1.. SPDX-License-Identifier: GPL-2.0 2 3=================================================================== 4The Definitive KVM (Kernel-based Virtual Machine) API Documentation 5=================================================================== 6 71. General description 8====================== 9 10The kvm API is a set of ioctls that are issued to control various aspects 11of a virtual machine. The ioctls belong to the following classes: 12 13 - System ioctls: These query and set global attributes which affect the 14 whole kvm subsystem. In addition a system ioctl is used to create 15 virtual machines. 16 17 - VM ioctls: These query and set attributes that affect an entire virtual 18 machine, for example memory layout. In addition a VM ioctl is used to 19 create virtual cpus (vcpus) and devices. 20 21 VM ioctls must be issued from the same process (address space) that was 22 used to create the VM. 23 24 - vcpu ioctls: These query and set attributes that control the operation 25 of a single virtual cpu. 26 27 vcpu ioctls should be issued from the same thread that was used to create 28 the vcpu, except for asynchronous vcpu ioctl that are marked as such in 29 the documentation. Otherwise, the first ioctl after switching threads 30 could see a performance impact. 31 32 - device ioctls: These query and set attributes that control the operation 33 of a single device. 34 35 device ioctls must be issued from the same process (address space) that 36 was used to create the VM. 37 382. File descriptors 39=================== 40 41The kvm API is centered around file descriptors. An initial 42open("/dev/kvm") obtains a handle to the kvm subsystem; this handle 43can be used to issue system ioctls. A KVM_CREATE_VM ioctl on this 44handle will create a VM file descriptor which can be used to issue VM 45ioctls. A KVM_CREATE_VCPU or KVM_CREATE_DEVICE ioctl on a VM fd will 46create a virtual cpu or device and return a file descriptor pointing to 47the new resource. Finally, ioctls on a vcpu or device fd can be used 48to control the vcpu or device. For vcpus, this includes the important 49task of actually running guest code. 50 51In general file descriptors can be migrated among processes by means 52of fork() and the SCM_RIGHTS facility of unix domain socket. These 53kinds of tricks are explicitly not supported by kvm. While they will 54not cause harm to the host, their actual behavior is not guaranteed by 55the API. See "General description" for details on the ioctl usage 56model that is supported by KVM. 57 58It is important to note that although VM ioctls may only be issued from 59the process that created the VM, a VM's lifecycle is associated with its 60file descriptor, not its creator (process). In other words, the VM and 61its resources, *including the associated address space*, are not freed 62until the last reference to the VM's file descriptor has been released. 63For example, if fork() is issued after ioctl(KVM_CREATE_VM), the VM will 64not be freed until both the parent (original) process and its child have 65put their references to the VM's file descriptor. 66 67Because a VM's resources are not freed until the last reference to its 68file descriptor is released, creating additional references to a VM 69via fork(), dup(), etc... without careful consideration is strongly 70discouraged and may have unwanted side effects, e.g. memory allocated 71by and on behalf of the VM's process may not be freed/unaccounted when 72the VM is shut down. 73 74 753. Extensions 76============= 77 78As of Linux 2.6.22, the KVM ABI has been stabilized: no backward 79incompatible change are allowed. However, there is an extension 80facility that allows backward-compatible extensions to the API to be 81queried and used. 82 83The extension mechanism is not based on the Linux version number. 84Instead, kvm defines extension identifiers and a facility to query 85whether a particular extension identifier is available. If it is, a 86set of ioctls is available for application use. 87 88 894. API description 90================== 91 92This section describes ioctls that can be used to control kvm guests. 93For each ioctl, the following information is provided along with a 94description: 95 96 Capability: 97 which KVM extension provides this ioctl. Can be 'basic', 98 which means that is will be provided by any kernel that supports 99 API version 12 (see section 4.1), a KVM_CAP_xyz constant, which 100 means availability needs to be checked with KVM_CHECK_EXTENSION 101 (see section 4.4), or 'none' which means that while not all kernels 102 support this ioctl, there's no capability bit to check its 103 availability: for kernels that don't support the ioctl, 104 the ioctl returns -ENOTTY. 105 106 Architectures: 107 which instruction set architectures provide this ioctl. 108 x86 includes both i386 and x86_64. 109 110 Type: 111 system, vm, or vcpu. 112 113 Parameters: 114 what parameters are accepted by the ioctl. 115 116 Returns: 117 the return value. General error numbers (EBADF, ENOMEM, EINVAL) 118 are not detailed, but errors with specific meanings are. 119 120 1214.1 KVM_GET_API_VERSION 122----------------------- 123 124:Capability: basic 125:Architectures: all 126:Type: system ioctl 127:Parameters: none 128:Returns: the constant KVM_API_VERSION (=12) 129 130This identifies the API version as the stable kvm API. It is not 131expected that this number will change. However, Linux 2.6.20 and 1322.6.21 report earlier versions; these are not documented and not 133supported. Applications should refuse to run if KVM_GET_API_VERSION 134returns a value other than 12. If this check passes, all ioctls 135described as 'basic' will be available. 136 137 1384.2 KVM_CREATE_VM 139----------------- 140 141:Capability: basic 142:Architectures: all 143:Type: system ioctl 144:Parameters: machine type identifier (KVM_VM_*) 145:Returns: a VM fd that can be used to control the new virtual machine. 146 147The new VM has no virtual cpus and no memory. 148You probably want to use 0 as machine type. 149 150In order to create user controlled virtual machines on S390, check 151KVM_CAP_S390_UCONTROL and use the flag KVM_VM_S390_UCONTROL as 152privileged user (CAP_SYS_ADMIN). 153 154On arm64, the physical address size for a VM (IPA Size limit) is limited 155to 40bits by default. The limit can be configured if the host supports the 156extension KVM_CAP_ARM_VM_IPA_SIZE. When supported, use 157KVM_VM_TYPE_ARM_IPA_SIZE(IPA_Bits) to set the size in the machine type 158identifier, where IPA_Bits is the maximum width of any physical 159address used by the VM. The IPA_Bits is encoded in bits[7-0] of the 160machine type identifier. 161 162e.g, to configure a guest to use 48bit physical address size:: 163 164 vm_fd = ioctl(dev_fd, KVM_CREATE_VM, KVM_VM_TYPE_ARM_IPA_SIZE(48)); 165 166The requested size (IPA_Bits) must be: 167 168 == ========================================================= 169 0 Implies default size, 40bits (for backward compatibility) 170 N Implies N bits, where N is a positive integer such that, 171 32 <= N <= Host_IPA_Limit 172 == ========================================================= 173 174Host_IPA_Limit is the maximum possible value for IPA_Bits on the host and 175is dependent on the CPU capability and the kernel configuration. The limit can 176be retrieved using KVM_CAP_ARM_VM_IPA_SIZE of the KVM_CHECK_EXTENSION 177ioctl() at run-time. 178 179Creation of the VM will fail if the requested IPA size (whether it is 180implicit or explicit) is unsupported on the host. 181 182Please note that configuring the IPA size does not affect the capability 183exposed by the guest CPUs in ID_AA64MMFR0_EL1[PARange]. It only affects 184size of the address translated by the stage2 level (guest physical to 185host physical address translations). 