24> (32bit Host) 2814 0x8030 0000 03 <index into the kvm_riscv_csr struct:24> (64bit Host) 2815 2816Following are the RISC-V csr registers: 2817 2818======================= ========= ============================================= 2819 Encoding Register Description 2820======================= ========= ============================================= 2821 0x80x0 0000 0300 0000 sstatus Supervisor status 2822 0x80x0 0000 0300 0001 sie Supervisor interrupt enable 2823 0x80x0 0000 0300 0002 stvec Supervisor trap vector base 2824 0x80x0 0000 0300 0003 sscratch Supervisor scratch register 2825 0x80x0 0000 0300 0004 sepc Supervisor exception program counter 2826 0x80x0 0000 0300 0005 scause Supervisor trap cause 2827 0x80x0 0000 0300 0006 stval Supervisor bad address or instruction 2828 0x80x0 0000 0300 0007 sip Supervisor interrupt pending 2829 0x80x0 0000 0300 0008 satp Supervisor address translation and protection 2830======================= ========= ============================================= 2831 2832RISC-V timer registers represent the timer state of a Guest VCPU and it has 2833the following id bit patterns:: 2834 2835 0x8030 0000 04 <index into the kvm_riscv_timer struct:24> 2836 2837Following are the RISC-V timer registers: 2838 2839======================= ========= ============================================= 2840 Encoding Register Description 2841======================= ========= ============================================= 2842 0x8030 0000 0400 0000 frequency Time base frequency (read-only) 2843 0x8030 0000 0400 0001 time Time value visible to Guest 2844 0x8030 0000 0400 0002 compare Time compare programmed by Guest 2845 0x8030 0000 0400 0003 state Time compare state (1 = ON or 0 = OFF) 2846======================= ========= ============================================= 2847 2848RISC-V F-extension registers represent the single precision floating point 2849state of a Guest VCPU and it has the following id bit patterns:: 2850 2851 0x8020 0000 05 <index into the __riscv_f_ext_state struct:24> 2852 2853Following are the RISC-V F-extension registers: 2854 2855======================= ========= ============================================= 2856 Encoding Register Description 2857======================= ========= ============================================= 2858 0x8020 0000 0500 0000 f[0] Floating point register 0 2859 ... 2860 0x8020 0000 0500 001f f[31] Floating point register 31 2861 0x8020 0000 0500 0020 fcsr Floating point control and status register 2862======================= ========= ============================================= 2863 2864RISC-V D-extension registers represent the double precision floating point 2865state of a Guest VCPU and it has the following id bit patterns:: 2866 2867 0x8020 0000 06 <index into the __riscv_d_ext_state struct:24> (fcsr) 2868 0x8030 0000 06 <index into the __riscv_d_ext_state struct:24> (non-fcsr) 2869 2870Following are the RISC-V D-extension registers: 2871 2872======================= ========= ============================================= 2873 Encoding Register Description 2874======================= ========= ============================================= 2875 0x8030 0000 0600 0000 f[0] Floating point register 0 2876 ... 2877 0x8030 0000 0600 001f f[31] Floating point register 31 2878 0x8020 0000 0600 0020 fcsr Floating point control and status register 2879======================= ========= ============================================= 2880 2881LoongArch registers are mapped using the lower 32 bits. The upper 16 bits of 2882that is the register group type. 2883 2884LoongArch csr registers are used to control guest cpu or get status of guest 2885cpu, and they have the following id bit patterns:: 2886 2887 0x9030 0000 0001 00 <reg:5> <sel:3> (64-bit) 2888 2889LoongArch KVM control registers are used to implement some new defined functions 2890such as set vcpu counter or reset vcpu, and they have the following id bit patterns:: 2891 2892 0x9030 0000 0002 <reg:16> 2893 2894 28954.69 KVM_GET_ONE_REG 2896-------------------- 2897 2898:Capability: KVM_CAP_ONE_REG 2899:Architectures: all 2900:Type: vcpu ioctl 2901:Parameters: struct kvm_one_reg (in and out) 2902:Returns: 0 on success, negative value on failure 2903 2904Errors include: 2905 2906 ======== ============================================================ 2907 ENOENT no such register 2908 EINVAL invalid register ID, or no such register or used with VMs in 2909 protected virtualization mode on s390 2910 EPERM (arm64) register access not allowed before vcpu finalization 2911 ======== ============================================================ 2912 2913(These error codes are indicative only: do not rely on a specific error 2914code being returned in a specific situation.) 2915 2916This ioctl allows to receive the value of a single register implemented 2917in a vcpu. The register to read is indicated by the "id" field of the 2918kvm_one_reg struct passed in. On success, the register value can be found 2919at the memory location pointed to by "addr". 2920 2921The list of registers accessible using this interface is identical to the 2922list in 4.68. 2923 2924 29254.70 KVM_KVMCLOCK_CTRL 2926---------------------- 2927 2928:Capability: KVM_CAP_KVMCLOCK_CTRL 2929:Architectures: Any that implement pvclocks (currently x86 only) 2930:Type: vcpu ioctl 2931:Parameters: None 2932:Returns: 0 on success, -1 on error 2933 2934This ioctl sets a flag accessible to the guest indicating that the specified 2935vCPU has been paused by the host userspace. 2936 2937The host will set a flag in the pvclock structure that is checked from the 2938soft lockup watchdog. The flag is part of the pvclock structure that is 2939shared between guest and host, specifically the second bit of the flags 2940field of the pvclock_vcpu_time_info structure. It will be set exclusively by 2941the host and read/cleared exclusively by the guest. The guest operation of 2942checking and clearing the flag must be an atomic operation so 2943load-link/store-conditional, or equivalent must be used. There are two cases 2944where the guest will clear the flag: when the soft lockup watchdog timer resets 2945itself or when a soft lockup is detected. This ioctl can be called any time 2946after pausing the vcpu, but before it is resumed. 2947 2948 29494.71 KVM_SIGNAL_MSI 2950------------------- 2951 2952:Capability: KVM_CAP_SIGNAL_MSI 2953:Architectures: x86 arm64 2954:Type: vm ioctl 2955:Parameters: struct kvm_msi (in) 2956:Returns: >0 on delivery, 0 if guest blocked the MSI, and -1 on error 2957 2958Directly inject a MSI message. Only valid with in-kernel irqchip that handles 2959MSI messages. 2960 2961:: 2962 2963 struct kvm_msi { 2964 __u32 address_lo; 2965 __u32 address_hi; 2966 __u32 data; 2967 __u32 flags; 2968 __u32 devid; 2969 __u8 pad[12]; 2970 }; 2971 2972flags: 2973 KVM_MSI_VALID_DEVID: devid contains a valid value. The per-VM 2974 KVM_CAP_MSI_DEVID capability advertises the requirement to provide 2975 the device ID. If this capability is not available, userspace 2976 should never set the KVM_MSI_VALID_DEVID flag as the ioctl might fail. 2977 2978If KVM_MSI_VALID_DEVID is set, devid contains a unique device identifier 2979for the device that wrote the MSI message. For PCI, this is usually a 2980BDF identifier in the lower 16 bits. 2981 2982On x86, address_hi is ignored unless the KVM_X2APIC_API_USE_32BIT_IDS 2983feature of KVM_CAP_X2APIC_API capability is enabled. If it is enabled, 2984address_hi bits 31-8 provide bits 31-8 of the destination id. Bits 7-0 of 2985address_hi must be zero. 2986 2987 29884.71 KVM_CREATE_PIT2 2989-------------------- 2990 2991:Capability: KVM_CAP_PIT2 2992:Architectures: x86 2993:Type: vm ioctl 2994:Parameters: struct kvm_pit_config (in) 2995:Returns: 0 on success, -1 on error 2996 2997Creates an in-kernel device model for the i8254 PIT. This call is only valid 2998after enabling in-kernel irqchip support via KVM_CREATE_IRQCHIP. The following 2999parameters have to be passed:: 3000 3001 struct kvm_pit_config { 3002 __u32 flags; 3003 __u32 pad[15]; 3004 }; 3005 3006Valid flags are:: 3007 3008 #define KVM_PIT_SPEAKER_DUMMY 1 /* emulate speaker port stub */ 3009 3010PIT timer interrupts may use a per-VM kernel thread for injection. If it 3011exists, this thread will have a name of the following pattern:: 3012 3013 kvm-pit/<owner-process-pid> 3014 3015When running a guest with elevated priorities, the scheduling parameters of 3016this thread may have to be adjusted accordingly. 3017 3018This IOCTL replaces the obsolete KVM_CREATE_PIT. 3019 3020 30214.72 KVM_GET_PIT2 3022----------------- 3023 3024:Capability: KVM_CAP_PIT_STATE2 3025:Architectures: x86 3026:Type: vm ioctl 3027:Parameters: struct kvm_pit_state2 (out) 3028:Returns: 0 on success, -1 on error 3029 3030Retrieves the state of the in-kernel PIT model. Only valid after 3031KVM_CREATE_PIT2. The state is returned in the following structure:: 3032 3033 struct kvm_pit_state2 { 3034 struct kvm_pit_channel_state channels[3]; 3035 __u32 flags; 3036 __u32 reserved[9]; 3037 }; 3038 3039Valid flags are:: 3040 3041 /* disable PIT in HPET legacy mode */ 3042 #define KVM_PIT_FLAGS_HPET_LEGACY 0x00000001 3043 /* speaker port data bit enabled */ 3044 #define KVM_PIT_FLAGS_SPEAKER_DATA_ON 0x00000002 3045 3046This IOCTL replaces the obsolete KVM_GET_PIT. 3047 3048 30494.73 KVM_SET_PIT2 3050----------------- 3051 3052:Capability: KVM_CAP_PIT_STATE2 3053:Architectures: x86 3054:Type: vm ioctl 3055:Parameters: struct kvm_pit_state2 (in) 3056:Returns: 0 on success, -1 on error 3057 3058Sets the state of the in-kernel PIT model. Only valid after KVM_CREATE_PIT2. 3059See KVM_GET_PIT2 for details on struct kvm_pit_state2. 3060 3061This IOCTL replaces the obsolete KVM_SET_PIT. 3062 3063 30644.74 KVM_PPC_GET_SMMU_INFO 3065-------------------------- 3066 3067:Capability: KVM_CAP_PPC_GET_SMMU_INFO 3068:Architectures: powerpc 3069:Type: vm ioctl 3070:Parameters: None 3071:Returns: 0 on success, -1 on error 3072 3073This populates and returns a structure describing the features of 3074the "Server" class MMU emulation supported by KVM. 3075This can in turn be used by userspace to generate the appropriate 3076device-tree properties for the guest operating system. 3077 3078The structure contains some global information, followed by an 3079array of supported segment page sizes:: 3080 3081 struct kvm_ppc_smmu_info { 3082 __u64 flags; 3083 __u32 slb_size; 3084 __u32 pad; 3085 struct kvm_ppc_one_seg_page_size sps[KVM_PPC_PAGE_SIZES_MAX_SZ]; 3086 }; 3087 3088The supported flags are: 3089 3090 - KVM_PPC_PAGE_SIZES_REAL: 3091 When that flag is set, guest page sizes must "fit" the backing 3092 store page sizes. When not set, any page size in the list can 3093 be used regardless of how they are backed by userspace. 3094 3095 - KVM_PPC_1T_SEGMENTS 3096 The emulated MMU supports 1T segments in addition to the 3097 standard 256M ones. 3098 3099 - KVM_PPC_NO_HASH 3100 This flag indicates that HPT guests are not supported by KVM, 3101 thus all guests must use radix MMU mode. 3102 3103The "slb_size" field indicates how many SLB entries are supported 3104 3105The "sps" array contains 8 entries indicating the supported base 3106page sizes for a segment in increasing order. Each entry is defined 3107as follow:: 3108 3109 struct kvm_ppc_one_seg_page_size { 3110 __u32 page_shift; /* Base page shift of segment (or 0) */ 3111 __u32 slb_enc; /* SLB encoding for BookS */ 3112 struct kvm_ppc_one_page_size enc[KVM_PPC_PAGE_SIZES_MAX_SZ]; 3113 }; 3114 3115An entry with a "page_shift" of 0 is unused. Because the array is 3116organized in increasing order, a lookup can stop when encountering 3117such an entry. 3118 3119The "slb_enc" field provides the encoding to use in the SLB for the 3120page size. The bits are in positions such as the value can directly 3121be OR'ed into the "vsid" argument of the slbmte instruction. 3122 3123The "enc" array is a list which for each of those segment base page 3124size provides the list of supported actual page sizes (which can be 3125only larger or equal to the base page size), along with the 3126corresponding encoding in the hash PTE. Similarly, the array is 31278 entries sorted by increasing sizes and an entry with a "0" shift 3128is an empty entry and a terminator:: 3129 3130 struct kvm_ppc_one_page_size { 3131 __u32 page_shift; /* Page shift (or 0) */ 3132 __u32 pte_enc; /* Encoding in the HPTE (>>12) */ 3133 }; 3134 3135The "pte_enc" field provides a value that can OR'ed into the hash 3136PTE's RPN field (ie, it needs to be shifted left by 12 to OR it 3137into the hash PTE second double word). 3138 31394.75 KVM_IRQFD 3140-------------- 3141 3142:Capability: KVM_CAP_IRQFD 3143:Architectures: x86 s390 arm64 3144:Type: vm ioctl 3145:Parameters: struct kvm_irqfd (in) 3146:Returns: 0 on success, -1 on error 3147 3148Allows setting an eventfd to directly trigger a guest interrupt. 3149kvm_irqfd.fd specifies the file descriptor to use as the eventfd and 3150kvm_irqfd.gsi specifies the irqchip pin toggled by this event. When 3151an event is triggered on the eventfd, an interrupt is injected into 3152the guest using the specified gsi pin. The irqfd is removed using 3153the KVM_IRQFD_FLAG_DEASSIGN flag, specifying both kvm_irqfd.fd 3154and kvm_irqfd.gsi. 3155 3156With KVM_CAP_IRQFD_RESAMPLE, KVM_IRQFD supports a de-assert and notify 3157mechanism allowing emulation of level-triggered, irqfd-based 3158interrupts. When KVM_IRQFD_FLAG_RESAMPLE is set the user must pass an 3159additional eventfd in the kvm_irqfd.resamplefd field. When operating 3160in resample mode, posting of an interrupt through kvm_irq.fd asserts 3161the specified gsi in the irqchip. When the irqchip is resampled, such 3162as from an EOI, the gsi is de-asserted and the user is notified via 3163kvm_irqfd.resamplefd. It is the user's responsibility to re-queue 3164the interrupt if the device making use of it still requires service. 3165Note that closing the resamplefd is not sufficient to disable the 3166irqfd. The KVM_IRQFD_FLAG_RESAMPLE is only necessary on assignment 3167and need not be specified with KVM_IRQFD_FLAG_DEASSIGN. 3168 3169On arm64, gsi routing being supported, the following can happen: 3170 3171- in case no routing entry is associated to this gsi, injection fails 3172- in case the gsi is associated to an irqchip routing entry, 3173 irqchip.pin + 32 corresponds to the injected SPI ID. 3174- in case the gsi is associated to an MSI routing entry, the MSI 3175 message and device ID are translated into an LPI (support restricted 3176 to GICv3 ITS in-kernel emulation). 3177 31784.76 KVM_PPC_ALLOCATE_HTAB 3179-------------------------- 3180 3181:Capability: KVM_CAP_PPC_ALLOC_HTAB 3182:Architectures: powerpc 3183:Type: vm ioctl 3184:Parameters: Pointer to u32 containing hash table order (in/out) 3185:Returns: 0 on success, -1 on error 3186 3187This requests the host kernel to allocate an MMU hash table for a 3188guest using the PAPR paravirtualization interface. This only does 3189anything if the kernel is configured to use the Book 3S HV style of 3190virtualization. Otherwise the capability doesn't exist and the ioctl 3191returns an ENOTTY error. The rest of this description assumes Book 3S 3192HV. 3193 3194There must be no vcpus running when this ioctl is called; if there 3195are, it will do nothing and return an EBUSY error. 3196 3197The parameter is a pointer to a 32-bit unsigned integer variable 3198containing the order (log base 2) of the desired size of the hash 3199table, which must be between 18 and 46. On successful return from the 3200ioctl, the value will not be changed by the kernel. 3201 3202If no hash table has been allocated when any vcpu is asked to run 3203(with the KVM_RUN ioctl), the host kernel will allocate a 3204default-sized hash table (16 MB). 3205 3206If this ioctl is called when a hash table has already been allocated, 3207with a different order from the existing hash table, the existing hash 3208table will be freed and a new one allocated. If this is ioctl is 3209called when a hash table has already been allocated of the same order 3210as specified, the kernel will clear out the existing hash table (zero 3211all HPTEs). In either case, if the guest is using the virtualized 3212real-mode area (VRMA) facility, the kernel will re-create the VMRA 3213HPTEs on the next KVM_RUN of any vcpu. 3214 32154.77 KVM_S390_INTERRUPT 3216----------------------- 3217 3218:Capability: basic 3219:Architectures: s390 3220:Type: vm ioctl, vcpu ioctl 3221:Parameters: struct kvm_s390_interrupt (in) 3222:Returns: 0 on success, -1 on error 3223 3224Allows to inject an interrupt to the guest. Interrupts can be floating 3225(vm ioctl) or per cpu (vcpu ioctl), depending on the interrupt type. 3226 3227Interrupt parameters are passed via kvm_s390_interrupt:: 3228 3229 struct kvm_s390_interrupt { 3230 __u32 type; 3231 __u32 parm; 3232 __u64 parm64; 3233 }; 3234 3235type can be one of the following: 3236 3237KVM_S390_SIGP_STOP (vcpu) 3238 - sigp stop; optional flags in parm 3239KVM_S390_PROGRAM_INT (vcpu) 3240 - program check; code in parm 3241KVM_S390_SIGP_SET_PREFIX (vcpu) 3242 - sigp set prefix; prefix address in parm 3243KVM_S390_RESTART (vcpu) 3244 - restart 3245KVM_S390_INT_CLOCK_COMP (vcpu) 3246 - clock comparator interrupt 3247KVM_S390_INT_CPU_TIMER (vcpu) 3248 - CPU timer interrupt 3249KVM_S390_INT_VIRTIO (vm) 3250 - virtio external interrupt; external interrupt 3251 parameters in parm and parm64 3252KVM_S390_INT_SERVICE (vm) 3253 - sclp external interrupt; sclp parameter in parm 3254KVM_S390_INT_EMERGENCY (vcpu) 3255 - sigp emergency; source cpu in parm 3256KVM_S390_INT_EXTERNAL_CALL (vcpu) 3257 - sigp external call; source cpu in parm 3258KVM_S390_INT_IO(ai,cssid,ssid,schid) (vm) 3259 - compound value to indicate an 3260 I/O interrupt (ai - adapter interrupt; cssid,ssid,schid - subchannel); 3261 I/O interruption parameters in parm (subchannel) and parm64 (intparm, 3262 interruption subclass) 3263KVM_S390_MCHK (vm, vcpu) 3264 - machine check interrupt; cr 14 bits in parm, machine check interrupt 3265 code in parm64 (note that machine checks needing further payload are not 3266 supported by this ioctl) 3267 3268This is an asynchronous vcpu ioctl and can be invoked from any thread. 3269 32704.78 KVM_PPC_GET_HTAB_FD 3271------------------------ 3272 3273:Capability: KVM_CAP_PPC_HTAB_FD 3274:Architectures: powerpc 3275:Type: vm ioctl 3276:Parameters: Pointer to struct kvm_get_htab_fd (in) 3277:Returns: file descriptor number (>= 0) on success, -1 on error 3278 3279This returns a file descriptor that can be used either to read out the 3280entries in the guest's hashed page table (HPT), or to write entries to 3281initialize the HPT. The returned fd can only be written to if the 3282KVM_GET_HTAB_WRITE bit is set in the flags field of the argument, and 3283can only be read if that bit is clear. The argument struct looks like 3284this:: 3285 3286 /* For KVM_PPC_GET_HTAB_FD */ 3287 struct kvm_get_htab_fd { 3288 __u64 flags; 3289 __u64 start_index; 3290 __u64 reserved[2]; 3291 }; 3292 3293 /* Values for kvm_get_htab_fd.flags */ 3294 #define KVM_GET_HTAB_BOLTED_ONLY ((__u64)0x1) 3295 #define KVM_GET_HTAB_WRITE ((__u64)0x2) 3296 3297The 'start_index' field gives the index in the HPT of the entry at 3298which to start reading. It is ignored when writing. 3299 3300Reads on the fd will initially supply information about all 3301"interesting" HPT entries. Interesting entries are those with the 3302bolted bit set, if the KVM_GET_HTAB_BOLTED_ONLY bit is set, otherwise 3303all entries. When the end of the HPT is reached, the read() will 3304return. If read() is called again on the fd, it will start again from 3305the beginning of the HPT, but will only return HPT entries that have 3306changed since they were last read. 3307 3308Data read or written is structured as a header (8 bytes) followed by a 3309series of valid HPT entries (16 bytes) each. The header indicates how 3310many valid HPT entries there are and how many invalid entries follow 3311the valid entries. The invalid entries are not represented explicitly 3312in the stream. The header format is:: 3313 3314 struct kvm_get_htab_header { 3315 __u32 index; 3316 __u16 n_valid; 3317 __u16 n_invalid; 3318 }; 3319 3320Writes to the fd create HPT entries starting at the index given in the 3321header; first 'n_valid' valid entries with contents from the data 3322written, then 'n_invalid' invalid entries, invalidating any previously 3323valid entries found. 3324 33254.79 KVM_CREATE_DEVICE 3326---------------------- 3327 3328:Capability: KVM_CAP_DEVICE_CTRL 3329:Architectures: all 3330:Type: vm ioctl 3331:Parameters: struct kvm_create_device (in/out) 3332:Returns: 0 on success, -1 on error 3333 3334Errors: 3335 3336 ====== ======================================================= 3337 ENODEV The device type is unknown or unsupported 3338 EEXIST Device already created, and this type of device may not 3339 be instantiated multiple times 3340 ====== ======================================================= 3341 3342 Other error conditions may be defined by individual device types or 3343 have their standard meanings. 3344 3345Creates an emulated device in the kernel. The file descriptor returned 3346in fd can be used with KVM_SET/GET/HAS_DEVICE_ATTR. 3347 3348If the KVM_CREATE_DEVICE_TEST flag is set, only test whether the 3349device type is supported (not necessarily whether it can be created 3350in the current vm). 3351 3352Individual devices should not define flags. Attributes should be used 3353for specifying any behavior that is not implied by the device type 3354number. 3355 3356:: 3357 3358 struct kvm_create_device { 3359 __u32 type; /* in: KVM_DEV_TYPE_xxx */ 3360 __u32 fd; /* out: device handle */ 3361 __u32 flags; /* in: KVM_CREATE_DEVICE_xxx */ 3362 }; 3363 33644.80 KVM_SET_DEVICE_ATTR/KVM_GET_DEVICE_ATTR 3365-------------------------------------------- 3366 3367:Capability: KVM_CAP_DEVICE_CTRL, KVM_CAP_VM_ATTRIBUTES for vm device, 3368 KVM_CAP_VCPU_ATTRIBUTES for vcpu device 3369 KVM_CAP_SYS_ATTRIBUTES for system (/dev/kvm) device (no set) 3370:Architectures: x86, arm64, s390 3371:Type: device ioctl, vm ioctl, vcpu ioctl 3372:Parameters: struct kvm_device_attr 3373:Returns: 0 on success, -1 on error 3374 3375Errors: 3376 3377 ===== ============================================================= 3378 ENXIO The group or attribute is unknown/unsupported for this device 3379 or hardware support is missing. 3380 EPERM The attribute cannot (currently) be accessed this way 3381 (e.g. read-only attribute, or attribute that only makes 3382 sense when the device is in a different state) 3383 ===== ============================================================= 3384 3385 Other error conditions may be defined by individual device types. 3386 3387Gets/sets a specified piece of device configuration and/or state. The 3388semantics are device-specific. See individual device documentation in 3389the "devices" directory. As with ONE_REG, the size of the data 3390transferred is defined by the particular attribute. 3391 3392:: 3393 3394 struct kvm_device_attr { 3395 __u32 flags; /* no flags currently defined */ 3396 __u32 group; /* device-defined */ 3397 __u64 attr; /* group-defined */ 3398 __u64 addr; /* userspace address of attr data */ 3399 }; 3400 34014.81 KVM_HAS_DEVICE_ATTR 3402------------------------ 3403 3404:Capability: KVM_CAP_DEVICE_CTRL, KVM_CAP_VM_ATTRIBUTES for vm device, 3405 KVM_CAP_VCPU_ATTRIBUTES for vcpu device 3406 KVM_CAP_SYS_ATTRIBUTES for system (/dev/kvm) device 3407:Type: device ioctl, vm ioctl, vcpu ioctl 3408:Parameters: struct kvm_device_attr 3409:Returns: 0 on success, -1 on error 3410 3411Errors: 3412 3413 ===== ============================================================= 3414 ENXIO The group or attribute is unknown/unsupported for this device 3415 or hardware support is missing. 3416 ===== ============================================================= 3417 3418Tests whether a device supports a particular attribute. A successful 3419return indicates the attribute is implemented. It does not necessarily 3420indicate that the attribute can be read or written in the device's 3421current state. "addr" is ignored. 3422 3423.. _KVM_ARM_VCPU_INIT: 3424 34254.82 KVM_ARM_VCPU_INIT 3426---------------------- 3427 3428:Capability: basic 3429:Architectures: arm64 3430:Type: vcpu ioctl 3431:Parameters: struct kvm_vcpu_init (in) 3432:Returns: 0 on success; -1 on error 3433 3434Errors: 3435 3436 ====== ================================================================= 3437 EINVAL the target is unknown, or the combination of features is invalid. 3438 ENOENT a features bit specified is unknown. 3439 ====== ================================================================= 3440 3441This tells KVM what type of CPU to present to the guest, and what 3442optional features it should have. This will cause a reset of the cpu 3443registers to their initial values. If this is not called, KVM_RUN will 3444return ENOEXEC for that vcpu. 3445 3446The initial values are defined as: 3447 - Processor state: 3448 * AArch64: EL1h, D, A, I and F bits set. All other bits 3449 are cleared. 3450 * AArch32: SVC, A, I and F bits set. All other bits are 3451 cleared. 3452 - General Purpose registers, including PC and SP: set to 0 3453 - FPSIMD/NEON registers: set to 0 3454 - SVE registers: set to 0 3455 - System registers: Reset to their architecturally defined 3456 values as for a warm reset to EL1 (resp. SVC) 3457 3458Note that because some registers reflect machine topology, all vcpus 3459should be created before this ioctl is invoked. 3460 3461Userspace can call this function multiple times for a given vcpu, including 3462after the vcpu has been run. This will reset the vcpu to its initial 3463state. All calls to this function after the initial call must use the same 3464target and same set of feature flags, otherwise EINVAL will be returned. 3465 3466Possible features: 3467 3468 - KVM_ARM_VCPU_POWER_OFF: Starts the CPU in a power-off state. 3469 Depends on KVM_CAP_ARM_PSCI. If not set, the CPU will be powered on 3470 and execute guest code when KVM_RUN is called. 3471 - KVM_ARM_VCPU_EL1_32BIT: Starts the CPU in a 32bit mode. 3472 Depends on KVM_CAP_ARM_EL1_32BIT (arm64 only). 3473 - KVM_ARM_VCPU_PSCI_0_2: Emulate PSCI v0.2 (or a future revision 3474 backward compatible with v0.2) for the CPU. 3475 Depends on KVM_CAP_ARM_PSCI_0_2. 3476 - KVM_ARM_VCPU_PMU_V3: Emulate PMUv3 for the CPU. 3477 Depends on KVM_CAP_ARM_PMU_V3. 3478 3479 - KVM_ARM_VCPU_PTRAUTH_ADDRESS: Enables Address Pointer authentication 3480 for arm64 only. 3481 Depends on KVM_CAP_ARM_PTRAUTH_ADDRESS. 3482 If KVM_CAP_ARM_PTRAUTH_ADDRESS and KVM_CAP_ARM_PTRAUTH_GENERIC are 3483 both present, then both KVM_ARM_VCPU_PTRAUTH_ADDRESS and 3484 KVM_ARM_VCPU_PTRAUTH_GENERIC must be requested or neither must be 3485 requested. 3486 3487 - KVM_ARM_VCPU_PTRAUTH_GENERIC: Enables Generic Pointer authentication 3488 for arm64 only. 3489 Depends on KVM_CAP_ARM_PTRAUTH_GENERIC. 3490 If KVM_CAP_ARM_PTRAUTH_ADDRESS and KVM_CAP_ARM_PTRAUTH_GENERIC are 3491 both present, then both KVM_ARM_VCPU_PTRAUTH_ADDRESS and 3492 KVM_ARM_VCPU_PTRAUTH_GENERIC must be requested or neither must be 3493 requested. 3494 3495 - KVM_ARM_VCPU_SVE: Enables SVE for the CPU (arm64 only). 3496 Depends on KVM_CAP_ARM_SVE. 3497 Requires KVM_ARM_VCPU_FINALIZE(KVM_ARM_VCPU_SVE): 3498 3499 * After KVM_ARM_VCPU_INIT: 3500 3501 - KVM_REG_ARM64_SVE_VLS may be read using KVM_GET_ONE_REG: the 3502 initial value of this pseudo-register indicates the best set of 3503 vector lengths possible for a vcpu on this host. 3504 3505 * Before KVM_ARM_VCPU_FINALIZE(KVM_ARM_VCPU_SVE): 3506 3507 - KVM_RUN and KVM_GET_REG_LIST are not available; 3508 3509 - KVM_GET_ONE_REG and KVM_SET_ONE_REG cannot be used to access 3510 the scalable architectural SVE registers 3511 KVM_REG_ARM64_SVE_ZREG(), KVM_REG_ARM64_SVE_PREG() or 3512 KVM_REG_ARM64_SVE_FFR; 3513 3514 - KVM_REG_ARM64_SVE_VLS may optionally be written using 3515 KVM_SET_ONE_REG, to modify the set of vector lengths available 3516 for the vcpu. 3517 3518 * After KVM_ARM_VCPU_FINALIZE(KVM_ARM_VCPU_SVE): 3519 3520 - the KVM_REG_ARM64_SVE_VLS pseudo-register is immutable, and can 3521 no longer be written using KVM_SET_ONE_REG. 3522 35234.83 KVM_ARM_PREFERRED_TARGET 3524----------------------------- 3525 3526:Capability: basic 3527:Architectures: arm64 3528:Type: vm ioctl 3529:Parameters: struct kvm_vcpu_init (out) 3530:Returns: 0 on success; -1 on error 3531 3532Errors: 3533 3534 ====== ========================================== 3535 ENODEV no preferred target available for the host 3536 ====== ========================================== 3537 3538This queries KVM for preferred CPU target type which can be emulated 3539by KVM on underlying host. 3540 3541The ioctl returns struct kvm_vcpu_init instance containing information 3542about preferred CPU target type and recommended features for it. The 3543kvm_vcpu_init->features bitmap returned will have feature bits set if 3544the preferred target recommends setting these features, but this is 3545not mandatory. 3546 3547The information returned by this ioctl can be used to prepare an instance 3548of struct kvm_vcpu_init for KVM_ARM_VCPU_INIT ioctl which will result in 3549VCPU matching underlying host. 3550 3551 35524.84 KVM_GET_REG_LIST 3553--------------------- 3554 3555:Capability: basic 3556:Architectures: arm64, mips, riscv 3557:Type: vcpu ioctl 3558:Parameters: struct kvm_reg_list (in/out) 3559:Returns: 0 on success; -1 on error 3560 3561Errors: 3562 3563 ===== ============================================================== 3564 E2BIG the reg index list is too big to fit in the array specified by 3565 the user (the number required will be written into n). 3566 ===== ============================================================== 3567 3568:: 3569 3570 struct kvm_reg_list { 3571 __u64 n; /* number of registers in reg[] */ 3572 __u64 reg[0]; 3573 }; 3574 3575This ioctl returns the guest registers that are supported for the 3576KVM_GET_ONE_REG/KVM_SET_ONE_REG calls. 3577 3578Note that s390 does not support KVM_GET_REG_LIST for historical reasons 3579(read: nobody cared). The set of registers in kernels 4.x and newer is: 3580 3581- KVM_REG_S390_TODPR 3582 3583- KVM_REG_S390_EPOCHDIFF 3584 3585- KVM_REG_S390_CPU_TIMER 3586 3587- KVM_REG_S390_CLOCK_COMP 3588 3589- KVM_REG_S390_PFTOKEN 3590 3591- KVM_REG_S390_PFCOMPARE 3592 3593- KVM_REG_S390_PFSELECT 3594 3595- KVM_REG_S390_PP 3596 3597- KVM_REG_S390_GBEA 3598 3599 36004.85 KVM_ARM_SET_DEVICE_ADDR (deprecated) 3601----------------------------------------- 3602 3603:Capability: KVM_CAP_ARM_SET_DEVICE_ADDR 3604:Architectures: arm64 3605:Type: vm ioctl 3606:Parameters: struct kvm_arm_device_address (in) 3607:Returns: 0 on success, -1 on error 3608 3609Errors: 3610 3611 ====== ============================================ 3612 ENODEV The device id is unknown 3613 ENXIO Device not supported on current system 3614 EEXIST Address already set 3615 E2BIG Address outside guest physical address space 3616 EBUSY Address overlaps with other device range 3617 ====== ============================================ 3618 3619:: 3620 3621 struct kvm_arm_device_addr { 3622 __u64 id; 3623 __u64 addr; 3624 }; 3625 3626Specify a device address in the guest's physical address space where guests 3627can access emulated or directly exposed devices, which the host kernel needs 3628to know about. The id field is an architecture specific identifier for a 3629specific device. 