186 187 1884.3 KVM_GET_MSR_INDEX_LIST, KVM_GET_MSR_FEATURE_INDEX_LIST 189---------------------------------------------------------- 190 191:Capability: basic, KVM_CAP_GET_MSR_FEATURES for KVM_GET_MSR_FEATURE_INDEX_LIST 192:Architectures: x86 193:Type: system ioctl 194:Parameters: struct kvm_msr_list (in/out) 195:Returns: 0 on success; -1 on error 196 197Errors: 198 199 ====== ============================================================ 200 EFAULT the msr index list cannot be read from or written to 201 E2BIG the msr index list is too big to fit in the array specified by 202 the user. 203 ====== ============================================================ 204 205:: 206 207 struct kvm_msr_list { 208 __u32 nmsrs; /* number of msrs in entries */ 209 __u32 indices[0]; 210 }; 211 212The user fills in the size of the indices array in nmsrs, and in return 213kvm adjusts nmsrs to reflect the actual number of msrs and fills in the 214indices array with their numbers. 215 216KVM_GET_MSR_INDEX_LIST returns the guest msrs that are supported. The list 217varies by kvm version and host processor, but does not change otherwise. 218 219Note: if kvm indicates supports MCE (KVM_CAP_MCE), then the MCE bank MSRs are 220not returned in the MSR list, as different vcpus can have a different number 221of banks, as set via the KVM_X86_SETUP_MCE ioctl. 222 223KVM_GET_MSR_FEATURE_INDEX_LIST returns the list of MSRs that can be passed 224to the KVM_GET_MSRS system ioctl. This lets userspace probe host capabilities 225and processor features that are exposed via MSRs (e.g., VMX capabilities). 226This list also varies by kvm version and host processor, but does not change 227otherwise. 228 229 2304.4 KVM_CHECK_EXTENSION 231----------------------- 232 233:Capability: basic, KVM_CAP_CHECK_EXTENSION_VM for vm ioctl 234:Architectures: all 235:Type: system ioctl, vm ioctl 236:Parameters: extension identifier (KVM_CAP_*) 237:Returns: 0 if unsupported; 1 (or some other positive integer) if supported 238 239The API allows the application to query about extensions to the core 240kvm API. Userspace passes an extension identifier (an integer) and 241receives an integer that describes the extension availability. 242Generally 0 means no and 1 means yes, but some extensions may report 243additional information in the integer return value. 244 245Based on their initialization different VMs may have different capabilities. 246It is thus encouraged to use the vm ioctl to query for capabilities (available 247with KVM_CAP_CHECK_EXTENSION_VM on the vm fd) 248 2494.5 KVM_GET_VCPU_MMAP_SIZE 250-------------------------- 251 252:Capability: basic 253:Architectures: all 254:Type: system ioctl 255:Parameters: none 256:Returns: size of vcpu mmap area, in bytes 257 258The KVM_RUN ioctl (cf.) communicates with userspace via a shared 259memory region. This ioctl returns the size of that region. See the 260KVM_RUN documentation for details. 261 262Besides the size of the KVM_RUN communication region, other areas of 263the VCPU file descriptor can be mmap-ed, including: 264 265- if KVM_CAP_COALESCED_MMIO is available, a page at 266 KVM_COALESCED_MMIO_PAGE_OFFSET * PAGE_SIZE; for historical reasons, 267 this page is included in the result of KVM_GET_VCPU_MMAP_SIZE. 268 KVM_CAP_COALESCED_MMIO is not documented yet. 269 270- if KVM_CAP_DIRTY_LOG_RING is available, a number of pages at 271 KVM_DIRTY_LOG_PAGE_OFFSET * PAGE_SIZE. For more information on 272 KVM_CAP_DIRTY_LOG_RING, see section 8.3. 273 274 2754.7 KVM_CREATE_VCPU 276------------------- 277 278:Capability: basic 279:Architectures: all 280:Type: vm ioctl 281:Parameters: vcpu id (apic id on x86) 282:Returns: vcpu fd on success, -1 on error 283 284This API adds a vcpu to a virtual machine. No more than max_vcpus may be added. 285The vcpu id is an integer in the range [0, max_vcpu_id). 286 287The recommended max_vcpus value can be retrieved using the KVM_CAP_NR_VCPUS of 288the KVM_CHECK_EXTENSION ioctl() at run-time. 289The maximum possible value for max_vcpus can be retrieved using the 290KVM_CAP_MAX_VCPUS of the KVM_CHECK_EXTENSION ioctl() at run-time. 291 292If the KVM_CAP_NR_VCPUS does not exist, you should assume that max_vcpus is 4 293cpus max. 294If the KVM_CAP_MAX_VCPUS does not exist, you should assume that max_vcpus is 295same as the value returned from KVM_CAP_NR_VCPUS. 296 297The maximum possible value for max_vcpu_id can be retrieved using the 298KVM_CAP_MAX_VCPU_ID of the KVM_CHECK_EXTENSION ioctl() at run-time. 299 300If the KVM_CAP_MAX_VCPU_ID does not exist, you should assume that max_vcpu_id 301is the same as the value returned from KVM_CAP_MAX_VCPUS. 302 303On powerpc using book3s_hv mode, the vcpus are mapped onto virtual 304threads in one or more virtual CPU cores. (This is because the 305hardware requires all the hardware threads in a CPU core to be in the 306same partition.) The KVM_CAP_PPC_SMT capability indicates the number 307of vcpus per virtual core (vcore). The vcore id is obtained by 308dividing the vcpu id by the number of vcpus per vcore. The vcpus in a 309given vcore will always be in the same physical core as each other 310(though that might be a different physical core from time to time). 311Userspace can control the threading (SMT) mode of the guest by its 312allocation of vcpu ids. For example, if userspace wants 313single-threaded guest vcpus, it should make all vcpu ids be a multiple 314of the number of vcpus per vcore. 315 316For virtual cpus that have been created with S390 user controlled virtual 317machines, the resulting vcpu fd can be memory mapped at page offset 318KVM_S390_SIE_PAGE_OFFSET in order to obtain a memory map of the virtual 319cpu's hardware control block. 320 321 3224.8 KVM_GET_DIRTY_LOG (vm ioctl) 323-------------------------------- 324 325:Capability: basic 326:Architectures: all 327:Type: vm ioctl 328:Parameters: struct kvm_dirty_log (in/out) 329:Returns: 0 on success, -1 on error 330 331:: 332 333 /* for KVM_GET_DIRTY_LOG */ 334 struct kvm_dirty_log { 335 __u32 slot; 336 __u32 padding; 337 union { 338 void __user *dirty_bitmap; /* one bit per page */ 339 __u64 padding; 340 }; 341 }; 342 343Given a memory slot, return a bitmap containing any pages dirtied 344since the last call to this ioctl. Bit 0 is the first page in the 345memory slot. Ensure the entire structure is cleared to avoid padding 346issues. 347 348If KVM_CAP_MULTI_ADDRESS_SPACE is available, bits 16-31 of slot field specifies 349the address space for which you want to return the dirty bitmap. See 350KVM_SET_USER_MEMORY_REGION for details on the usage of slot field. 351 352The bits in the dirty bitmap are cleared before the ioctl returns, unless 353KVM_CAP_MANUAL_DIRTY_LOG_PROTECT2 is enabled. For more information, 354see the description of the capability. 355 356Note that the Xen shared info page, if configured, shall always be assumed 357to be dirty. KVM will not explicitly mark it such. 358 359 3604.10 KVM_RUN 361------------ 362 363:Capability: basic 364:Architectures: all 365:Type: vcpu ioctl 366:Parameters: none 367:Returns: 0 on success, -1 on error 368 369Errors: 370 371 ======= ============================================================== 372 EINTR an unmasked signal is pending 373 ENOEXEC the vcpu hasn't been initialized or the guest tried to execute 374 instructions from device memory (arm64) 375 ENOSYS data abort outside memslots with no syndrome info and 376 KVM_CAP_ARM_NISV_TO_USER not enabled (arm64) 377 EPERM SVE feature set but not finalized (arm64) 378 ======= ============================================================== 379 380This ioctl is used to run a guest virtual cpu. While there are no 381explicit parameters, there is an implicit parameter block that can be 382obtained by mmap()ing the vcpu fd at offset 0, with the size given by 383KVM_GET_VCPU_MMAP_SIZE. The parameter block is formatted as a 'struct 384kvm_run' (see below). 385 386 3874.11 KVM_GET_REGS 388----------------- 389 390:Capability: basic 391:Architectures: all except arm64 392:Type: vcpu ioctl 393:Parameters: struct kvm_regs (out) 394:Returns: 0 on success, -1 on error 395 396Reads the general purpose registers from the vcpu. 397 398:: 399 400 /* x86 */ 401 struct kvm_regs { 402 /* out (KVM_GET_REGS) / in (KVM_SET_REGS) */ 403 __u64 rax, rbx, rcx, rdx; 404 __u64 rsi, rdi, rsp, rbp; 405 __u64 r8, r9, r10, r11; 406 __u64 r12, r13, r14, r15; 407 __u64 rip, rflags; 408 }; 409 410 /* mips */ 411 struct kvm_regs { 412 /* out (KVM_GET_REGS) / in (KVM_SET_REGS) */ 413 __u64 gpr[32]; 414 __u64 hi; 415 __u64 lo; 416 __u64 pc; 417 }; 418 419 4204.12 KVM_SET_REGS 421----------------- 422 423:Capability: basic 424:Architectures: all except arm64 425:Type: vcpu ioctl 426:Parameters: struct kvm_regs (in) 427:Returns: 0 on success, -1 on error 428 429Writes the general purpose registers into the vcpu. 430 431See KVM_GET_REGS for the data structure. 432 433 4344.13 KVM_GET_SREGS 435------------------ 436 437:Capability: basic 438:Architectures: x86, ppc 439:Type: vcpu ioctl 440:Parameters: struct kvm_sregs (out) 441:Returns: 0 on success, -1 on error 442 443Reads special registers from the vcpu. 444 445:: 446 447 /* x86 */ 448 struct kvm_sregs { 449 struct kvm_segment cs, ds, es, fs, gs, ss; 450 struct kvm_segment tr, ldt; 451 struct kvm_dtable gdt, idt; 452 __u64 cr0, cr2, cr3, cr4, cr8; 453 __u64 efer; 454 __u64 apic_base; 455 __u64 interrupt_bitmap[(KVM_NR_INTERRUPTS + 63) / 64]; 456 }; 457 458 /* ppc -- see arch/powerpc/include/uapi/asm/kvm.h */ 459 460interrupt_bitmap is a bitmap of pending external interrupts. At most 461one bit may be set. This interrupt has been acknowledged by the APIC 462but not yet injected into the cpu core. 463 464 4654.14 KVM_SET_SREGS 466------------------ 467 468:Capability: basic 469:Architectures: x86, ppc 470:Type: vcpu ioctl 471:Parameters: struct kvm_sregs (in) 472:Returns: 0 on success, -1 on error 473 474Writes special registers into the vcpu. See KVM_GET_SREGS for the 475data structures. 476 477 4784.15 KVM_TRANSLATE 479------------------ 480 481:Capability: basic 482:Architectures: x86 483:Type: vcpu ioctl 484:Parameters: struct kvm_translation (in/out) 485:Returns: 0 on success, -1 on error 486 487Translates a virtual address according to the vcpu's current address 488translation mode. 489 490:: 491 492 struct kvm_translation { 493 /* in */ 494 __u64 linear_address; 495 496 /* out */ 497 __u64 physical_address; 498 __u8 valid; 499 __u8 writeable; 500 __u8 usermode; 501 __u8 pad[5]; 502 }; 503 504 5054.16 KVM_INTERRUPT 506------------------ 507 508:Capability: basic 509:Architectures: x86, ppc, mips, riscv 510:Type: vcpu ioctl 511:Parameters: struct kvm_interrupt (in) 512:Returns: 0 on success, negative on failure. 513 514Queues a hardware interrupt vector to be injected. 515 516:: 517 518 /* for KVM_INTERRUPT */ 519 struct kvm_interrupt { 520 /* in */ 521 __u32 irq; 522 }; 523 524X86: 525^^^^ 526 527:Returns: 528 529 ========= =================================== 530 0 on success, 531 -EEXIST if an interrupt is already enqueued 532 -EINVAL the irq number is invalid 533 -ENXIO if the PIC is in the kernel 534 -EFAULT if the pointer is invalid 535 ========= =================================== 536 537Note 'irq' is an interrupt vector, not an interrupt pin or line. This 538ioctl is useful if the in-kernel PIC is not used. 539 540PPC: 541^^^^ 542 543Queues an external interrupt to be injected. This ioctl is overleaded 544with 3 different irq values: 545 546a) KVM_INTERRUPT_SET 547 548 This injects an edge type external interrupt into the guest once it's ready 549 to receive interrupts. When injected, the interrupt is done. 550 551b) KVM_INTERRUPT_UNSET 552 553 This unsets any pending interrupt. 554 555 Only available with KVM_CAP_PPC_UNSET_IRQ. 556 557c) KVM_INTERRUPT_SET_LEVEL 558 559 This injects a level type external interrupt into the guest context. The 560 interrupt stays pending until a specific ioctl with KVM_INTERRUPT_UNSET 561 is triggered. 562 563 Only available with KVM_CAP_PPC_IRQ_LEVEL. 564 565Note that any value for 'irq' other than the ones stated above is invalid 566and incurs unexpected behavior. 567 568This is an asynchronous vcpu ioctl and can be invoked from any thread. 569 570MIPS: 571^^^^^ 572 573Queues an external interrupt to be injected into the virtual CPU. A negative 574interrupt number dequeues the interrupt. 575 576This is an asynchronous vcpu ioctl and can be invoked from any thread. 577 578RISC-V: 579^^^^^^^ 580 581Queues an external interrupt to be injected into the virutal CPU. This ioctl 582is overloaded with 2 different irq values: 583 584a) KVM_INTERRUPT_SET 585 586 This sets external interrupt for a virtual CPU and it will receive 587 once it is ready. 588 589b) KVM_INTERRUPT_UNSET 590 591 This clears pending external interrupt for a virtual CPU. 592 593This is an asynchronous vcpu ioctl and can be invoked from any thread. 594 595 5964.17 KVM_DEBUG_GUEST 597-------------------- 598 599:Capability: basic 600:Architectures: none 601:Type: vcpu ioctl 602:Parameters: none) 603:Returns: -1 on error 604 605Support for this has been removed. Use KVM_SET_GUEST_DEBUG instead. 606 607 6084.18 KVM_GET_MSRS 609----------------- 610 611:Capability: basic (vcpu), KVM_CAP_GET_MSR_FEATURES (system) 612:Architectures: x86 613:Type: system ioctl, vcpu ioctl 614:Parameters: struct kvm_msrs (in/out) 615:Returns: number of msrs successfully returned; 616 -1 on error 617 618When used as a system ioctl: 619Reads the values of MSR-based features that are available for the VM. This 620is similar to KVM_GET_SUPPORTED_CPUID, but it returns MSR indices and values. 621The list of msr-based features can be obtained using KVM_GET_MSR_FEATURE_INDEX_LIST 622in a system ioctl. 623 624When used as a vcpu ioctl: 625Reads model-specific registers from the vcpu. Supported msr indices can 626be obtained using KVM_GET_MSR_INDEX_LIST in a system ioctl. 627 628:: 629 630 struct kvm_msrs { 631 __u32 nmsrs; /* number of msrs in entries */ 632 __u32 pad; 633 634 struct kvm_msr_entry entries[0]; 635 }; 636 637 struct kvm_msr_entry { 638 __u32 index; 639 __u32 reserved; 640 __u64 data; 641 }; 642 643Application code should set the 'nmsrs' member (which indicates the 644size of the entries array) and the 'index' member of each array entry. 645kvm will fill in the 'data' member. 646 647 6484.19 KVM_SET_MSRS 649----------------- 650 651:Capability: basic 652:Architectures: x86 653:Type: vcpu ioctl 654:Parameters: struct kvm_msrs (in) 655:Returns: number of msrs successfully set (see below), -1 on error 656 657Writes model-specific registers to the vcpu. See KVM_GET_MSRS for the 658data structures. 659 660Application code should set the 'nmsrs' member (which indicates the 661size of the entries array), and the 'index' and 'data' members of each 662array entry. 663 664It tries to set the MSRs in array entries[] one by one. If setting an MSR 665fails, e.g., due to setting reserved bits, the MSR isn't supported/emulated 666by KVM, etc..., it stops processing the MSR list and returns the number of 667MSRs that have been set successfully. 668 669 6704.20 KVM_SET_CPUID 671------------------ 672 673:Capability: basic 674:Architectures: x86 675:Type: vcpu ioctl 676:Parameters: struct kvm_cpuid (in) 677:Returns: 0 on success, -1 on error 678 679Defines the vcpu responses to the cpuid instruction. Applications 680should use the KVM_SET_CPUID2 ioctl if available. 681 682Caveat emptor: 683 - If this IOCTL fails, KVM gives no guarantees that previous valid CPUID 684 configuration (if there is) is not corrupted. Userspace can get a copy 685 of the resulting CPUID configuration through KVM_GET_CPUID2 in case. 686 - Using KVM_SET_CPUID{,2} after KVM_RUN, i.e. changing the guest vCPU model 687 after running the guest, may cause guest instability. 688 - Using heterogeneous CPUID configurations, modulo APIC IDs, topology, etc... 689 may cause guest instability. 690 691:: 692 693 struct kvm_cpuid_entry { 694 __u32 function; 695 __u32 eax; 696 __u32 ebx; 697 __u32 ecx; 698 __u32 edx; 699 __u32 padding; 700 }; 701 702 /* for KVM_SET_CPUID */ 703 struct kvm_cpuid { 704 __u32 nent; 705 __u32 padding; 706 struct kvm_cpuid_entry entries[0]; 707 }; 708 709 7104.21 KVM_SET_SIGNAL_MASK 711------------------------ 712 713:Capability: basic 714:Architectures: all 715:Type: vcpu ioctl 716:Parameters: struct kvm_signal_mask (in) 717:Returns: 0 on success, -1 on error 718 719Defines which signals are blocked during execution of KVM_RUN. This 720signal mask temporarily overrides the threads signal mask. Any 721unblocked signal received (except SIGKILL and SIGSTOP, which retain 722their traditional behaviour) will cause KVM_RUN to return with -EINTR. 