3630 3631arm64 divides the id field into two parts, a device id and an 3632address type id specific to the individual device:: 3633 3634 bits: | 63 ... 32 | 31 ... 16 | 15 ... 0 | 3635 field: | 0x00000000 | device id | addr type id | 3636 3637arm64 currently only require this when using the in-kernel GIC 3638support for the hardware VGIC features, using KVM_ARM_DEVICE_VGIC_V2 3639as the device id. When setting the base address for the guest's 3640mapping of the VGIC virtual CPU and distributor interface, the ioctl 3641must be called after calling KVM_CREATE_IRQCHIP, but before calling 3642KVM_RUN on any of the VCPUs. Calling this ioctl twice for any of the 3643base addresses will return -EEXIST. 3644 3645Note, this IOCTL is deprecated and the more flexible SET/GET_DEVICE_ATTR API 3646should be used instead. 3647 3648 36494.86 KVM_PPC_RTAS_DEFINE_TOKEN 3650------------------------------ 3651 3652:Capability: KVM_CAP_PPC_RTAS 3653:Architectures: ppc 3654:Type: vm ioctl 3655:Parameters: struct kvm_rtas_token_args 3656:Returns: 0 on success, -1 on error 3657 3658Defines a token value for a RTAS (Run Time Abstraction Services) 3659service in order to allow it to be handled in the kernel. The 3660argument struct gives the name of the service, which must be the name 3661of a service that has a kernel-side implementation. If the token 3662value is non-zero, it will be associated with that service, and 3663subsequent RTAS calls by the guest specifying that token will be 3664handled by the kernel. If the token value is 0, then any token 3665associated with the service will be forgotten, and subsequent RTAS 3666calls by the guest for that service will be passed to userspace to be 3667handled. 3668 36694.87 KVM_SET_GUEST_DEBUG 3670------------------------ 3671 3672:Capability: KVM_CAP_SET_GUEST_DEBUG 3673:Architectures: x86, s390, ppc, arm64 3674:Type: vcpu ioctl 3675:Parameters: struct kvm_guest_debug (in) 3676:Returns: 0 on success; -1 on error 3677 3678:: 3679 3680 struct kvm_guest_debug { 3681 __u32 control; 3682 __u32 pad; 3683 struct kvm_guest_debug_arch arch; 3684 }; 3685 3686Set up the processor specific debug registers and configure vcpu for 3687handling guest debug events. There are two parts to the structure, the 3688first a control bitfield indicates the type of debug events to handle 3689when running. Common control bits are: 3690 3691 - KVM_GUESTDBG_ENABLE: guest debugging is enabled 3692 - KVM_GUESTDBG_SINGLESTEP: the next run should single-step 3693 3694The top 16 bits of the control field are architecture specific control 3695flags which can include the following: 3696 3697 - KVM_GUESTDBG_USE_SW_BP: using software breakpoints [x86, arm64] 3698 - KVM_GUESTDBG_USE_HW_BP: using hardware breakpoints [x86, s390] 3699 - KVM_GUESTDBG_USE_HW: using hardware debug events [arm64] 3700 - KVM_GUESTDBG_INJECT_DB: inject DB type exception [x86] 3701 - KVM_GUESTDBG_INJECT_BP: inject BP type exception [x86] 3702 - KVM_GUESTDBG_EXIT_PENDING: trigger an immediate guest exit [s390] 3703 - KVM_GUESTDBG_BLOCKIRQ: avoid injecting interrupts/NMI/SMI [x86] 3704 3705For example KVM_GUESTDBG_USE_SW_BP indicates that software breakpoints 3706are enabled in memory so we need to ensure breakpoint exceptions are 3707correctly trapped and the KVM run loop exits at the breakpoint and not 3708running off into the normal guest vector. For KVM_GUESTDBG_USE_HW_BP 3709we need to ensure the guest vCPUs architecture specific registers are 3710updated to the correct (supplied) values. 3711 3712The second part of the structure is architecture specific and 3713typically contains a set of debug registers. 3714 3715For arm64 the number of debug registers is implementation defined and 3716can be determined by querying the KVM_CAP_GUEST_DEBUG_HW_BPS and 3717KVM_CAP_GUEST_DEBUG_HW_WPS capabilities which return a positive number 3718indicating the number of supported registers. 3719 3720For ppc, the KVM_CAP_PPC_GUEST_DEBUG_SSTEP capability indicates whether 3721the single-step debug event (KVM_GUESTDBG_SINGLESTEP) is supported. 3722 3723Also when supported, KVM_CAP_SET_GUEST_DEBUG2 capability indicates the 3724supported KVM_GUESTDBG_* bits in the control field. 3725 3726When debug events exit the main run loop with the reason 3727KVM_EXIT_DEBUG with the kvm_debug_exit_arch part of the kvm_run 3728structure containing architecture specific debug information. 3729 37304.88 KVM_GET_EMULATED_CPUID 3731--------------------------- 3732 3733:Capability: KVM_CAP_EXT_EMUL_CPUID 3734:Architectures: x86 3735:Type: system ioctl 3736:Parameters: struct kvm_cpuid2 (in/out) 3737:Returns: 0 on success, -1 on error 3738 3739:: 3740 3741 struct kvm_cpuid2 { 3742 __u32 nent; 3743 __u32 flags; 3744 struct kvm_cpuid_entry2 entries[0]; 3745 }; 3746 3747The member 'flags' is used for passing flags from userspace. 3748 3749:: 3750 3751 #define KVM_CPUID_FLAG_SIGNIFCANT_INDEX BIT(0) 3752 #define KVM_CPUID_FLAG_STATEFUL_FUNC BIT(1) /* deprecated */ 3753 #define KVM_CPUID_FLAG_STATE_READ_NEXT BIT(2) /* deprecated */ 3754 3755 struct kvm_cpuid_entry2 { 3756 __u32 function; 3757 __u32 index; 3758 __u32 flags; 3759 __u32 eax; 3760 __u32 ebx; 3761 __u32 ecx; 3762 __u32 edx; 3763 __u32 padding[3]; 3764 }; 3765 3766This ioctl returns x86 cpuid features which are emulated by 3767kvm.Userspace can use the information returned by this ioctl to query 3768which features are emulated by kvm instead of being present natively. 3769 3770Userspace invokes KVM_GET_EMULATED_CPUID by passing a kvm_cpuid2 3771structure with the 'nent' field indicating the number of entries in 3772the variable-size array 'entries'. If the number of entries is too low 3773to describe the cpu capabilities, an error (E2BIG) is returned. If the 3774number is too high, the 'nent' field is adjusted and an error (ENOMEM) 3775is returned. If the number is just right, the 'nent' field is adjusted 3776to the number of valid entries in the 'entries' array, which is then 3777filled. 3778 3779The entries returned are the set CPUID bits of the respective features 3780which kvm emulates, as returned by the CPUID instruction, with unknown 3781or unsupported feature bits cleared. 3782 3783Features like x2apic, for example, may not be present in the host cpu 3784but are exposed by kvm in KVM_GET_SUPPORTED_CPUID because they can be 3785emulated efficiently and thus not included here. 3786 3787The fields in each entry are defined as follows: 3788 3789 function: 3790 the eax value used to obtain the entry 3791 index: 3792 the ecx value used to obtain the entry (for entries that are 3793 affected by ecx) 3794 flags: 3795 an OR of zero or more of the following: 3796 3797 KVM_CPUID_FLAG_SIGNIFCANT_INDEX: 3798 if the index field is valid 3799 3800 eax, ebx, ecx, edx: 3801 3802 the values returned by the cpuid instruction for 3803 this function/index combination 3804 38054.89 KVM_S390_MEM_OP 3806-------------------- 3807 3808:Capability: KVM_CAP_S390_MEM_OP, KVM_CAP_S390_PROTECTED, KVM_CAP_S390_MEM_OP_EXTENSION 3809:Architectures: s390 3810:Type: vm ioctl, vcpu ioctl 3811:Parameters: struct kvm_s390_mem_op (in) 3812:Returns: = 0 on success, 3813 < 0 on generic error (e.g. -EFAULT or -ENOMEM), 3814 16 bit program exception code if the access causes such an exception 3815 3816Read or write data from/to the VM's memory. 3817The KVM_CAP_S390_MEM_OP_EXTENSION capability specifies what functionality is 3818supported. 3819 3820Parameters are specified via the following structure:: 3821 3822 struct kvm_s390_mem_op { 3823 __u64 gaddr; /* the guest address */ 3824 __u64 flags; /* flags */ 3825 __u32 size; /* amount of bytes */ 3826 __u32 op; /* type of operation */ 3827 __u64 buf; /* buffer in userspace */ 3828 union { 3829 struct { 3830 __u8 ar; /* the access register number */ 3831 __u8 key; /* access key, ignored if flag unset */ 3832 __u8 pad1[6]; /* ignored */ 3833 __u64 old_addr; /* ignored if flag unset */ 3834 }; 3835 __u32 sida_offset; /* offset into the sida */ 3836 __u8 reserved[32]; /* ignored */ 3837 }; 3838 }; 3839 3840The start address of the memory region has to be specified in the "gaddr" 3841field, and the length of the region in the "size" field (which must not 3842be 0). The maximum value for "size" can be obtained by checking the 3843KVM_CAP_S390_MEM_OP capability. "buf" is the buffer supplied by the 3844userspace application where the read data should be written to for 3845a read access, or where the data that should be written is stored for 3846a write access. The "reserved" field is meant for future extensions. 3847Reserved and unused values are ignored. Future extension that add members must 3848introduce new flags. 3849 3850The type of operation is specified in the "op" field. Flags modifying 3851their behavior can be set in the "flags" field. Undefined flag bits must 3852be set to 0. 3853 3854Possible operations are: 3855 * ``KVM_S390_MEMOP_LOGICAL_READ`` 3856 * ``KVM_S390_MEMOP_LOGICAL_WRITE`` 3857 * ``KVM_S390_MEMOP_ABSOLUTE_READ`` 3858 * ``KVM_S390_MEMOP_ABSOLUTE_WRITE`` 3859 * ``KVM_S390_MEMOP_SIDA_READ`` 3860 * ``KVM_S390_MEMOP_SIDA_WRITE`` 3861 * ``KVM_S390_MEMOP_ABSOLUTE_CMPXCHG`` 3862 3863Logical read/write: 3864^^^^^^^^^^^^^^^^^^^ 3865 3866Access logical memory, i.e. translate the given guest address to an absolute 3867address given the state of the VCPU and use the absolute address as target of 3868the access. "ar" designates the access register number to be used; the valid 3869range is 0..15. 3870Logical accesses are permitted for the VCPU ioctl only. 3871Logical accesses are permitted for non-protected guests only. 3872 3873Supported flags: 3874 * ``KVM_S390_MEMOP_F_CHECK_ONLY`` 3875 * ``KVM_S390_MEMOP_F_INJECT_EXCEPTION`` 3876 * ``KVM_S390_MEMOP_F_SKEY_PROTECTION`` 3877 3878The KVM_S390_MEMOP_F_CHECK_ONLY flag can be set to check whether the 3879corresponding memory access would cause an access exception; however, 3880no actual access to the data in memory at the destination is performed. 3881In this case, "buf" is unused and can be NULL. 3882 3883In case an access exception occurred during the access (or would occur 3884in case of KVM_S390_MEMOP_F_CHECK_ONLY), the ioctl returns a positive 3885error number indicating the type of exception. This exception is also 3886raised directly at the corresponding VCPU if the flag 3887KVM_S390_MEMOP_F_INJECT_EXCEPTION is set. 3888On protection exceptions, unless specified otherwise, the injected 3889translation-exception identifier (TEID) indicates suppression. 3890 3891If the KVM_S390_MEMOP_F_SKEY_PROTECTION flag is set, storage key 3892protection is also in effect and may cause exceptions if accesses are 3893prohibited given the access key designated by "key"; the valid range is 0..15. 3894KVM_S390_MEMOP_F_SKEY_PROTECTION is available if KVM_CAP_S390_MEM_OP_EXTENSION 3895is > 0. 3896Since the accessed memory may span multiple pages and those pages might have 3897different storage keys, it is possible that a protection exception occurs 3898after memory has been modified. In this case, if the exception is injected, 3899the TEID does not indicate suppression. 3900 3901Absolute read/write: 3902^^^^^^^^^^^^^^^^^^^^ 3903 3904Access absolute memory. This operation is intended to be used with the 3905KVM_S390_MEMOP_F_SKEY_PROTECTION flag, to allow accessing memory and performing 3906the checks required for storage key protection as one operation (as opposed to 3907user space getting the storage keys, performing the checks, and accessing 3908memory thereafter, which could lead to a delay between check and access). 3909Absolute accesses are permitted for the VM ioctl if KVM_CAP_S390_MEM_OP_EXTENSION 3910has the KVM_S390_MEMOP_EXTENSION_CAP_BASE bit set. 3911Currently absolute accesses are not permitted for VCPU ioctls. 3912Absolute accesses are permitted for non-protected guests only. 3913 3914Supported flags: 3915 * ``KVM_S390_MEMOP_F_CHECK_ONLY`` 3916 * ``KVM_S390_MEMOP_F_SKEY_PROTECTION`` 3917 3918The semantics of the flags common with logical accesses are as for logical 3919accesses. 3920 3921Absolute cmpxchg: 3922^^^^^^^^^^^^^^^^^ 3923 3924Perform cmpxchg on absolute guest memory. Intended for use with the 3925KVM_S390_MEMOP_F_SKEY_PROTECTION flag. 3926Instead of doing an unconditional write, the access occurs only if the target 3927location contains the value pointed to by "old_addr". 3928This is performed as an atomic cmpxchg with the length specified by the "size" 3929parameter. "size" must be a power of two up to and including 16. 3930If the exchange did not take place because the target value doesn't match the 3931old value, the value "old_addr" points to is replaced by the target value. 3932User space can tell if an exchange took place by checking if this replacement 3933occurred. The cmpxchg op is permitted for the VM ioctl if 3934KVM_CAP_S390_MEM_OP_EXTENSION has flag KVM_S390_MEMOP_EXTENSION_CAP_CMPXCHG set. 3935 3936Supported flags: 3937 * ``KVM_S390_MEMOP_F_SKEY_PROTECTION`` 3938 3939SIDA read/write: 3940^^^^^^^^^^^^^^^^ 3941 3942Access the secure instruction data area which contains memory operands necessary 3943for instruction emulation for protected guests. 3944SIDA accesses are available if the KVM_CAP_S390_PROTECTED capability is available. 3945SIDA accesses are permitted for the VCPU ioctl only. 3946SIDA accesses are permitted for protected guests only. 3947 3948No flags are supported. 3949 39504.90 KVM_S390_GET_SKEYS 3951----------------------- 3952 3953:Capability: KVM_CAP_S390_SKEYS 3954:Architectures: s390 3955:Type: vm ioctl 3956:Parameters: struct kvm_s390_skeys 3957:Returns: 0 on success, KVM_S390_GET_SKEYS_NONE if guest is not using storage 3958 keys, negative value on error 3959 3960This ioctl is used to get guest storage key values on the s390 3961architecture. The ioctl takes parameters via the kvm_s390_skeys struct:: 3962 3963 struct kvm_s390_skeys { 3964 __u64 start_gfn; 3965 __u64 count; 3966 __u64 skeydata_addr; 3967 __u32 flags; 3968 __u32 reserved[9]; 3969 }; 3970 3971The start_gfn field is the number of the first guest frame whose storage keys 3972you want to get. 3973 3974The count field is the number of consecutive frames (starting from start_gfn) 3975whose storage keys to get. The count field must be at least 1 and the maximum 3976allowed value is defined as KVM_S390_SKEYS_MAX. Values outside this range 3977will cause the ioctl to return -EINVAL. 3978 3979The skeydata_addr field is the address to a buffer large enough to hold count 3980bytes. This buffer will be filled with storage key data by the ioctl. 3981 39824.91 KVM_S390_SET_SKEYS 3983----------------------- 3984 3985:Capability: KVM_CAP_S390_SKEYS 3986:Architectures: s390 3987:Type: vm ioctl 3988:Parameters: struct kvm_s390_skeys 3989:Returns: 0 on success, negative value on error 3990 3991This ioctl is used to set guest storage key values on the s390 3992architecture. The ioctl takes parameters via the kvm_s390_skeys struct. 3993See section on KVM_S390_GET_SKEYS for struct definition. 3994 3995The start_gfn field is the number of the first guest frame whose storage keys 3996you want to set. 3997 3998The count field is the number of consecutive frames (starting from start_gfn) 3999whose storage keys to get. The count field must be at least 1 and the maximum 4000allowed value is defined as KVM_S390_SKEYS_MAX. Values outside this range 4001will cause the ioctl to return -EINVAL. 4002 4003The skeydata_addr field is the address to a buffer containing count bytes of 4004storage keys. Each byte in the buffer will be set as the storage key for a 4005single frame starting at start_gfn for count frames. 4006 4007Note: If any architecturally invalid key value is found in the given data then 4008the ioctl will return -EINVAL. 4009 40104.92 KVM_S390_IRQ 4011----------------- 4012 4013:Capability: KVM_CAP_S390_INJECT_IRQ 4014:Architectures: s390 4015:Type: vcpu ioctl 4016:Parameters: struct kvm_s390_irq (in) 4017:Returns: 0 on success, -1 on error 4018 4019Errors: 4020 4021 4022 ====== ================================================================= 4023 EINVAL interrupt type is invalid 4024 type is KVM_S390_SIGP_STOP and flag parameter is invalid value, 4025 type is KVM_S390_INT_EXTERNAL_CALL and code is bigger 4026 than the maximum of VCPUs 4027 EBUSY type is KVM_S390_SIGP_SET_PREFIX and vcpu is not stopped, 4028 type is KVM_S390_SIGP_STOP and a stop irq is already pending, 4029 type is KVM_S390_INT_EXTERNAL_CALL and an external call interrupt 4030 is already pending 4031 ====== ================================================================= 4032 4033Allows to inject an interrupt to the guest. 4034 4035Using struct kvm_s390_irq as a parameter allows 4036to inject additional payload which is not 4037possible via KVM_S390_INTERRUPT. 4038 4039Interrupt parameters are passed via kvm_s390_irq:: 4040 4041 struct kvm_s390_irq { 4042 __u64 type; 4043 union { 4044 struct kvm_s390_io_info io; 4045 struct kvm_s390_ext_info ext; 4046 struct kvm_s390_pgm_info pgm; 4047 struct kvm_s390_emerg_info emerg; 4048 struct kvm_s390_extcall_info extcall; 4049 struct kvm_s390_prefix_info prefix; 4050 struct kvm_s390_stop_info stop; 4051 struct kvm_s390_mchk_info mchk; 4052 char reserved[64]; 4053 } u; 4054 }; 4055 4056type can be one of the following: 4057 4058- KVM_S390_SIGP_STOP - sigp stop; parameter in .stop 4059- KVM_S390_PROGRAM_INT - program check; parameters in .pgm 4060- KVM_S390_SIGP_SET_PREFIX - sigp set prefix; parameters in .prefix 4061- KVM_S390_RESTART - restart; no parameters 4062- KVM_S390_INT_CLOCK_COMP - clock comparator interrupt; no parameters 4063- KVM_S390_INT_CPU_TIMER - CPU timer interrupt; no parameters 4064- KVM_S390_INT_EMERGENCY - sigp emergency; parameters in .emerg 4065- KVM_S390_INT_EXTERNAL_CALL - sigp external call; parameters in .extcall 4066- KVM_S390_MCHK - machine check interrupt; parameters in .mchk 4067 4068This is an asynchronous vcpu ioctl and can be invoked from any thread. 4069 40704.94 KVM_S390_GET_IRQ_STATE 4071--------------------------- 4072 4073:Capability: KVM_CAP_S390_IRQ_STATE 4074:Architectures: s390 4075:Type: vcpu ioctl 4076:Parameters: struct kvm_s390_irq_state (out) 4077:Returns: >= number of bytes copied into buffer, 4078 -EINVAL if buffer size is 0, 4079 -ENOBUFS if buffer size is too small to fit all pending interrupts, 4080 -EFAULT if the buffer address was invalid 4081 4082This ioctl allows userspace to retrieve the complete state of all currently 4083pending interrupts in a single buffer. Use cases include migration 4084and introspection. The parameter structure contains the address of a 4085userspace buffer and its length:: 4086 4087 struct kvm_s390_irq_state { 4088 __u64 buf; 4089 __u32 flags; /* will stay unused for compatibility reasons */ 4090 __u32 len; 4091 __u32 reserved[4]; /* will stay unused for compatibility reasons */ 4092 }; 4093 4094Userspace passes in the above struct and for each pending interrupt a 4095struct kvm_s390_irq is copied to the provided buffer. 4096 4097The structure contains a flags and a reserved field for future extensions. As 4098the kernel never checked for flags == 0 and QEMU never pre-zeroed flags and 4099reserved, these fields can not be used in the future without breaking 4100compatibility. 4101 4102If -ENOBUFS is returned the buffer provided was too small and userspace 4103may retry with a bigger buffer. 4104 41054.95 KVM_S390_SET_IRQ_STATE 4106--------------------------- 4107 4108:Capability: KVM_CAP_S390_IRQ_STATE 4109:Architectures: s390 4110:Type: vcpu ioctl 4111:Parameters: struct kvm_s390_irq_state (in) 4112:Returns: 0 on success, 4113 -EFAULT if the buffer address was invalid, 4114 -EINVAL for an invalid buffer length (see below), 4115 -EBUSY if there were already interrupts pending, 4116 errors occurring when actually injecting the 4117 interrupt. See KVM_S390_IRQ. 4118 4119This ioctl allows userspace to set the complete state of all cpu-local 4120interrupts currently pending for the vcpu. It is intended for restoring 4121interrupt state after a migration. The input parameter is a userspace buffer 4122containing a struct kvm_s390_irq_state:: 4123 4124 struct kvm_s390_irq_state { 4125 __u64 buf; 4126 __u32 flags; /* will stay unused for compatibility reasons */ 4127 __u32 len; 4128 __u32 reserved[4]; /* will stay unused for compatibility reasons */ 4129 }; 4130 4131The restrictions for flags and reserved apply as well. 4132(see KVM_S390_GET_IRQ_STATE) 4133 4134The userspace memory referenced by buf contains a struct kvm_s390_irq 4135for each interrupt to be injected into the guest. 4136If one of the interrupts could not be injected for some reason the 4137ioctl aborts. 4138 4139len must be a multiple of sizeof(struct kvm_s390_irq). It must be > 0 4140and it must not exceed (max_vcpus + 32) * sizeof(struct kvm_s390_irq), 4141which is the maximum number of possibly pending cpu-local interrupts. 4142 41434.96 KVM_SMI 4144------------ 4145 4146:Capability: KVM_CAP_X86_SMM 4147:Architectures: x86 4148:Type: vcpu ioctl 4149:Parameters: none 4150:Returns: 0 on success, -1 on error 4151 4152Queues an SMI on the thread's vcpu. 4153 41544.97 KVM_X86_SET_MSR_FILTER 4155---------------------------- 4156 4157:Capability: KVM_CAP_X86_MSR_FILTER 4158:Architectures: x86 4159:Type: vm ioctl 4160:Parameters: struct kvm_msr_filter 4161:Returns: 0 on success, < 0 on error 4162 4163:: 4164 4165 struct kvm_msr_filter_range { 4166 #define KVM_MSR_FILTER_READ (1 << 0) 4167 #define KVM_MSR_FILTER_WRITE (1 << 1) 4168 __u32 flags; 4169 __u32 nmsrs; /* number of msrs in bitmap */ 4170 __u32 base; /* MSR index the bitmap starts at */ 4171 __u8 *bitmap; /* a 1 bit allows the operations in flags, 0 denies */ 4172 }; 4173 4174 #define KVM_MSR_FILTER_MAX_RANGES 16 4175 struct kvm_msr_filter { 4176 #define KVM_MSR_FILTER_DEFAULT_ALLOW (0 << 0) 4177 #define KVM_MSR_FILTER_DEFAULT_DENY (1 << 0) 4178 __u32 flags; 4179 struct kvm_msr_filter_range ranges[KVM_MSR_FILTER_MAX_RANGES]; 4180 }; 4181 4182flags values for ``struct kvm_msr_filter_range``: 4183 4184``KVM_MSR_FILTER_READ`` 4185 4186 Filter read accesses to MSRs using the given bitmap. A 0 in the bitmap 4187 indicates that read accesses should be denied, while a 1 indicates that 4188 a read for a particular MSR should be allowed regardless of the default 4189 filter action. 4190 4191``KVM_MSR_FILTER_WRITE`` 4192 4193 Filter write accesses to MSRs using the given bitmap. A 0 in the bitmap 4194 indicates that write accesses should be denied, while a 1 indicates that 4195 a write for a particular MSR should be allowed regardless of the default 4196 filter action. 4197 4198flags values for ``struct kvm_msr_filter``: 4199 4200``KVM_MSR_FILTER_DEFAULT_ALLOW`` 4201 4202 If no filter range matches an MSR index that is getting accessed, KVM will 4203 allow accesses to all MSRs by default. 4204 4205``KVM_MSR_FILTER_DEFAULT_DENY`` 4206 4207 If no filter range matches an MSR index that is getting accessed, KVM will 4208 deny accesses to all MSRs by default. 4209 4210This ioctl allows userspace to define up to 16 bitmaps of MSR ranges to deny 4211guest MSR accesses that would normally be allowed by KVM. If an MSR is not 4212covered by a specific range, the "default" filtering behavior applies. Each 4213bitmap range covers MSRs from [base .. base+nmsrs). 4214 4215If an MSR access is denied by userspace, the resulting KVM behavior depends on 4216whether or not KVM_CAP_X86_USER_SPACE_MSR's KVM_MSR_EXIT_REASON_FILTER is 4217enabled. If KVM_MSR_EXIT_REASON_FILTER is enabled, KVM will exit to userspace 4218on denied accesses, i.e. userspace effectively intercepts the MSR access. If 4219KVM_MSR_EXIT_REASON_FILTER is not enabled, KVM will inject a #GP into the guest 4220on denied accesses. Note, if an MSR access is denied during emulation of MSR 4221load/stores during VMX transitions, KVM ignores KVM_MSR_EXIT_REASON_FILTER. 4222See the below warning for full details. 4223 4224If an MSR access is allowed by userspace, KVM will emulate and/or virtualize 4225the access in accordance with the vCPU model. Note, KVM may still ultimately 4226inject a #GP if an access is allowed by userspace, e.g. if KVM doesn't support 4227the MSR, or to follow architectural behavior for the MSR. 4228 4229By default, KVM operates in KVM_MSR_FILTER_DEFAULT_ALLOW mode with no MSR range 4230filters. 4231 4232Calling this ioctl with an empty set of ranges (all nmsrs == 0) disables MSR 4233filtering. In that mode, ``KVM_MSR_FILTER_DEFAULT_DENY`` is invalid and causes 4234an error. 4235 4236.. warning:: 4237 MSR accesses that are side effects of instruction execution (emulated or 4238 native) are not filtered as hardware does not honor MSR bitmaps outside of 4239 RDMSR and WRMSR, and KVM mimics that behavior when emulating instructions 4240 to avoid pointless divergence from hardware. E.g. RDPID reads MSR_TSC_AUX, 4241 SYSENTER reads the SYSENTER MSRs, etc. 4242 4243 MSRs that are loaded/stored via dedicated VMCS fields are not filtered as 4244 part of VM-Enter/VM-Exit emulation. 4245 4246 MSRs that are loaded/store via VMX's load/store lists _are_ filtered as part 4247 of VM-Enter/VM-Exit emulation. If an MSR access is denied on VM-Enter, KVM 4248 synthesizes a consistency check VM-Exit(EXIT_REASON_MSR_LOAD_FAIL). If an 4249 MSR access is denied on VM-Exit, KVM synthesizes a VM-Abort. In short, KVM 4250 extends Intel's architectural list of MSRs that cannot be loaded/saved via 4251 the VM-Enter/VM-Exit MSR list. It is platform owner's responsibility to 4252 to communicate any such restrictions to their end users. 4253 4254 x2APIC MSR accesses cannot be filtered (KVM silently ignores filters that 4255 cover any x2APIC MSRs). 4256 4257Note, invoking this ioctl while a vCPU is running is inherently racy. However, 4258KVM does guarantee that vCPUs will see either the previous filter or the new 4259filter, e.g. MSRs with identical settings in both the old and new filter will 4260have deterministic behavior. 4261 4262Similarly, if userspace wishes to intercept on denied accesses, 4263KVM_MSR_EXIT_REASON_FILTER must be enabled before activating any filters, and 4264left enabled until after all filters are deactivated. Failure to do so may 4265result in KVM injecting a #GP instead of exiting to userspace. 4266 42674.98 KVM_CREATE_SPAPR_TCE_64 4268---------------------------- 4269 4270:Capability: KVM_CAP_SPAPR_TCE_64 4271:Architectures: powerpc 4272:Type: vm ioctl 4273:Parameters: struct kvm_create_spapr_tce_64 (in) 4274:Returns: file descriptor for manipulating the created TCE table 4275 4276This is an extension for KVM_CAP_SPAPR_TCE which only supports 32bit 4277windows, described in 4.62 KVM_CREATE_SPAPR_TCE 4278 4279This capability uses extended struct in ioctl interface:: 4280 4281 /* for KVM_CAP_SPAPR_TCE_64 */ 4282 struct kvm_create_spapr_tce_64 { 4283 __u64 liobn; 4284 __u32 page_shift; 4285 __u32 flags; 4286 __u64 offset; /* in pages */ 4287 __u64 size; /* in pages */ 4288 }; 4289 4290The aim of extension is to support an additional bigger DMA window with 4291a variable page size. 4292KVM_CREATE_SPAPR_TCE_64 receives a 64bit window size, an IOMMU page shift and 4293a bus offset of the corresponding DMA window, @size and @offset are numbers 4294of IOMMU pages. 4295 4296@flags are not used at the moment. 4297 4298The rest of functionality is identical to KVM_CREATE_SPAPR_TCE. 4299 43004.99 KVM_REINJECT_CONTROL 4301------------------------- 4302 4303:Capability: KVM_CAP_REINJECT_CONTROL 4304:Architectures: x86 4305:Type: vm ioctl 4306:Parameters: struct kvm_reinject_control (in) 4307:Returns: 0 on success, 4308 -EFAULT if struct kvm_reinject_control cannot be read, 4309 -ENXIO if KVM_CREATE_PIT or KVM_CREATE_PIT2 didn't succeed earlier. 4310 4311i8254 (PIT) has two modes, reinject and !reinject. The default is reinject, 4312where KVM queues elapsed i8254 ticks and monitors completion of interrupt from 4313vector(s) that i8254 injects. Reinject mode dequeues a tick and injects its 4314interrupt whenever there isn't a pending interrupt from i8254. 4315!reinject mode injects an interrupt as soon as a tick arrives. 4316 4317:: 4318 4319 struct kvm_reinject_control { 4320 __u8 pit_reinject; 4321 __u8 reserved[31]; 4322 }; 4323 4324pit_reinject = 0 (!reinject mode) is recommended, unless running an old 4325operating system that uses the PIT for timing (e.g. Linux 2.4.x). 4326 43274.100 KVM_PPC_CONFIGURE_V3_MMU 4328------------------------------ 4329 4330:Capability: KVM_CAP_PPC_MMU_RADIX or KVM_CAP_PPC_MMU_HASH_V3 4331:Architectures: ppc 4332:Type: vm ioctl 4333:Parameters: struct kvm_ppc_mmuv3_cfg (in) 4334:Returns: 0 on success, 4335 -EFAULT if struct kvm_ppc_mmuv3_cfg cannot be read, 4336 -EINVAL if the configuration is invalid 4337 4338This ioctl controls whether the guest will use radix or HPT (hashed 4339page table) translation, and sets the pointer to the process table for 4340the guest. 4341 4342:: 4343 4344 struct kvm_ppc_mmuv3_cfg { 4345 __u64 flags; 4346 __u64 process_table; 4347 }; 4348 4349There are two bits that can be set in flags; KVM_PPC_MMUV3_RADIX and 4350KVM_PPC_MMUV3_GTSE. KVM_PPC_MMUV3_RADIX, if set, configures the guest 4351to use radix tree translation, and if clear, to use HPT translation. 4352KVM_PPC_MMUV3_GTSE, if set and if KVM permits it, configures the guest 4353to be able to use the global TLB and SLB invalidation instructions; 4354if clear, the guest may not use these instructions. 4355 4356The process_table field specifies the address and size of the guest 4357process table, which is in the guest's space. This field is formatted 4358as the second doubleword of the partition table entry, as defined in 4359the Power ISA V3.00, Book III section 5.7.6.1. 4360 43614.101 KVM_PPC_GET_RMMU_INFO 4362--------------------------- 4363 4364:Capability: KVM_CAP_PPC_MMU_RADIX 4365:Architectures: ppc 4366:Type: vm ioctl 4367:Parameters: struct kvm_ppc_rmmu_info (out) 4368:Returns: 0 on success, 4369 -EFAULT if struct kvm_ppc_rmmu_info cannot be written, 4370 -EINVAL if no useful information can be returned 4371 4372This ioctl returns a structure containing two things: (a) a list 4373containing supported radix tree geometries, and (b) a list that maps 4374page sizes to put in the "AP" (actual page size) field for the tlbie 4375(TLB invalidate entry) instruction. 4376 4377:: 4378 4379 struct kvm_ppc_rmmu_info { 4380 struct kvm_ppc_radix_geom { 4381 __u8 page_shift; 4382 __u8 level_bits[4]; 4383 __u8 pad[3]; 4384 } geometries[8]; 4385 __u32 ap_encodings[8]; 4386 }; 4387 4388The geometries[] field gives up to 8 supported geometries for the 4389radix page table, in terms of the log base 2 of the smallest page 4390size, and the number of bits indexed at each level of the tree, from 4391the PTE level up to the PGD level in that order. Any unused entries 4392will have 0 in the page_shift field. 