723 724Note the signal will only be delivered if not blocked by the original 725signal mask. 726 727:: 728 729 /* for KVM_SET_SIGNAL_MASK */ 730 struct kvm_signal_mask { 731 __u32 len; 732 __u8 sigset[0]; 733 }; 734 735 7364.22 KVM_GET_FPU 737---------------- 738 739:Capability: basic 740:Architectures: x86 741:Type: vcpu ioctl 742:Parameters: struct kvm_fpu (out) 743:Returns: 0 on success, -1 on error 744 745Reads the floating point state from the vcpu. 746 747:: 748 749 /* for KVM_GET_FPU and KVM_SET_FPU */ 750 struct kvm_fpu { 751 __u8 fpr[8][16]; 752 __u16 fcw; 753 __u16 fsw; 754 __u8 ftwx; /* in fxsave format */ 755 __u8 pad1; 756 __u16 last_opcode; 757 __u64 last_ip; 758 __u64 last_dp; 759 __u8 xmm[16][16]; 760 __u32 mxcsr; 761 __u32 pad2; 762 }; 763 764 7654.23 KVM_SET_FPU 766---------------- 767 768:Capability: basic 769:Architectures: x86 770:Type: vcpu ioctl 771:Parameters: struct kvm_fpu (in) 772:Returns: 0 on success, -1 on error 773 774Writes the floating point state to the vcpu. 775 776:: 777 778 /* for KVM_GET_FPU and KVM_SET_FPU */ 779 struct kvm_fpu { 780 __u8 fpr[8][16]; 781 __u16 fcw; 782 __u16 fsw; 783 __u8 ftwx; /* in fxsave format */ 784 __u8 pad1; 785 __u16 last_opcode; 786 __u64 last_ip; 787 __u64 last_dp; 788 __u8 xmm[16][16]; 789 __u32 mxcsr; 790 __u32 pad2; 791 }; 792 793 7944.24 KVM_CREATE_IRQCHIP 795----------------------- 796 797:Capability: KVM_CAP_IRQCHIP, KVM_CAP_S390_IRQCHIP (s390) 798:Architectures: x86, arm64, s390 799:Type: vm ioctl 800:Parameters: none 801:Returns: 0 on success, -1 on error 802 803Creates an interrupt controller model in the kernel. 804On x86, creates a virtual ioapic, a virtual PIC (two PICs, nested), and sets up 805future vcpus to have a local APIC. IRQ routing for GSIs 0-15 is set to both 806PIC and IOAPIC; GSI 16-23 only go to the IOAPIC. 807On arm64, a GICv2 is created. Any other GIC versions require the usage of 808KVM_CREATE_DEVICE, which also supports creating a GICv2. Using 809KVM_CREATE_DEVICE is preferred over KVM_CREATE_IRQCHIP for GICv2. 810On s390, a dummy irq routing table is created. 811 812Note that on s390 the KVM_CAP_S390_IRQCHIP vm capability needs to be enabled 813before KVM_CREATE_IRQCHIP can be used. 814 815 8164.25 KVM_IRQ_LINE 817----------------- 818 819:Capability: KVM_CAP_IRQCHIP 820:Architectures: x86, arm64 821:Type: vm ioctl 822:Parameters: struct kvm_irq_level 823:Returns: 0 on success, -1 on error 824 825Sets the level of a GSI input to the interrupt controller model in the kernel. 826On some architectures it is required that an interrupt controller model has 827been previously created with KVM_CREATE_IRQCHIP. Note that edge-triggered 828interrupts require the level to be set to 1 and then back to 0. 829 830On real hardware, interrupt pins can be active-low or active-high. This 831does not matter for the level field of struct kvm_irq_level: 1 always 832means active (asserted), 0 means inactive (deasserted). 833 834x86 allows the operating system to program the interrupt polarity 835(active-low/active-high) for level-triggered interrupts, and KVM used 836to consider the polarity. However, due to bitrot in the handling of 837active-low interrupts, the above convention is now valid on x86 too. 838This is signaled by KVM_CAP_X86_IOAPIC_POLARITY_IGNORED. Userspace 839should not present interrupts to the guest as active-low unless this 840capability is present (or unless it is not using the in-kernel irqchip, 841of course). 842 843 844arm64 can signal an interrupt either at the CPU level, or at the 845in-kernel irqchip (GIC), and for in-kernel irqchip can tell the GIC to 846use PPIs designated for specific cpus. The irq field is interpreted 847like this:: 848 849 bits: | 31 ... 28 | 27 ... 24 | 23 ... 16 | 15 ... 0 | 850 field: | vcpu2_index | irq_type | vcpu_index | irq_id | 851 852The irq_type field has the following values: 853 854- irq_type[0]: 855 out-of-kernel GIC: irq_id 0 is IRQ, irq_id 1 is FIQ 856- irq_type[1]: 857 in-kernel GIC: SPI, irq_id between 32 and 1019 (incl.) 858 (the vcpu_index field is ignored) 859- irq_type[2]: 860 in-kernel GIC: PPI, irq_id between 16 and 31 (incl.) 861 862(The irq_id field thus corresponds nicely to the IRQ ID in the ARM GIC specs) 863 864In both cases, level is used to assert/deassert the line. 865 866When KVM_CAP_ARM_IRQ_LINE_LAYOUT_2 is supported, the target vcpu is 867identified as (256 * vcpu2_index + vcpu_index). Otherwise, vcpu2_index 868must be zero. 869 870Note that on arm64, the KVM_CAP_IRQCHIP capability only conditions 871injection of interrupts for the in-kernel irqchip. KVM_IRQ_LINE can always 872be used for a userspace interrupt controller. 873 874:: 875 876 struct kvm_irq_level { 877 union { 878 __u32 irq; /* GSI */ 879 __s32 status; /* not used for KVM_IRQ_LEVEL */ 880 }; 881 __u32 level; /* 0 or 1 */ 882 }; 883 884 8854.26 KVM_GET_IRQCHIP 886-------------------- 887 888:Capability: KVM_CAP_IRQCHIP 889:Architectures: x86 890:Type: vm ioctl 891:Parameters: struct kvm_irqchip (in/out) 892:Returns: 0 on success, -1 on error 893 894Reads the state of a kernel interrupt controller created with 895KVM_CREATE_IRQCHIP into a buffer provided by the caller. 896 897:: 898 899 struct kvm_irqchip { 900 __u32 chip_id; /* 0 = PIC1, 1 = PIC2, 2 = IOAPIC */ 901 __u32 pad; 902 union { 903 char dummy[512]; /* reserving space */ 904 struct kvm_pic_state pic; 905 struct kvm_ioapic_state ioapic; 906 } chip; 907 }; 908 909 9104.27 KVM_SET_IRQCHIP 911-------------------- 912 913:Capability: KVM_CAP_IRQCHIP 914:Architectures: x86 915:Type: vm ioctl 916:Parameters: struct kvm_irqchip (in) 917:Returns: 0 on success, -1 on error 918 919Sets the state of a kernel interrupt controller created with 920KVM_CREATE_IRQCHIP from a buffer provided by the caller. 921 922:: 923 924 struct kvm_irqchip { 925 __u32 chip_id; /* 0 = PIC1, 1 = PIC2, 2 = IOAPIC */ 926 __u32 pad; 927 union { 928 char dummy[512]; /* reserving space */ 929 struct kvm_pic_state pic; 930 struct kvm_ioapic_state ioapic; 931 } chip; 932 }; 933 934 9354.28 KVM_XEN_HVM_CONFIG 936----------------------- 937 938:Capability: KVM_CAP_XEN_HVM 939:Architectures: x86 940:Type: vm ioctl 941:Parameters: struct kvm_xen_hvm_config (in) 942:Returns: 0 on success, -1 on error 943 944Sets the MSR that the Xen HVM guest uses to initialize its hypercall 945page, and provides the starting address and size of the hypercall 946blobs in userspace. When the guest writes the MSR, kvm copies one 947page of a blob (32- or 64-bit, depending on the vcpu mode) to guest 948memory. 949 950:: 951 952 struct kvm_xen_hvm_config { 953 __u32 flags; 954 __u32 msr; 955 __u64 blob_addr_32; 956 __u64 blob_addr_64; 957 __u8 blob_size_32; 958 __u8 blob_size_64; 959 __u8 pad2[30]; 960 }; 961 962If certain flags are returned from the KVM_CAP_XEN_HVM check, they may 963be set in the flags field of this ioctl: 964 965The KVM_XEN_HVM_CONFIG_INTERCEPT_HCALL flag requests KVM to generate 966the contents of the hypercall page automatically; hypercalls will be 967intercepted and passed to userspace through KVM_EXIT_XEN. In this 968ase, all of the blob size and address fields must be zero. 969 970The KVM_XEN_HVM_CONFIG_EVTCHN_SEND flag indicates to KVM that userspace 971will always use the KVM_XEN_HVM_EVTCHN_SEND ioctl to deliver event 972channel interrupts rather than manipulating the guest's shared_info 973structures directly. This, in turn, may allow KVM to enable features 974such as intercepting the SCHEDOP_poll hypercall to accelerate PV 975spinlock operation for the guest. Userspace may still use the ioctl 976to deliver events if it was advertised, even if userspace does not 977send this indication that it will always do so 978 979No other flags are currently valid in the struct kvm_xen_hvm_config. 980 9814.29 KVM_GET_CLOCK 982------------------ 983 984:Capability: KVM_CAP_ADJUST_CLOCK 985:Architectures: x86 986:Type: vm ioctl 987:Parameters: struct kvm_clock_data (out) 988:Returns: 0 on success, -1 on error 989 990Gets the current timestamp of kvmclock as seen by the current guest. In 991conjunction with KVM_SET_CLOCK, it is used to ensure monotonicity on scenarios 992such as migration. 993 994When KVM_CAP_ADJUST_CLOCK is passed to KVM_CHECK_EXTENSION, it returns the 995set of bits that KVM can return in struct kvm_clock_data's flag member. 996 997The following flags are defined: 998 999KVM_CLOCK_TSC_STABLE 1000 If set, the returned value is the exact kvmclock 1001 value seen by all VCPUs at the instant when KVM_GET_CLOCK was called. 1002 If clear, the returned value is simply CLOCK_MONOTONIC plus a constant 1003 offset; the offset can be modified with KVM_SET_CLOCK. KVM will try 1004 to make all VCPUs follow this clock, but the exact value read by each 1005 VCPU could differ, because the host TSC is not stable. 1006 1007KVM_CLOCK_REALTIME 1008 If set, the `realtime` field in the kvm_clock_data 1009 structure is populated with the value of the host's real time 1010 clocksource at the instant when KVM_GET_CLOCK was called. If clear, 1011 the `realtime` field does not contain a value. 1012 1013KVM_CLOCK_HOST_TSC 1014 If set, the `host_tsc` field in the kvm_clock_data 1015 structure is populated with the value of the host's timestamp counter (TSC) 1016 at the instant when KVM_GET_CLOCK was called. If clear, the `host_tsc` field 1017 does not contain a value. 1018 1019:: 1020 1021 struct kvm_clock_data { 1022 __u64 clock; /* kvmclock current value */ 1023 __u32 flags; 1024 __u32 pad0; 1025 __u64 realtime; 1026 __u64 host_tsc; 1027 __u32 pad[4]; 1028 }; 1029 1030 10314.30 KVM_SET_CLOCK 1032------------------ 1033 1034:Capability: KVM_CAP_ADJUST_CLOCK 1035:Architectures: x86 1036:Type: vm ioctl 1037:Parameters: struct kvm_clock_data (in) 1038:Returns: 0 on success, -1 on error 1039 1040Sets the current timestamp of kvmclock to the value specified in its parameter. 1041In conjunction with KVM_GET_CLOCK, it is used to ensure monotonicity on scenarios 1042such as migration. 1043 1044The following flags can be passed: 1045 1046KVM_CLOCK_REALTIME 1047 If set, KVM will compare the value of the `realtime` field 1048 with the value of the host's real time clocksource at the instant when 1049 KVM_SET_CLOCK was called. The difference in elapsed time is added to the final 1050 kvmclock value that will be provided to guests. 1051 1052Other flags returned by ``KVM_GET_CLOCK`` are accepted but ignored. 1053 1054:: 1055 1056 struct kvm_clock_data { 1057 __u64 clock; /* kvmclock current value */ 1058 __u32 flags; 1059 __u32 pad0; 1060 __u64 realtime; 1061 __u64 host_tsc; 1062 __u32 pad[4]; 1063 }; 1064 1065 10664.31 KVM_GET_VCPU_EVENTS 1067------------------------ 1068 1069:Capability: KVM_CAP_VCPU_EVENTS 1070:Extended by: KVM_CAP_INTR_SHADOW 1071:Architectures: x86, arm64 1072:Type: vcpu ioctl 1073:Parameters: struct kvm_vcpu_event (out) 1074:Returns: 0 on success, -1 on error 1075 1076X86: 1077^^^^ 1078 1079Gets currently pending exceptions, interrupts, and NMIs as well as related 1080states of the vcpu. 1081 1082:: 1083 1084 struct kvm_vcpu_events { 1085 struct { 1086 __u8 injected; 1087 __u8 nr; 1088 __u8 has_error_code; 1089 __u8 pending; 1090 __u32 error_code; 1091 } exception; 1092 struct { 1093 __u8 injected; 1094 __u8 nr; 1095 __u8 soft; 1096 __u8 shadow; 1097 } interrupt; 1098 struct { 1099 __u8 injected; 1100 __u8 pending; 1101 __u8 masked; 1102 __u8 pad; 1103 } nmi; 1104 __u32 sipi_vector; 1105 __u32 flags; 1106 struct { 1107 __u8 smm; 1108 __u8 pending; 1109 __u8 smm_inside_nmi; 1110 __u8 latched_init; 1111 } smi; 1112 __u8 reserved[27]; 1113 __u8 exception_has_payload; 1114 __u64 exception_payload; 1115 }; 1116 1117The following bits are defined in the flags field: 1118 1119- KVM_VCPUEVENT_VALID_SHADOW may be set to signal that 1120 interrupt.shadow contains a valid state. 1121 1122- KVM_VCPUEVENT_VALID_SMM may be set to signal that smi contains a 1123 valid state. 1124 1125- KVM_VCPUEVENT_VALID_PAYLOAD may be set to signal that the 1126 exception_has_payload, exception_payload, and exception.pending 1127 fields contain a valid state. This bit will be set whenever 1128 KVM_CAP_EXCEPTION_PAYLOAD is enabled. 1129 1130- KVM_VCPUEVENT_VALID_TRIPLE_FAULT may be set to signal that the 1131 triple_fault_pending field contains a valid state. This bit will 1132 be set whenever KVM_CAP_X86_TRIPLE_FAULT_EVENT is enabled. 1133 1134ARM64: 1135^^^^^^ 1136 1137If the guest accesses a device that is being emulated by the host kernel in 1138such a way that a real device would generate a physical SError, KVM may make 1139a virtual SError pending for that VCPU. This system error interrupt remains 1140pending until the guest takes the exception by unmasking PSTATE.A. 1141 1142Running the VCPU may cause it to take a pending SError, or make an access that 1143causes an SError to become pending. The event's description is only valid while 1144the VPCU is not running. 1145 1146This API provides a way to read and write the pending 'event' state that is not 1147visible to the guest. To save, restore or migrate a VCPU the struct representing 1148the state can be read then written using this GET/SET API, along with the other 1149guest-visible registers. It is not possible to 'cancel' an SError that has been 1150made pending. 1151 1152A device being emulated in user-space may also wish to generate an SError. To do 1153this the events structure can be populated by user-space. The current state 1154should be read first, to ensure no existing SError is pending. If an existing 1155SError is pending, the architecture's 'Multiple SError interrupts' rules should 1156be followed. (2.5.3 of DDI0587.a "ARM Reliability, Availability, and 1157Serviceability (RAS) Specification"). 1158 1159SError exceptions always have an ESR value. Some CPUs have the ability to 1160specify what the virtual SError's ESR value should be. These systems will 1161advertise KVM_CAP_ARM_INJECT_SERROR_ESR. In this case exception.has_esr will 1162always have a non-zero value when read, and the agent making an SError pending 1163should specify the ISS field in the lower 24 bits of exception.serror_esr. If 1164the system supports KVM_CAP_ARM_INJECT_SERROR_ESR, but user-space sets the events 1165with exception.has_esr as zero, KVM will choose an ESR. 1166 1167Specifying exception.has_esr on a system that does not support it will return 1168-EINVAL. Setting anything other than the lower 24bits of exception.serror_esr 1169will return -EINVAL. 1170 1171It is not possible to read back a pending external abort (injected via 1172KVM_SET_VCPU_EVENTS or otherwise) because such an exception is always delivered 1173directly to the virtual CPU). 1174 1175:: 1176 1177 struct kvm_vcpu_events { 1178 struct { 1179 __u8 serror_pending; 1180 __u8 serror_has_esr; 1181 __u8 ext_dabt_pending; 1182 /* Align it to 8 bytes */ 1183 __u8 pad[5]; 1184 __u64 serror_esr; 1185 } exception; 1186 __u32 reserved[12]; 1187 }; 1188 11894.32 KVM_SET_VCPU_EVENTS 1190------------------------ 1191 1192:Capability: KVM_CAP_VCPU_EVENTS 1193:Extended by: KVM_CAP_INTR_SHADOW 1194:Architectures: x86, arm64 1195:Type: vcpu ioctl 1196:Parameters: struct kvm_vcpu_event (in) 1197:Returns: 0 on success, -1 on error 1198 1199X86: 1200^^^^ 1201 1202Set pending exceptions, interrupts, and NMIs as well as related states of the 1203vcpu. 1204 1205See KVM_GET_VCPU_EVENTS for the data structure. 1206 1207Fields that may be modified asynchronously by running VCPUs can be excluded 1208from the update. These fields are nmi.pending, sipi_vector, smi.smm, 1209smi.pending. Keep the corresponding bits in the flags field cleared to 1210suppress overwriting the current in-kernel state. The bits are: 1211 1212=============================== ================================== 1213KVM_VCPUEVENT_VALID_NMI_PENDING transfer nmi.pending to the kernel 1214KVM_VCPUEVENT_VALID_SIPI_VECTOR transfer sipi_vector 1215KVM_VCPUEVENT_VALID_SMM transfer the smi sub-struct. 1216=============================== ================================== 1217 1218If KVM_CAP_INTR_SHADOW is available, KVM_VCPUEVENT_VALID_SHADOW can be set in 1219the flags field to signal that interrupt.shadow contains a valid state and 1220shall be written into the VCPU. 1221 1222KVM_VCPUEVENT_VALID_SMM can only be set if KVM_CAP_X86_SMM is available. 