4393 4394The ap_encodings gives the supported page sizes and their AP field 4395encodings, encoded with the AP value in the top 3 bits and the log 4396base 2 of the page size in the bottom 6 bits. 4397 43984.102 KVM_PPC_RESIZE_HPT_PREPARE 4399-------------------------------- 4400 4401:Capability: KVM_CAP_SPAPR_RESIZE_HPT 4402:Architectures: powerpc 4403:Type: vm ioctl 4404:Parameters: struct kvm_ppc_resize_hpt (in) 4405:Returns: 0 on successful completion, 4406 >0 if a new HPT is being prepared, the value is an estimated 4407 number of milliseconds until preparation is complete, 4408 -EFAULT if struct kvm_reinject_control cannot be read, 4409 -EINVAL if the supplied shift or flags are invalid, 4410 -ENOMEM if unable to allocate the new HPT, 4411 4412Used to implement the PAPR extension for runtime resizing of a guest's 4413Hashed Page Table (HPT). Specifically this starts, stops or monitors 4414the preparation of a new potential HPT for the guest, essentially 4415implementing the H_RESIZE_HPT_PREPARE hypercall. 4416 4417:: 4418 4419 struct kvm_ppc_resize_hpt { 4420 __u64 flags; 4421 __u32 shift; 4422 __u32 pad; 4423 }; 4424 4425If called with shift > 0 when there is no pending HPT for the guest, 4426this begins preparation of a new pending HPT of size 2^(shift) bytes. 4427It then returns a positive integer with the estimated number of 4428milliseconds until preparation is complete. 4429 4430If called when there is a pending HPT whose size does not match that 4431requested in the parameters, discards the existing pending HPT and 4432creates a new one as above. 4433 4434If called when there is a pending HPT of the size requested, will: 4435 4436 * If preparation of the pending HPT is already complete, return 0 4437 * If preparation of the pending HPT has failed, return an error 4438 code, then discard the pending HPT. 4439 * If preparation of the pending HPT is still in progress, return an 4440 estimated number of milliseconds until preparation is complete. 4441 4442If called with shift == 0, discards any currently pending HPT and 4443returns 0 (i.e. cancels any in-progress preparation). 4444 4445flags is reserved for future expansion, currently setting any bits in 4446flags will result in an -EINVAL. 4447 4448Normally this will be called repeatedly with the same parameters until 4449it returns <= 0. The first call will initiate preparation, subsequent 4450ones will monitor preparation until it completes or fails. 4451 44524.103 KVM_PPC_RESIZE_HPT_COMMIT 4453------------------------------- 4454 4455:Capability: KVM_CAP_SPAPR_RESIZE_HPT 4456:Architectures: powerpc 4457:Type: vm ioctl 4458:Parameters: struct kvm_ppc_resize_hpt (in) 4459:Returns: 0 on successful completion, 4460 -EFAULT if struct kvm_reinject_control cannot be read, 4461 -EINVAL if the supplied shift or flags are invalid, 4462 -ENXIO is there is no pending HPT, or the pending HPT doesn't 4463 have the requested size, 4464 -EBUSY if the pending HPT is not fully prepared, 4465 -ENOSPC if there was a hash collision when moving existing 4466 HPT entries to the new HPT, 4467 -EIO on other error conditions 4468 4469Used to implement the PAPR extension for runtime resizing of a guest's 4470Hashed Page Table (HPT). Specifically this requests that the guest be 4471transferred to working with the new HPT, essentially implementing the 4472H_RESIZE_HPT_COMMIT hypercall. 4473 4474:: 4475 4476 struct kvm_ppc_resize_hpt { 4477 __u64 flags; 4478 __u32 shift; 4479 __u32 pad; 4480 }; 4481 4482This should only be called after KVM_PPC_RESIZE_HPT_PREPARE has 4483returned 0 with the same parameters. In other cases 4484KVM_PPC_RESIZE_HPT_COMMIT will return an error (usually -ENXIO or 4485-EBUSY, though others may be possible if the preparation was started, 4486but failed). 4487 4488This will have undefined effects on the guest if it has not already 4489placed itself in a quiescent state where no vcpu will make MMU enabled 4490memory accesses. 4491 4492On successful completion, the pending HPT will become the guest's active 4493HPT and the previous HPT will be discarded. 4494 4495On failure, the guest will still be operating on its previous HPT. 4496 44974.104 KVM_X86_GET_MCE_CAP_SUPPORTED 4498----------------------------------- 4499 4500:Capability: KVM_CAP_MCE 4501:Architectures: x86 4502:Type: system ioctl 4503:Parameters: u64 mce_cap (out) 4504:Returns: 0 on success, -1 on error 4505 4506Returns supported MCE capabilities. The u64 mce_cap parameter 4507has the same format as the MSR_IA32_MCG_CAP register. Supported 4508capabilities will have the corresponding bits set. 4509 45104.105 KVM_X86_SETUP_MCE 4511----------------------- 4512 4513:Capability: KVM_CAP_MCE 4514:Architectures: x86 4515:Type: vcpu ioctl 4516:Parameters: u64 mcg_cap (in) 4517:Returns: 0 on success, 4518 -EFAULT if u64 mcg_cap cannot be read, 4519 -EINVAL if the requested number of banks is invalid, 4520 -EINVAL if requested MCE capability is not supported. 4521 4522Initializes MCE support for use. The u64 mcg_cap parameter 4523has the same format as the MSR_IA32_MCG_CAP register and 4524specifies which capabilities should be enabled. The maximum 4525supported number of error-reporting banks can be retrieved when 4526checking for KVM_CAP_MCE. The supported capabilities can be 4527retrieved with KVM_X86_GET_MCE_CAP_SUPPORTED. 4528 45294.106 KVM_X86_SET_MCE 4530--------------------- 4531 4532:Capability: KVM_CAP_MCE 4533:Architectures: x86 4534:Type: vcpu ioctl 4535:Parameters: struct kvm_x86_mce (in) 4536:Returns: 0 on success, 4537 -EFAULT if struct kvm_x86_mce cannot be read, 4538 -EINVAL if the bank number is invalid, 4539 -EINVAL if VAL bit is not set in status field. 4540 4541Inject a machine check error (MCE) into the guest. The input 4542parameter is:: 4543 4544 struct kvm_x86_mce { 4545 __u64 status; 4546 __u64 addr; 4547 __u64 misc; 4548 __u64 mcg_status; 4549 __u8 bank; 4550 __u8 pad1[7]; 4551 __u64 pad2[3]; 4552 }; 4553 4554If the MCE being reported is an uncorrected error, KVM will 4555inject it as an MCE exception into the guest. If the guest 4556MCG_STATUS register reports that an MCE is in progress, KVM 4557causes an KVM_EXIT_SHUTDOWN vmexit. 4558 4559Otherwise, if the MCE is a corrected error, KVM will just 4560store it in the corresponding bank (provided this bank is 4561not holding a previously reported uncorrected error). 4562 45634.107 KVM_S390_GET_CMMA_BITS 4564---------------------------- 4565 4566:Capability: KVM_CAP_S390_CMMA_MIGRATION 4567:Architectures: s390 4568:Type: vm ioctl 4569:Parameters: struct kvm_s390_cmma_log (in, out) 4570:Returns: 0 on success, a negative value on error 4571 4572Errors: 4573 4574 ====== ============================================================= 4575 ENOMEM not enough memory can be allocated to complete the task 4576 ENXIO if CMMA is not enabled 4577 EINVAL if KVM_S390_CMMA_PEEK is not set but migration mode was not enabled 4578 EINVAL if KVM_S390_CMMA_PEEK is not set but dirty tracking has been 4579 disabled (and thus migration mode was automatically disabled) 4580 EFAULT if the userspace address is invalid or if no page table is 4581 present for the addresses (e.g. when using hugepages). 4582 ====== ============================================================= 4583 4584This ioctl is used to get the values of the CMMA bits on the s390 4585architecture. It is meant to be used in two scenarios: 4586 4587- During live migration to save the CMMA values. Live migration needs 4588 to be enabled via the KVM_REQ_START_MIGRATION VM property. 4589- To non-destructively peek at the CMMA values, with the flag 4590 KVM_S390_CMMA_PEEK set. 4591 4592The ioctl takes parameters via the kvm_s390_cmma_log struct. The desired 4593values are written to a buffer whose location is indicated via the "values" 4594member in the kvm_s390_cmma_log struct. The values in the input struct are 4595also updated as needed. 4596 4597Each CMMA value takes up one byte. 4598 4599:: 4600 4601 struct kvm_s390_cmma_log { 4602 __u64 start_gfn; 4603 __u32 count; 4604 __u32 flags; 4605 union { 4606 __u64 remaining; 4607 __u64 mask; 4608 }; 4609 __u64 values; 4610 }; 4611 4612start_gfn is the number of the first guest frame whose CMMA values are 4613to be retrieved, 4614 4615count is the length of the buffer in bytes, 4616 4617values points to the buffer where the result will be written to. 4618 4619If count is greater than KVM_S390_SKEYS_MAX, then it is considered to be 4620KVM_S390_SKEYS_MAX. KVM_S390_SKEYS_MAX is re-used for consistency with 4621other ioctls. 4622 4623The result is written in the buffer pointed to by the field values, and 4624the values of the input parameter are updated as follows. 4625 4626Depending on the flags, different actions are performed. The only 4627supported flag so far is KVM_S390_CMMA_PEEK. 4628 4629The default behaviour if KVM_S390_CMMA_PEEK is not set is: 4630start_gfn will indicate the first page frame whose CMMA bits were dirty. 4631It is not necessarily the same as the one passed as input, as clean pages 4632are skipped. 4633 4634count will indicate the number of bytes actually written in the buffer. 4635It can (and very often will) be smaller than the input value, since the 4636buffer is only filled until 16 bytes of clean values are found (which 4637are then not copied in the buffer). Since a CMMA migration block needs 4638the base address and the length, for a total of 16 bytes, we will send 4639back some clean data if there is some dirty data afterwards, as long as 4640the size of the clean data does not exceed the size of the header. This 4641allows to minimize the amount of data to be saved or transferred over 4642the network at the expense of more roundtrips to userspace. The next 4643invocation of the ioctl will skip over all the clean values, saving 4644potentially more than just the 16 bytes we found. 4645 4646If KVM_S390_CMMA_PEEK is set: 4647the existing storage attributes are read even when not in migration 4648mode, and no other action is performed; 4649 4650the output start_gfn will be equal to the input start_gfn, 4651 4652the output count will be equal to the input count, except if the end of 4653memory has been reached. 4654 4655In both cases: 4656the field "remaining" will indicate the total number of dirty CMMA values 4657still remaining, or 0 if KVM_S390_CMMA_PEEK is set and migration mode is 4658not enabled. 4659 4660mask is unused. 4661 4662values points to the userspace buffer where the result will be stored. 4663 46644.108 KVM_S390_SET_CMMA_BITS 4665---------------------------- 4666 4667:Capability: KVM_CAP_S390_CMMA_MIGRATION 4668:Architectures: s390 4669:Type: vm ioctl 4670:Parameters: struct kvm_s390_cmma_log (in) 4671:Returns: 0 on success, a negative value on error 4672 4673This ioctl is used to set the values of the CMMA bits on the s390 4674architecture. It is meant to be used during live migration to restore 4675the CMMA values, but there are no restrictions on its use. 4676The ioctl takes parameters via the kvm_s390_cmma_values struct. 4677Each CMMA value takes up one byte. 4678 4679:: 4680 4681 struct kvm_s390_cmma_log { 4682 __u64 start_gfn; 4683 __u32 count; 4684 __u32 flags; 4685 union { 4686 __u64 remaining; 4687 __u64 mask; 4688 }; 4689 __u64 values; 4690 }; 4691 4692start_gfn indicates the starting guest frame number, 4693 4694count indicates how many values are to be considered in the buffer, 4695 4696flags is not used and must be 0. 4697 4698mask indicates which PGSTE bits are to be considered. 4699 4700remaining is not used. 4701 4702values points to the buffer in userspace where to store the values. 4703 4704This ioctl can fail with -ENOMEM if not enough memory can be allocated to 4705complete the task, with -ENXIO if CMMA is not enabled, with -EINVAL if 4706the count field is too large (e.g. more than KVM_S390_CMMA_SIZE_MAX) or 4707if the flags field was not 0, with -EFAULT if the userspace address is 4708invalid, if invalid pages are written to (e.g. after the end of memory) 4709or if no page table is present for the addresses (e.g. when using 4710hugepages). 4711 47124.109 KVM_PPC_GET_CPU_CHAR 4713-------------------------- 4714 4715:Capability: KVM_CAP_PPC_GET_CPU_CHAR 4716:Architectures: powerpc 4717:Type: vm ioctl 4718:Parameters: struct kvm_ppc_cpu_char (out) 4719:Returns: 0 on successful completion, 4720 -EFAULT if struct kvm_ppc_cpu_char cannot be written 4721 4722This ioctl gives userspace information about certain characteristics 4723of the CPU relating to speculative execution of instructions and 4724possible information leakage resulting from speculative execution (see 4725CVE-2017-5715, CVE-2017-5753 and CVE-2017-5754). The information is 4726returned in struct kvm_ppc_cpu_char, which looks like this:: 4727 4728 struct kvm_ppc_cpu_char { 4729 __u64 character; /* characteristics of the CPU */ 4730 __u64 behaviour; /* recommended software behaviour */ 4731 __u64 character_mask; /* valid bits in character */ 4732 __u64 behaviour_mask; /* valid bits in behaviour */ 4733 }; 4734 4735For extensibility, the character_mask and behaviour_mask fields 4736indicate which bits of character and behaviour have been filled in by 4737the kernel. If the set of defined bits is extended in future then 4738userspace will be able to tell whether it is running on a kernel that 4739knows about the new bits. 4740 4741The character field describes attributes of the CPU which can help 4742with preventing inadvertent information disclosure - specifically, 4743whether there is an instruction to flash-invalidate the L1 data cache 4744(ori 30,30,0 or mtspr SPRN_TRIG2,rN), whether the L1 data cache is set 4745to a mode where entries can only be used by the thread that created 4746them, whether the bcctr[l] instruction prevents speculation, and 4747whether a speculation barrier instruction (ori 31,31,0) is provided. 4748 4749The behaviour field describes actions that software should take to 4750prevent inadvertent information disclosure, and thus describes which 4751vulnerabilities the hardware is subject to; specifically whether the 4752L1 data cache should be flushed when returning to user mode from the 4753kernel, and whether a speculation barrier should be placed between an 4754array bounds check and the array access. 4755 4756These fields use the same bit definitions as the new 4757H_GET_CPU_CHARACTERISTICS hypercall. 4758 47594.110 KVM_MEMORY_ENCRYPT_OP 4760--------------------------- 4761 4762:Capability: basic 4763:Architectures: x86 4764:Type: vm 4765:Parameters: an opaque platform specific structure (in/out) 4766:Returns: 0 on success; -1 on error 4767 4768If the platform supports creating encrypted VMs then this ioctl can be used 4769for issuing platform-specific memory encryption commands to manage those 4770encrypted VMs. 4771 4772Currently, this ioctl is used for issuing Secure Encrypted Virtualization 4773(SEV) commands on AMD Processors. The SEV commands are defined in 4774Documentation/virt/kvm/x86/amd-memory-encryption.rst. 4775 47764.111 KVM_MEMORY_ENCRYPT_REG_REGION 4777----------------------------------- 4778 4779:Capability: basic 4780:Architectures: x86 4781:Type: system 4782:Parameters: struct kvm_enc_region (in) 4783:Returns: 0 on success; -1 on error 4784 4785This ioctl can be used to register a guest memory region which may 4786contain encrypted data (e.g. guest RAM, SMRAM etc). 4787 4788It is used in the SEV-enabled guest. When encryption is enabled, a guest 4789memory region may contain encrypted data. The SEV memory encryption 4790engine uses a tweak such that two identical plaintext pages, each at 4791different locations will have differing ciphertexts. So swapping or 4792moving ciphertext of those pages will not result in plaintext being 4793swapped. So relocating (or migrating) physical backing pages for the SEV 4794guest will require some additional steps. 4795 4796Note: The current SEV key management spec does not provide commands to 4797swap or migrate (move) ciphertext pages. Hence, for now we pin the guest 4798memory region registered with the ioctl. 4799 48004.112 KVM_MEMORY_ENCRYPT_UNREG_REGION 4801------------------------------------- 4802 4803:Capability: basic 4804:Architectures: x86 4805:Type: system 4806:Parameters: struct kvm_enc_region (in) 4807:Returns: 0 on success; -1 on error 4808 4809This ioctl can be used to unregister the guest memory region registered 4810with KVM_MEMORY_ENCRYPT_REG_REGION ioctl above. 4811 48124.113 KVM_HYPERV_EVENTFD 4813------------------------ 4814 4815:Capability: KVM_CAP_HYPERV_EVENTFD 4816:Architectures: x86 4817:Type: vm ioctl 4818:Parameters: struct kvm_hyperv_eventfd (in) 4819 4820This ioctl (un)registers an eventfd to receive notifications from the guest on 4821the specified Hyper-V connection id through the SIGNAL_EVENT hypercall, without 4822causing a user exit. SIGNAL_EVENT hypercall with non-zero event flag number 4823(bits 24-31) still triggers a KVM_EXIT_HYPERV_HCALL user exit. 4824 4825:: 4826 4827 struct kvm_hyperv_eventfd { 4828 __u32 conn_id; 4829 __s32 fd; 4830 __u32 flags; 4831 __u32 padding[3]; 4832 }; 4833 4834The conn_id field should fit within 24 bits:: 4835 4836 #define KVM_HYPERV_CONN_ID_MASK 0x00ffffff 4837 4838The acceptable values for the flags field are:: 4839 4840 #define KVM_HYPERV_EVENTFD_DEASSIGN (1 << 0) 4841 4842:Returns: 0 on success, 4843 -EINVAL if conn_id or flags is outside the allowed range, 4844 -ENOENT on deassign if the conn_id isn't registered, 4845 -EEXIST on assign if the conn_id is already registered 4846 48474.114 KVM_GET_NESTED_STATE 4848-------------------------- 4849 4850:Capability: KVM_CAP_NESTED_STATE 4851:Architectures: x86 4852:Type: vcpu ioctl 4853:Parameters: struct kvm_nested_state (in/out) 4854:Returns: 0 on success, -1 on error 4855 4856Errors: 4857 4858 ===== ============================================================= 4859 E2BIG the total state size exceeds the value of 'size' specified by 4860 the user; the size required will be written into size. 4861 ===== ============================================================= 4862 4863:: 4864 4865 struct kvm_nested_state { 4866 __u16 flags; 4867 __u16 format; 4868 __u32 size; 4869 4870 union { 4871 struct kvm_vmx_nested_state_hdr vmx; 4872 struct kvm_svm_nested_state_hdr svm; 4873 4874 /* Pad the header to 128 bytes. */ 4875 __u8 pad[120]; 4876 } hdr; 4877 4878 union { 4879 struct kvm_vmx_nested_state_data vmx[0]; 4880 struct kvm_svm_nested_state_data svm[0]; 4881 } data; 4882 }; 4883 4884 #define KVM_STATE_NESTED_GUEST_MODE 0x00000001 4885 #define KVM_STATE_NESTED_RUN_PENDING 0x00000002 4886 #define KVM_STATE_NESTED_EVMCS 0x00000004 4887 4888 #define KVM_STATE_NESTED_FORMAT_VMX 0 4889 #define KVM_STATE_NESTED_FORMAT_SVM 1 4890 4891 #define KVM_STATE_NESTED_VMX_VMCS_SIZE 0x1000 4892 4893 #define KVM_STATE_NESTED_VMX_SMM_GUEST_MODE 0x00000001 4894 #define KVM_STATE_NESTED_VMX_SMM_VMXON 0x00000002 4895 4896 #define KVM_STATE_VMX_PREEMPTION_TIMER_DEADLINE 0x00000001 4897 4898 struct kvm_vmx_nested_state_hdr { 4899 __u64 vmxon_pa; 4900 __u64 vmcs12_pa; 4901 4902 struct { 4903 __u16 flags; 4904 } smm; 4905 4906 __u32 flags; 4907 __u64 preemption_timer_deadline; 4908 }; 4909 4910 struct kvm_vmx_nested_state_data { 4911 __u8 vmcs12[KVM_STATE_NESTED_VMX_VMCS_SIZE]; 4912 __u8 shadow_vmcs12[KVM_STATE_NESTED_VMX_VMCS_SIZE]; 4913 }; 4914 4915This ioctl copies the vcpu's nested virtualization state from the kernel to 4916userspace. 4917 4918The maximum size of the state can be retrieved by passing KVM_CAP_NESTED_STATE 4919to the KVM_CHECK_EXTENSION ioctl(). 4920 49214.115 KVM_SET_NESTED_STATE 4922-------------------------- 4923 4924:Capability: KVM_CAP_NESTED_STATE 4925:Architectures: x86 4926:Type: vcpu ioctl 4927:Parameters: struct kvm_nested_state (in) 4928:Returns: 0 on success, -1 on error 4929 4930This copies the vcpu's kvm_nested_state struct from userspace to the kernel. 4931For the definition of struct kvm_nested_state, see KVM_GET_NESTED_STATE. 4932 49334.116 KVM_(UN)REGISTER_COALESCED_MMIO 4934------------------------------------- 4935 4936:Capability: KVM_CAP_COALESCED_MMIO (for coalesced mmio) 4937 KVM_CAP_COALESCED_PIO (for coalesced pio) 4938:Architectures: all 4939:Type: vm ioctl 4940:Parameters: struct kvm_coalesced_mmio_zone 4941:Returns: 0 on success, < 0 on error 4942 4943Coalesced I/O is a performance optimization that defers hardware 4944register write emulation so that userspace exits are avoided. It is 4945typically used to reduce the overhead of emulating frequently accessed 4946hardware registers. 4947 4948When a hardware register is configured for coalesced I/O, write accesses 4949do not exit to userspace and their value is recorded in a ring buffer 4950that is shared between kernel and userspace. 4951 4952Coalesced I/O is used if one or more write accesses to a hardware 4953register can be deferred until a read or a write to another hardware 4954register on the same device. This last access will cause a vmexit and 4955userspace will process accesses from the ring buffer before emulating 4956it. That will avoid exiting to userspace on repeated writes. 4957 4958Coalesced pio is based on coalesced mmio. There is little difference 4959between coalesced mmio and pio except that coalesced pio records accesses 4960to I/O ports. 4961 49624.117 KVM_CLEAR_DIRTY_LOG 4963------------------------- 4964 4965:Capability: KVM_CAP_MANUAL_DIRTY_LOG_PROTECT2 4966:Architectures: x86, arm64, mips 4967:Type: vm ioctl 4968:Parameters: struct kvm_clear_dirty_log (in) 4969:Returns: 0 on success, -1 on error 4970 4971:: 4972 4973 /* for KVM_CLEAR_DIRTY_LOG */ 4974 struct kvm_clear_dirty_log { 4975 __u32 slot; 4976 __u32 num_pages; 4977 __u64 first_page; 4978 union { 4979 void __user *dirty_bitmap; /* one bit per page */ 4980 __u64 padding; 4981 }; 4982 }; 4983 4984The ioctl clears the dirty status of pages in a memory slot, according to 4985the bitmap that is passed in struct kvm_clear_dirty_log's dirty_bitmap 4986field. Bit 0 of the bitmap corresponds to page "first_page" in the 4987memory slot, and num_pages is the size in bits of the input bitmap. 4988first_page must be a multiple of 64; num_pages must also be a multiple of 498964 unless first_page + num_pages is the size of the memory slot. For each 4990bit that is set in the input bitmap, the corresponding page is marked "clean" 4991in KVM's dirty bitmap, and dirty tracking is re-enabled for that page 4992(for example via write-protection, or by clearing the dirty bit in 4993a page table entry). 4994 4995If KVM_CAP_MULTI_ADDRESS_SPACE is available, bits 16-31 of slot field specifies 4996the address space for which you want to clear the dirty status. See 4997KVM_SET_USER_MEMORY_REGION for details on the usage of slot field. 4998 4999This ioctl is mostly useful when KVM_CAP_MANUAL_DIRTY_LOG_PROTECT2 5000is enabled; for more information, see the description of the capability. 5001However, it can always be used as long as KVM_CHECK_EXTENSION confirms 5002that KVM_CAP_MANUAL_DIRTY_LOG_PROTECT2 is present. 5003 50044.118 KVM_GET_SUPPORTED_HV_CPUID 5005-------------------------------- 5006 5007:Capability: KVM_CAP_HYPERV_CPUID (vcpu), KVM_CAP_SYS_HYPERV_CPUID (system) 5008:Architectures: x86 5009:Type: system ioctl, vcpu ioctl 5010:Parameters: struct kvm_cpuid2 (in/out) 5011:Returns: 0 on success, -1 on error 5012 5013:: 5014 5015 struct kvm_cpuid2 { 5016 __u32 nent; 5017 __u32 padding; 5018 struct kvm_cpuid_entry2 entries[0]; 5019 }; 5020 5021 struct kvm_cpuid_entry2 { 5022 __u32 function; 5023 __u32 index; 5024 __u32 flags; 5025 __u32 eax; 5026 __u32 ebx; 5027 __u32 ecx; 5028 __u32 edx; 5029 __u32 padding[3]; 5030 }; 5031 5032This ioctl returns x86 cpuid features leaves related to Hyper-V emulation in 5033KVM. Userspace can use the information returned by this ioctl to construct 5034cpuid information presented to guests consuming Hyper-V enlightenments (e.g. 5035Windows or Hyper-V guests). 5036 5037CPUID feature leaves returned by this ioctl are defined by Hyper-V Top Level 5038Functional Specification (TLFS). These leaves can't be obtained with 5039KVM_GET_SUPPORTED_CPUID ioctl because some of them intersect with KVM feature 5040leaves (0x40000000, 0x40000001). 5041 5042Currently, the following list of CPUID leaves are returned: 5043 5044 - HYPERV_CPUID_VENDOR_AND_MAX_FUNCTIONS 5045 - HYPERV_CPUID_INTERFACE 5046 - HYPERV_CPUID_VERSION 5047 - HYPERV_CPUID_FEATURES 5048 - HYPERV_CPUID_ENLIGHTMENT_INFO 5049 - HYPERV_CPUID_IMPLEMENT_LIMITS 5050 - HYPERV_CPUID_NESTED_FEATURES 5051 - HYPERV_CPUID_SYNDBG_VENDOR_AND_MAX_FUNCTIONS 5052 - HYPERV_CPUID_SYNDBG_INTERFACE 5053 - HYPERV_CPUID_SYNDBG_PLATFORM_CAPABILITIES 5054 5055Userspace invokes KVM_GET_SUPPORTED_HV_CPUID by passing a kvm_cpuid2 structure 5056with the 'nent' field indicating the number of entries in the variable-size 5057array 'entries'. If the number of entries is too low to describe all Hyper-V 5058feature leaves, an error (E2BIG) is returned. If the number is more or equal 5059to the number of Hyper-V feature leaves, the 'nent' field is adjusted to the 5060number of valid entries in the 'entries' array, which is then filled. 5061 5062'index' and 'flags' fields in 'struct kvm_cpuid_entry2' are currently reserved, 5063userspace should not expect to get any particular value there. 5064 5065Note, vcpu version of KVM_GET_SUPPORTED_HV_CPUID is currently deprecated. Unlike 5066system ioctl which exposes all supported feature bits unconditionally, vcpu 5067version has the following quirks: 5068 5069- HYPERV_CPUID_NESTED_FEATURES leaf and HV_X64_ENLIGHTENED_VMCS_RECOMMENDED 5070 feature bit are only exposed when Enlightened VMCS was previously enabled 5071 on the corresponding vCPU (KVM_CAP_HYPERV_ENLIGHTENED_VMCS). 5072- HV_STIMER_DIRECT_MODE_AVAILABLE bit is only exposed with in-kernel LAPIC. 5073 (presumes KVM_CREATE_IRQCHIP has already been called). 5074 50754.119 KVM_ARM_VCPU_FINALIZE 5076--------------------------- 5077 5078:Architectures: arm64 5079:Type: vcpu ioctl 5080:Parameters: int feature (in) 5081:Returns: 0 on success, -1 on error 5082 5083Errors: 5084 5085 ====== ============================================================== 5086 EPERM feature not enabled, needs configuration, or already finalized 5087 EINVAL feature unknown or not present 5088 ====== ============================================================== 5089 5090Recognised values for feature: 5091 5092 ===== =========================================== 5093 arm64 KVM_ARM_VCPU_SVE (requires KVM_CAP_ARM_SVE) 5094 ===== =========================================== 5095 5096Finalizes the configuration of the specified vcpu feature. 5097 5098The vcpu must already have been initialised, enabling the affected feature, by 5099means of a successful :ref:`KVM_ARM_VCPU_INIT <KVM_ARM_VCPU_INIT>` call with the 5100appropriate flag set in features[]. 5101 5102For affected vcpu features, this is a mandatory step that must be performed 5103before the vcpu is fully usable. 5104 5105Between KVM_ARM_VCPU_INIT and KVM_ARM_VCPU_FINALIZE, the feature may be 5106configured by use of ioctls such as KVM_SET_ONE_REG. The exact configuration 5107that should be performed and how to do it are feature-dependent. 5108 5109Other calls that depend on a particular feature being finalized, such as 5110KVM_RUN, KVM_GET_REG_LIST, KVM_GET_ONE_REG and KVM_SET_ONE_REG, will fail with 5111-EPERM unless the feature has already been finalized by means of a 5112KVM_ARM_VCPU_FINALIZE call. 5113 5114See KVM_ARM_VCPU_INIT for details of vcpu features that require finalization 5115using this ioctl. 5116 51174.120 KVM_SET_PMU_EVENT_FILTER 5118------------------------------ 5119 5120:Capability: KVM_CAP_PMU_EVENT_FILTER 5121:Architectures: x86 5122:Type: vm ioctl 5123:Parameters: struct kvm_pmu_event_filter (in) 5124:Returns: 0 on success, -1 on error 5125 5126Errors: 5127 5128 ====== ============================================================ 5129 EFAULT args[0] cannot be accessed 5130 EINVAL args[0] contains invalid data in the filter or filter events 5131 E2BIG nevents is too large 5132 EBUSY not enough memory to allocate the filter 5133 ====== ============================================================ 5134 5135:: 5136 5137 struct kvm_pmu_event_filter { 5138 __u32 action; 5139 __u32 nevents; 5140 __u32 fixed_counter_bitmap; 5141 __u32 flags; 5142 __u32 pad[4]; 5143 __u64 events[0]; 5144 }; 5145 5146This ioctl restricts the set of PMU events the guest can program by limiting 5147which event select and unit mask combinations are permitted. 5148 5149The argument holds a list of filter events which will be allowed or denied. 5150 5151Filter events only control general purpose counters; fixed purpose counters 5152are controlled by the fixed_counter_bitmap. 5153 5154Valid values for 'flags':: 5155 5156``0`` 5157 5158To use this mode, clear the 'flags' field. 5159 5160In this mode each event will contain an event select + unit mask. 5161 5162When the guest attempts to program the PMU the guest's event select + 5163unit mask is compared against the filter events to determine whether the 5164guest should have access. 5165 5166``KVM_PMU_EVENT_FLAG_MASKED_EVENTS`` 5167:Capability: KVM_CAP_PMU_EVENT_MASKED_EVENTS 5168 5169In this mode each filter event will contain an event select, mask, match, and 5170exclude value. To encode a masked event use:: 5171 5172 KVM_PMU_ENCODE_MASKED_ENTRY() 5173 5174An encoded event will follow this layout:: 5175 5176 Bits Description 5177 ---- ----------- 5178 7:0 event select (low bits) 5179 15:8 umask match 5180 31:16 unused 5181 35:32 event select (high bits) 5182 36:54 unused 5183 55 exclude bit 5184 63:56 umask mask 5185 5186When the guest attempts to program the PMU, these steps are followed in 5187determining if the guest should have access: 5188 5189 1. Match the event select from the guest against the filter events. 5190 2. If a match is found, match the guest's unit mask to the mask and match 5191 values of the included filter events. 5192 I.e. (unit mask & mask) == match && !exclude. 5193 3. If a match is found, match the guest's unit mask to the mask and match 5194 values of the excluded filter events. 5195 I.e. (unit mask & mask) == match && exclude. 5196 4. 5197 a. If an included match is found and an excluded match is not found, filter 5198 the event. 5199 b. For everything else, do not filter the event. 5200 5. 5201 a. If the event is filtered and it's an allow list, allow the guest to 5202 program the event. 5203 b. If the event is filtered and it's a deny list, do not allow the guest to 5204 program the event. 5205 5206When setting a new pmu event filter, -EINVAL will be returned if any of the 5207unused fields are set or if any of the high bits (35:32) in the event 5208select are set when called on Intel. 