1223 1224If KVM_CAP_EXCEPTION_PAYLOAD is enabled, KVM_VCPUEVENT_VALID_PAYLOAD 1225can be set in the flags field to signal that the 1226exception_has_payload, exception_payload, and exception.pending fields 1227contain a valid state and shall be written into the VCPU. 1228 1229If KVM_CAP_X86_TRIPLE_FAULT_EVENT is enabled, KVM_VCPUEVENT_VALID_TRIPLE_FAULT 1230can be set in flags field to signal that the triple_fault field contains 1231a valid state and shall be written into the VCPU. 1232 1233ARM64: 1234^^^^^^ 1235 1236User space may need to inject several types of events to the guest. 1237 1238Set the pending SError exception state for this VCPU. It is not possible to 1239'cancel' an Serror that has been made pending. 1240 1241If the guest performed an access to I/O memory which could not be handled by 1242userspace, for example because of missing instruction syndrome decode 1243information or because there is no device mapped at the accessed IPA, then 1244userspace can ask the kernel to inject an external abort using the address 1245from the exiting fault on the VCPU. It is a programming error to set 1246ext_dabt_pending after an exit which was not either KVM_EXIT_MMIO or 1247KVM_EXIT_ARM_NISV. This feature is only available if the system supports 1248KVM_CAP_ARM_INJECT_EXT_DABT. This is a helper which provides commonality in 1249how userspace reports accesses for the above cases to guests, across different 1250userspace implementations. Nevertheless, userspace can still emulate all Arm 1251exceptions by manipulating individual registers using the KVM_SET_ONE_REG API. 1252 1253See KVM_GET_VCPU_EVENTS for the data structure. 1254 1255 12564.33 KVM_GET_DEBUGREGS 1257---------------------- 1258 1259:Capability: KVM_CAP_DEBUGREGS 1260:Architectures: x86 1261:Type: vm ioctl 1262:Parameters: struct kvm_debugregs (out) 1263:Returns: 0 on success, -1 on error 1264 1265Reads debug registers from the vcpu. 1266 1267:: 1268 1269 struct kvm_debugregs { 1270 __u64 db[4]; 1271 __u64 dr6; 1272 __u64 dr7; 1273 __u64 flags; 1274 __u64 reserved[9]; 1275 }; 1276 1277 12784.34 KVM_SET_DEBUGREGS 1279---------------------- 1280 1281:Capability: KVM_CAP_DEBUGREGS 1282:Architectures: x86 1283:Type: vm ioctl 1284:Parameters: struct kvm_debugregs (in) 1285:Returns: 0 on success, -1 on error 1286 1287Writes debug registers into the vcpu. 1288 1289See KVM_GET_DEBUGREGS for the data structure. The flags field is unused 1290yet and must be cleared on entry. 1291 1292 12934.35 KVM_SET_USER_MEMORY_REGION 1294------------------------------- 1295 1296:Capability: KVM_CAP_USER_MEMORY 1297:Architectures: all 1298:Type: vm ioctl 1299:Parameters: struct kvm_userspace_memory_region (in) 1300:Returns: 0 on success, -1 on error 1301 1302:: 1303 1304 struct kvm_userspace_memory_region { 1305 __u32 slot; 1306 __u32 flags; 1307 __u64 guest_phys_addr; 1308 __u64 memory_size; /* bytes */ 1309 __u64 userspace_addr; /* start of the userspace allocated memory */ 1310 }; 1311 1312 /* for kvm_userspace_memory_region::flags */ 1313 #define KVM_MEM_LOG_DIRTY_PAGES (1UL << 0) 1314 #define KVM_MEM_READONLY (1UL << 1) 1315 1316This ioctl allows the user to create, modify or delete a guest physical 1317memory slot. Bits 0-15 of "slot" specify the slot id and this value 1318should be less than the maximum number of user memory slots supported per 1319VM. The maximum allowed slots can be queried using KVM_CAP_NR_MEMSLOTS. 1320Slots may not overlap in guest physical address space. 1321 1322If KVM_CAP_MULTI_ADDRESS_SPACE is available, bits 16-31 of "slot" 1323specifies the address space which is being modified. They must be 1324less than the value that KVM_CHECK_EXTENSION returns for the 1325KVM_CAP_MULTI_ADDRESS_SPACE capability. Slots in separate address spaces 1326are unrelated; the restriction on overlapping slots only applies within 1327each address space. 1328 1329Deleting a slot is done by passing zero for memory_size. When changing 1330an existing slot, it may be moved in the guest physical memory space, 1331or its flags may be modified, but it may not be resized. 1332 1333Memory for the region is taken starting at the address denoted by the 1334field userspace_addr, which must point at user addressable memory for 1335the entire memory slot size. Any object may back this memory, including 1336anonymous memory, ordinary files, and hugetlbfs. 1337 1338On architectures that support a form of address tagging, userspace_addr must 1339be an untagged address. 1340 1341It is recommended that the lower 21 bits of guest_phys_addr and userspace_addr 1342be identical. This allows large pages in the guest to be backed by large 1343pages in the host. 1344 1345The flags field supports two flags: KVM_MEM_LOG_DIRTY_PAGES and 1346KVM_MEM_READONLY. The former can be set to instruct KVM to keep track of 1347writes to memory within the slot. See KVM_GET_DIRTY_LOG ioctl to know how to 1348use it. The latter can be set, if KVM_CAP_READONLY_MEM capability allows it, 1349to make a new slot read-only. In this case, writes to this memory will be 1350posted to userspace as KVM_EXIT_MMIO exits. 1351 1352When the KVM_CAP_SYNC_MMU capability is available, changes in the backing of 1353the memory region are automatically reflected into the guest. For example, an 1354mmap() that affects the region will be made visible immediately. Another 1355example is madvise(MADV_DROP). 1356 1357 13584.36 KVM_SET_TSS_ADDR 1359--------------------- 1360 1361:Capability: KVM_CAP_SET_TSS_ADDR 1362:Architectures: x86 1363:Type: vm ioctl 1364:Parameters: unsigned long tss_address (in) 1365:Returns: 0 on success, -1 on error 1366 1367This ioctl defines the physical address of a three-page region in the guest 1368physical address space. The region must be within the first 4GB of the 1369guest physical address space and must not conflict with any memory slot 1370or any mmio address. The guest may malfunction if it accesses this memory 1371region. 1372 1373This ioctl is required on Intel-based hosts. This is needed on Intel hardware 1374because of a quirk in the virtualization implementation (see the internals 1375documentation when it pops into existence). 1376 1377 13784.37 KVM_ENABLE_CAP 1379------------------- 1380 1381:Capability: KVM_CAP_ENABLE_CAP 1382:Architectures: mips, ppc, s390, x86 1383:Type: vcpu ioctl 1384:Parameters: struct kvm_enable_cap (in) 1385:Returns: 0 on success; -1 on error 1386 1387:Capability: KVM_CAP_ENABLE_CAP_VM 1388:Architectures: all 1389:Type: vm ioctl 1390:Parameters: struct kvm_enable_cap (in) 1391:Returns: 0 on success; -1 on error 1392 1393.. note:: 1394 1395 Not all extensions are enabled by default. Using this ioctl the application 1396 can enable an extension, making it available to the guest. 1397 1398On systems that do not support this ioctl, it always fails. On systems that 1399do support it, it only works for extensions that are supported for enablement. 1400 1401To check if a capability can be enabled, the KVM_CHECK_EXTENSION ioctl should 1402be used. 1403 1404:: 1405 1406 struct kvm_enable_cap { 1407 /* in */ 1408 __u32 cap; 1409 1410The capability that is supposed to get enabled. 1411 1412:: 1413 1414 __u32 flags; 1415 1416A bitfield indicating future enhancements. Has to be 0 for now. 1417 1418:: 1419 1420 __u64 args[4]; 1421 1422Arguments for enabling a feature. If a feature needs initial values to 1423function properly, this is the place to put them. 1424 1425:: 1426 1427 __u8 pad[64]; 1428 }; 1429 1430The vcpu ioctl should be used for vcpu-specific capabilities, the vm ioctl 1431for vm-wide capabilities. 1432 14334.38 KVM_GET_MP_STATE 1434--------------------- 1435 1436:Capability: KVM_CAP_MP_STATE 1437:Architectures: x86, s390, arm64, riscv 1438:Type: vcpu ioctl 1439:Parameters: struct kvm_mp_state (out) 1440:Returns: 0 on success; -1 on error 1441 1442:: 1443 1444 struct kvm_mp_state { 1445 __u32 mp_state; 1446 }; 1447 1448Returns the vcpu's current "multiprocessing state" (though also valid on 1449uniprocessor guests). 