5209 5210Valid values for 'action':: 5211 5212 #define KVM_PMU_EVENT_ALLOW 0 5213 #define KVM_PMU_EVENT_DENY 1 5214 5215Via this API, KVM userspace can also control the behavior of the VM's fixed 5216counters (if any) by configuring the "action" and "fixed_counter_bitmap" fields. 5217 5218Specifically, KVM follows the following pseudo-code when determining whether to 5219allow the guest FixCtr[i] to count its pre-defined fixed event:: 5220 5221 FixCtr[i]_is_allowed = (action == ALLOW) && (bitmap & BIT(i)) || 5222 (action == DENY) && !(bitmap & BIT(i)); 5223 FixCtr[i]_is_denied = !FixCtr[i]_is_allowed; 5224 5225KVM always consumes fixed_counter_bitmap, it's userspace's responsibility to 5226ensure fixed_counter_bitmap is set correctly, e.g. if userspace wants to define 5227a filter that only affects general purpose counters. 5228 5229Note, the "events" field also applies to fixed counters' hardcoded event_select 5230and unit_mask values. "fixed_counter_bitmap" has higher priority than "events" 5231if there is a contradiction between the two. 5232 52334.121 KVM_PPC_SVM_OFF 5234--------------------- 5235 5236:Capability: basic 5237:Architectures: powerpc 5238:Type: vm ioctl 5239:Parameters: none 5240:Returns: 0 on successful completion, 5241 5242Errors: 5243 5244 ====== ================================================================ 5245 EINVAL if ultravisor failed to terminate the secure guest 5246 ENOMEM if hypervisor failed to allocate new radix page tables for guest 5247 ====== ================================================================ 5248 5249This ioctl is used to turn off the secure mode of the guest or transition 5250the guest from secure mode to normal mode. This is invoked when the guest 5251is reset. This has no effect if called for a normal guest. 5252 5253This ioctl issues an ultravisor call to terminate the secure guest, 5254unpins the VPA pages and releases all the device pages that are used to 5255track the secure pages by hypervisor. 5256 52574.122 KVM_S390_NORMAL_RESET 5258--------------------------- 5259 5260:Capability: KVM_CAP_S390_VCPU_RESETS 5261:Architectures: s390 5262:Type: vcpu ioctl 5263:Parameters: none 5264:Returns: 0 5265 5266This ioctl resets VCPU registers and control structures according to 5267the cpu reset definition in the POP (Principles Of Operation). 5268 52694.123 KVM_S390_INITIAL_RESET 5270---------------------------- 5271 5272:Capability: basic 5273:Architectures: s390 5274:Type: vcpu ioctl 5275:Parameters: none 5276:Returns: 0 5277 5278This ioctl resets VCPU registers and control structures according to 5279the initial cpu reset definition in the POP. However, the cpu is not 5280put into ESA mode. This reset is a superset of the normal reset. 5281 52824.124 KVM_S390_CLEAR_RESET 5283-------------------------- 5284 5285:Capability: KVM_CAP_S390_VCPU_RESETS 5286:Architectures: s390 5287:Type: vcpu ioctl 5288:Parameters: none 5289:Returns: 0 5290 5291This ioctl resets VCPU registers and control structures according to 5292the clear cpu reset definition in the POP. However, the cpu is not put 5293into ESA mode. This reset is a superset of the initial reset. 5294 5295 52964.125 KVM_S390_PV_COMMAND 5297------------------------- 5298 5299:Capability: KVM_CAP_S390_PROTECTED 5300:Architectures: s390 5301:Type: vm ioctl 5302:Parameters: struct kvm_pv_cmd 5303:Returns: 0 on success, < 0 on error 5304 5305:: 5306 5307 struct kvm_pv_cmd { 5308 __u32 cmd; /* Command to be executed */ 5309 __u16 rc; /* Ultravisor return code */ 5310 __u16 rrc; /* Ultravisor return reason code */ 5311 __u64 data; /* Data or address */ 5312 __u32 flags; /* flags for future extensions. Must be 0 for now */ 5313 __u32 reserved[3]; 5314 }; 5315 5316**Ultravisor return codes** 5317The Ultravisor return (reason) codes are provided by the kernel if a 5318Ultravisor call has been executed to achieve the results expected by 5319the command. Therefore they are independent of the IOCTL return 5320code. If KVM changes `rc`, its value will always be greater than 0 5321hence setting it to 0 before issuing a PV command is advised to be 5322able to detect a change of `rc`. 5323 5324**cmd values:** 5325 5326KVM_PV_ENABLE 5327 Allocate memory and register the VM with the Ultravisor, thereby 5328 donating memory to the Ultravisor that will become inaccessible to 5329 KVM. All existing CPUs are converted to protected ones. After this 5330 command has succeeded, any CPU added via hotplug will become 5331 protected during its creation as well. 5332 5333 Errors: 5334 5335 ===== ============================= 5336 EINTR an unmasked signal is pending 5337 ===== ============================= 5338 5339KVM_PV_DISABLE 5340 Deregister the VM from the Ultravisor and reclaim the memory that had 5341 been donated to the Ultravisor, making it usable by the kernel again. 5342 All registered VCPUs are converted back to non-protected ones. If a 5343 previous protected VM had been prepared for asynchronous teardown with 5344 KVM_PV_ASYNC_CLEANUP_PREPARE and not subsequently torn down with 5345 KVM_PV_ASYNC_CLEANUP_PERFORM, it will be torn down in this call 5346 together with the current protected VM. 5347 5348KVM_PV_VM_SET_SEC_PARMS 5349 Pass the image header from VM memory to the Ultravisor in 5350 preparation of image unpacking and verification. 5351 5352KVM_PV_VM_UNPACK 5353 Unpack (protect and decrypt) a page of the encrypted boot image. 5354 5355KVM_PV_VM_VERIFY 5356 Verify the integrity of the unpacked image. Only if this succeeds, 5357 KVM is allowed to start protected VCPUs. 5358 5359KVM_PV_INFO 5360 :Capability: KVM_CAP_S390_PROTECTED_DUMP 5361 5362 Presents an API that provides Ultravisor related data to userspace 5363 via subcommands. len_max is the size of the user space buffer, 5364 len_written is KVM's indication of how much bytes of that buffer 5365 were actually written to. len_written can be used to determine the 5366 valid fields if more response fields are added in the future. 5367 5368 :: 5369 5370 enum pv_cmd_info_id { 5371 KVM_PV_INFO_VM, 5372 KVM_PV_INFO_DUMP, 5373 }; 5374 5375 struct kvm_s390_pv_info_header { 5376 __u32 id; 5377 __u32 len_max; 5378 __u32 len_written; 5379 __u32 reserved; 5380 }; 5381 5382 struct kvm_s390_pv_info { 5383 struct kvm_s390_pv_info_header header; 5384 struct kvm_s390_pv_info_dump dump; 5385 struct kvm_s390_pv_info_vm vm; 5386 }; 5387 5388**subcommands:** 5389 5390 KVM_PV_INFO_VM 5391 This subcommand provides basic Ultravisor information for PV 5392 hosts. These values are likely also exported as files in the sysfs 5393 firmware UV query interface but they are more easily available to 5394 programs in this API. 5395 5396 The installed calls and feature_indication members provide the 5397 installed UV calls and the UV's other feature indications. 5398 5399 The max_* members provide information about the maximum number of PV 5400 vcpus, PV guests and PV guest memory size. 5401 5402 :: 5403 5404 struct kvm_s390_pv_info_vm { 5405 __u64 inst_calls_list[4]; 5406 __u64 max_cpus; 5407 __u64 max_guests; 5408 __u64 max_guest_addr; 5409 __u64 feature_indication; 5410 }; 5411 5412 5413 KVM_PV_INFO_DUMP 5414 This subcommand provides information related to dumping PV guests. 5415 5416 :: 5417 5418 struct kvm_s390_pv_info_dump { 5419 __u64 dump_cpu_buffer_len; 5420 __u64 dump_config_mem_buffer_per_1m; 5421 __u64 dump_config_finalize_len; 5422 }; 5423 5424KVM_PV_DUMP 5425 :Capability: KVM_CAP_S390_PROTECTED_DUMP 5426 5427 Presents an API that provides calls which facilitate dumping a 5428 protected VM. 5429 5430 :: 5431 5432 struct kvm_s390_pv_dmp { 5433 __u64 subcmd; 5434 __u64 buff_addr; 5435 __u64 buff_len; 5436 __u64 gaddr; /* For dump storage state */ 5437 }; 5438 5439 **subcommands:** 5440 5441 KVM_PV_DUMP_INIT 5442 Initializes the dump process of a protected VM. If this call does 5443 not succeed all other subcommands will fail with -EINVAL. This 5444 subcommand will return -EINVAL if a dump process has not yet been 5445 completed. 5446 5447 Not all PV vms can be dumped, the owner needs to set `dump 5448 allowed` PCF bit 34 in the SE header to allow dumping. 5449 5450 KVM_PV_DUMP_CONFIG_STOR_STATE 5451 Stores `buff_len` bytes of tweak component values starting with 5452 the 1MB block specified by the absolute guest address 5453 (`gaddr`). `buff_len` needs to be `conf_dump_storage_state_len` 5454 aligned and at least >= the `conf_dump_storage_state_len` value 5455 provided by the dump uv_info data. buff_user might be written to 5456 even if an error rc is returned. For instance if we encounter a 5457 fault after writing the first page of data. 5458 5459 KVM_PV_DUMP_COMPLETE 5460 If the subcommand succeeds it completes the dump process and lets 5461 KVM_PV_DUMP_INIT be called again. 5462 5463 On success `conf_dump_finalize_len` bytes of completion data will be 5464 stored to the `buff_addr`. The completion data contains a key 5465 derivation seed, IV, tweak nonce and encryption keys as well as an 5466 authentication tag all of which are needed to decrypt the dump at a 5467 later time. 5468 5469KVM_PV_ASYNC_CLEANUP_PREPARE 5470 :Capability: KVM_CAP_S390_PROTECTED_ASYNC_DISABLE 5471 5472 Prepare the current protected VM for asynchronous teardown. Most 5473 resources used by the current protected VM will be set aside for a 5474 subsequent asynchronous teardown. The current protected VM will then 5475 resume execution immediately as non-protected. There can be at most 5476 one protected VM prepared for asynchronous teardown at any time. If 5477 a protected VM had already been prepared for teardown without 5478 subsequently calling KVM_PV_ASYNC_CLEANUP_PERFORM, this call will 5479 fail. In that case, the userspace process should issue a normal 5480 KVM_PV_DISABLE. The resources set aside with this call will need to 5481 be cleaned up with a subsequent call to KVM_PV_ASYNC_CLEANUP_PERFORM 5482 or KVM_PV_DISABLE, otherwise they will be cleaned up when KVM 5483 terminates. KVM_PV_ASYNC_CLEANUP_PREPARE can be called again as soon 5484 as cleanup starts, i.e. before KVM_PV_ASYNC_CLEANUP_PERFORM finishes. 5485 5486KVM_PV_ASYNC_CLEANUP_PERFORM 5487 :Capability: KVM_CAP_S390_PROTECTED_ASYNC_DISABLE 5488 5489 Tear down the protected VM previously prepared for teardown with 5490 KVM_PV_ASYNC_CLEANUP_PREPARE. The resources that had been set aside 5491 will be freed during the execution of this command. This PV command 5492 should ideally be issued by userspace from a separate thread. If a 5493 fatal signal is received (or the process terminates naturally), the 5494 command will terminate immediately without completing, and the normal 5495 KVM shutdown procedure will take care of cleaning up all remaining 5496 protected VMs, including the ones whose teardown was interrupted by 5497 process termination. 5498 54994.126 KVM_XEN_HVM_SET_ATTR 5500-------------------------- 5501 5502:Capability: KVM_CAP_XEN_HVM / KVM_XEN_HVM_CONFIG_SHARED_INFO 5503:Architectures: x86 5504:Type: vm ioctl 5505:Parameters: struct kvm_xen_hvm_attr 5506:Returns: 0 on success, < 0 on error 5507 5508:: 5509 5510 struct kvm_xen_hvm_attr { 5511 __u16 type; 5512 __u16 pad[3]; 5513 union { 5514 __u8 long_mode; 5515 __u8 vector; 5516 __u8 runstate_update_flag; 5517 union { 5518 __u64 gfn; 5519 __u64 hva; 5520 } shared_info; 5521 struct { 5522 __u32 send_port; 5523 __u32 type; /* EVTCHNSTAT_ipi / EVTCHNSTAT_interdomain */ 5524 __u32 flags; 5525 union { 5526 struct { 5527 __u32 port; 5528 __u32 vcpu; 5529 __u32 priority; 5530 } port; 5531 struct { 5532 __u32 port; /* Zero for eventfd */ 5533 __s32 fd; 5534 } eventfd; 5535 __u32 padding[4]; 5536 } deliver; 5537 } evtchn; 5538 __u32 xen_version; 5539 __u64 pad[8]; 5540 } u; 5541 }; 5542 5543type values: 5544 5545KVM_XEN_ATTR_TYPE_LONG_MODE 5546 Sets the ABI mode of the VM to 32-bit or 64-bit (long mode). This 5547 determines the layout of the shared_info page exposed to the VM. 5548 5549KVM_XEN_ATTR_TYPE_SHARED_INFO 5550 Sets the guest physical frame number at which the Xen shared_info 5551 page resides. Note that although Xen places vcpu_info for the first 5552 32 vCPUs in the shared_info page, KVM does not automatically do so 5553 and instead requires that KVM_XEN_VCPU_ATTR_TYPE_VCPU_INFO or 5554 KVM_XEN_VCPU_ATTR_TYPE_VCPU_INFO_HVA be used explicitly even when 5555 the vcpu_info for a given vCPU resides at the "default" location 5556 in the shared_info page. This is because KVM may not be aware of 5557 the Xen CPU id which is used as the index into the vcpu_info[] 5558 array, so may know the correct default location. 5559 5560 Note that the shared_info page may be constantly written to by KVM; 5561 it contains the event channel bitmap used to deliver interrupts to 5562 a Xen guest, amongst other things. It is exempt from dirty tracking 5563 mechanisms — KVM will not explicitly mark the page as dirty each 5564 time an event channel interrupt is delivered to the guest! Thus, 5565 userspace should always assume that the designated GFN is dirty if 5566 any vCPU has been running or any event channel interrupts can be 5567 routed to the guest. 5568 5569 Setting the gfn to KVM_XEN_INVALID_GFN will disable the shared_info 5570 page. 5571 5572KVM_XEN_ATTR_TYPE_SHARED_INFO_HVA 5573 If the KVM_XEN_HVM_CONFIG_SHARED_INFO_HVA flag is also set in the 5574 Xen capabilities, then this attribute may be used to set the 5575 userspace address at which the shared_info page resides, which 5576 will always be fixed in the VMM regardless of where it is mapped 5577 in guest physical address space. This attribute should be used in 5578 preference to KVM_XEN_ATTR_TYPE_SHARED_INFO as it avoids 5579 unnecessary invalidation of an internal cache when the page is 5580 re-mapped in guest physcial address space. 5581 5582 Setting the hva to zero will disable the shared_info page. 5583 5584KVM_XEN_ATTR_TYPE_UPCALL_VECTOR 5585 Sets the exception vector used to deliver Xen event channel upcalls. 5586 This is the HVM-wide vector injected directly by the hypervisor 5587 (not through the local APIC), typically configured by a guest via 5588 HVM_PARAM_CALLBACK_IRQ. This can be disabled again (e.g. for guest 5589 SHUTDOWN_soft_reset) by setting it to zero. 5590 5591KVM_XEN_ATTR_TYPE_EVTCHN 5592 This attribute is available when the KVM_CAP_XEN_HVM ioctl indicates 5593 support for KVM_XEN_HVM_CONFIG_EVTCHN_SEND features. It configures 5594 an outbound port number for interception of EVTCHNOP_send requests 5595 from the guest. A given sending port number may be directed back to 5596 a specified vCPU (by APIC ID) / port / priority on the guest, or to 5597 trigger events on an eventfd. The vCPU and priority can be changed 5598 by setting KVM_XEN_EVTCHN_UPDATE in a subsequent call, but other 5599 fields cannot change for a given sending port. A port mapping is 5600 removed by using KVM_XEN_EVTCHN_DEASSIGN in the flags field. Passing 5601 KVM_XEN_EVTCHN_RESET in the flags field removes all interception of 5602 outbound event channels. The values of the flags field are mutually 5603 exclusive and cannot be combined as a bitmask. 5604 5605KVM_XEN_ATTR_TYPE_XEN_VERSION 5606 This attribute is available when the KVM_CAP_XEN_HVM ioctl indicates 5607 support for KVM_XEN_HVM_CONFIG_EVTCHN_SEND features. It configures 5608 the 32-bit version code returned to the guest when it invokes the 5609 XENVER_version call; typically (XEN_MAJOR << 16 | XEN_MINOR). PV 5610 Xen guests will often use this to as a dummy hypercall to trigger 5611 event channel delivery, so responding within the kernel without 5612 exiting to userspace is beneficial. 5613 5614KVM_XEN_ATTR_TYPE_RUNSTATE_UPDATE_FLAG 5615 This attribute is available when the KVM_CAP_XEN_HVM ioctl indicates 5616 support for KVM_XEN_HVM_CONFIG_RUNSTATE_UPDATE_FLAG. It enables the 5617 XEN_RUNSTATE_UPDATE flag which allows guest vCPUs to safely read 5618 other vCPUs' vcpu_runstate_info. Xen guests enable this feature via 5619 the VMASST_TYPE_runstate_update_flag of the HYPERVISOR_vm_assist 5620 hypercall. 5621 56224.127 KVM_XEN_HVM_GET_ATTR 5623-------------------------- 5624 5625:Capability: KVM_CAP_XEN_HVM / KVM_XEN_HVM_CONFIG_SHARED_INFO 5626:Architectures: x86 5627:Type: vm ioctl 5628:Parameters: struct kvm_xen_hvm_attr 5629:Returns: 0 on success, < 0 on error 5630 5631Allows Xen VM attributes to be read. For the structure and types, 5632see KVM_XEN_HVM_SET_ATTR above. The KVM_XEN_ATTR_TYPE_EVTCHN 5633attribute cannot be read. 5634 56354.128 KVM_XEN_VCPU_SET_ATTR 5636--------------------------- 5637 5638:Capability: KVM_CAP_XEN_HVM / KVM_XEN_HVM_CONFIG_SHARED_INFO 5639:Architectures: x86 5640:Type: vcpu ioctl 5641:Parameters: struct kvm_xen_vcpu_attr 5642:Returns: 0 on success, < 0 on error 5643 5644:: 5645 5646 struct kvm_xen_vcpu_attr { 5647 __u16 type; 5648 __u16 pad[3]; 5649 union { 5650 __u64 gpa; 5651 __u64 pad[4]; 5652 struct { 5653 __u64 state; 5654 __u64 state_entry_time; 5655 __u64 time_running; 5656 __u64 time_runnable; 5657 __u64 time_blocked; 5658 __u64 time_offline; 5659 } runstate; 5660 __u32 vcpu_id; 5661 struct { 5662 __u32 port; 5663 __u32 priority; 5664 __u64 expires_ns; 5665 } timer; 5666 __u8 vector; 5667 } u; 5668 }; 5669 5670type values: 5671 5672KVM_XEN_VCPU_ATTR_TYPE_VCPU_INFO 5673 Sets the guest physical address of the vcpu_info for a given vCPU. 5674 As with the shared_info page for the VM, the corresponding page may be 5675 dirtied at any time if event channel interrupt delivery is enabled, so 5676 userspace should always assume that the page is dirty without relying 5677 on dirty logging. Setting the gpa to KVM_XEN_INVALID_GPA will disable 5678 the vcpu_info. 5679 5680KVM_XEN_VCPU_ATTR_TYPE_VCPU_INFO_HVA 5681 If the KVM_XEN_HVM_CONFIG_SHARED_INFO_HVA flag is also set in the 5682 Xen capabilities, then this attribute may be used to set the 5683 userspace address of the vcpu_info for a given vCPU. It should 5684 only be used when the vcpu_info resides at the "default" location 5685 in the shared_info page. In this case it is safe to assume the 5686 userspace address will not change, because the shared_info page is 5687 an overlay on guest memory and remains at a fixed host address 5688 regardless of where it is mapped in guest physical address space 5689 and hence unnecessary invalidation of an internal cache may be 5690 avoided if the guest memory layout is modified. 5691 If the vcpu_info does not reside at the "default" location then 5692 it is not guaranteed to remain at the same host address and 5693 hence the aforementioned cache invalidation is required. 5694 5695KVM_XEN_VCPU_ATTR_TYPE_VCPU_TIME_INFO 5696 Sets the guest physical address of an additional pvclock structure 5697 for a given vCPU. This is typically used for guest vsyscall support. 5698 Setting the gpa to KVM_XEN_INVALID_GPA will disable the structure. 5699 5700KVM_XEN_VCPU_ATTR_TYPE_RUNSTATE_ADDR 5701 Sets the guest physical address of the vcpu_runstate_info for a given 5702 vCPU. This is how a Xen guest tracks CPU state such as steal time. 5703 Setting the gpa to KVM_XEN_INVALID_GPA will disable the runstate area. 5704 5705KVM_XEN_VCPU_ATTR_TYPE_RUNSTATE_CURRENT 5706 Sets the runstate (RUNSTATE_running/_runnable/_blocked/_offline) of 5707 the given vCPU from the .u.runstate.state member of the structure. 5708 KVM automatically accounts running and runnable time but blocked 5709 and offline states are only entered explicitly. 5710 5711KVM_XEN_VCPU_ATTR_TYPE_RUNSTATE_DATA 5712 Sets all fields of the vCPU runstate data from the .u.runstate member 5713 of the structure, including the current runstate. The state_entry_time 5714 must equal the sum of the other four times. 5715 5716KVM_XEN_VCPU_ATTR_TYPE_RUNSTATE_ADJUST 5717 This *adds* the contents of the .u.runstate members of the structure 5718 to the corresponding members of the given vCPU's runstate data, thus 5719 permitting atomic adjustments to the runstate times. The adjustment 5720 to the state_entry_time must equal the sum of the adjustments to the 5721 other four times. The state field must be set to -1, or to a valid 5722 runstate value (RUNSTATE_running, RUNSTATE_runnable, RUNSTATE_blocked 5723 or RUNSTATE_offline) to set the current accounted state as of the 5724 adjusted state_entry_time. 5725 5726KVM_XEN_VCPU_ATTR_TYPE_VCPU_ID 5727 This attribute is available when the KVM_CAP_XEN_HVM ioctl indicates 5728 support for KVM_XEN_HVM_CONFIG_EVTCHN_SEND features. It sets the Xen 5729 vCPU ID of the given vCPU, to allow timer-related VCPU operations to 5730 be intercepted by KVM. 5731 5732KVM_XEN_VCPU_ATTR_TYPE_TIMER 5733 This attribute is available when the KVM_CAP_XEN_HVM ioctl indicates 5734 support for KVM_XEN_HVM_CONFIG_EVTCHN_SEND features. It sets the 5735 event channel port/priority for the VIRQ_TIMER of the vCPU, as well 5736 as allowing a pending timer to be saved/restored. Setting the timer 5737 port to zero disables kernel handling of the singleshot timer. 5738 5739KVM_XEN_VCPU_ATTR_TYPE_UPCALL_VECTOR 5740 This attribute is available when the KVM_CAP_XEN_HVM ioctl indicates 5741 support for KVM_XEN_HVM_CONFIG_EVTCHN_SEND features. It sets the 5742 per-vCPU local APIC upcall vector, configured by a Xen guest with 5743 the HVMOP_set_evtchn_upcall_vector hypercall. This is typically 5744 used by Windows guests, and is distinct from the HVM-wide upcall 5745 vector configured with HVM_PARAM_CALLBACK_IRQ. It is disabled by 5746 setting the vector to zero. 5747 5748 57494.129 KVM_XEN_VCPU_GET_ATTR 5750--------------------------- 5751 5752:Capability: KVM_CAP_XEN_HVM / KVM_XEN_HVM_CONFIG_SHARED_INFO 5753:Architectures: x86 5754:Type: vcpu ioctl 5755:Parameters: struct kvm_xen_vcpu_attr 5756:Returns: 0 on success, < 0 on error 5757 5758Allows Xen vCPU attributes to be read. For the structure and types, 5759see KVM_XEN_VCPU_SET_ATTR above. 5760 5761The KVM_XEN_VCPU_ATTR_TYPE_RUNSTATE_ADJUST type may not be used 5762with the KVM_XEN_VCPU_GET_ATTR ioctl. 5763 57644.130 KVM_ARM_MTE_COPY_TAGS 5765--------------------------- 5766 5767:Capability: KVM_CAP_ARM_MTE 5768:Architectures: arm64 5769:Type: vm ioctl 5770:Parameters: struct kvm_arm_copy_mte_tags 5771:Returns: number of bytes copied, < 0 on error (-EINVAL for incorrect 5772 arguments, -EFAULT if memory cannot be accessed). 5773 5774:: 5775 5776 struct kvm_arm_copy_mte_tags { 5777 __u64 guest_ipa; 5778 __u64 length; 5779 void __user *addr; 5780 __u64 flags; 5781 __u64 reserved[2]; 5782 }; 5783 5784Copies Memory Tagging Extension (MTE) tags to/from guest tag memory. The 5785``guest_ipa`` and ``length`` fields must be ``PAGE_SIZE`` aligned. 5786``length`` must not be bigger than 2^31 - PAGE_SIZE bytes. The ``addr`` 5787field must point to a buffer which the tags will be copied to or from. 5788 5789``flags`` specifies the direction of copy, either ``KVM_ARM_TAGS_TO_GUEST`` or 5790``KVM_ARM_TAGS_FROM_GUEST``. 5791 5792The size of the buffer to store the tags is ``(length / 16)`` bytes 5793(granules in MTE are 16 bytes long). Each byte contains a single tag 5794value. This matches the format of ``PTRACE_PEEKMTETAGS`` and 5795``PTRACE_POKEMTETAGS``. 5796 5797If an error occurs before any data is copied then a negative error code is 5798returned. If some tags have been copied before an error occurs then the number 5799of bytes successfully copied is returned. If the call completes successfully 5800then ``length`` is returned. 5801 58024.131 KVM_GET_SREGS2 5803-------------------- 5804 5805:Capability: KVM_CAP_SREGS2 5806:Architectures: x86 5807:Type: vcpu ioctl 5808:Parameters: struct kvm_sregs2 (out) 5809:Returns: 0 on success, -1 on error 5810 5811Reads special registers from the vcpu. 5812This ioctl (when supported) replaces the KVM_GET_SREGS. 5813 5814:: 5815 5816 struct kvm_sregs2 { 5817 /* out (KVM_GET_SREGS2) / in (KVM_SET_SREGS2) */ 5818 struct kvm_segment cs, ds, es, fs, gs, ss; 5819 struct kvm_segment tr, ldt; 5820 struct kvm_dtable gdt, idt; 5821 __u64 cr0, cr2, cr3, cr4, cr8; 5822 __u64 efer; 5823 __u64 apic_base; 5824 __u64 flags; 5825 __u64 pdptrs[4]; 5826 }; 5827 5828flags values for ``kvm_sregs2``: 5829 5830``KVM_SREGS2_FLAGS_PDPTRS_VALID`` 5831 5832 Indicates that the struct contains valid PDPTR values. 5833 5834 58354.132 KVM_SET_SREGS2 5836-------------------- 5837 5838:Capability: KVM_CAP_SREGS2 5839:Architectures: x86 5840:Type: vcpu ioctl 5841:Parameters: struct kvm_sregs2 (in) 5842:Returns: 0 on success, -1 on error 5843 5844Writes special registers into the vcpu. 5845See KVM_GET_SREGS2 for the data structures. 5846This ioctl (when supported) replaces the KVM_SET_SREGS. 5847 58484.133 KVM_GET_STATS_FD 5849---------------------- 5850 5851:Capability: KVM_CAP_STATS_BINARY_FD 5852:Architectures: all 5853:Type: vm ioctl, vcpu ioctl 5854:Parameters: none 5855:Returns: statistics file descriptor on success, < 0 on error 5856 5857Errors: 5858 5859 ====== ====================================================== 5860 ENOMEM if the fd could not be created due to lack of memory 5861 EMFILE if the number of opened files exceeds the limit 5862 ====== ====================================================== 5863 5864The returned file descriptor can be used to read VM/vCPU statistics data in 5865binary format. The data in the file descriptor consists of four blocks 5866organized as follows: 5867 5868+-------------+ 5869| Header | 5870+-------------+ 5871| id string | 5872+-------------+ 5873| Descriptors | 5874+-------------+ 5875| Stats Data | 5876+-------------+ 5877 5878Apart from the header starting at offset 0, please be aware that it is 5879not guaranteed that the four blocks are adjacent or in the above order; 5880the offsets of the id, descriptors and data blocks are found in the 5881header. However, all four blocks are aligned to 64 bit offsets in the 5882file and they do not overlap. 5883 5884All blocks except the data block are immutable. Userspace can read them 5885only one time after retrieving the file descriptor, and then use ``pread`` or 5886``lseek`` to read the statistics repeatedly. 5887 5888All data is in system endianness. 5889 5890The format of the header is as follows:: 5891 5892 struct kvm_stats_header { 5893 __u32 flags; 5894 __u32 name_size; 5895 __u32 num_desc; 5896 __u32 id_offset; 5897 __u32 desc_offset; 5898 __u32 data_offset; 5899 }; 5900 5901The ``flags`` field is not used at the moment. It is always read as 0. 5902 5903The ``name_size`` field is the size (in byte) of the statistics name string 5904(including trailing '\0') which is contained in the "id string" block and 5905appended at the end of every descriptor. 5906 5907The ``num_desc`` field is the number of descriptors that are included in the 5908descriptor block. (The actual number of values in the data block may be 5909larger, since each descriptor may comprise more than one value). 5910 5911The ``id_offset`` field is the offset of the id string from the start of the 5912file indicated by the file descriptor. It is a multiple of 8. 5913 5914The ``desc_offset`` field is the offset of the Descriptors block from the start 5915of the file indicated by the file descriptor. It is a multiple of 8. 5916 5917The ``data_offset`` field is the offset of the Stats Data block from the start 5918of the file indicated by the file descriptor. It is a multiple of 8. 5919 5920The id string block contains a string which identifies the file descriptor on 5921which KVM_GET_STATS_FD was invoked. The size of the block, including the 5922trailing ``'\0'``, is indicated by the ``name_size`` field in the header. 5923 5924The descriptors block is only needed to be read once for the lifetime of the 5925file descriptor contains a sequence of ``struct kvm_stats_desc``, each followed 5926by a string of size ``name_size``. 5927:: 5928 5929 #define KVM_STATS_TYPE_SHIFT 0 5930 #define KVM_STATS_TYPE_MASK (0xF << KVM_STATS_TYPE_SHIFT) 5931 #define KVM_STATS_TYPE_CUMULATIVE (0x0 << KVM_STATS_TYPE_SHIFT) 5932 #define KVM_STATS_TYPE_INSTANT (0x1 << KVM_STATS_TYPE_SHIFT) 5933 #define KVM_STATS_TYPE_PEAK (0x2 << KVM_STATS_TYPE_SHIFT) 5934 #define KVM_STATS_TYPE_LINEAR_HIST (0x3 << KVM_STATS_TYPE_SHIFT) 5935 #define KVM_STATS_TYPE_LOG_HIST (0x4 << KVM_STATS_TYPE_SHIFT) 5936 #define KVM_STATS_TYPE_MAX KVM_STATS_TYPE_LOG_HIST 5937 5938 #define KVM_STATS_UNIT_SHIFT 4 5939 #define KVM_STATS_UNIT_MASK (0xF << KVM_STATS_UNIT_SHIFT) 5940 #define KVM_STATS_UNIT_NONE (0x0 << KVM_STATS_UNIT_SHIFT) 5941 #define KVM_STATS_UNIT_BYTES (0x1 << KVM_STATS_UNIT_SHIFT) 5942 #define KVM_STATS_UNIT_SECONDS (0x2 << KVM_STATS_UNIT_SHIFT) 5943 #define KVM_STATS_UNIT_CYCLES (0x3 << KVM_STATS_UNIT_SHIFT) 5944 #define KVM_STATS_UNIT_BOOLEAN (0x4 << KVM_STATS_UNIT_SHIFT) 5945 #define KVM_STATS_UNIT_MAX KVM_STATS_UNIT_BOOLEAN 5946 5947 #define KVM_STATS_BASE_SHIFT 8 5948 #define KVM_STATS_BASE_MASK (0xF << KVM_STATS_BASE_SHIFT) 5949 #define KVM_STATS_BASE_POW10 (0x0 << KVM_STATS_BASE_SHIFT) 5950 #define KVM_STATS_BASE_POW2 (0x1 << KVM_STATS_BASE_SHIFT) 5951 #define KVM_STATS_BASE_MAX KVM_STATS_BASE_POW2 5952 5953 struct kvm_stats_desc { 5954 __u32 flags; 5955 __s16 exponent; 5956 __u16 size; 5957 __u32 offset; 5958 __u32 bucket_size; 5959 char name[]; 5960 }; 5961 5962The ``flags`` field contains the type and unit of the statistics data described 5963by this descriptor. Its endianness is CPU native. 5964The following flags are supported: 5965 5966Bits 0-3 of ``flags`` encode the type: 5967 5968 * ``KVM_STATS_TYPE_CUMULATIVE`` 5969 The statistics reports a cumulative count. The value of data can only be increased. 5970 Most of the counters used in KVM are of this type. 5971 The corresponding ``size`` field for this type is always 1. 5972 All cumulative statistics data are read/write. 5973 * ``KVM_STATS_TYPE_INSTANT`` 5974 The statistics reports an instantaneous value. Its value can be increased or 5975 decreased. This type is usually used as a measurement of some resources, 5976 like the number of dirty pages, the number of large pages, etc. 5977 All instant statistics are read only. 5978 The corresponding ``size`` field for this type is always 1. 5979 * ``KVM_STATS_TYPE_PEAK`` 5980 The statistics data reports a peak value, for example the maximum number 5981 of items in a hash table bucket, the longest time waited and so on. 5982 The value of data can only be increased. 5983 The corresponding ``size`` field for this type is always 1. 5984 * ``KVM_STATS_TYPE_LINEAR_HIST`` 5985 The statistic is reported as a linear histogram. The number of 5986 buckets is specified by the ``size`` field. The size of buckets is specified 5987 by the ``hist_param`` field. The range of the Nth bucket (1 <= N < ``size``) 5988 is [``hist_param``*(N-1), ``hist_param``*N), while the range of the last 5989 bucket is [``hist_param``*(``size``-1), +INF). (+INF means positive infinity 5990 value.) 5991 * ``KVM_STATS_TYPE_LOG_HIST`` 5992 The statistic is reported as a logarithmic histogram. The number of 5993 buckets is specified by the ``size`` field. The range of the first bucket is 5994 [0, 1), while the range of the last bucket is [pow(2, ``size``-2), +INF). 5995 Otherwise, The Nth bucket (1 < N < ``size``) covers 5996 [pow(2, N-2), pow(2, N-1)). 5997 5998Bits 4-7 of ``flags`` encode the unit: 5999 6000 * ``KVM_STATS_UNIT_NONE`` 6001 There is no unit for the value of statistics data. This usually means that 6002 the value is a simple counter of an event. 6003 * ``KVM_STATS_UNIT_BYTES`` 6004 It indicates that the statistics data is used to measure memory size, in the 6005 unit of Byte, KiByte, MiByte, GiByte, etc. The unit of the data is 6006 determined by the ``exponent`` field in the descriptor. 6007 * ``KVM_STATS_UNIT_SECONDS`` 6008 It indicates that the statistics data is used to measure time or latency. 6009 * ``KVM_STATS_UNIT_CYCLES`` 6010 It indicates that the statistics data is used to measure CPU clock cycles. 6011 * ``KVM_STATS_UNIT_BOOLEAN`` 6012 It indicates that the statistic will always be either 0 or 1. Boolean 6013 statistics of "peak" type will never go back from 1 to 0. Boolean 6014 statistics can be linear histograms (with two buckets) but not logarithmic 6015 histograms. 6016 6017Note that, in the case of histograms, the unit applies to the bucket 6018ranges, while the bucket value indicates how many samples fell in the 6019bucket's range. 6020 6021Bits 8-11 of ``flags``, together with ``exponent``, encode the scale of the 6022unit: 6023 6024 * ``KVM_STATS_BASE_POW10`` 6025 The scale is based on power of 10. It is used for measurement of time and 6026 CPU clock cycles. For example, an exponent of -9 can be used with 6027 ``KVM_STATS_UNIT_SECONDS`` to express that the unit is nanoseconds. 6028 * ``KVM_STATS_BASE_POW2`` 6029 The scale is based on power of 2. It is used for measurement of memory size. 6030 For example, an exponent of 20 can be used with ``KVM_STATS_UNIT_BYTES`` to 6031 express that the unit is MiB. 6032 6033The ``size`` field is the number of values of this statistics data. Its 6034value is usually 1 for most of simple statistics. 1 means it contains an 6035unsigned 64bit data. 6036 6037The ``offset`` field is the offset from the start of Data Block to the start of 6038the corresponding statistics data. 6039 6040The ``bucket_size`` field is used as a parameter for histogram statistics data. 6041It is only used by linear histogram statistics data, specifying the size of a 6042bucket in the unit expressed by bits 4-11 of ``flags`` together with ``exponent``. 6043 6044The ``name`` field is the name string of the statistics data. The name string 6045starts at the end of ``struct kvm_stats_desc``. The maximum length including 6046the trailing ``'\0'``, is indicated by ``name_size`` in the header. 6047 6048The Stats Data block contains an array of 64-bit values in the same order 6049as the descriptors in Descriptors block. 6050 60514.134 KVM_GET_XSAVE2 6052-------------------- 6053 6054:Capability: KVM_CAP_XSAVE2 6055:Architectures: x86 6056:Type: vcpu ioctl 6057:Parameters: struct kvm_xsave (out) 6058:Returns: 0 on success, -1 on error 6059 6060 6061:: 6062 6063 struct kvm_xsave { 6064 __u32 region[1024]; 6065 __u32 extra[0]; 6066 }; 6067 6068This ioctl would copy current vcpu's xsave struct to the userspace. It 6069copies as many bytes as are returned by KVM_CHECK_EXTENSION(KVM_CAP_XSAVE2) 6070when invoked on the vm file descriptor. The size value returned by 6071KVM_CHECK_EXTENSION(KVM_CAP_XSAVE2) will always be at least 4096. 6072Currently, it is only greater than 4096 if a dynamic feature has been 6073enabled with ``arch_prctl()``, but this may change in the future. 6074 6075The offsets of the state save areas in struct kvm_xsave follow the contents 6076of CPUID leaf 0xD on the host. 6077 60784.135 KVM_XEN_HVM_EVTCHN_SEND 6079----------------------------- 6080 6081:Capability: KVM_CAP_XEN_HVM / KVM_XEN_HVM_CONFIG_EVTCHN_SEND 6082:Architectures: x86 6083:Type: vm ioctl 6084:Parameters: struct kvm_irq_routing_xen_evtchn 6085:Returns: 0 on success, < 0 on error 6086 6087 6088:: 6089 6090 struct kvm_irq_routing_xen_evtchn { 6091 __u32 port; 6092 __u32 vcpu; 6093 __u32 priority; 6094 }; 6095 6096This ioctl injects an event channel interrupt directly to the guest vCPU. 6097 60984.136 KVM_S390_PV_CPU_COMMAND 6099----------------------------- 6100 6101:Capability: KVM_CAP_S390_PROTECTED_DUMP 6102:Architectures: s390 6103:Type: vcpu ioctl 6104:Parameters: none 6105:Returns: 0 on success, < 0 on error 6106 6107This ioctl closely mirrors `KVM_S390_PV_COMMAND` but handles requests 6108for vcpus. It re-uses the kvm_s390_pv_dmp struct and hence also shares 6109the command ids. 6110 6111**command:** 6112 6113KVM_PV_DUMP 6114 Presents an API that provides calls which facilitate dumping a vcpu 6115 of a protected VM. 6116 6117**subcommand:** 6118 6119KVM_PV_DUMP_CPU 6120 Provides encrypted dump data like register values. 6121 The length of the returned data is provided by uv_info.guest_cpu_stor_len. 6122 61234.137 KVM_S390_ZPCI_OP 6124---------------------- 6125 6126:Capability: KVM_CAP_S390_ZPCI_OP 6127:Architectures: s390 6128:Type: vm ioctl 6129:Parameters: struct kvm_s390_zpci_op (in) 6130:Returns: 0 on success, <0 on error 6131 6132Used to manage hardware-assisted virtualization features for zPCI devices. 6133 6134Parameters are specified via the following structure:: 6135 6136 struct kvm_s390_zpci_op { 6137 /* in */ 6138 __u32 fh; /* target device */ 6139 __u8 op; /* operation to perform */ 6140 __u8 pad[3]; 6141 union { 6142 /* for KVM_S390_ZPCIOP_REG_AEN */ 6143 struct { 6144 __u64 ibv; /* Guest addr of interrupt bit vector */ 6145 __u64 sb; /* Guest addr of summary bit */ 6146 __u32 flags; 6147 __u32 noi; /* Number of interrupts */ 6148 __u8 isc; /* Guest interrupt subclass */ 6149 __u8 sbo; /* Offset of guest summary bit vector */ 6150 __u16 pad; 6151 } reg_aen; 6152 __u64 reserved[8]; 6153 } u; 6154 }; 6155 6156The type of operation is specified in the "op" field. 6157KVM_S390_ZPCIOP_REG_AEN is used to register the VM for adapter event 6158notification interpretation, which will allow firmware delivery of adapter 6159events directly to the vm, with KVM providing a backup delivery mechanism; 6160KVM_S390_ZPCIOP_DEREG_AEN is used to subsequently disable interpretation of 6161adapter event notifications. 6162 6163The target zPCI function must also be specified via the "fh" field. For the 6164KVM_S390_ZPCIOP_REG_AEN operation, additional information to establish firmware 6165delivery must be provided via the "reg_aen" struct. 6166 6167The "pad" and "reserved" fields may be used for future extensions and should be 6168set to 0s by userspace. 6169 61704.138 KVM_ARM_SET_COUNTER_OFFSET 6171-------------------------------- 6172 6173:Capability: KVM_CAP_COUNTER_OFFSET 6174:Architectures: arm64 6175:Type: vm ioctl 6176:Parameters: struct kvm_arm_counter_offset (in) 6177:Returns: 0 on success, < 0 on error 6178 6179This capability indicates that userspace is able to apply a single VM-wide 6180offset to both the virtual and physical counters as viewed by the guest 6181using the KVM_ARM_SET_CNT_OFFSET ioctl and the following data structure: 6182 6183:: 6184 6185 struct kvm_arm_counter_offset { 6186 __u64 counter_offset; 6187 __u64 reserved; 6188 }; 6189 6190The offset describes a number of counter cycles that are subtracted from 6191both virtual and physical counter views (similar to the effects of the 6192CNTVOFF_EL2 and CNTPOFF_EL2 system registers, but only global). The offset 6193always applies to all vcpus (already created or created after this ioctl) 6194for this VM. 6195 6196It is userspace's responsibility to compute the offset based, for example, 6197on previous values of the guest counters. 6198 6199Any value other than 0 for the "reserved" field may result in an error 6200(-EINVAL) being returned. This ioctl can also return -EBUSY if any vcpu 6201ioctl is issued concurrently. 6202 6203Note that using this ioctl results in KVM ignoring subsequent userspace 6204writes to the CNTVCT_EL0 and CNTPCT_EL0 registers using the SET_ONE_REG 6205interface. No error will be returned, but the resulting offset will not be 6206applied. 6207 6208.. _KVM_ARM_GET_REG_WRITABLE_MASKS: 6209 62104.139 KVM_ARM_GET_REG_WRITABLE_MASKS 6211------------------------------------ 6212 6213:Capability: KVM_CAP_ARM_SUPPORTED_REG_MASK_RANGES 6214:Architectures: arm64 6215:Type: vm ioctl 6216:Parameters: struct reg_mask_range (in/out) 6217:Returns: 0 on success, < 0 on error 6218 6219 6220:: 6221 6222 #define KVM_ARM_FEATURE_ID_RANGE 0 6223 #define KVM_ARM_FEATURE_ID_RANGE_SIZE (3 * 8 * 8) 6224 6225 struct reg_mask_range { 6226 __u64 addr; /* Pointer to mask array */ 6227 __u32 range; /* Requested range */ 6228 __u32 reserved[13]; 6229 }; 6230 6231This ioctl copies the writable masks for a selected range of registers to 6232userspace. 6233 6234The ``addr`` field is a pointer to the destination array where KVM copies 6235the writable masks. 6236 6237The ``range`` field indicates the requested range of registers. 6238``KVM_CHECK_EXTENSION`` for the ``KVM_CAP_ARM_SUPPORTED_REG_MASK_RANGES`` 6239capability returns the supported ranges, expressed as a set of flags. Each 6240flag's bit index represents a possible value for the ``range`` field. 6241All other values are reserved for future use and KVM may return an error. 6242 6243The ``reserved[13]`` array is reserved for future use and should be 0, or 6244KVM may return an error. 6245 6246KVM_ARM_FEATURE_ID_RANGE (0) 6247^^^^^^^^^^^^^^^^^^^^^^^^^^^^ 6248 6249The Feature ID range is defined as the AArch64 System register space with 6250op0==3, op1=={0, 1, 3}, CRn==0, CRm=={0-7}, op2=={0-7}. 6251 6252The mask returned array pointed to by ``addr`` is indexed by the macro 6253``ARM64_FEATURE_ID_RANGE_IDX(op0, op1, crn, crm, op2)``, allowing userspace 6254to know what fields can be changed for the system register described by 6255``op0, op1, crn, crm, op2``. KVM rejects ID register values that describe a 6256superset of the features supported by the system. 6257 62584.140 KVM_SET_USER_MEMORY_REGION2 6259--------------------------------- 6260 6261:Capability: KVM_CAP_USER_MEMORY2 6262:Architectures: all 6263:Type: vm ioctl 6264:Parameters: struct kvm_userspace_memory_region2 (in) 6265:Returns: 0 on success, -1 on error 6266 6267KVM_SET_USER_MEMORY_REGION2 is an extension to KVM_SET_USER_MEMORY_REGION that 6268allows mapping guest_memfd memory into a guest. All fields shared with 6269KVM_SET_USER_MEMORY_REGION identically. Userspace can set KVM_MEM_GUEST_MEMFD 6270in flags to have KVM bind the memory region to a given guest_memfd range of 6271[guest_memfd_offset, guest_memfd_offset + memory_size]. The target guest_memfd 6272must point at a file created via KVM_CREATE_GUEST_MEMFD on the current VM, and 6273the target range must not be bound to any other memory region. All standard 6274bounds checks apply (use common sense). 6275 6276:: 6277 6278 struct kvm_userspace_memory_region2 { 6279 __u32 slot; 6280 __u32 flags; 6281 __u64 guest_phys_addr; 6282 __u64 memory_size; /* bytes */ 6283 __u64 userspace_addr; /* start of the userspace allocated memory */ 6284 __u64 guest_memfd_offset; 6285 __u32 guest_memfd; 6286 __u32 pad1; 6287 __u64 pad2[14]; 6288 }; 6289 6290A KVM_MEM_GUEST_MEMFD region _must_ have a valid guest_memfd (private memory) and 6291userspace_addr (shared memory). However, "valid" for userspace_addr simply 6292means that the address itself must be a legal userspace address. The backing 6293mapping for userspace_addr is not required to be valid/populated at the time of 6294KVM_SET_USER_MEMORY_REGION2, e.g. shared memory can be lazily mapped/allocated 6295on-demand. 6296 6297When mapping a gfn into the guest, KVM selects shared vs. private, i.e consumes 6298userspace_addr vs. guest_memfd, based on the gfn's KVM_MEMORY_ATTRIBUTE_PRIVATE 6299state. At VM creation time, all memory is shared, i.e. the PRIVATE attribute 6300is '0' for all gfns. Userspace can control whether memory is shared/private by 6301toggling KVM_MEMORY_ATTRIBUTE_PRIVATE via KVM_SET_MEMORY_ATTRIBUTES as needed. 6302 6303S390: 6304^^^^^ 6305 6306Returns -EINVAL if the VM has the KVM_VM_S390_UCONTROL flag set. 6307Returns -EINVAL if called on a protected VM. 6308 63094.141 KVM_SET_MEMORY_ATTRIBUTES 6310------------------------------- 6311 6312:Capability: KVM_CAP_MEMORY_ATTRIBUTES 6313:Architectures: x86 6314:Type: vm ioctl 6315:Parameters: struct kvm_memory_attributes (in) 6316:Returns: 0 on success, <0 on error 6317 6318KVM_SET_MEMORY_ATTRIBUTES allows userspace to set memory attributes for a range 6319of guest physical memory. 6320 6321:: 6322 6323 struct kvm_memory_attributes { 6324 __u64 address; 6325 __u64 size; 6326 __u64 attributes; 6327 __u64 flags; 6328 }; 6329 6330 #define KVM_MEMORY_ATTRIBUTE_PRIVATE (1ULL << 3) 6331 6332The address and size must be page aligned. The supported attributes can be 6333retrieved via ioctl(KVM_CHECK_EXTENSION) on KVM_CAP_MEMORY_ATTRIBUTES. If 6334executed on a VM, KVM_CAP_MEMORY_ATTRIBUTES precisely returns the attributes 6335supported by that VM. If executed at system scope, KVM_CAP_MEMORY_ATTRIBUTES 6336returns all attributes supported by KVM. The only attribute defined at this 6337time is KVM_MEMORY_ATTRIBUTE_PRIVATE, which marks the associated gfn as being 6338guest private memory. 6339 6340Note, there is no "get" API. Userspace is responsible for explicitly tracking 6341the state of a gfn/page as needed. 6342 6343The "flags" field is reserved for future extensions and must be '0'. 6344 63454.142 KVM_CREATE_GUEST_MEMFD 6346---------------------------- 6347 6348:Capability: KVM_CAP_GUEST_MEMFD 6349:Architectures: none 6350:Type: vm ioctl 6351:Parameters: struct kvm_create_guest_memfd(in) 6352:Returns: A file descriptor on success, <0 on error 6353 6354KVM_CREATE_GUEST_MEMFD creates an anonymous file and returns a file descriptor 6355that refers to it. guest_memfd files are roughly analogous to files created 6356via memfd_create(), e.g. guest_memfd files live in RAM, have volatile storage, 6357and are automatically released when the last reference is dropped. Unlike 6358"regular" memfd_create() files, guest_memfd files are bound to their owning 6359virtual machine (see below), cannot be mapped, read, or written by userspace, 6360and cannot be resized (guest_memfd files do however support PUNCH_HOLE). 6361 6362:: 6363 6364 struct kvm_create_guest_memfd { 6365 __u64 size; 6366 __u64 flags; 6367 __u64 reserved[6]; 6368 }; 6369 6370Conceptually, the inode backing a guest_memfd file represents physical memory, 6371i.e. is coupled to the virtual machine as a thing, not to a "struct kvm". The 6372file itself, which is bound to a "struct kvm", is that instance's view of the 6373underlying memory, e.g. effectively provides the translation of guest addresses 6374to host memory. This allows for use cases where multiple KVM structures are 6375used to manage a single virtual machine, e.g. when performing intrahost 6376migration of a virtual machine. 6377 6378KVM currently only supports mapping guest_memfd via KVM_SET_USER_MEMORY_REGION2, 6379and more specifically via the guest_memfd and guest_memfd_offset fields in 6380"struct kvm_userspace_memory_region2", where guest_memfd_offset is the offset 6381into the guest_memfd instance. For a given guest_memfd file, there can be at 6382most one mapping per page, i.e. binding multiple memory regions to a single 6383guest_memfd range is not allowed (any number of memory regions can be bound to 6384a single guest_memfd file, but the bound ranges must not overlap). 6385 6386See KVM_SET_USER_MEMORY_REGION2 for additional details. 6387 63884.143 KVM_PRE_FAULT_MEMORY 6389--------------------------- 6390 6391:Capability: KVM_CAP_PRE_FAULT_MEMORY 6392:Architectures: none 6393:Type: vcpu ioctl 6394:Parameters: struct kvm_pre_fault_memory (in/out) 6395:Returns: 0 if at least one page is processed, < 0 on error 6396 6397Errors: 6398 6399 ========== =============================================================== 6400 EINVAL The specified `gpa` and `size` were invalid (e.g. not 6401 page aligned, causes an overflow, or size is zero). 6402 ENOENT The specified `gpa` is outside defined memslots. 6403 EINTR An unmasked signal is pending and no page was processed. 6404 EFAULT The parameter address was invalid. 6405 EOPNOTSUPP Mapping memory for a GPA is unsupported by the 6406 hypervisor, and/or for the current vCPU state/mode. 6407 EIO unexpected error conditions (also causes a WARN) 6408 ========== =============================================================== 6409 6410:: 6411 6412 struct kvm_pre_fault_memory { 6413 /* in/out */ 6414 __u64 gpa; 6415 __u64 size; 6416 /* in */ 6417 __u64 flags; 6418 __u64 padding[5]; 6419 }; 6420 6421KVM_PRE_FAULT_MEMORY populates KVM's stage-2 page tables used to map memory 6422for the current vCPU state. KVM maps memory as if the vCPU generated a 6423stage-2 read page fault, e.g. faults in memory as needed, but doesn't break 6424CoW. However, KVM does not mark any newly created stage-2 PTE as Accessed. 6425 6426In the case of confidential VM types where there is an initial set up of 6427private guest memory before the guest is 'finalized'/measured, this ioctl 6428should only be issued after completing all the necessary setup to put the 6429guest into a 'finalized' state so that the above semantics can be reliably 6430ensured. 6431 6432In some cases, multiple vCPUs might share the page tables. In this 6433case, the ioctl can be called in parallel. 6434 6435When the ioctl returns, the input values are updated to point to the 6436remaining range. If `size` > 0 on return, the caller can just issue 6437the ioctl again with the same `struct kvm_map_memory` argument. 6438 6439Shadow page tables cannot support this ioctl because they 6440are indexed by virtual address or nested guest physical address. 6441Calling this ioctl when the guest is using shadow page tables (for 6442example because it is running a nested guest with nested page tables) 6443will fail with `EOPNOTSUPP` even if `KVM_CHECK_EXTENSION` reports 6444the capability to be present. 6445 6446`flags` must currently be zero. 6447 6448 6449.. _kvm_run: 6450 64515. The kvm_run structure 6452======================== 6453 6454Application code obtains a pointer to the kvm_run structure by 6455mmap()ing a vcpu fd. From that point, application code can control 6456execution by changing fields in kvm_run prior to calling the KVM_RUN 6457ioctl, and obtain information about the reason KVM_RUN returned by 6458looking up structure members. 6459 6460:: 6461 6462 struct kvm_run { 6463 /* in */ 6464 __u8 request_interrupt_window; 6465 6466Request that KVM_RUN return when it becomes possible to inject external 6467interrupts into the guest. Useful in conjunction with KVM_INTERRUPT. 6468 6469:: 6470 6471 __u8 immediate_exit; 6472 6473This field is polled once when KVM_RUN starts; if non-zero, KVM_RUN 6474exits immediately, returning -EINTR. In the common scenario where a 6475signal is used to "kick" a VCPU out of KVM_RUN, this field can be used 6476to avoid usage of KVM_SET_SIGNAL_MASK, which has worse scalability. 6477Rather than blocking the signal outside KVM_RUN, userspace can set up 6478a signal handler that sets run->immediate_exit to a non-zero value. 6479 6480This field is ignored if KVM_CAP_IMMEDIATE_EXIT is not available. 6481 6482:: 6483 6484 __u8 padding1[6]; 6485 6486 /* out */ 6487 __u32 exit_reason; 6488 6489When KVM_RUN has returned successfully (return value 0), this informs 6490application code why KVM_RUN has returned. Allowable values for this 6491field are detailed below. 6492 6493:: 6494 6495 __u8 ready_for_interrupt_injection; 6496 6497If request_interrupt_window has been specified, this field indicates 6498an interrupt can be injected now with KVM_INTERRUPT. 6499 6500:: 6501 6502 __u8 if_flag; 6503 6504The value of the current interrupt flag. Only valid if in-kernel 6505local APIC is not used. 6506 6507:: 6508 6509 __u16 flags; 6510 6511More architecture-specific flags detailing state of the VCPU that may 6512affect the device's behavior. Current defined flags:: 6513 6514 /* x86, set if the VCPU is in system management mode */ 6515 #define KVM_RUN_X86_SMM (1 << 0) 6516 /* x86, set if bus lock detected in VM */ 6517 #define KVM_RUN_X86_BUS_LOCK (1 << 1) 6518 /* x86, set if the VCPU is executing a nested (L2) guest */ 6519 #define KVM_RUN_X86_GUEST_MODE (1 << 2) 6520 6521 /* arm64, set for KVM_EXIT_DEBUG */ 6522 #define KVM_DEBUG_ARCH_HSR_HIGH_VALID (1 << 0) 6523 6524:: 6525 6526 /* in (pre_kvm_run), out (post_kvm_run) */ 6527 __u64 cr8; 6528 6529The value of the cr8 register. Only valid if in-kernel local APIC is 6530not used. Both input and output. 6531 6532:: 6533 6534 __u64 apic_base; 6535 6536The value of the APIC BASE msr. Only valid if in-kernel local 6537APIC is not used. Both input and output. 6538 6539:: 6540 6541 union { 6542 /* KVM_EXIT_UNKNOWN */ 6543 struct { 6544 __u64 hardware_exit_reason; 6545 } hw; 6546 6547If exit_reason is KVM_EXIT_UNKNOWN, the vcpu has exited due to unknown 6548reasons. Further architecture-specific information is available in 6549hardware_exit_reason. 6550 6551:: 6552 6553 /* KVM_EXIT_FAIL_ENTRY */ 6554 struct { 6555 __u64 hardware_entry_failure_reason; 6556 __u32 cpu; /* if KVM_LAST_CPU */ 6557 } fail_entry; 6558 6559If exit_reason is KVM_EXIT_FAIL_ENTRY, the vcpu could not be run due 6560to unknown reasons. Further architecture-specific information is 6561available in hardware_entry_failure_reason. 6562 6563:: 6564 6565 /* KVM_EXIT_EXCEPTION */ 6566 struct { 6567 __u32 exception; 6568 __u32 error_code; 6569 } ex; 6570 6571Unused. 6572 6573:: 6574 6575 /* KVM_EXIT_IO */ 6576 struct { 6577 #define KVM_EXIT_IO_IN 0 6578 #define KVM_EXIT_IO_OUT 1 6579 __u8 direction; 6580 __u8 size; /* bytes */ 6581 __u16 port; 6582 __u32 count; 6583 __u64 data_offset; /* relative to kvm_run start */ 6584 } io; 6585 6586If exit_reason is KVM_EXIT_IO, then the vcpu has 6587executed a port I/O instruction which could not be satisfied by kvm. 6588data_offset describes where the data is located (KVM_EXIT_IO_OUT) or 6589where kvm expects application code to place the data for the next 6590KVM_RUN invocation (KVM_EXIT_IO_IN). Data format is a packed array. 6591 6592:: 6593 6594 /* KVM_EXIT_DEBUG */ 6595 struct { 6596 struct kvm_debug_exit_arch arch; 6597 } debug; 6598 6599If the exit_reason is KVM_EXIT_DEBUG, then a vcpu is processing a debug event 6600for which architecture specific information is returned. 6601 6602:: 6603 6604 /* KVM_EXIT_MMIO */ 6605 struct { 6606 __u64 phys_addr; 6607 __u8 data[8]; 6608 __u32 len; 6609 __u8 is_write; 6610 } mmio; 6611 6612If exit_reason is KVM_EXIT_MMIO, then the vcpu has 6613executed a memory-mapped I/O instruction which could not be satisfied 6614by kvm. The 'data' member contains the written data if 'is_write' is 6615true, and should be filled by application code otherwise. 6616 6617The 'data' member contains, in its first 'len' bytes, the value as it would 6618appear if the VCPU performed a load or store of the appropriate width directly 6619to the byte array. 6620 6621.. note:: 6622 6623 For KVM_EXIT_IO, KVM_EXIT_MMIO, KVM_EXIT_OSI, KVM_EXIT_PAPR, KVM_EXIT_XEN, 6624 KVM_EXIT_EPR, KVM_EXIT_X86_RDMSR and KVM_EXIT_X86_WRMSR the corresponding 6625 operations are complete (and guest state is consistent) only after userspace 6626 has re-entered the kernel with KVM_RUN. The kernel side will first finish 6627 incomplete operations and then check for pending signals. 6628 6629 The pending state of the operation is not preserved in state which is 6630 visible to userspace, thus userspace should ensure that the operation is 6631 completed before performing a live migration. Userspace can re-enter the 6632 guest with an unmasked signal pending or with the immediate_exit field set 6633 to complete pending operations without allowing any further instructions 6634 to be executed. 6635 6636:: 6637 6638 /* KVM_EXIT_HYPERCALL */ 6639 struct { 6640 __u64 nr; 6641 __u64 args[6]; 6642 __u64 ret; 6643 __u64 flags; 6644 } hypercall; 6645 6646 6647It is strongly recommended that userspace use ``KVM_EXIT_IO`` (x86) or 6648``KVM_EXIT_MMIO`` (all except s390) to implement functionality that 6649requires a guest to interact with host userspace. 6650 6651.. note:: KVM_EXIT_IO is significantly faster than KVM_EXIT_MMIO. 6652 6653For arm64: 6654---------- 6655 6656SMCCC exits can be enabled depending on the configuration of the SMCCC 6657filter. See the Documentation/virt/kvm/devices/vm.rst 6658``KVM_ARM_SMCCC_FILTER`` for more details. 6659 6660``nr`` contains the function ID of the guest's SMCCC call. Userspace is 6661expected to use the ``KVM_GET_ONE_REG`` ioctl to retrieve the call 6662parameters from the vCPU's GPRs. 6663 6664Definition of ``flags``: 6665 - ``KVM_HYPERCALL_EXIT_SMC``: Indicates that the guest used the SMC 6666 conduit to initiate the SMCCC call. If this bit is 0 then the guest 6667 used the HVC conduit for the SMCCC call. 6668 6669 - ``KVM_HYPERCALL_EXIT_16BIT``: Indicates that the guest used a 16bit 6670 instruction to initiate the SMCCC call. If this bit is 0 then the 6671 guest used a 32bit instruction. An AArch64 guest always has this 6672 bit set to 0. 6673 6674At the point of exit, PC points to the instruction immediately following 6675the trapping instruction. 6676 6677:: 6678 6679 /* KVM_EXIT_TPR_ACCESS */ 6680 struct { 6681 __u64 rip; 6682 __u32 is_write; 6683 __u32 pad; 6684 } tpr_access; 6685 6686To be documented (KVM_TPR_ACCESS_REPORTING). 6687 6688:: 6689 6690 /* KVM_EXIT_S390_SIEIC */ 6691 struct { 6692 __u8 icptcode; 6693 __u64 mask; /* psw upper half */ 6694 __u64 addr; /* psw lower half */ 6695 __u16 ipa; 6696 __u32 ipb; 6697 } s390_sieic; 6698 6699s390 specific. 6700 6701:: 6702 6703 /* KVM_EXIT_S390_RESET */ 6704 #define KVM_S390_RESET_POR 1 6705 #define KVM_S390_RESET_CLEAR 2 6706 #define KVM_S390_RESET_SUBSYSTEM 4 6707 #define KVM_S390_RESET_CPU_INIT 8 6708 #define KVM_S390_RESET_IPL 16 6709 __u64 s390_reset_flags; 6710 6711s390 specific. 6712 6713:: 6714 6715 /* KVM_EXIT_S390_UCONTROL */ 6716 struct { 6717 __u64 trans_exc_code; 6718 __u32 pgm_code; 6719 } s390_ucontrol; 6720 6721s390 specific. A page fault has occurred for a user controlled virtual 6722machine (KVM_VM_S390_UNCONTROL) on its host page table that cannot be 6723resolved by the kernel. 6724The program code and the translation exception code that were placed 6725in the cpu's lowcore are presented here as defined by the z Architecture 6726Principles of Operation Book in the Chapter for Dynamic Address Translation 6727(DAT) 6728 6729:: 6730 6731 /* KVM_EXIT_DCR */ 6732 struct { 6733 __u32 dcrn; 6734 __u32 data; 6735 __u8 is_write; 6736 } dcr; 6737 6738Deprecated - was used for 440 KVM. 6739 6740:: 6741 6742 /* KVM_EXIT_OSI */ 6743 struct { 6744 __u64 gprs[32]; 6745 } osi; 6746 6747MOL uses a special hypercall interface it calls 'OSI'. To enable it, we catch 6748hypercalls and exit with this exit struct that contains all the guest gprs. 6749 6750If exit_reason is KVM_EXIT_OSI, then the vcpu has triggered such a hypercall. 6751Userspace can now handle the hypercall and when it's done modify the gprs as 6752necessary. Upon guest entry all guest GPRs will then be replaced by the values 6753in this struct. 6754 6755:: 6756 6757 /* KVM_EXIT_PAPR_HCALL */ 6758 struct { 6759 __u64 nr; 6760 __u64 ret; 6761 __u64 args[9]; 6762 } papr_hcall; 6763 6764This is used on 64-bit PowerPC when emulating a pSeries partition, 6765e.g. with the 'pseries' machine type in qemu. It occurs when the 6766guest does a hypercall using the 'sc 1' instruction. The 'nr' field 6767contains the hypercall number (from the guest R3), and 'args' contains 6768the arguments (from the guest R4 - R12). Userspace should put the 6769return code in 'ret' and any extra returned values in args[]. 6770The possible hypercalls are defined in the Power Architecture Platform 6771Requirements (PAPR) document available from www.power.org (free 6772developer registration required to access it). 6773 6774:: 6775 6776 /* KVM_EXIT_S390_TSCH */ 6777 struct { 6778 __u16 subchannel_id; 6779 __u16 subchannel_nr; 6780 __u32 io_int_parm; 6781 __u32 io_int_word; 6782 __u32 ipb; 6783 __u8 dequeued; 6784 } s390_tsch; 6785 6786s390 specific. This exit occurs when KVM_CAP_S390_CSS_SUPPORT has been enabled 6787and TEST SUBCHANNEL was intercepted. If dequeued is set, a pending I/O 6788interrupt for the target subchannel has been dequeued and subchannel_id, 6789subchannel_nr, io_int_parm and io_int_word contain the parameters for that 6790interrupt. ipb is needed for instruction parameter decoding. 6791 6792:: 6793 6794 /* KVM_EXIT_EPR */ 6795 struct { 6796 __u32 epr; 6797 } epr; 6798 6799On FSL BookE PowerPC chips, the interrupt controller has a fast patch 6800interrupt acknowledge path to the core. When the core successfully 6801delivers an interrupt, it automatically populates the EPR register with 6802the interrupt vector number and acknowledges the interrupt inside 6803the interrupt controller. 6804 6805In case the interrupt controller lives in user space, we need to do 6806the interrupt acknowledge cycle through it to fetch the next to be 6807delivered interrupt vector using this exit. 6808 6809It gets triggered whenever both KVM_CAP_PPC_EPR are enabled and an 6810external interrupt has just been delivered into the guest. User space 6811should put the acknowledged interrupt vector into the 'epr' field. 6812 6813:: 6814 6815 /* KVM_EXIT_SYSTEM_EVENT */ 6816 struct { 6817 #define KVM_SYSTEM_EVENT_SHUTDOWN 1 6818 #define KVM_SYSTEM_EVENT_RESET 2 6819 #define KVM_SYSTEM_EVENT_CRASH 3 6820 #define KVM_SYSTEM_EVENT_WAKEUP 4 6821 #define KVM_SYSTEM_EVENT_SUSPEND 5 6822 #define KVM_SYSTEM_EVENT_SEV_TERM 6 6823 __u32 type; 6824 __u32 ndata; 6825 __u64 data[16]; 6826 } system_event; 6827 6828If exit_reason is KVM_EXIT_SYSTEM_EVENT then the vcpu has triggered 6829a system-level event using some architecture specific mechanism (hypercall 6830or some special instruction). In case of ARM64, this is triggered using 6831HVC instruction based PSCI call from the vcpu. 6832 6833The 'type' field describes the system-level event type. 6834Valid values for 'type' are: 6835 6836 - KVM_SYSTEM_EVENT_SHUTDOWN -- the guest has requested a shutdown of the 6837 VM. Userspace is not obliged to honour this, and if it does honour 6838 this does not need to destroy the VM synchronously (ie it may call 6839 KVM_RUN again before shutdown finally occurs). 6840 - KVM_SYSTEM_EVENT_RESET -- the guest has requested a reset of the VM. 6841 As with SHUTDOWN, userspace can choose to ignore the request, or 6842 to schedule the reset to occur in the future and may call KVM_RUN again. 6843 - KVM_SYSTEM_EVENT_CRASH -- the guest crash occurred and the guest 6844 has requested a crash condition maintenance. Userspace can choose 6845 to ignore the request, or to gather VM memory core dump and/or 6846 reset/shutdown of the VM. 6847 - KVM_SYSTEM_EVENT_SEV_TERM -- an AMD SEV guest requested termination. 6848 The guest physical address of the guest's GHCB is stored in `data[0]`. 6849 - KVM_SYSTEM_EVENT_WAKEUP -- the exiting vCPU is in a suspended state and 6850 KVM has recognized a wakeup event. Userspace may honor this event by 6851 marking the exiting vCPU as runnable, or deny it and call KVM_RUN again. 