1450 1451Possible values are: 1452 1453 ========================== =============================================== 1454 KVM_MP_STATE_RUNNABLE the vcpu is currently running 1455 [x86,arm64,riscv] 1456 KVM_MP_STATE_UNINITIALIZED the vcpu is an application processor (AP) 1457 which has not yet received an INIT signal [x86] 1458 KVM_MP_STATE_INIT_RECEIVED the vcpu has received an INIT signal, and is 1459 now ready for a SIPI [x86] 1460 KVM_MP_STATE_HALTED the vcpu has executed a HLT instruction and 1461 is waiting for an interrupt [x86] 1462 KVM_MP_STATE_SIPI_RECEIVED the vcpu has just received a SIPI (vector 1463 accessible via KVM_GET_VCPU_EVENTS) [x86] 1464 KVM_MP_STATE_STOPPED the vcpu is stopped [s390,arm64,riscv] 1465 KVM_MP_STATE_CHECK_STOP the vcpu is in a special error state [s390] 1466 KVM_MP_STATE_OPERATING the vcpu is operating (running or halted) 1467 [s390] 1468 KVM_MP_STATE_LOAD the vcpu is in a special load/startup state 1469 [s390] 1470 KVM_MP_STATE_SUSPENDED the vcpu is in a suspend state and is waiting 1471 for a wakeup event [arm64] 1472 ========================== =============================================== 1473 1474On x86, this ioctl is only useful after KVM_CREATE_IRQCHIP. Without an 1475in-kernel irqchip, the multiprocessing state must be maintained by userspace on 1476these architectures. 1477 1478For arm64: 1479^^^^^^^^^^ 1480 1481If a vCPU is in the KVM_MP_STATE_SUSPENDED state, KVM will emulate the 1482architectural execution of a WFI instruction. 1483 1484If a wakeup event is recognized, KVM will exit to userspace with a 1485KVM_SYSTEM_EVENT exit, where the event type is KVM_SYSTEM_EVENT_WAKEUP. If 1486userspace wants to honor the wakeup, it must set the vCPU's MP state to 1487KVM_MP_STATE_RUNNABLE. If it does not, KVM will continue to await a wakeup 1488event in subsequent calls to KVM_RUN. 1489 1490.. warning:: 1491 1492 If userspace intends to keep the vCPU in a SUSPENDED state, it is 1493 strongly recommended that userspace take action to suppress the 1494 wakeup event (such as masking an interrupt). Otherwise, subsequent 1495 calls to KVM_RUN will immediately exit with a KVM_SYSTEM_EVENT_WAKEUP 1496 event and inadvertently waste CPU cycles. 1497 1498 Additionally, if userspace takes action to suppress a wakeup event, 1499 it is strongly recommended that it also restores the vCPU to its 1500 original state when the vCPU is made RUNNABLE again. For example, 1501 if userspace masked a pending interrupt to suppress the wakeup, 1502 the interrupt should be unmasked before returning control to the 1503 guest. 1504 1505For riscv: 1506^^^^^^^^^^ 1507 1508The only states that are valid are KVM_MP_STATE_STOPPED and 1509KVM_MP_STATE_RUNNABLE which reflect if the vcpu is paused or not. 1510 15114.39 KVM_SET_MP_STATE 1512--------------------- 1513 1514:Capability: KVM_CAP_MP_STATE 1515:Architectures: x86, s390, arm64, riscv 1516:Type: vcpu ioctl 1517:Parameters: struct kvm_mp_state (in) 1518:Returns: 0 on success; -1 on error 1519 1520Sets the vcpu's current "multiprocessing state"; see KVM_GET_MP_STATE for 1521arguments. 1522 1523On x86, this ioctl is only useful after KVM_CREATE_IRQCHIP. Without an 1524in-kernel irqchip, the multiprocessing state must be maintained by userspace on 1525these architectures. 1526 1527For arm64/riscv: 1528^^^^^^^^^^^^^^^^ 1529 1530The only states that are valid are KVM_MP_STATE_STOPPED and 1531KVM_MP_STATE_RUNNABLE which reflect if the vcpu should be paused or not. 1532 15334.40 KVM_SET_IDENTITY_MAP_ADDR 1534------------------------------ 1535 1536:Capability: KVM_CAP_SET_IDENTITY_MAP_ADDR 1537:Architectures: x86 1538:Type: vm ioctl 1539:Parameters: unsigned long identity (in) 1540:Returns: 0 on success, -1 on error 1541 1542This ioctl defines the physical address of a one-page region in the guest 1543physical address space. The region must be within the first 4GB of the 1544guest physical address space and must not conflict with any memory slot 1545or any mmio address. The guest may malfunction if it accesses this memory 1546region. 1547 1548Setting the address to 0 will result in resetting the address to its default 1549(0xfffbc000). 1550 1551This ioctl is required on Intel-based hosts. This is needed on Intel hardware 1552because of a quirk in the virtualization implementation (see the internals 1553documentation when it pops into existence). 1554 1555Fails if any VCPU has already been created. 1556 15574.41 KVM_SET_BOOT_CPU_ID 1558------------------------ 1559 1560:Capability: KVM_CAP_SET_BOOT_CPU_ID 1561:Architectures: x86 1562:Type: vm ioctl 1563:Parameters: unsigned long vcpu_id 1564:Returns: 0 on success, -1 on error 1565 1566Define which vcpu is the Bootstrap Processor (BSP). Values are the same 1567as the vcpu id in KVM_CREATE_VCPU. If this ioctl is not called, the default 1568is vcpu 0. This ioctl has to be called before vcpu creation, 1569otherwise it will return EBUSY error. 1570 1571 15724.42 KVM_GET_XSAVE 1573------------------ 1574 1575:Capability: KVM_CAP_XSAVE 1576:Architectures: x86 1577:Type: vcpu ioctl 1578:Parameters: struct kvm_xsave (out) 1579:Returns: 0 on success, -1 on error 1580 1581 1582:: 1583 1584 struct kvm_xsave { 1585 __u32 region[1024]; 1586 __u32 extra[0]; 1587 }; 1588 1589This ioctl would copy current vcpu's xsave struct to the userspace. 1590 1591 15924.43 KVM_SET_XSAVE 1593------------------ 1594 1595:Capability: KVM_CAP_XSAVE and KVM_CAP_XSAVE2 1596:Architectures: x86 1597:Type: vcpu ioctl 1598:Parameters: struct kvm_xsave (in) 1599:Returns: 0 on success, -1 on error 1600 1601:: 1602 1603 1604 struct kvm_xsave { 1605 __u32 region[1024]; 1606 __u32 extra[0]; 1607 }; 1608 1609This ioctl would copy userspace's xsave struct to the kernel. It copies 1610as many bytes as are returned by KVM_CHECK_EXTENSION(KVM_CAP_XSAVE2), 1611when invoked on the vm file descriptor. The size value returned by 1612KVM_CHECK_EXTENSION(KVM_CAP_XSAVE2) will always be at least 4096. 1613Currently, it is only greater than 4096 if a dynamic feature has been 1614enabled with ``arch_prctl()``, but this may change in the future. 1615 1616The offsets of the state save areas in struct kvm_xsave follow the 1617contents of CPUID leaf 0xD on the host. 1618 1619 16204.44 KVM_GET_XCRS 1621----------------- 1622 1623:Capability: KVM_CAP_XCRS 1624:Architectures: x86 1625:Type: vcpu ioctl 1626:Parameters: struct kvm_xcrs (out) 1627:Returns: 0 on success, -1 on error 1628 1629:: 1630 1631 struct kvm_xcr { 1632 __u32 xcr; 1633 __u32 reserved; 1634 __u64 value; 1635 }; 1636 1637 struct kvm_xcrs { 1638 __u32 nr_xcrs; 1639 __u32 flags; 1640 struct kvm_xcr xcrs[KVM_MAX_XCRS]; 1641 __u64 padding[16]; 1642 }; 1643 1644This ioctl would copy current vcpu's xcrs to the userspace. 1645 1646 16474.45 KVM_SET_XCRS 1648----------------- 1649 1650:Capability: KVM_CAP_XCRS 1651:Architectures: x86 1652:Type: vcpu ioctl 1653:Parameters: struct kvm_xcrs (in) 1654:Returns: 0 on success, -1 on error 1655 1656:: 1657 1658 struct kvm_xcr { 1659 __u32 xcr; 1660 __u32 reserved; 1661 __u64 value; 1662 }; 1663 1664 struct kvm_xcrs { 1665 __u32 nr_xcrs; 1666 __u32 flags; 1667 struct kvm_xcr xcrs[KVM_MAX_XCRS]; 1668 __u64 padding[16]; 1669 }; 1670 1671This ioctl would set vcpu's xcr to the value userspace specified. 1672 1673 16744.46 KVM_GET_SUPPORTED_CPUID 1675---------------------------- 1676 1677:Capability: KVM_CAP_EXT_CPUID 1678:Architectures: x86 1679:Type: system ioctl 1680:Parameters: struct kvm_cpuid2 (in/out) 1681:Returns: 0 on success, -1 on error 1682 1683:: 1684 1685 struct kvm_cpuid2 { 1686 __u32 nent; 1687 __u32 padding; 1688 struct kvm_cpuid_entry2 entries[0]; 1689 }; 1690 1691 #define KVM_CPUID_FLAG_SIGNIFCANT_INDEX BIT(0) 1692 #define KVM_CPUID_FLAG_STATEFUL_FUNC BIT(1) /* deprecated */ 1693 #define KVM_CPUID_FLAG_STATE_READ_NEXT BIT(2) /* deprecated */ 1694 1695 struct kvm_cpuid_entry2 { 1696 __u32 function; 1697 __u32 index; 1698 __u32 flags; 1699 __u32 eax; 1700 __u32 ebx; 1701 __u32 ecx; 1702 __u32 edx; 1703 __u32 padding[3]; 1704 }; 1705 1706This ioctl returns x86 cpuid features which are supported by both the 1707hardware and kvm in its default configuration. Userspace can use the 1708information returned by this ioctl to construct cpuid information (for 1709KVM_SET_CPUID2) that is consistent with hardware, kernel, and 1710userspace capabilities, and with user requirements (for example, the 1711user may wish to constrain cpuid to emulate older hardware, or for 1712feature consistency across a cluster). 