6852 - KVM_SYSTEM_EVENT_SUSPEND -- the guest has requested a suspension of 6853 the VM. 6854 6855If KVM_CAP_SYSTEM_EVENT_DATA is present, the 'data' field can contain 6856architecture specific information for the system-level event. Only 6857the first `ndata` items (possibly zero) of the data array are valid. 6858 6859 - for arm64, data[0] is set to KVM_SYSTEM_EVENT_RESET_FLAG_PSCI_RESET2 if 6860 the guest issued a SYSTEM_RESET2 call according to v1.1 of the PSCI 6861 specification. 6862 6863 - for arm64, data[0] is set to KVM_SYSTEM_EVENT_SHUTDOWN_FLAG_PSCI_OFF2 6864 if the guest issued a SYSTEM_OFF2 call according to v1.3 of the PSCI 6865 specification. 6866 6867 - for RISC-V, data[0] is set to the value of the second argument of the 6868 ``sbi_system_reset`` call. 6869 6870Previous versions of Linux defined a `flags` member in this struct. The 6871field is now aliased to `data[0]`. Userspace can assume that it is only 6872written if ndata is greater than 0. 6873 6874For arm/arm64: 6875-------------- 6876 6877KVM_SYSTEM_EVENT_SUSPEND exits are enabled with the 6878KVM_CAP_ARM_SYSTEM_SUSPEND VM capability. If a guest invokes the PSCI 6879SYSTEM_SUSPEND function, KVM will exit to userspace with this event 6880type. 6881 6882It is the sole responsibility of userspace to implement the PSCI 6883SYSTEM_SUSPEND call according to ARM DEN0022D.b 5.19 "SYSTEM_SUSPEND". 6884KVM does not change the vCPU's state before exiting to userspace, so 6885the call parameters are left in-place in the vCPU registers. 6886 6887Userspace is _required_ to take action for such an exit. It must 6888either: 6889 6890 - Honor the guest request to suspend the VM. Userspace can request 6891 in-kernel emulation of suspension by setting the calling vCPU's 6892 state to KVM_MP_STATE_SUSPENDED. Userspace must configure the vCPU's 6893 state according to the parameters passed to the PSCI function when 6894 the calling vCPU is resumed. See ARM DEN0022D.b 5.19.1 "Intended use" 6895 for details on the function parameters. 6896 6897 - Deny the guest request to suspend the VM. See ARM DEN0022D.b 5.19.2 6898 "Caller responsibilities" for possible return values. 6899 6900Hibernation using the PSCI SYSTEM_OFF2 call is enabled when PSCI v1.3 6901is enabled. If a guest invokes the PSCI SYSTEM_OFF2 function, KVM will 6902exit to userspace with the KVM_SYSTEM_EVENT_SHUTDOWN event type and with 6903data[0] set to KVM_SYSTEM_EVENT_SHUTDOWN_FLAG_PSCI_OFF2. The only 6904supported hibernate type for the SYSTEM_OFF2 function is HIBERNATE_OFF. 6905 6906:: 6907 6908 /* KVM_EXIT_IOAPIC_EOI */ 6909 struct { 6910 __u8 vector; 6911 } eoi; 6912 6913Indicates that the VCPU's in-kernel local APIC received an EOI for a 6914level-triggered IOAPIC interrupt. This exit only triggers when the 6915IOAPIC is implemented in userspace (i.e. KVM_CAP_SPLIT_IRQCHIP is enabled); 6916the userspace IOAPIC should process the EOI and retrigger the interrupt if 6917it is still asserted. Vector is the LAPIC interrupt vector for which the 6918EOI was received. 6919 6920:: 6921 6922 struct kvm_hyperv_exit { 6923 #define KVM_EXIT_HYPERV_SYNIC 1 6924 #define KVM_EXIT_HYPERV_HCALL 2 6925 #define KVM_EXIT_HYPERV_SYNDBG 3 6926 __u32 type; 6927 __u32 pad1; 6928 union { 6929 struct { 6930 __u32 msr; 6931 __u32 pad2; 6932 __u64 control; 6933 __u64 evt_page; 6934 __u64 msg_page; 6935 } synic; 6936 struct { 6937 __u64 input; 6938 __u64 result; 6939 __u64 params[2]; 6940 } hcall; 6941 struct { 6942 __u32 msr; 6943 __u32 pad2; 6944 __u64 control; 6945 __u64 status; 6946 __u64 send_page; 6947 __u64 recv_page; 6948 __u64 pending_page; 6949 } syndbg; 6950 } u; 6951 }; 6952 /* KVM_EXIT_HYPERV */ 6953 struct kvm_hyperv_exit hyperv; 6954 6955Indicates that the VCPU exits into userspace to process some tasks 6956related to Hyper-V emulation. 6957 6958Valid values for 'type' are: 6959 6960 - KVM_EXIT_HYPERV_SYNIC -- synchronously notify user-space about 6961 6962Hyper-V SynIC state change. Notification is used to remap SynIC 6963event/message pages and to enable/disable SynIC messages/events processing 6964in userspace. 6965 6966 - KVM_EXIT_HYPERV_SYNDBG -- synchronously notify user-space about 6967 6968Hyper-V Synthetic debugger state change. Notification is used to either update 6969the pending_page location or to send a control command (send the buffer located 6970in send_page or recv a buffer to recv_page). 6971 6972:: 6973 6974 /* KVM_EXIT_ARM_NISV */ 6975 struct { 6976 __u64 esr_iss; 6977 __u64 fault_ipa; 6978 } arm_nisv; 6979 6980Used on arm64 systems. If a guest accesses memory not in a memslot, 6981KVM will typically return to userspace and ask it to do MMIO emulation on its 6982behalf. However, for certain classes of instructions, no instruction decode 6983(direction, length of memory access) is provided, and fetching and decoding 6984the instruction from the VM is overly complicated to live in the kernel. 6985 6986Historically, when this situation occurred, KVM would print a warning and kill 6987the VM. KVM assumed that if the guest accessed non-memslot memory, it was 6988trying to do I/O, which just couldn't be emulated, and the warning message was 6989phrased accordingly. However, what happened more often was that a guest bug 6990caused access outside the guest memory areas which should lead to a more 6991meaningful warning message and an external abort in the guest, if the access 6992did not fall within an I/O window. 6993 6994Userspace implementations can query for KVM_CAP_ARM_NISV_TO_USER, and enable 6995this capability at VM creation. Once this is done, these types of errors will 6996instead return to userspace with KVM_EXIT_ARM_NISV, with the valid bits from 6997the ESR_EL2 in the esr_iss field, and the faulting IPA in the fault_ipa field. 6998Userspace can either fix up the access if it's actually an I/O access by 6999decoding the instruction from guest memory (if it's very brave) and continue 7000executing the guest, or it can decide to suspend, dump, or restart the guest. 7001 7002Note that KVM does not skip the faulting instruction as it does for 7003KVM_EXIT_MMIO, but userspace has to emulate any change to the processing state 7004if it decides to decode and emulate the instruction. 7005 7006This feature isn't available to protected VMs, as userspace does not 7007have access to the state that is required to perform the emulation. 7008Instead, a data abort exception is directly injected in the guest. 7009Note that although KVM_CAP_ARM_NISV_TO_USER will be reported if 7010queried outside of a protected VM context, the feature will not be 7011exposed if queried on a protected VM file descriptor. 7012 7013:: 7014 7015 /* KVM_EXIT_X86_RDMSR / KVM_EXIT_X86_WRMSR */ 7016 struct { 7017 __u8 error; /* user -> kernel */ 7018 __u8 pad[7]; 7019 __u32 reason; /* kernel -> user */ 7020 __u32 index; /* kernel -> user */ 7021 __u64 data; /* kernel <-> user */ 7022 } msr; 7023 7024Used on x86 systems. When the VM capability KVM_CAP_X86_USER_SPACE_MSR is 7025enabled, MSR accesses to registers that would invoke a #GP by KVM kernel code 7026may instead trigger a KVM_EXIT_X86_RDMSR exit for reads and KVM_EXIT_X86_WRMSR 7027exit for writes. 7028 7029The "reason" field specifies why the MSR interception occurred. Userspace will 7030only receive MSR exits when a particular reason was requested during through 7031ENABLE_CAP. Currently valid exit reasons are: 7032 7033============================ ======================================== 7034 KVM_MSR_EXIT_REASON_UNKNOWN access to MSR that is unknown to KVM 7035 KVM_MSR_EXIT_REASON_INVAL access to invalid MSRs or reserved bits 7036 KVM_MSR_EXIT_REASON_FILTER access blocked by KVM_X86_SET_MSR_FILTER 7037============================ ======================================== 7038 7039For KVM_EXIT_X86_RDMSR, the "index" field tells userspace which MSR the guest 7040wants to read. To respond to this request with a successful read, userspace 7041writes the respective data into the "data" field and must continue guest 7042execution to ensure the read data is transferred into guest register state. 7043 7044If the RDMSR request was unsuccessful, userspace indicates that with a "1" in 7045the "error" field. This will inject a #GP into the guest when the VCPU is 7046executed again. 7047 7048For KVM_EXIT_X86_WRMSR, the "index" field tells userspace which MSR the guest 7049wants to write. Once finished processing the event, userspace must continue 7050vCPU execution. If the MSR write was unsuccessful, userspace also sets the 7051"error" field to "1". 7052 7053See KVM_X86_SET_MSR_FILTER for details on the interaction with MSR filtering. 7054 7055:: 7056 7057 7058 struct kvm_xen_exit { 7059 #define KVM_EXIT_XEN_HCALL 1 7060 __u32 type; 7061 union { 7062 struct { 7063 __u32 longmode; 7064 __u32 cpl; 7065 __u64 input; 7066 __u64 result; 7067 __u64 params[6]; 7068 } hcall; 7069 } u; 7070 }; 7071 /* KVM_EXIT_XEN */ 7072 struct kvm_hyperv_exit xen; 7073 7074Indicates that the VCPU exits into userspace to process some tasks 7075related to Xen emulation. 7076 7077Valid values for 'type' are: 7078 7079 - KVM_EXIT_XEN_HCALL -- synchronously notify user-space about Xen hypercall. 7080 Userspace is expected to place the hypercall result into the appropriate 7081 field before invoking KVM_RUN again. 7082 7083:: 7084 7085 /* KVM_EXIT_RISCV_SBI */ 7086 struct { 7087 unsigned long extension_id; 7088 unsigned long function_id; 7089 unsigned long args[6]; 7090 unsigned long ret[2]; 7091 } riscv_sbi; 7092 7093If exit reason is KVM_EXIT_RISCV_SBI then it indicates that the VCPU has 7094done a SBI call which is not handled by KVM RISC-V kernel module. The details 7095of the SBI call are available in 'riscv_sbi' member of kvm_run structure. The 7096'extension_id' field of 'riscv_sbi' represents SBI extension ID whereas the 7097'function_id' field represents function ID of given SBI extension. The 'args' 7098array field of 'riscv_sbi' represents parameters for the SBI call and 'ret' 7099array field represents return values. The userspace should update the return 7100values of SBI call before resuming the VCPU. For more details on RISC-V SBI 7101spec refer, https://github.com/riscv/riscv-sbi-doc. 7102 7103:: 7104 7105 /* KVM_EXIT_MEMORY_FAULT */ 7106 struct { 7107 #define KVM_MEMORY_EXIT_FLAG_PRIVATE (1ULL << 3) 7108 __u64 flags; 7109 __u64 gpa; 7110 __u64 size; 7111 } memory_fault; 7112 7113KVM_EXIT_MEMORY_FAULT indicates the vCPU has encountered a memory fault that 7114could not be resolved by KVM. The 'gpa' and 'size' (in bytes) describe the 7115guest physical address range [gpa, gpa + size) of the fault. The 'flags' field 7116describes properties of the faulting access that are likely pertinent: 7117 7118 - KVM_MEMORY_EXIT_FLAG_PRIVATE - When set, indicates the memory fault occurred 7119 on a private memory access. When clear, indicates the fault occurred on a 7120 shared access. 7121 7122Note! KVM_EXIT_MEMORY_FAULT is unique among all KVM exit reasons in that it 7123accompanies a return code of '-1', not '0'! errno will always be set to EFAULT 7124or EHWPOISON when KVM exits with KVM_EXIT_MEMORY_FAULT, userspace should assume 7125kvm_run.exit_reason is stale/undefined for all other error numbers. 7126 7127:: 7128 7129 /* KVM_EXIT_NOTIFY */ 7130 struct { 7131 #define KVM_NOTIFY_CONTEXT_INVALID (1 << 0) 7132 __u32 flags; 7133 } notify; 7134 7135Used on x86 systems. When the VM capability KVM_CAP_X86_NOTIFY_VMEXIT is 7136enabled, a VM exit generated if no event window occurs in VM non-root mode 7137for a specified amount of time. Once KVM_X86_NOTIFY_VMEXIT_USER is set when 7138enabling the cap, it would exit to userspace with the exit reason 7139KVM_EXIT_NOTIFY for further handling. The "flags" field contains more 7140detailed info. 7141 7142The valid value for 'flags' is: 7143 7144 - KVM_NOTIFY_CONTEXT_INVALID -- the VM context is corrupted and not valid 7145 in VMCS. It would run into unknown result if resume the target VM. 7146 7147:: 7148 7149 /* Fix the size of the union. */ 7150 char padding[256]; 7151 }; 7152 7153 /* 7154 * shared registers between kvm and userspace. 7155 * kvm_valid_regs specifies the register classes set by the host 7156 * kvm_dirty_regs specified the register classes dirtied by userspace 7157 * struct kvm_sync_regs is architecture specific, as well as the 7158 * bits for kvm_valid_regs and kvm_dirty_regs 7159 */ 7160 __u64 kvm_valid_regs; 7161 __u64 kvm_dirty_regs; 7162 union { 7163 struct kvm_sync_regs regs; 7164 char padding[SYNC_REGS_SIZE_BYTES]; 7165 } s; 7166 7167If KVM_CAP_SYNC_REGS is defined, these fields allow userspace to access 7168certain guest registers without having to call SET/GET_*REGS. Thus we can 7169avoid some system call overhead if userspace has to handle the exit. 7170Userspace can query the validity of the structure by checking 7171kvm_valid_regs for specific bits. These bits are architecture specific 7172and usually define the validity of a groups of registers. (e.g. one bit 7173for general purpose registers) 7174 7175Please note that the kernel is allowed to use the kvm_run structure as the 7176primary storage for certain register types. Therefore, the kernel may use the 7177values in kvm_run even if the corresponding bit in kvm_dirty_regs is not set. 7178 7179 7180.. _cap_enable: 7181 71826. Capabilities that can be enabled on vCPUs 7183============================================ 7184 7185There are certain capabilities that change the behavior of the virtual CPU or 7186the virtual machine when enabled. To enable them, please see 7187:ref:`KVM_ENABLE_CAP`. 7188 7189Below you can find a list of capabilities and what their effect on the vCPU or 7190the virtual machine is when enabling them. 7191 7192The following information is provided along with the description: 7193 7194 Architectures: 7195 which instruction set architectures provide this ioctl. 7196 x86 includes both i386 and x86_64. 7197 7198 Target: 7199 whether this is a per-vcpu or per-vm capability. 7200 7201 Parameters: 7202 what parameters are accepted by the capability. 7203 7204 Returns: 7205 the return value. General error numbers (EBADF, ENOMEM, EINVAL) 7206 are not detailed, but errors with specific meanings are. 7207 7208 72096.1 KVM_CAP_PPC_OSI 7210------------------- 7211 7212:Architectures: ppc 7213:Target: vcpu 7214:Parameters: none 7215:Returns: 0 on success; -1 on error 7216 7217This capability enables interception of OSI hypercalls that otherwise would 7218be treated as normal system calls to be injected into the guest. OSI hypercalls 7219were invented by Mac-on-Linux to have a standardized communication mechanism 7220between the guest and the host. 7221 7222When this capability is enabled, KVM_EXIT_OSI can occur. 7223 7224 72256.2 KVM_CAP_PPC_PAPR 7226-------------------- 7227 7228:Architectures: ppc 7229:Target: vcpu 7230:Parameters: none 7231:Returns: 0 on success; -1 on error 7232 7233This capability enables interception of PAPR hypercalls. PAPR hypercalls are 7234done using the hypercall instruction "sc 1". 7235 7236It also sets the guest privilege level to "supervisor" mode. Usually the guest 7237runs in "hypervisor" privilege mode with a few missing features. 7238 7239In addition to the above, it changes the semantics of SDR1. In this mode, the 7240HTAB address part of SDR1 contains an HVA instead of a GPA, as PAPR keeps the 7241HTAB invisible to the guest. 7242 7243When this capability is enabled, KVM_EXIT_PAPR_HCALL can occur. 7244 7245 72466.3 KVM_CAP_SW_TLB 7247------------------ 7248 7249:Architectures: ppc 7250:Target: vcpu 7251:Parameters: args[0] is the address of a struct kvm_config_tlb 7252:Returns: 0 on success; -1 on error 7253 7254:: 7255 7256 struct kvm_config_tlb { 7257 __u64 params; 7258 __u64 array; 7259 __u32 mmu_type; 7260 __u32 array_len; 7261 }; 7262 7263Configures the virtual CPU's TLB array, establishing a shared memory area 7264between userspace and KVM. The "params" and "array" fields are userspace 7265addresses of mmu-type-specific data structures. The "array_len" field is an 7266safety mechanism, and should be set to the size in bytes of the memory that 7267userspace has reserved for the array. It must be at least the size dictated 7268by "mmu_type" and "params". 7269 7270While KVM_RUN is active, the shared region is under control of KVM. Its 7271contents are undefined, and any modification by userspace results in 7272boundedly undefined behavior. 7273 7274On return from KVM_RUN, the shared region will reflect the current state of 7275the guest's TLB. If userspace makes any changes, it must call KVM_DIRTY_TLB 7276to tell KVM which entries have been changed, prior to calling KVM_RUN again 7277on this vcpu. 7278 7279For mmu types KVM_MMU_FSL_BOOKE_NOHV and KVM_MMU_FSL_BOOKE_HV: 7280 7281 - The "params" field is of type "struct kvm_book3e_206_tlb_params". 7282 - The "array" field points to an array of type "struct 7283 kvm_book3e_206_tlb_entry". 7284 - The array consists of all entries in the first TLB, followed by all 7285 entries in the second TLB. 7286 - Within a TLB, entries are ordered first by increasing set number. Within a 7287 set, entries are ordered by way (increasing ESEL). 7288 - The hash for determining set number in TLB0 is: (MAS2 >> 12) & (num_sets - 1) 7289 where "num_sets" is the tlb_sizes[] value divided by the tlb_ways[] value. 7290 - The tsize field of mas1 shall be set to 4K on TLB0, even though the 7291 hardware ignores this value for TLB0. 7292 72936.4 KVM_CAP_S390_CSS_SUPPORT 7294---------------------------- 7295 7296:Architectures: s390 7297:Target: vcpu 7298:Parameters: none 7299:Returns: 0 on success; -1 on error 7300 7301This capability enables support for handling of channel I/O instructions. 7302 7303TEST PENDING INTERRUPTION and the interrupt portion of TEST SUBCHANNEL are 7304handled in-kernel, while the other I/O instructions are passed to userspace. 7305 7306When this capability is enabled, KVM_EXIT_S390_TSCH will occur on TEST 7307SUBCHANNEL intercepts. 7308 7309Note that even though this capability is enabled per-vcpu, the complete 7310virtual machine is affected. 7311 73126.5 KVM_CAP_PPC_EPR 7313------------------- 7314 7315:Architectures: ppc 7316:Target: vcpu 7317:Parameters: args[0] defines whether the proxy facility is active 7318:Returns: 0 on success; -1 on error 7319 7320This capability enables or disables the delivery of interrupts through the 7321external proxy facility. 7322 7323When enabled (args[0] != 0), every time the guest gets an external interrupt 7324delivered, it automatically exits into user space with a KVM_EXIT_EPR exit 7325to receive the topmost interrupt vector. 7326 7327When disabled (args[0] == 0), behavior is as if this facility is unsupported. 7328 7329When this capability is enabled, KVM_EXIT_EPR can occur. 7330 73316.6 KVM_CAP_IRQ_MPIC 7332-------------------- 7333 7334:Architectures: ppc 7335:Parameters: args[0] is the MPIC device fd; 7336 args[1] is the MPIC CPU number for this vcpu 7337 7338This capability connects the vcpu to an in-kernel MPIC device. 7339 73406.7 KVM_CAP_IRQ_XICS 7341-------------------- 7342 7343:Architectures: ppc 7344:Target: vcpu 7345:Parameters: args[0] is the XICS device fd; 7346 args[1] is the XICS CPU number (server ID) for this vcpu 7347 7348This capability connects the vcpu to an in-kernel XICS device. 7349 73506.8 KVM_CAP_S390_IRQCHIP 7351------------------------ 7352 7353:Architectures: s390 7354:Target: vm 7355:Parameters: none 7356 7357This capability enables the in-kernel irqchip for s390. Please refer to 7358"4.24 KVM_CREATE_IRQCHIP" for details. 7359 73606.9 KVM_CAP_MIPS_FPU 7361-------------------- 7362 7363:Architectures: mips 7364:Target: vcpu 7365:Parameters: args[0] is reserved for future use (should be 0). 7366 7367This capability allows the use of the host Floating Point Unit by the guest. It 7368allows the Config1.FP bit to be set to enable the FPU in the guest. Once this is 7369done the ``KVM_REG_MIPS_FPR_*`` and ``KVM_REG_MIPS_FCR_*`` registers can be 7370accessed (depending on the current guest FPU register mode), and the Status.FR, 7371Config5.FRE bits are accessible via the KVM API and also from the guest, 7372depending on them being supported by the FPU. 7373 73746.10 KVM_CAP_MIPS_MSA 7375--------------------- 7376 7377:Architectures: mips 7378:Target: vcpu 7379:Parameters: args[0] is reserved for future use (should be 0). 7380 7381This capability allows the use of the MIPS SIMD Architecture (MSA) by the guest. 7382It allows the Config3.MSAP bit to be set to enable the use of MSA by the guest. 7383Once this is done the ``KVM_REG_MIPS_VEC_*`` and ``KVM_REG_MIPS_MSA_*`` 7384registers can be accessed, and the Config5.MSAEn bit is accessible via the 7385KVM API and also from the guest. 7386 73876.74 KVM_CAP_SYNC_REGS 7388---------------------- 7389 7390:Architectures: s390, x86 7391:Target: s390: always enabled, x86: vcpu 7392:Parameters: none 7393:Returns: x86: KVM_CHECK_EXTENSION returns a bit-array indicating which register 7394 sets are supported 7395 (bitfields defined in arch/x86/include/uapi/asm/kvm.h). 7396 7397As described above in the kvm_sync_regs struct info in section :ref:`kvm_run`, 7398KVM_CAP_SYNC_REGS "allow[s] userspace to access certain guest registers 7399without having to call SET/GET_*REGS". This reduces overhead by eliminating 7400repeated ioctl calls for setting and/or getting register values. This is 7401particularly important when userspace is making synchronous guest state 7402modifications, e.g. when emulating and/or intercepting instructions in 7403userspace. 7404 7405For s390 specifics, please refer to the source code. 7406 7407For x86: 7408 7409- the register sets to be copied out to kvm_run are selectable 7410 by userspace (rather that all sets being copied out for every exit). 7411- vcpu_events are available in addition to regs and sregs. 7412 7413For x86, the 'kvm_valid_regs' field of struct kvm_run is overloaded to 7414function as an input bit-array field set by userspace to indicate the 7415specific register sets to be copied out on the next exit. 7416 7417To indicate when userspace has modified values that should be copied into 7418the vCPU, the all architecture bitarray field, 'kvm_dirty_regs' must be set. 7419This is done using the same bitflags as for the 'kvm_valid_regs' field. 7420If the dirty bit is not set, then the register set values will not be copied 7421into the vCPU even if they've been modified. 7422 7423Unused bitfields in the bitarrays must be set to zero. 7424 7425:: 7426 7427 struct kvm_sync_regs { 7428 struct kvm_regs regs; 7429 struct kvm_sregs sregs; 7430 struct kvm_vcpu_events events; 7431 }; 7432 74336.75 KVM_CAP_PPC_IRQ_XIVE 7434------------------------- 7435 7436:Architectures: ppc 7437:Target: vcpu 7438:Parameters: args[0] is the XIVE device fd; 7439 args[1] is the XIVE CPU number (server ID) for this vcpu 7440 7441This capability connects the vcpu to an in-kernel XIVE device. 7442 7443.. _cap_enable_vm: 7444 74457. Capabilities that can be enabled on VMs 7446========================================== 7447 7448There are certain capabilities that change the behavior of the virtual 7449machine when enabled. To enable them, please see section 7450:ref:`KVM_ENABLE_CAP`. Below you can find a list of capabilities and 7451what their effect on the VM is when enabling them. 7452 7453The following information is provided along with the description: 7454 7455 Architectures: 7456 which instruction set architectures provide this ioctl. 7457 x86 includes both i386 and x86_64. 7458 7459 Parameters: 7460 what parameters are accepted by the capability. 7461 7462 Returns: 7463 the return value. General error numbers (EBADF, ENOMEM, EINVAL) 7464 are not detailed, but errors with specific meanings are. 7465 7466 74677.1 KVM_CAP_PPC_ENABLE_HCALL 7468---------------------------- 7469 7470:Architectures: ppc 7471:Parameters: args[0] is the sPAPR hcall number; 7472 args[1] is 0 to disable, 1 to enable in-kernel handling 7473 7474This capability controls whether individual sPAPR hypercalls (hcalls) 7475get handled by the kernel or not. Enabling or disabling in-kernel 7476handling of an hcall is effective across the VM. On creation, an 7477initial set of hcalls are enabled for in-kernel handling, which 7478consists of those hcalls for which in-kernel handlers were implemented 7479before this capability was implemented. If disabled, the kernel will 7480not to attempt to handle the hcall, but will always exit to userspace 7481to handle it. Note that it may not make sense to enable some and 7482disable others of a group of related hcalls, but KVM does not prevent 7483userspace from doing that. 7484 7485If the hcall number specified is not one that has an in-kernel 7486implementation, the KVM_ENABLE_CAP ioctl will fail with an EINVAL 7487error. 7488 74897.2 KVM_CAP_S390_USER_SIGP 7490-------------------------- 7491 7492:Architectures: s390 7493:Parameters: none 7494 7495This capability controls which SIGP orders will be handled completely in user 7496space. With this capability enabled, all fast orders will be handled completely 7497in the kernel: 7498 7499- SENSE 7500- SENSE RUNNING 7501- EXTERNAL CALL 7502- EMERGENCY SIGNAL 7503- CONDITIONAL EMERGENCY SIGNAL 7504 7505All other orders will be handled completely in user space. 7506 7507Only privileged operation exceptions will be checked for in the kernel (or even 7508in the hardware prior to interception). If this capability is not enabled, the 7509old way of handling SIGP orders is used (partially in kernel and user space). 7510 75117.3 KVM_CAP_S390_VECTOR_REGISTERS 7512--------------------------------- 7513 7514:Architectures: s390 7515:Parameters: none 7516:Returns: 0 on success, negative value on error 7517 7518Allows use of the vector registers introduced with z13 processor, and 7519provides for the synchronization between host and user space. Will 7520return -EINVAL if the machine does not support vectors. 7521 75227.4 KVM_CAP_S390_USER_STSI 7523-------------------------- 7524 7525:Architectures: s390 7526:Parameters: none 7527 7528This capability allows post-handlers for the STSI instruction. After 7529initial handling in the kernel, KVM exits to user space with 7530KVM_EXIT_S390_STSI to allow user space to insert further data. 7531 7532Before exiting to userspace, kvm handlers should fill in s390_stsi field of 7533vcpu->run:: 7534 7535 struct { 7536 __u64 addr; 7537 __u8 ar; 7538 __u8 reserved; 7539 __u8 fc; 7540 __u8 sel1; 7541 __u16 sel2; 7542 } s390_stsi; 7543 7544 @addr - guest address of STSI SYSIB 7545 @fc - function code 7546 @sel1 - selector 1 7547 @sel2 - selector 2 7548 @ar - access register number 7549 7550KVM handlers should exit to userspace with rc = -EREMOTE. 7551 75527.5 KVM_CAP_SPLIT_IRQCHIP 7553------------------------- 7554 7555:Architectures: x86 7556:Parameters: args[0] - number of routes reserved for userspace IOAPICs 7557:Returns: 0 on success, -1 on error 7558 7559Create a local apic for each processor in the kernel. This can be used 7560instead of KVM_CREATE_IRQCHIP if the userspace VMM wishes to emulate the 7561IOAPIC and PIC (and also the PIT, even though this has to be enabled 7562separately). 7563 7564This capability also enables in kernel routing of interrupt requests; 7565when KVM_CAP_SPLIT_IRQCHIP only routes of KVM_IRQ_ROUTING_MSI type are 7566used in the IRQ routing table. The first args[0] MSI routes are reserved 7567for the IOAPIC pins. Whenever the LAPIC receives an EOI for these routes, 7568a KVM_EXIT_IOAPIC_EOI vmexit will be reported to userspace. 7569 7570Fails if VCPU has already been created, or if the irqchip is already in the 7571kernel (i.e. KVM_CREATE_IRQCHIP has already been called). 7572 75737.6 KVM_CAP_S390_RI 7574------------------- 7575 7576:Architectures: s390 7577:Parameters: none 7578 7579Allows use of runtime-instrumentation introduced with zEC12 processor. 7580Will return -EINVAL if the machine does not support runtime-instrumentation. 7581Will return -EBUSY if a VCPU has already been created. 7582 75837.7 KVM_CAP_X2APIC_API 7584---------------------- 7585 7586:Architectures: x86 7587:Parameters: args[0] - features that should be enabled 7588:Returns: 0 on success, -EINVAL when args[0] contains invalid features 7589 7590Valid feature flags in args[0] are:: 7591 7592 #define KVM_X2APIC_API_USE_32BIT_IDS (1ULL << 0) 7593 #define KVM_X2APIC_API_DISABLE_BROADCAST_QUIRK (1ULL << 1) 7594 7595Enabling KVM_X2APIC_API_USE_32BIT_IDS changes the behavior of 7596KVM_SET_GSI_ROUTING, KVM_SIGNAL_MSI, KVM_SET_LAPIC, and KVM_GET_LAPIC, 7597allowing the use of 32-bit APIC IDs. See KVM_CAP_X2APIC_API in their 7598respective sections. 7599 7600KVM_X2APIC_API_DISABLE_BROADCAST_QUIRK must be enabled for x2APIC to work 7601in logical mode or with more than 255 VCPUs. Otherwise, KVM treats 0xff 7602as a broadcast even in x2APIC mode in order to support physical x2APIC 7603without interrupt remapping. This is undesirable in logical mode, 7604where 0xff represents CPUs 0-7 in cluster 0. 7605 76067.8 KVM_CAP_S390_USER_INSTR0 7607---------------------------- 7608 7609:Architectures: s390 7610:Parameters: none 7611 7612With this capability enabled, all illegal instructions 0x0000 (2 bytes) will 7613be intercepted and forwarded to user space. User space can use this 7614mechanism e.g. to realize 2-byte software breakpoints. The kernel will 7615not inject an operating exception for these instructions, user space has 7616to take care of that. 7617 7618This capability can be enabled dynamically even if VCPUs were already 7619created and are running. 7620 76217.9 KVM_CAP_S390_GS 7622------------------- 7623 7624:Architectures: s390 7625:Parameters: none 7626:Returns: 0 on success; -EINVAL if the machine does not support 7627 guarded storage; -EBUSY if a VCPU has already been created. 7628 7629Allows use of guarded storage for the KVM guest. 7630 76317.10 KVM_CAP_S390_AIS 7632--------------------- 7633 7634:Architectures: s390 7635:Parameters: none 7636 7637Allow use of adapter-interruption suppression. 7638:Returns: 0 on success; -EBUSY if a VCPU has already been created. 7639 76407.11 KVM_CAP_PPC_SMT 7641-------------------- 7642 7643:Architectures: ppc 7644:Parameters: vsmt_mode, flags 7645 7646Enabling this capability on a VM provides userspace with a way to set 7647the desired virtual SMT mode (i.e. the number of virtual CPUs per 7648virtual core). The virtual SMT mode, vsmt_mode, must be a power of 2 7649between 1 and 8. On POWER8, vsmt_mode must also be no greater than 7650the number of threads per subcore for the host. Currently flags must 7651be 0. A successful call to enable this capability will result in 7652vsmt_mode being returned when the KVM_CAP_PPC_SMT capability is 7653subsequently queried for the VM. This capability is only supported by 7654HV KVM, and can only be set before any VCPUs have been created. 7655The KVM_CAP_PPC_SMT_POSSIBLE capability indicates which virtual SMT 7656modes are available. 7657 76587.12 KVM_CAP_PPC_FWNMI 7659---------------------- 7660 7661:Architectures: ppc 7662:Parameters: none 7663 7664With this capability a machine check exception in the guest address 7665space will cause KVM to exit the guest with NMI exit reason. This 7666enables QEMU to build error log and branch to guest kernel registered 7667machine check handling routine. Without this capability KVM will 7668branch to guests' 0x200 interrupt vector. 7669 76707.13 KVM_CAP_X86_DISABLE_EXITS 7671------------------------------ 7672 7673:Architectures: x86 7674:Parameters: args[0] defines which exits are disabled 7675:Returns: 0 on success, -EINVAL when args[0] contains invalid exits 7676 7677Valid bits in args[0] are:: 7678 7679 #define KVM_X86_DISABLE_EXITS_MWAIT (1 << 0) 7680 #define KVM_X86_DISABLE_EXITS_HLT (1 << 1) 7681 #define KVM_X86_DISABLE_EXITS_PAUSE (1 << 2) 7682 #define KVM_X86_DISABLE_EXITS_CSTATE (1 << 3) 7683 7684Enabling this capability on a VM provides userspace with a way to no 7685longer intercept some instructions for improved latency in some 7686workloads, and is suggested when vCPUs are associated to dedicated 7687physical CPUs. More bits can be added in the future; userspace can 7688just pass the KVM_CHECK_EXTENSION result to KVM_ENABLE_CAP to disable 7689all such vmexits. 7690 7691Do not enable KVM_FEATURE_PV_UNHALT if you disable HLT exits. 7692 76937.14 KVM_CAP_S390_HPAGE_1M 7694-------------------------- 7695 7696:Architectures: s390 7697:Parameters: none 7698:Returns: 0 on success, -EINVAL if hpage module parameter was not set 7699 or cmma is enabled, or the VM has the KVM_VM_S390_UCONTROL 7700 flag set 7701 7702With this capability the KVM support for memory backing with 1m pages 7703through hugetlbfs can be enabled for a VM. After the capability is 7704enabled, cmma can't be enabled anymore and pfmfi and the storage key 7705interpretation are disabled. If cmma has already been enabled or the 7706hpage module parameter is not set to 1, -EINVAL is returned. 7707 7708While it is generally possible to create a huge page backed VM without 7709this capability, the VM will not be able to run. 7710 77117.15 KVM_CAP_MSR_PLATFORM_INFO 7712------------------------------ 7713 7714:Architectures: x86 7715:Parameters: args[0] whether feature should be enabled or not 7716 7717With this capability, a guest may read the MSR_PLATFORM_INFO MSR. Otherwise, 7718a #GP would be raised when the guest tries to access. Currently, this 7719capability does not enable write permissions of this MSR for the guest. 7720 77217.16 KVM_CAP_PPC_NESTED_HV 7722-------------------------- 7723 7724:Architectures: ppc 7725:Parameters: none 7726:Returns: 0 on success, -EINVAL when the implementation doesn't support 7727 nested-HV virtualization. 7728 7729HV-KVM on POWER9 and later systems allows for "nested-HV" 7730virtualization, which provides a way for a guest VM to run guests that 7731can run using the CPU's supervisor mode (privileged non-hypervisor 7732state). Enabling this capability on a VM depends on the CPU having 7733the necessary functionality and on the facility being enabled with a 7734kvm-hv module parameter. 7735 77367.17 KVM_CAP_EXCEPTION_PAYLOAD 7737------------------------------ 7738 7739:Architectures: x86 7740:Parameters: args[0] whether feature should be enabled or not 7741 7742With this capability enabled, CR2 will not be modified prior to the 7743emulated VM-exit when L1 intercepts a #PF exception that occurs in 7744L2. Similarly, for kvm-intel only, DR6 will not be modified prior to 7745the emulated VM-exit when L1 intercepts a #DB exception that occurs in 7746L2. As a result, when KVM_GET_VCPU_EVENTS reports a pending #PF (or 7747#DB) exception for L2, exception.has_payload will be set and the 7748faulting address (or the new DR6 bits*) will be reported in the 7749exception_payload field. Similarly, when userspace injects a #PF (or 7750#DB) into L2 using KVM_SET_VCPU_EVENTS, it is expected to set 7751exception.has_payload and to put the faulting address - or the new DR6 7752bits\ [#]_ - in the exception_payload field. 7753 7754This capability also enables exception.pending in struct 7755kvm_vcpu_events, which allows userspace to distinguish between pending 7756and injected exceptions. 7757 7758 7759.. [#] For the new DR6 bits, note that bit 16 is set iff the #DB exception 7760 will clear DR6.RTM. 7761 77627.18 KVM_CAP_MANUAL_DIRTY_LOG_PROTECT2 7763-------------------------------------- 7764 7765:Architectures: x86, arm64, mips 7766:Parameters: args[0] whether feature should be enabled or not 7767 7768Valid flags are:: 7769 7770 #define KVM_DIRTY_LOG_MANUAL_PROTECT_ENABLE (1 << 0) 7771 #define KVM_DIRTY_LOG_INITIALLY_SET (1 << 1) 7772 7773With KVM_DIRTY_LOG_MANUAL_PROTECT_ENABLE is set, KVM_GET_DIRTY_LOG will not 7774automatically clear and write-protect all pages that are returned as dirty. 7775Rather, userspace will have to do this operation separately using 7776KVM_CLEAR_DIRTY_LOG. 7777 7778At the cost of a slightly more complicated operation, this provides better 7779scalability and responsiveness for two reasons. First, 7780KVM_CLEAR_DIRTY_LOG ioctl can operate on a 64-page granularity rather 7781than requiring to sync a full memslot; this ensures that KVM does not 7782take spinlocks for an extended period of time. Second, in some cases a 7783large amount of time can pass between a call to KVM_GET_DIRTY_LOG and 7784userspace actually using the data in the page. Pages can be modified 7785during this time, which is inefficient for both the guest and userspace: 7786the guest will incur a higher penalty due to write protection faults, 7787while userspace can see false reports of dirty pages. Manual reprotection 7788helps reducing this time, improving guest performance and reducing the 7789number of dirty log false positives. 7790 7791With KVM_DIRTY_LOG_INITIALLY_SET set, all the bits of the dirty bitmap 7792will be initialized to 1 when created. This also improves performance because 7793dirty logging can be enabled gradually in small chunks on the first call 7794to KVM_CLEAR_DIRTY_LOG. KVM_DIRTY_LOG_INITIALLY_SET depends on 7795KVM_DIRTY_LOG_MANUAL_PROTECT_ENABLE (it is also only available on 7796x86 and arm64 for now). 7797 7798KVM_CAP_MANUAL_DIRTY_LOG_PROTECT2 was previously available under the name 7799KVM_CAP_MANUAL_DIRTY_LOG_PROTECT, but the implementation had bugs that make 7800it hard or impossible to use it correctly. The availability of 7801KVM_CAP_MANUAL_DIRTY_LOG_PROTECT2 signals that those bugs are fixed. 7802Userspace should not try to use KVM_CAP_MANUAL_DIRTY_LOG_PROTECT. 7803 78047.19 KVM_CAP_PPC_SECURE_GUEST 7805------------------------------ 7806 7807:Architectures: ppc 7808 7809This capability indicates that KVM is running on a host that has 7810ultravisor firmware and thus can support a secure guest. On such a 7811system, a guest can ask the ultravisor to make it a secure guest, 7812one whose memory is inaccessible to the host except for pages which 7813are explicitly requested to be shared with the host. The ultravisor 7814notifies KVM when a guest requests to become a secure guest, and KVM 7815has the opportunity to veto the transition. 7816 7817If present, this capability can be enabled for a VM, meaning that KVM 7818will allow the transition to secure guest mode. Otherwise KVM will 7819veto the transition. 7820 78217.20 KVM_CAP_HALT_POLL 7822---------------------- 7823 7824:Architectures: all 7825:Target: VM 7826:Parameters: args[0] is the maximum poll time in nanoseconds 7827:Returns: 0 on success; -1 on error 7828 7829KVM_CAP_HALT_POLL overrides the kvm.halt_poll_ns module parameter to set the 7830maximum halt-polling time for all vCPUs in the target VM. This capability can 7831be invoked at any time and any number of times to dynamically change the 7832maximum halt-polling time. 7833 7834See Documentation/virt/kvm/halt-polling.rst for more information on halt 7835polling. 7836 78377.21 KVM_CAP_X86_USER_SPACE_MSR 7838------------------------------- 7839 7840:Architectures: x86 7841:Target: VM 7842:Parameters: args[0] contains the mask of KVM_MSR_EXIT_REASON_* events to report 7843:Returns: 0 on success; -1 on error 7844 7845This capability allows userspace to intercept RDMSR and WRMSR instructions if 7846access to an MSR is denied. By default, KVM injects #GP on denied accesses. 7847 7848When a guest requests to read or write an MSR, KVM may not implement all MSRs 7849that are relevant to a respective system. It also does not differentiate by 7850CPU type. 7851 7852To allow more fine grained control over MSR handling, userspace may enable 7853this capability. With it enabled, MSR accesses that match the mask specified in 7854args[0] and would trigger a #GP inside the guest will instead trigger 7855KVM_EXIT_X86_RDMSR and KVM_EXIT_X86_WRMSR exit notifications. Userspace 7856can then implement model specific MSR handling and/or user notifications 7857to inform a user that an MSR was not emulated/virtualized by KVM. 7858 7859The valid mask flags are: 7860 7861============================ =============================================== 7862 KVM_MSR_EXIT_REASON_UNKNOWN intercept accesses to unknown (to KVM) MSRs 7863 KVM_MSR_EXIT_REASON_INVAL intercept accesses that are architecturally 7864 invalid according to the vCPU model and/or mode 7865 KVM_MSR_EXIT_REASON_FILTER intercept accesses that are denied by userspace 7866 via KVM_X86_SET_MSR_FILTER 7867============================ =============================================== 7868 78697.22 KVM_CAP_X86_BUS_LOCK_EXIT 7870------------------------------- 7871 7872:Architectures: x86 7873:Target: VM 7874:Parameters: args[0] defines the policy used when bus locks detected in guest 7875:Returns: 0 on success, -EINVAL when args[0] contains invalid bits 7876 7877Valid bits in args[0] are:: 7878 7879 #define KVM_BUS_LOCK_DETECTION_OFF (1 << 0) 7880 #define KVM_BUS_LOCK_DETECTION_EXIT (1 << 1) 7881 7882Enabling this capability on a VM provides userspace with a way to select a 7883policy to handle the bus locks detected in guest. Userspace can obtain the 7884supported modes from the result of KVM_CHECK_EXTENSION and define it through 7885the KVM_ENABLE_CAP. The supported modes are mutually-exclusive. 7886 7887This capability allows userspace to force VM exits on bus locks detected in the 7888guest, irrespective whether or not the host has enabled split-lock detection 7889(which triggers an #AC exception that KVM intercepts). This capability is 7890intended to mitigate attacks where a malicious/buggy guest can exploit bus 7891locks to degrade the performance of the whole system. 7892 7893If KVM_BUS_LOCK_DETECTION_OFF is set, KVM doesn't force guest bus locks to VM 7894exit, although the host kernel's split-lock #AC detection still applies, if 7895enabled. 7896 7897If KVM_BUS_LOCK_DETECTION_EXIT is set, KVM enables a CPU feature that ensures 7898bus locks in the guest trigger a VM exit, and KVM exits to userspace for all 7899such VM exits, e.g. to allow userspace to throttle the offending guest and/or 7900apply some other policy-based mitigation. When exiting to userspace, KVM sets 7901KVM_RUN_X86_BUS_LOCK in vcpu-run->flags, and conditionally sets the exit_reason 7902to KVM_EXIT_X86_BUS_LOCK. 7903 7904Note! Detected bus locks may be coincident with other exits to userspace, i.e. 7905KVM_RUN_X86_BUS_LOCK should be checked regardless of the primary exit reason if 7906userspace wants to take action on all detected bus locks. 7907 79087.23 KVM_CAP_PPC_DAWR1 7909---------------------- 7910 7911:Architectures: ppc 7912:Parameters: none 7913:Returns: 0 on success, -EINVAL when CPU doesn't support 2nd DAWR 7914 7915This capability can be used to check / enable 2nd DAWR feature provided 7916by POWER10 processor. 7917 7918 79197.24 KVM_CAP_VM_COPY_ENC_CONTEXT_FROM 7920------------------------------------- 7921 7922Architectures: x86 SEV enabled 7923Type: vm 7924Parameters: args[0] is the fd of the source vm 7925Returns: 0 on success; ENOTTY on error 7926 7927This capability enables userspace to copy encryption context from the vm 7928indicated by the fd to the vm this is called on. 7929 7930This is intended to support in-guest workloads scheduled by the host. This 7931allows the in-guest workload to maintain its own NPTs and keeps the two vms 7932from accidentally clobbering each other with interrupts and the like (separate 7933APIC/MSRs/etc). 7934 79357.25 KVM_CAP_SGX_ATTRIBUTE 7936-------------------------- 7937 7938:Architectures: x86 7939:Target: VM 7940:Parameters: args[0] is a file handle of a SGX attribute file in securityfs 7941:Returns: 0 on success, -EINVAL if the file handle is invalid or if a requested 7942 attribute is not supported by KVM. 7943 7944KVM_CAP_SGX_ATTRIBUTE enables a userspace VMM to grant a VM access to one or 7945more privileged enclave attributes. args[0] must hold a file handle to a valid 7946SGX attribute file corresponding to an attribute that is supported/restricted 7947by KVM (currently only PROVISIONKEY). 7948 7949The SGX subsystem restricts access to a subset of enclave attributes to provide 7950additional security for an uncompromised kernel, e.g. use of the PROVISIONKEY 7951is restricted to deter malware from using the PROVISIONKEY to obtain a stable 7952system fingerprint. To prevent userspace from circumventing such restrictions 7953by running an enclave in a VM, KVM prevents access to privileged attributes by 7954default. 7955 7956See Documentation/arch/x86/sgx.rst for more details. 7957 79587.26 KVM_CAP_PPC_RPT_INVALIDATE 7959------------------------------- 7960 7961:Capability: KVM_CAP_PPC_RPT_INVALIDATE 7962:Architectures: ppc 7963:Type: vm 7964 7965This capability indicates that the kernel is capable of handling 7966H_RPT_INVALIDATE hcall. 7967 7968In order to enable the use of H_RPT_INVALIDATE in the guest, 7969user space might have to advertise it for the guest. For example, 7970IBM pSeries (sPAPR) guest starts using it if "hcall-rpt-invalidate" is 7971present in the "ibm,hypertas-functions" device-tree property. 7972 7973This capability is enabled for hypervisors on platforms like POWER9 7974that support radix MMU. 7975 79767.27 KVM_CAP_EXIT_ON_EMULATION_FAILURE 7977-------------------------------------- 7978 7979:Architectures: x86 7980:Parameters: args[0] whether the feature should be enabled or not 7981 7982When this capability is enabled, an emulation failure will result in an exit 7983to userspace with KVM_INTERNAL_ERROR (except when the emulator was invoked 7984to handle a VMware backdoor instruction). Furthermore, KVM will now provide up 7985to 15 instruction bytes for any exit to userspace resulting from an emulation 7986failure. When these exits to userspace occur use the emulation_failure struct 7987instead of the internal struct. They both have the same layout, but the 7988emulation_failure struct matches the content better. It also explicitly 7989defines the 'flags' field which is used to describe the fields in the struct 7990that are valid (ie: if KVM_INTERNAL_ERROR_EMULATION_FLAG_INSTRUCTION_BYTES is 7991set in the 'flags' field then both 'insn_size' and 'insn_bytes' have valid data 7992in them.) 7993 79947.28 KVM_CAP_ARM_MTE 7995-------------------- 7996 7997:Architectures: arm64 7998:Parameters: none 7999 8000This capability indicates that KVM (and the hardware) supports exposing the 8001Memory Tagging Extensions (MTE) to the guest. It must also be enabled by the 8002VMM before creating any VCPUs to allow the guest access. Note that MTE is only 8003available to a guest running in AArch64 mode and enabling this capability will 8004cause attempts to create AArch32 VCPUs to fail. 8005 8006When enabled the guest is able to access tags associated with any memory given 8007to the guest. KVM will ensure that the tags are maintained during swap or 8008hibernation of the host; however the VMM needs to manually save/restore the 8009tags as appropriate if the VM is migrated. 8010 8011When this capability is enabled all memory in memslots must be mapped as 8012``MAP_ANONYMOUS`` or with a RAM-based file mapping (``tmpfs``, ``memfd``), 8013attempts to create a memslot with an invalid mmap will result in an 8014-EINVAL return. 8015 8016When enabled the VMM may make use of the ``KVM_ARM_MTE_COPY_TAGS`` ioctl to 8017perform a bulk copy of tags to/from the guest. 8018 80197.29 KVM_CAP_VM_MOVE_ENC_CONTEXT_FROM 8020------------------------------------- 8021 8022:Architectures: x86 SEV enabled 8023:Type: vm 8024:Parameters: args[0] is the fd of the source vm 8025:Returns: 0 on success 8026 8027This capability enables userspace to migrate the encryption context from the VM 8028indicated by the fd to the VM this is called on. 8029 8030This is intended to support intra-host migration of VMs between userspace VMMs, 8031upgrading the VMM process without interrupting the guest. 8032 80337.30 KVM_CAP_PPC_AIL_MODE_3 8034------------------------------- 8035 8036:Capability: KVM_CAP_PPC_AIL_MODE_3 8037:Architectures: ppc 8038:Type: vm 8039 8040This capability indicates that the kernel supports the mode 3 setting for the 8041"Address Translation Mode on Interrupt" aka "Alternate Interrupt Location" 8042resource that is controlled with the H_SET_MODE hypercall. 8043 8044This capability allows a guest kernel to use a better-performance mode for 8045handling interrupts and system calls. 8046 80477.31 KVM_CAP_DISABLE_QUIRKS2 8048---------------------------- 8049 8050:Capability: KVM_CAP_DISABLE_QUIRKS2 8051:Parameters: args[0] - set of KVM quirks to disable 8052:Architectures: x86 8053:Type: vm 8054 8055This capability, if enabled, will cause KVM to disable some behavior 8056quirks. 8057 8058Calling KVM_CHECK_EXTENSION for this capability returns a bitmask of 8059quirks that can be disabled in KVM. 8060 8061The argument to KVM_ENABLE_CAP for this capability is a bitmask of 8062quirks to disable, and must be a subset of the bitmask returned by 8063KVM_CHECK_EXTENSION. 8064 8065The valid bits in cap.args[0] are: 8066 8067=================================== ============================================ 8068 KVM_X86_QUIRK_LINT0_REENABLED By default, the reset value for the LVT 8069 LINT0 register is 0x700 (APIC_MODE_EXTINT). 8070 When this quirk is disabled, the reset value 8071 is 0x10000 (APIC_LVT_MASKED). 8072 8073 KVM_X86_QUIRK_CD_NW_CLEARED By default, KVM clears CR0.CD and CR0.NW on 8074 AMD CPUs to workaround buggy guest firmware 8075 that runs in perpetuity with CR0.CD, i.e. 8076 with caches in "no fill" mode. 8077 8078 When this quirk is disabled, KVM does not 8079 change the value of CR0.CD and CR0.NW. 8080 8081 KVM_X86_QUIRK_LAPIC_MMIO_HOLE By default, the MMIO LAPIC interface is 8082 available even when configured for x2APIC 8083 mode. When this quirk is disabled, KVM 8084 disables the MMIO LAPIC interface if the 8085 LAPIC is in x2APIC mode. 8086 8087 KVM_X86_QUIRK_OUT_7E_INC_RIP By default, KVM pre-increments %rip before 8088 exiting to userspace for an OUT instruction 8089 to port 0x7e. When this quirk is disabled, 8090 KVM does not pre-increment %rip before 8091 exiting to userspace. 8092 8093 KVM_X86_QUIRK_MISC_ENABLE_NO_MWAIT When this quirk is disabled, KVM sets 8094 CPUID.01H:ECX[bit 3] (MONITOR/MWAIT) if 8095 IA32_MISC_ENABLE[bit 18] (MWAIT) is set. 8096 Additionally, when this quirk is disabled, 8097 KVM clears CPUID.01H:ECX[bit 3] if 8098 IA32_MISC_ENABLE[bit 18] is cleared. 8099 8100 KVM_X86_QUIRK_FIX_HYPERCALL_INSN By default, KVM rewrites guest 8101 VMMCALL/VMCALL instructions to match the 8102 vendor's hypercall instruction for the 8103 system. When this quirk is disabled, KVM 8104 will no longer rewrite invalid guest 8105 hypercall instructions. Executing the 8106 incorrect hypercall instruction will 8107 generate a #UD within the guest. 8108 8109KVM_X86_QUIRK_MWAIT_NEVER_UD_FAULTS By default, KVM emulates MONITOR/MWAIT (if 8110 they are intercepted) as NOPs regardless of 8111 whether or not MONITOR/MWAIT are supported 8112 according to guest CPUID. When this quirk 8113 is disabled and KVM_X86_DISABLE_EXITS_MWAIT 8114 is not set (MONITOR/MWAIT are intercepted), 8115 KVM will inject a #UD on MONITOR/MWAIT if 8116 they're unsupported per guest CPUID. Note, 8117 KVM will modify MONITOR/MWAIT support in 8118 guest CPUID on writes to MISC_ENABLE if 8119 KVM_X86_QUIRK_MISC_ENABLE_NO_MWAIT is 8120 disabled. 8121 8122KVM_X86_QUIRK_SLOT_ZAP_ALL By default, for KVM_X86_DEFAULT_VM VMs, KVM 8123 invalidates all SPTEs in all memslots and 8124 address spaces when a memslot is deleted or 8125 moved. When this quirk is disabled (or the 8126 VM type isn't KVM_X86_DEFAULT_VM), KVM only 8127 ensures the backing memory of the deleted 8128 or moved memslot isn't reachable, i.e KVM 8129 _may_ invalidate only SPTEs related to the 8130 memslot. 8131 8132KVM_X86_QUIRK_STUFF_FEATURE_MSRS By default, at vCPU creation, KVM sets the 8133 vCPU's MSR_IA32_PERF_CAPABILITIES (0x345), 8134 MSR_IA32_ARCH_CAPABILITIES (0x10a), 8135 MSR_PLATFORM_INFO (0xce), and all VMX MSRs 8136 (0x480..0x492) to the maximal capabilities 8137 supported by KVM. KVM also sets 8138 MSR_IA32_UCODE_REV (0x8b) to an arbitrary 8139 value (which is different for Intel vs. 8140 AMD). Lastly, when guest CPUID is set (by 8141 userspace), KVM modifies select VMX MSR 8142 fields to force consistency between guest 8143 CPUID and L2's effective ISA. When this 8144 quirk is disabled, KVM zeroes the vCPU's MSR 8145 values (with two exceptions, see below), 8146 i.e. treats the feature MSRs like CPUID 8147 leaves and gives userspace full control of 8148 the vCPU model definition. This quirk does 8149 not affect VMX MSRs CR0/CR4_FIXED1 (0x487 8150 and 0x489), as KVM does now allow them to 8151 be set by userspace (KVM sets them based on 8152 guest CPUID, for safety purposes). 8153=================================== ============================================ 8154 81557.32 KVM_CAP_MAX_VCPU_ID 8156------------------------ 8157 8158:Architectures: x86 8159:Target: VM 8160:Parameters: args[0] - maximum APIC ID value set for current VM 8161:Returns: 0 on success, -EINVAL if args[0] is beyond KVM_MAX_VCPU_IDS 8162 supported in KVM or if it has been set. 8163 8164This capability allows userspace to specify maximum possible APIC ID 8165assigned for current VM session prior to the creation of vCPUs, saving 8166memory for data structures indexed by the APIC ID. Userspace is able 8167to calculate the limit to APIC ID values from designated 8168CPU topology. 8169 8170The value can be changed only until KVM_ENABLE_CAP is set to a nonzero 8171value or until a vCPU is created. Upon creation of the first vCPU, 8172if the value was set to zero or KVM_ENABLE_CAP was not invoked, KVM 8173uses the return value of KVM_CHECK_EXTENSION(KVM_CAP_MAX_VCPU_ID) as 8174the maximum APIC ID. 8175 81767.33 KVM_CAP_X86_NOTIFY_VMEXIT 8177------------------------------ 8178 8179:Architectures: x86 8180:Target: VM 8181:Parameters: args[0] is the value of notify window as well as some flags 8182:Returns: 0 on success, -EINVAL if args[0] contains invalid flags or notify 8183 VM exit is unsupported. 8184 8185Bits 63:32 of args[0] are used for notify window. 8186Bits 31:0 of args[0] are for some flags. Valid bits are:: 8187 8188 #define KVM_X86_NOTIFY_VMEXIT_ENABLED (1 << 0) 8189 #define KVM_X86_NOTIFY_VMEXIT_USER (1 << 1) 8190 8191This capability allows userspace to configure the notify VM exit on/off 8192in per-VM scope during VM creation. Notify VM exit is disabled by default. 8193When userspace sets KVM_X86_NOTIFY_VMEXIT_ENABLED bit in args[0], VMM will 8194enable this feature with the notify window provided, which will generate 8195a VM exit if no event window occurs in VM non-root mode for a specified of 8196time (notify window). 8197 8198If KVM_X86_NOTIFY_VMEXIT_USER is set in args[0], upon notify VM exits happen, 8199KVM would exit to userspace for handling. 8200 8201This capability is aimed to mitigate the threat that malicious VMs can 8202cause CPU stuck (due to event windows don't open up) and make the CPU 8203unavailable to host or other VMs. 8204 82057.34 KVM_CAP_MEMORY_FAULT_INFO 8206------------------------------ 8207 8208:Architectures: x86 8209:Returns: Informational only, -EINVAL on direct KVM_ENABLE_CAP. 8210 8211The presence of this capability indicates that KVM_RUN will fill 8212kvm_run.memory_fault if KVM cannot resolve a guest page fault VM-Exit, e.g. if 8213there is a valid memslot but no backing VMA for the corresponding host virtual 8214address. 8215 8216The information in kvm_run.memory_fault is valid if and only if KVM_RUN returns 8217an error with errno=EFAULT or errno=EHWPOISON *and* kvm_run.exit_reason is set 8218to KVM_EXIT_MEMORY_FAULT. 8219 8220Note: Userspaces which attempt to resolve memory faults so that they can retry 8221KVM_RUN are encouraged to guard against repeatedly receiving the same 8222error/annotated fault. 8223 8224See KVM_EXIT_MEMORY_FAULT for more information. 8225 82267.35 KVM_CAP_X86_APIC_BUS_CYCLES_NS 8227----------------------------------- 8228 8229:Architectures: x86 8230:Target: VM 8231:Parameters: args[0] is the desired APIC bus clock rate, in nanoseconds 8232:Returns: 0 on success, -EINVAL if args[0] contains an invalid value for the 8233 frequency or if any vCPUs have been created, -ENXIO if a virtual 8234 local APIC has not been created using KVM_CREATE_IRQCHIP. 8235 8236This capability sets the VM's APIC bus clock frequency, used by KVM's in-kernel 8237virtual APIC when emulating APIC timers. KVM's default value can be retrieved 8238by KVM_CHECK_EXTENSION. 8239 8240Note: Userspace is responsible for correctly configuring CPUID 0x15, a.k.a. the 8241core crystal clock frequency, if a non-zero CPUID 0x15 is exposed to the guest. 8242 82437.36 KVM_CAP_X86_GUEST_MODE 8244------------------------------ 8245 8246:Architectures: x86 8247:Returns: Informational only, -EINVAL on direct KVM_ENABLE_CAP. 8248 8249The presence of this capability indicates that KVM_RUN will update the 8250KVM_RUN_X86_GUEST_MODE bit in kvm_run.flags to indicate whether the 8251vCPU was executing nested guest code when it exited. 8252 8253KVM exits with the register state of either the L1 or L2 guest 8254depending on which executed at the time of an exit. Userspace must 8255take care to differentiate between these cases. 8256 82578. Other capabilities. 8258====================== 8259 8260This section lists capabilities that give information about other 8261features of the KVM implementation. 8262 82638.1 KVM_CAP_PPC_HWRNG 8264--------------------- 8265 8266:Architectures: ppc 8267 8268This capability, if KVM_CHECK_EXTENSION indicates that it is 8269available, means that the kernel has an implementation of the 8270H_RANDOM hypercall backed by a hardware random-number generator. 8271If present, the kernel H_RANDOM handler can be enabled for guest use 8272with the KVM_CAP_PPC_ENABLE_HCALL capability. 8273 82748.2 KVM_CAP_HYPERV_SYNIC 8275------------------------ 8276 8277:Architectures: x86 8278 8279This capability, if KVM_CHECK_EXTENSION indicates that it is 8280available, means that the kernel has an implementation of the 8281Hyper-V Synthetic interrupt controller(SynIC). Hyper-V SynIC is 8282used to support Windows Hyper-V based guest paravirt drivers(VMBus). 8283 8284In order to use SynIC, it has to be activated by setting this 8285capability via KVM_ENABLE_CAP ioctl on the vcpu fd. Note that this 8286will disable the use of APIC hardware virtualization even if supported 8287by the CPU, as it's incompatible with SynIC auto-EOI behavior. 8288 82898.3 KVM_CAP_PPC_MMU_RADIX 8290------------------------- 8291 8292:Architectures: ppc 8293 8294This capability, if KVM_CHECK_EXTENSION indicates that it is 8295available, means that the kernel can support guests using the 8296radix MMU defined in Power ISA V3.00 (as implemented in the POWER9 8297processor). 8298 82998.4 KVM_CAP_PPC_MMU_HASH_V3 8300--------------------------- 8301 8302:Architectures: ppc 8303 8304This capability, if KVM_CHECK_EXTENSION indicates that it is 8305available, means that the kernel can support guests using the 8306hashed page table MMU defined in Power ISA V3.00 (as implemented in 8307the POWER9 processor), including in-memory segment tables. 8308 83098.5 KVM_CAP_MIPS_VZ 8310------------------- 8311 8312:Architectures: mips 8313 8314This capability, if KVM_CHECK_EXTENSION on the main kvm handle indicates that 8315it is available, means that full hardware assisted virtualization capabilities 8316of the hardware are available for use through KVM. An appropriate 8317KVM_VM_MIPS_* type must be passed to KVM_CREATE_VM to create a VM which 8318utilises it. 8319 8320If KVM_CHECK_EXTENSION on a kvm VM handle indicates that this capability is 8321available, it means that the VM is using full hardware assisted virtualization 8322capabilities of the hardware. This is useful to check after creating a VM with 8323KVM_VM_MIPS_DEFAULT. 8324 8325The value returned by KVM_CHECK_EXTENSION should be compared against known 8326values (see below). All other values are reserved. This is to allow for the 8327possibility of other hardware assisted virtualization implementations which 8328may be incompatible with the MIPS VZ ASE. 8329 8330== ========================================================================== 8331 0 The trap & emulate implementation is in use to run guest code in user 8332 mode. Guest virtual memory segments are rearranged to fit the guest in the 8333 user mode address space. 8334 8335 1 The MIPS VZ ASE is in use, providing full hardware assisted 8336 virtualization, including standard guest virtual memory segments. 8337== ========================================================================== 8338 83398.6 KVM_CAP_MIPS_TE 8340------------------- 8341 8342:Architectures: mips 8343 8344This capability, if KVM_CHECK_EXTENSION on the main kvm handle indicates that 8345it is available, means that the trap & emulate implementation is available to 8346run guest code in user mode, even if KVM_CAP_MIPS_VZ indicates that hardware 8347assisted virtualisation is also available. KVM_VM_MIPS_TE (0) must be passed 8348to KVM_CREATE_VM to create a VM which utilises it. 8349 8350If KVM_CHECK_EXTENSION on a kvm VM handle indicates that this capability is 8351available, it means that the VM is using trap & emulate. 8352 83538.7 KVM_CAP_MIPS_64BIT 8354---------------------- 8355 8356:Architectures: mips 8357 8358This capability indicates the supported architecture type of the guest, i.e. the 8359supported register and address width. 8360 8361The values returned when this capability is checked by KVM_CHECK_EXTENSION on a 8362kvm VM handle correspond roughly to the CP0_Config.AT register field, and should 8363be checked specifically against known values (see below). All other values are 8364reserved. 8365 8366== ======================================================================== 8367 0 MIPS32 or microMIPS32. 8368 Both registers and addresses are 32-bits wide. 8369 It will only be possible to run 32-bit guest code. 8370 8371 1 MIPS64 or microMIPS64 with access only to 32-bit compatibility segments. 8372 Registers are 64-bits wide, but addresses are 32-bits wide. 8373 64-bit guest code may run but cannot access MIPS64 memory segments. 8374 It will also be possible to run 32-bit guest code. 8375 8376 2 MIPS64 or microMIPS64 with access to all address segments. 8377 Both registers and addresses are 64-bits wide. 8378 It will be possible to run 64-bit or 32-bit guest code. 8379== ======================================================================== 8380 83818.9 KVM_CAP_ARM_USER_IRQ 8382------------------------ 8383 8384:Architectures: arm64 8385 8386This capability, if KVM_CHECK_EXTENSION indicates that it is available, means 8387that if userspace creates a VM without an in-kernel interrupt controller, it 8388will be notified of changes to the output level of in-kernel emulated devices, 8389which can generate virtual interrupts, presented to the VM. 8390For such VMs, on every return to userspace, the kernel 8391updates the vcpu's run->s.regs.device_irq_level field to represent the actual 8392output level of the device. 8393 8394Whenever kvm detects a change in the device output level, kvm guarantees at 8395least one return to userspace before running the VM. This exit could either 8396be a KVM_EXIT_INTR or any other exit event, like KVM_EXIT_MMIO. This way, 8397userspace can always sample the device output level and re-compute the state of 8398the userspace interrupt controller. Userspace should always check the state 8399of run->s.regs.device_irq_level on every kvm exit. 8400The value in run->s.regs.device_irq_level can represent both level and edge 8401triggered interrupt signals, depending on the device. Edge triggered interrupt 8402signals will exit to userspace with the bit in run->s.regs.device_irq_level 8403set exactly once per edge signal. 8404 8405The field run->s.regs.device_irq_level is available independent of 8406run->kvm_valid_regs or run->kvm_dirty_regs bits. 8407 8408If KVM_CAP_ARM_USER_IRQ is supported, the KVM_CHECK_EXTENSION ioctl returns a 8409number larger than 0 indicating the version of this capability is implemented 8410and thereby which bits in run->s.regs.device_irq_level can signal values. 8411 8412Currently the following bits are defined for the device_irq_level bitmap:: 8413 8414 KVM_CAP_ARM_USER_IRQ >= 1: 8415 8416 KVM_ARM_DEV_EL1_VTIMER - EL1 virtual timer 8417 KVM_ARM_DEV_EL1_PTIMER - EL1 physical timer 8418 KVM_ARM_DEV_PMU - ARM PMU overflow interrupt signal 8419 8420Future versions of kvm may implement additional events. These will get 8421indicated by returning a higher number from KVM_CHECK_EXTENSION and will be 8422listed above. 8423 84248.10 KVM_CAP_PPC_SMT_POSSIBLE 8425----------------------------- 8426 8427:Architectures: ppc 8428 8429Querying this capability returns a bitmap indicating the possible 8430virtual SMT modes that can be set using KVM_CAP_PPC_SMT. If bit N 8431(counting from the right) is set, then a virtual SMT mode of 2^N is 8432available. 8433 84348.11 KVM_CAP_HYPERV_SYNIC2 8435-------------------------- 8436 8437:Architectures: x86 8438 8439This capability enables a newer version of Hyper-V Synthetic interrupt 8440controller (SynIC). The only difference with KVM_CAP_HYPERV_SYNIC is that KVM 8441doesn't clear SynIC message and event flags pages when they are enabled by 8442writing to the respective MSRs. 8443 84448.12 KVM_CAP_HYPERV_VP_INDEX 8445---------------------------- 8446 8447:Architectures: x86 8448 8449This capability indicates that userspace can load HV_X64_MSR_VP_INDEX msr. Its 8450value is used to denote the target vcpu for a SynIC interrupt. For 8451compatibility, KVM initializes this msr to KVM's internal vcpu index. When this 8452capability is absent, userspace can still query this msr's value. 8453 84548.13 KVM_CAP_S390_AIS_MIGRATION 8455------------------------------- 8456 8457:Architectures: s390 8458:Parameters: none 8459 8460This capability indicates if the flic device will be able to get/set the 8461AIS states for migration via the KVM_DEV_FLIC_AISM_ALL attribute and allows 8462to discover this without having to create a flic device. 8463 84648.14 KVM_CAP_S390_PSW 8465--------------------- 8466 8467:Architectures: s390 8468 8469This capability indicates that the PSW is exposed via the kvm_run structure. 8470 84718.15 KVM_CAP_S390_GMAP 8472---------------------- 8473 8474:Architectures: s390 8475 8476This capability indicates that the user space memory used as guest mapping can 8477be anywhere in the user memory address space, as long as the memory slots are 8478aligned and sized to a segment (1MB) boundary. 8479 84808.16 KVM_CAP_S390_COW 8481--------------------- 8482 8483:Architectures: s390 8484 8485This capability indicates that the user space memory used as guest mapping can 8486use copy-on-write semantics as well as dirty pages tracking via read-only page 8487tables. 8488 84898.17 KVM_CAP_S390_BPB 8490--------------------- 8491 8492:Architectures: s390 8493 8494This capability indicates that kvm will implement the interfaces to handle 8495reset, migration and nested KVM for branch prediction blocking. The stfle 8496facility 82 should not be provided to the guest without this capability. 8497 84988.18 KVM_CAP_HYPERV_TLBFLUSH 8499---------------------------- 8500 8501:Architectures: x86 8502 8503This capability indicates that KVM supports paravirtualized Hyper-V TLB Flush 8504hypercalls: 8505HvFlushVirtualAddressSpace, HvFlushVirtualAddressSpaceEx, 8506HvFlushVirtualAddressList, HvFlushVirtualAddressListEx. 8507 85088.19 KVM_CAP_ARM_INJECT_SERROR_ESR 8509---------------------------------- 8510 8511:Architectures: arm64 8512 8513This capability indicates that userspace can specify (via the 8514KVM_SET_VCPU_EVENTS ioctl) the syndrome value reported to the guest when it 8515takes a virtual SError interrupt exception. 8516If KVM advertises this capability, userspace can only specify the ISS field for 8517the ESR syndrome. Other parts of the ESR, such as the EC are generated by the 8518CPU when the exception is taken. If this virtual SError is taken to EL1 using 8519AArch64, this value will be reported in the ISS field of ESR_ELx. 8520 8521See KVM_CAP_VCPU_EVENTS for more details. 8522 85238.20 KVM_CAP_HYPERV_SEND_IPI 8524---------------------------- 8525 8526:Architectures: x86 8527 8528This capability indicates that KVM supports paravirtualized Hyper-V IPI send 8529hypercalls: 8530HvCallSendSyntheticClusterIpi, HvCallSendSyntheticClusterIpiEx. 8531 85328.21 KVM_CAP_HYPERV_DIRECT_TLBFLUSH 8533----------------------------------- 8534 8535:Architectures: x86 8536 8537This capability indicates that KVM running on top of Hyper-V hypervisor 8538enables Direct TLB flush for its guests meaning that TLB flush 8539hypercalls are handled by Level 0 hypervisor (Hyper-V) bypassing KVM. 8540Due to the different ABI for hypercall parameters between Hyper-V and 8541KVM, enabling this capability effectively disables all hypercall 8542handling by KVM (as some KVM hypercall may be mistakenly treated as TLB 8543flush hypercalls by Hyper-V) so userspace should disable KVM identification 8544in CPUID and only exposes Hyper-V identification. In this case, guest 8545thinks it's running on Hyper-V and only use Hyper-V hypercalls. 8546 85478.22 KVM_CAP_S390_VCPU_RESETS 8548----------------------------- 8549 8550:Architectures: s390 8551 8552This capability indicates that the KVM_S390_NORMAL_RESET and 8553KVM_S390_CLEAR_RESET ioctls are available. 8554 85558.23 KVM_CAP_S390_PROTECTED 8556--------------------------- 8557 8558:Architectures: s390 8559 8560This capability indicates that the Ultravisor has been initialized and 8561KVM can therefore start protected VMs. 8562This capability governs the KVM_S390_PV_COMMAND ioctl and the 8563KVM_MP_STATE_LOAD MP_STATE. KVM_SET_MP_STATE can fail for protected 8564guests when the state change is invalid. 8565 85668.24 KVM_CAP_STEAL_TIME 8567----------------------- 8568 8569:Architectures: arm64, x86 8570 8571This capability indicates that KVM supports steal time accounting. 8572When steal time accounting is supported it may be enabled with 8573architecture-specific interfaces. This capability and the architecture- 8574specific interfaces must be consistent, i.e. if one says the feature 8575is supported, than the other should as well and vice versa. For arm64 8576see Documentation/virt/kvm/devices/vcpu.rst "KVM_ARM_VCPU_PVTIME_CTRL". 8577For x86 see Documentation/virt/kvm/x86/msr.rst "MSR_KVM_STEAL_TIME". 8578 85798.25 KVM_CAP_S390_DIAG318 8580------------------------- 8581 8582:Architectures: s390 8583 8584This capability enables a guest to set information about its control program 8585(i.e. guest kernel type and version). The information is helpful during 8586system/firmware service events, providing additional data about the guest 8587environments running on the machine. 8588 8589The information is associated with the DIAGNOSE 0x318 instruction, which sets 8590an 8-byte value consisting of a one-byte Control Program Name Code (CPNC) and 8591a 7-byte Control Program Version Code (CPVC). The CPNC determines what 8592environment the control program is running in (e.g. Linux, z/VM...), and the 8593CPVC is used for information specific to OS (e.g. Linux version, Linux 8594distribution...) 8595 8596If this capability is available, then the CPNC and CPVC can be synchronized 8597between KVM and userspace via the sync regs mechanism (KVM_SYNC_DIAG318). 8598 85998.26 KVM_CAP_X86_USER_SPACE_MSR 8600------------------------------- 8601 8602:Architectures: x86 8603 8604This capability indicates that KVM supports deflection of MSR reads and 8605writes to user space. It can be enabled on a VM level. If enabled, MSR 8606accesses that would usually trigger a #GP by KVM into the guest will 8607instead get bounced to user space through the KVM_EXIT_X86_RDMSR and 8608KVM_EXIT_X86_WRMSR exit notifications. 8609 86108.27 KVM_CAP_X86_MSR_FILTER 8611--------------------------- 8612 8613:Architectures: x86 8614 8615This capability indicates that KVM supports that accesses to user defined MSRs 8616may be rejected. With this capability exposed, KVM exports new VM ioctl 8617KVM_X86_SET_MSR_FILTER which user space can call to specify bitmaps of MSR 8618ranges that KVM should deny access to. 8619 8620In combination with KVM_CAP_X86_USER_SPACE_MSR, this allows user space to 8621trap and emulate MSRs that are outside of the scope of KVM as well as 8622limit the attack surface on KVM's MSR emulation code. 8623 86248.28 KVM_CAP_ENFORCE_PV_FEATURE_CPUID 8625------------------------------------- 8626 8627Architectures: x86 8628 8629When enabled, KVM will disable paravirtual features provided to the 8630guest according to the bits in the KVM_CPUID_FEATURES CPUID leaf 8631(0x40000001). Otherwise, a guest may use the paravirtual features 8632regardless of what has actually been exposed through the CPUID leaf. 8633 8634.. _KVM_CAP_DIRTY_LOG_RING: 8635 86368.29 KVM_CAP_DIRTY_LOG_RING/KVM_CAP_DIRTY_LOG_RING_ACQ_REL 8637---------------------------------------------------------- 8638 8639:Architectures: x86, arm64 8640:Parameters: args[0] - size of the dirty log ring 8641 8642KVM is capable of tracking dirty memory using ring buffers that are 8643mmapped into userspace; there is one dirty ring per vcpu. 8644 8645The dirty ring is available to userspace as an array of 8646``struct kvm_dirty_gfn``. Each dirty entry is defined as:: 8647 8648 struct kvm_dirty_gfn { 8649 __u32 flags; 8650 __u32 slot; /* as_id | slot_id */ 8651 __u64 offset; 8652 }; 8653 8654The following values are defined for the flags field to define the 8655current state of the entry:: 8656 8657 #define KVM_DIRTY_GFN_F_DIRTY BIT(0) 8658 #define KVM_DIRTY_GFN_F_RESET BIT(1) 8659 #define KVM_DIRTY_GFN_F_MASK 0x3 8660 8661Userspace should call KVM_ENABLE_CAP ioctl right after KVM_CREATE_VM 8662ioctl to enable this capability for the new guest and set the size of 8663the rings. Enabling the capability is only allowed before creating any 8664vCPU, and the size of the ring must be a power of two. The larger the 8665ring buffer, the less likely the ring is full and the VM is forced to 8666exit to userspace. The optimal size depends on the workload, but it is 8667recommended that it be at least 64 KiB (4096 entries). 8668 8669Just like for dirty page bitmaps, the buffer tracks writes to 8670all user memory regions for which the KVM_MEM_LOG_DIRTY_PAGES flag was 8671set in KVM_SET_USER_MEMORY_REGION. Once a memory region is registered 8672with the flag set, userspace can start harvesting dirty pages from the 8673ring buffer. 8674 8675An entry in the ring buffer can be unused (flag bits ``00``), 8676dirty (flag bits ``01``) or harvested (flag bits ``1X``). The 8677state machine for the entry is as follows:: 8678 8679 dirtied harvested reset 8680 00 -----------> 01 -------------> 1X -------+ 8681 ^ | 8682 | | 8683 +------------------------------------------+ 8684 8685To harvest the dirty pages, userspace accesses the mmapped ring buffer 8686to read the dirty GFNs. If the flags has the DIRTY bit set (at this stage 8687the RESET bit must be cleared), then it means this GFN is a dirty GFN. 8688The userspace should harvest this GFN and mark the flags from state 8689``01b`` to ``1Xb`` (bit 0 will be ignored by KVM, but bit 1 must be set 8690to show that this GFN is harvested and waiting for a reset), and move 8691on to the next GFN. The userspace should continue to do this until the 8692flags of a GFN have the DIRTY bit cleared, meaning that it has harvested 8693all the dirty GFNs that were available. 8694 8695Note that on weakly ordered architectures, userspace accesses to the 8696ring buffer (and more specifically the 'flags' field) must be ordered, 8697using load-acquire/store-release accessors when available, or any 8698other memory barrier that will ensure this ordering. 8699 8700It's not necessary for userspace to harvest the all dirty GFNs at once. 8701However it must collect the dirty GFNs in sequence, i.e., the userspace 8702program cannot skip one dirty GFN to collect the one next to it. 8703 8704After processing one or more entries in the ring buffer, userspace 8705calls the VM ioctl KVM_RESET_DIRTY_RINGS to notify the kernel about 8706it, so that the kernel will reprotect those collected GFNs. 8707Therefore, the ioctl must be called *before* reading the content of 8708the dirty pages. 8709 8710The dirty ring can get full. When it happens, the KVM_RUN of the 8711vcpu will return with exit reason KVM_EXIT_DIRTY_LOG_FULL. 8712 8713The dirty ring interface has a major difference comparing to the 8714KVM_GET_DIRTY_LOG interface in that, when reading the dirty ring from 8715userspace, it's still possible that the kernel has not yet flushed the 8716processor's dirty page buffers into the kernel buffer (with dirty bitmaps, the 8717flushing is done by the KVM_GET_DIRTY_LOG ioctl). To achieve that, one 8718needs to kick the vcpu out of KVM_RUN using a signal. The resulting 8719vmexit ensures that all dirty GFNs are flushed to the dirty rings. 8720 8721NOTE: KVM_CAP_DIRTY_LOG_RING_ACQ_REL is the only capability that 8722should be exposed by weakly ordered architecture, in order to indicate 8723the additional memory ordering requirements imposed on userspace when 8724reading the state of an entry and mutating it from DIRTY to HARVESTED. 8725Architecture with TSO-like ordering (such as x86) are allowed to 8726expose both KVM_CAP_DIRTY_LOG_RING and KVM_CAP_DIRTY_LOG_RING_ACQ_REL 8727to userspace. 8728 8729After enabling the dirty rings, the userspace needs to detect the 8730capability of KVM_CAP_DIRTY_LOG_RING_WITH_BITMAP to see whether the 8731ring structures can be backed by per-slot bitmaps. With this capability 8732advertised, it means the architecture can dirty guest pages without 8733vcpu/ring context, so that some of the dirty information will still be 8734maintained in the bitmap structure. KVM_CAP_DIRTY_LOG_RING_WITH_BITMAP 8735can't be enabled if the capability of KVM_CAP_DIRTY_LOG_RING_ACQ_REL 8736hasn't been enabled, or any memslot has been existing. 8737 8738Note that the bitmap here is only a backup of the ring structure. The 8739use of the ring and bitmap combination is only beneficial if there is 8740only a very small amount of memory that is dirtied out of vcpu/ring 8741context. Otherwise, the stand-alone per-slot bitmap mechanism needs to 8742be considered. 8743 8744To collect dirty bits in the backup bitmap, userspace can use the same 8745KVM_GET_DIRTY_LOG ioctl. KVM_CLEAR_DIRTY_LOG isn't needed as long as all 8746the generation of the dirty bits is done in a single pass. Collecting 8747the dirty bitmap should be the very last thing that the VMM does before 8748considering the state as complete. VMM needs to ensure that the dirty 8749state is final and avoid missing dirty pages from another ioctl ordered 8750after the bitmap collection. 8751 8752NOTE: Multiple examples of using the backup bitmap: (1) save vgic/its 8753tables through command KVM_DEV_ARM_{VGIC_GRP_CTRL, ITS_SAVE_TABLES} on 8754KVM device "kvm-arm-vgic-its". (2) restore vgic/its tables through 8755command KVM_DEV_ARM_{VGIC_GRP_CTRL, ITS_RESTORE_TABLES} on KVM device 8756"kvm-arm-vgic-its". VGICv3 LPI pending status is restored. (3) save 8757vgic3 pending table through KVM_DEV_ARM_VGIC_{GRP_CTRL, SAVE_PENDING_TABLES} 8758command on KVM device "kvm-arm-vgic-v3". 8759 87608.30 KVM_CAP_XEN_HVM 8761-------------------- 8762 8763:Architectures: x86 8764 8765This capability indicates the features that Xen supports for hosting Xen 8766PVHVM guests. Valid flags are:: 8767 8768 #define KVM_XEN_HVM_CONFIG_HYPERCALL_MSR (1 << 0) 8769 #define KVM_XEN_HVM_CONFIG_INTERCEPT_HCALL (1 << 1) 8770 #define KVM_XEN_HVM_CONFIG_SHARED_INFO (1 << 2) 8771 #define KVM_XEN_HVM_CONFIG_RUNSTATE (1 << 3) 8772 #define KVM_XEN_HVM_CONFIG_EVTCHN_2LEVEL (1 << 4) 8773 #define KVM_XEN_HVM_CONFIG_EVTCHN_SEND (1 << 5) 8774 #define KVM_XEN_HVM_CONFIG_RUNSTATE_UPDATE_FLAG (1 << 6) 8775 #define KVM_XEN_HVM_CONFIG_PVCLOCK_TSC_UNSTABLE (1 << 7) 8776 8777The KVM_XEN_HVM_CONFIG_HYPERCALL_MSR flag indicates that the KVM_XEN_HVM_CONFIG 8778ioctl is available, for the guest to set its hypercall page. 8779 8780If KVM_XEN_HVM_CONFIG_INTERCEPT_HCALL is also set, the same flag may also be 8781provided in the flags to KVM_XEN_HVM_CONFIG, without providing hypercall page 8782contents, to request that KVM generate hypercall page content automatically 8783and also enable interception of guest hypercalls with KVM_EXIT_XEN. 8784 8785The KVM_XEN_HVM_CONFIG_SHARED_INFO flag indicates the availability of the 8786KVM_XEN_HVM_SET_ATTR, KVM_XEN_HVM_GET_ATTR, KVM_XEN_VCPU_SET_ATTR and 8787KVM_XEN_VCPU_GET_ATTR ioctls, as well as the delivery of exception vectors 8788for event channel upcalls when the evtchn_upcall_pending field of a vcpu's 8789vcpu_info is set. 8790 8791The KVM_XEN_HVM_CONFIG_RUNSTATE flag indicates that the runstate-related 8792features KVM_XEN_VCPU_ATTR_TYPE_RUNSTATE_ADDR/_CURRENT/_DATA/_ADJUST are 8793supported by the KVM_XEN_VCPU_SET_ATTR/KVM_XEN_VCPU_GET_ATTR ioctls. 8794 8795The KVM_XEN_HVM_CONFIG_EVTCHN_2LEVEL flag indicates that IRQ routing entries 8796of the type KVM_IRQ_ROUTING_XEN_EVTCHN are supported, with the priority 8797field set to indicate 2 level event channel delivery. 8798 8799The KVM_XEN_HVM_CONFIG_EVTCHN_SEND flag indicates that KVM supports 8800injecting event channel events directly into the guest with the 8801KVM_XEN_HVM_EVTCHN_SEND ioctl. It also indicates support for the 8802KVM_XEN_ATTR_TYPE_EVTCHN/XEN_VERSION HVM attributes and the 8803KVM_XEN_VCPU_ATTR_TYPE_VCPU_ID/TIMER/UPCALL_VECTOR vCPU attributes. 8804related to event channel delivery, timers, and the XENVER_version 8805interception. 8806 8807The KVM_XEN_HVM_CONFIG_RUNSTATE_UPDATE_FLAG flag indicates that KVM supports 8808the KVM_XEN_ATTR_TYPE_RUNSTATE_UPDATE_FLAG attribute in the KVM_XEN_SET_ATTR 8809and KVM_XEN_GET_ATTR ioctls. This controls whether KVM will set the 8810XEN_RUNSTATE_UPDATE flag in guest memory mapped vcpu_runstate_info during 8811updates of the runstate information. Note that versions of KVM which support 8812the RUNSTATE feature above, but not the RUNSTATE_UPDATE_FLAG feature, will 8813always set the XEN_RUNSTATE_UPDATE flag when updating the guest structure, 8814which is perhaps counterintuitive. When this flag is advertised, KVM will 8815behave more correctly, not using the XEN_RUNSTATE_UPDATE flag until/unless 8816specifically enabled (by the guest making the hypercall, causing the VMM 8817to enable the KVM_XEN_ATTR_TYPE_RUNSTATE_UPDATE_FLAG attribute). 8818 8819The KVM_XEN_HVM_CONFIG_PVCLOCK_TSC_UNSTABLE flag indicates that KVM supports 8820clearing the PVCLOCK_TSC_STABLE_BIT flag in Xen pvclock sources. This will be 8821done when the KVM_CAP_XEN_HVM ioctl sets the 8822KVM_XEN_HVM_CONFIG_PVCLOCK_TSC_UNSTABLE flag. 8823 88248.31 KVM_CAP_PPC_MULTITCE 8825------------------------- 8826 8827:Capability: KVM_CAP_PPC_MULTITCE 8828:Architectures: ppc 8829:Type: vm 8830 8831This capability means the kernel is capable of handling hypercalls 8832H_PUT_TCE_INDIRECT and H_STUFF_TCE without passing those into the user 8833space. This significantly accelerates DMA operations for PPC KVM guests. 8834User space should expect that its handlers for these hypercalls 8835are not going to be called if user space previously registered LIOBN 8836in KVM (via KVM_CREATE_SPAPR_TCE or similar calls). 8837 8838In order to enable H_PUT_TCE_INDIRECT and H_STUFF_TCE use in the guest, 8839user space might have to advertise it for the guest. For example, 8840IBM pSeries (sPAPR) guest starts using them if "hcall-multi-tce" is 8841present in the "ibm,hypertas-functions" device-tree property. 8842 8843The hypercalls mentioned above may or may not be processed successfully 8844in the kernel based fast path. If they can not be handled by the kernel, 8845they will get passed on to user space. So user space still has to have 8846an implementation for these despite the in kernel acceleration. 8847 8848This capability is always enabled. 8849 88508.32 KVM_CAP_PTP_KVM 8851-------------------- 8852 8853:Architectures: arm64 8854 8855This capability indicates that the KVM virtual PTP service is 8856supported in the host. A VMM can check whether the service is 8857available to the guest on migration. 8858 88598.33 KVM_CAP_HYPERV_ENFORCE_CPUID 8860--------------------------------- 8861 8862Architectures: x86 8863 8864When enabled, KVM will disable emulated Hyper-V features provided to the 8865guest according to the bits Hyper-V CPUID feature leaves. Otherwise, all 8866currently implemented Hyper-V features are provided unconditionally when 8867Hyper-V identification is set in the HYPERV_CPUID_INTERFACE (0x40000001) 8868leaf. 8869 88708.34 KVM_CAP_EXIT_HYPERCALL 8871--------------------------- 8872 8873:Capability: KVM_CAP_EXIT_HYPERCALL 8874:Architectures: x86 8875:Type: vm 8876 8877This capability, if enabled, will cause KVM to exit to userspace 8878with KVM_EXIT_HYPERCALL exit reason to process some hypercalls. 8879 8880Calling KVM_CHECK_EXTENSION for this capability will return a bitmask 8881of hypercalls that can be configured to exit to userspace. 8882Right now, the only such hypercall is KVM_HC_MAP_GPA_RANGE. 8883 8884The argument to KVM_ENABLE_CAP is also a bitmask, and must be a subset 8885of the result of KVM_CHECK_EXTENSION. KVM will forward to userspace 8886the hypercalls whose corresponding bit is in the argument, and return 8887ENOSYS for the others. 8888 88898.35 KVM_CAP_PMU_CAPABILITY 8890--------------------------- 8891 8892:Capability: KVM_CAP_PMU_CAPABILITY 8893:Architectures: x86 8894:Type: vm 8895:Parameters: arg[0] is bitmask of PMU virtualization capabilities. 8896:Returns: 0 on success, -EINVAL when arg[0] contains invalid bits 8897 8898This capability alters PMU virtualization in KVM. 8899 8900Calling KVM_CHECK_EXTENSION for this capability returns a bitmask of 8901PMU virtualization capabilities that can be adjusted on a VM. 8902 8903The argument to KVM_ENABLE_CAP is also a bitmask and selects specific 8904PMU virtualization capabilities to be applied to the VM. This can 8905only be invoked on a VM prior to the creation of VCPUs. 8906 8907At this time, KVM_PMU_CAP_DISABLE is the only capability. Setting 8908this capability will disable PMU virtualization for that VM. Usermode 8909should adjust CPUID leaf 0xA to reflect that the PMU is disabled. 8910 89118.36 KVM_CAP_ARM_SYSTEM_SUSPEND 8912------------------------------- 8913 8914:Capability: KVM_CAP_ARM_SYSTEM_SUSPEND 8915:Architectures: arm64 8916:Type: vm 8917 8918When enabled, KVM will exit to userspace with KVM_EXIT_SYSTEM_EVENT of 8919type KVM_SYSTEM_EVENT_SUSPEND to process the guest suspend request. 8920 89218.37 KVM_CAP_S390_PROTECTED_DUMP 8922-------------------------------- 8923 8924:Capability: KVM_CAP_S390_PROTECTED_DUMP 8925:Architectures: s390 8926:Type: vm 8927 8928This capability indicates that KVM and the Ultravisor support dumping 8929PV guests. The `KVM_PV_DUMP` command is available for the 8930`KVM_S390_PV_COMMAND` ioctl and the `KVM_PV_INFO` command provides 8931dump related UV data. Also the vcpu ioctl `KVM_S390_PV_CPU_COMMAND` is 8932available and supports the `KVM_PV_DUMP_CPU` subcommand. 8933 89348.38 KVM_CAP_VM_DISABLE_NX_HUGE_PAGES 8935------------------------------------- 8936 8937:Capability: KVM_CAP_VM_DISABLE_NX_HUGE_PAGES 8938:Architectures: x86 8939:Type: vm 8940:Parameters: arg[0] must be 0. 8941:Returns: 0 on success, -EPERM if the userspace process does not 8942 have CAP_SYS_BOOT, -EINVAL if args[0] is not 0 or any vCPUs have been 8943 created. 8944 8945This capability disables the NX huge pages mitigation for iTLB MULTIHIT. 8946 8947The capability has no effect if the nx_huge_pages module parameter is not set. 8948 8949This capability may only be set before any vCPUs are created. 8950 89518.39 KVM_CAP_S390_CPU_TOPOLOGY 8952------------------------------ 8953 8954:Capability: KVM_CAP_S390_CPU_TOPOLOGY 8955:Architectures: s390 8956:Type: vm 8957 8958This capability indicates that KVM will provide the S390 CPU Topology 8959facility which consist of the interpretation of the PTF instruction for 8960the function code 2 along with interception and forwarding of both the 8961PTF instruction with function codes 0 or 1 and the STSI(15,1,x) 8962instruction to the userland hypervisor. 8963 8964The stfle facility 11, CPU Topology facility, should not be indicated 8965to the guest without this capability. 8966 8967When this capability is present, KVM provides a new attribute group 8968on vm fd, KVM_S390_VM_CPU_TOPOLOGY. 8969This new attribute allows to get, set or clear the Modified Change 8970Topology Report (MTCR) bit of the SCA through the kvm_device_attr 8971structure. 8972 8973When getting the Modified Change Topology Report value, the attr->addr 8974must point to a byte where the value will be stored or retrieved from. 8975 89768.40 KVM_CAP_ARM_EAGER_SPLIT_CHUNK_SIZE 8977--------------------------------------- 8978 8979:Capability: KVM_CAP_ARM_EAGER_SPLIT_CHUNK_SIZE 8980:Architectures: arm64 8981:Type: vm 8982:Parameters: arg[0] is the new split chunk size. 8983:Returns: 0 on success, -EINVAL if any memslot was already created. 8984 8985This capability sets the chunk size used in Eager Page Splitting. 8986 8987Eager Page Splitting improves the performance of dirty-logging (used 8988in live migrations) when guest memory is backed by huge-pages. It 8989avoids splitting huge-pages (into PAGE_SIZE pages) on fault, by doing 8990it eagerly when enabling dirty logging (with the 8991KVM_MEM_LOG_DIRTY_PAGES flag for a memory region), or when using 8992KVM_CLEAR_DIRTY_LOG. 8993 8994The chunk size specifies how many pages to break at a time, using a 8995single allocation for each chunk. Bigger the chunk size, more pages 8996need to be allocated ahead of time. 8997 8998The chunk size needs to be a valid block size. The list of acceptable 8999block sizes is exposed in KVM_CAP_ARM_SUPPORTED_BLOCK_SIZES as a 900064-bit bitmap (each bit describing a block size). The default value is 90010, to disable the eager page splitting. 9002 90038.41 KVM_CAP_VM_TYPES 9004--------------------- 9005 9006:Capability: KVM_CAP_MEMORY_ATTRIBUTES 9007:Architectures: x86 9008:Type: system ioctl 9009 9010This capability returns a bitmap of support VM types. The 1-setting of bit @n 9011means the VM type with value @n is supported. Possible values of @n are:: 9012 9013 #define KVM_X86_DEFAULT_VM 0 9014 #define KVM_X86_SW_PROTECTED_VM 1 9015 #define KVM_X86_SEV_VM 2 9016 #define KVM_X86_SEV_ES_VM 3 9017 9018Note, KVM_X86_SW_PROTECTED_VM is currently only for development and testing. 9019Do not use KVM_X86_SW_PROTECTED_VM for "real" VMs, and especially not in 9020production. The behavior and effective ABI for software-protected VMs is 9021unstable. 9022 90239. Known KVM API problems 9024========================= 9025 9026In some cases, KVM's API has some inconsistencies or common pitfalls 9027that userspace need to be aware of. This section details some of 9028these issues. 9029 9030Most of them are architecture specific, so the section is split by 9031architecture. 9032 90339.1. x86 9034-------- 9035 9036``KVM_GET_SUPPORTED_CPUID`` issues 9037^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ 9038 9039In general, ``KVM_GET_SUPPORTED_CPUID`` is designed so that it is possible 9040to take its result and pass it directly to ``KVM_SET_CPUID2``. This section 9041documents some cases in which that requires some care. 9042 9043Local APIC features 9044~~~~~~~~~~~~~~~~~~~ 9045 9046CPU[EAX=1]:ECX[21] (X2APIC) is reported by ``KVM_GET_SUPPORTED_CPUID``, 9047but it can only be enabled if ``KVM_CREATE_IRQCHIP`` or 9048``KVM_ENABLE_CAP(KVM_CAP_IRQCHIP_SPLIT)`` are used to enable in-kernel emulation of 9049the local APIC. 9050 9051The same is true for the ``KVM_FEATURE_PV_UNHALT`` paravirtualized feature. 9052 9053CPU[EAX=1]:ECX[24] (TSC_DEADLINE) is not reported by ``KVM_GET_SUPPORTED_CPUID``. 9054It can be enabled if ``KVM_CAP_TSC_DEADLINE_TIMER`` is present and the kernel 9055has enabled in-kernel emulation of the local APIC. 9056 9057CPU topology 9058~~~~~~~~~~~~ 9059 9060Several CPUID values include topology information for the host CPU: 90610x0b and 0x1f for Intel systems, 0x8000001e for AMD systems. Different 9062versions of KVM return different values for this information and userspace 9063should not rely on it. Currently they return all zeroes. 9064 9065If userspace wishes to set up a guest topology, it should be careful that 9066the values of these three leaves differ for each CPU. In particular, 9067the APIC ID is found in EDX for all subleaves of 0x0b and 0x1f, and in EAX 9068for 0x8000001e; the latter also encodes the core id and node id in bits 90697:0 of EBX and ECX respectively. 9070 9071Obsolete ioctls and capabilities 9072^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ 9073 9074KVM_CAP_DISABLE_QUIRKS does not let userspace know which quirks are actually 9075available. Use ``KVM_CHECK_EXTENSION(KVM_CAP_DISABLE_QUIRKS2)`` instead if 9076available. 9077 9078Ordering of KVM_GET_*/KVM_SET_* ioctls 9079^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ 9080 9081TBD 9082