1713 1714Dynamically-enabled feature bits need to be requested with 1715``arch_prctl()`` before calling this ioctl. Feature bits that have not 1716been requested are excluded from the result. 1717 1718Note that certain capabilities, such as KVM_CAP_X86_DISABLE_EXITS, may 1719expose cpuid features (e.g. MONITOR) which are not supported by kvm in 1720its default configuration. If userspace enables such capabilities, it 1721is responsible for modifying the results of this ioctl appropriately. 1722 1723Userspace invokes KVM_GET_SUPPORTED_CPUID by passing a kvm_cpuid2 structure 1724with the 'nent' field indicating the number of entries in the variable-size 1725array 'entries'. If the number of entries is too low to describe the cpu 1726capabilities, an error (E2BIG) is returned. If the number is too high, 1727the 'nent' field is adjusted and an error (ENOMEM) is returned. If the 1728number is just right, the 'nent' field is adjusted to the number of valid 1729entries in the 'entries' array, which is then filled. 1730 1731The entries returned are the host cpuid as returned by the cpuid instruction, 1732with unknown or unsupported features masked out. Some features (for example, 1733x2apic), may not be present in the host cpu, but are exposed by kvm if it can 1734emulate them efficiently. The fields in each entry are defined as follows: 1735 1736 function: 1737 the eax value used to obtain the entry 1738 1739 index: 1740 the ecx value used to obtain the entry (for entries that are 1741 affected by ecx) 1742 1743 flags: 1744 an OR of zero or more of the following: 1745 1746 KVM_CPUID_FLAG_SIGNIFCANT_INDEX: 1747 if the index field is valid 1748 1749 eax, ebx, ecx, edx: 1750 the values returned by the cpuid instruction for 1751 this function/index combination 1752 1753The TSC deadline timer feature (CPUID leaf 1, ecx[24]) is always returned 1754as false, since the feature depends on KVM_CREATE_IRQCHIP for local APIC 1755support. Instead it is reported via:: 1756 1757 ioctl(KVM_CHECK_EXTENSION, KVM_CAP_TSC_DEADLINE_TIMER) 1758 1759if that returns true and you use KVM_CREATE_IRQCHIP, or if you emulate the 1760feature in userspace, then you can enable the feature for KVM_SET_CPUID2. 1761 1762 17634.47 KVM_PPC_GET_PVINFO 1764----------------------- 1765 1766:Capability: KVM_CAP_PPC_GET_PVINFO 1767:Architectures: ppc 1768:Type: vm ioctl 1769:Parameters: struct kvm_ppc_pvinfo (out) 1770:Returns: 0 on success, !0 on error 1771 1772:: 1773 1774 struct kvm_ppc_pvinfo { 1775 __u32 flags; 1776 __u32 hcall[4]; 1777 __u8 pad[108]; 1778 }; 1779 1780This ioctl fetches PV specific information that need to be passed to the guest 1781using the device tree or other means from vm context. 1782 1783The hcall array defines 4 instructions that make up a hypercall. 1784 1785If any additional field gets added to this structure later on, a bit for that 1786additional piece of information will be set in the flags bitmap. 1787 1788The flags bitmap is defined as:: 1789 1790 /* the host supports the ePAPR idle hcall 1791 #define KVM_PPC_PVINFO_FLAGS_EV_IDLE (1<<0) 1792 17934.52 KVM_SET_GSI_ROUTING 1794------------------------ 1795 1796:Capability: KVM_CAP_IRQ_ROUTING 1797:Architectures: x86 s390 arm64 1798:Type: vm ioctl 1799:Parameters: struct kvm_irq_routing (in) 1800:Returns: 0 on success, -1 on error 1801 1802Sets the GSI routing table entries, overwriting any previously set entries. 1803 1804On arm64, GSI routing has the following limitation: 1805 1806- GSI routing does not apply to KVM_IRQ_LINE but only to KVM_IRQFD. 1807 1808:: 1809 1810 struct kvm_irq_routing { 1811 __u32 nr; 1812 __u32 flags; 1813 struct kvm_irq_routing_entry entries[0]; 1814 }; 1815 1816No flags are specified so far, the corresponding field must be set to zero. 1817 1818:: 1819 1820 struct kvm_irq_routing_entry { 1821 __u32 gsi; 1822 __u32 type; 1823 __u32 flags; 1824 __u32 pad; 1825 union { 1826 struct kvm_irq_routing_irqchip irqchip; 1827 struct kvm_irq_routing_msi msi; 1828 struct kvm_irq_routing_s390_adapter adapter; 1829 struct kvm_irq_routing_hv_sint hv_sint; 1830 struct kvm_irq_routing_xen_evtchn xen_evtchn; 1831 __u32 pad[8]; 1832 } u; 1833 }; 1834 1835 /* gsi routing entry types */ 1836 #define KVM_IRQ_ROUTING_IRQCHIP 1 1837 #define KVM_IRQ_ROUTING_MSI 2 1838 #define KVM_IRQ_ROUTING_S390_ADAPTER 3 1839 #define KVM_IRQ_ROUTING_HV_SINT 4 1840 #define KVM_IRQ_ROUTING_XEN_EVTCHN 5 1841 1842flags: 1843 1844- KVM_MSI_VALID_DEVID: used along with KVM_IRQ_ROUTING_MSI routing entry 1845 type, specifies that the devid field contains a valid value. The per-VM 1846 KVM_CAP_MSI_DEVID capability advertises the requirement to provide 1847 the device ID. If this capability is not available, userspace should 1848 never set the KVM_MSI_VALID_DEVID flag as the ioctl might fail. 1849- zero otherwise 1850 1851:: 1852 1853 struct kvm_irq_routing_irqchip { 1854 __u32 irqchip; 1855 __u32 pin; 1856 }; 1857 1858 struct kvm_irq_routing_msi { 1859 __u32 address_lo; 1860 __u32 address_hi; 1861 __u32 data; 1862 union { 1863 __u32 pad; 1864 __u32 devid; 1865 }; 1866 }; 1867 1868If KVM_MSI_VALID_DEVID is set, devid contains a unique device identifier 1869for the device that wrote the MSI message. For PCI, this is usually a 1870BFD identifier in the lower 16 bits. 1871 1872On x86, address_hi is ignored unless the KVM_X2APIC_API_USE_32BIT_IDS 1873feature of KVM_CAP_X2APIC_API capability is enabled. If it is enabled, 1874address_hi bits 31-8 provide bits 31-8 of the destination id. Bits 7-0 of 1875address_hi must be zero. 1876 1877:: 1878 1879 struct kvm_irq_routing_s390_adapter { 1880 __u64 ind_addr; 1881 __u64 summary_addr; 1882 __u64 ind_offset; 1883 __u32 summary_offset; 1884 __u32 adapter_id; 1885 }; 1886 1887 struct kvm_irq_routing_hv_sint { 1888 __u32 vcpu; 1889 __u32 sint; 1890 }; 1891 1892 struct kvm_irq_routing_xen_evtchn { 1893 __u32 port; 1894 __u32 vcpu; 1895 __u32 priority; 1896 }; 1897 1898 1899When KVM_CAP_XEN_HVM includes the KVM_XEN_HVM_CONFIG_EVTCHN_2LEVEL bit 1900in its indication of supported features, routing to Xen event channels 1901is supported. Although the priority field is present, only the value 1902KVM_XEN_HVM_CONFIG_EVTCHN_2LEVEL is supported, which means delivery by 19032 level event channels. FIFO event channel support may be added in 1904the future. 1905 1906 19074.55 KVM_SET_TSC_KHZ 1908-------------------- 1909 1910:Capability: KVM_CAP_TSC_CONTROL / KVM_CAP_VM_TSC_CONTROL 1911:Architectures: x86 1912:Type: vcpu ioctl / vm ioctl 1913:Parameters: virtual tsc_khz 1914:Returns: 0 on success, -1 on error 1915 1916Specifies the tsc frequency for the virtual machine. The unit of the 1917frequency is KHz. 1918 1919If the KVM_CAP_VM_TSC_CONTROL capability is advertised, this can also 1920be used as a vm ioctl to set the initial tsc frequency of subsequently 1921created vCPUs. 1922 19234.56 KVM_GET_TSC_KHZ 1924-------------------- 1925 1926:Capability: KVM_CAP_GET_TSC_KHZ / KVM_CAP_VM_TSC_CONTROL 1927:Architectures: x86 1928:Type: vcpu ioctl / vm ioctl 1929:Parameters: none 1930:Returns: virtual tsc-khz on success, negative value on error 1931 1932Returns the tsc frequency of the guest. The unit of the return value is 1933KHz. If the host has unstable tsc this ioctl returns -EIO instead as an 1934error. 1935 1936 19374.57 KVM_GET_LAPIC 1938------------------ 1939 1940:Capability: KVM_CAP_IRQCHIP 1941:Architectures: x86 1942:Type: vcpu ioctl 1943:Parameters: struct kvm_lapic_state (out) 1944:Returns: 0 on success, -1 on error 1945 1946:: 1947 1948 #define KVM_APIC_REG_SIZE 0x400 1949 struct kvm_lapic_state { 1950 char regs[KVM_APIC_REG_SIZE]; 1951 }; 1952 1953Reads the Local APIC registers and copies them into the input argument. The 1954data format and layout are the same as documented in the architecture manual. 1955 1956If KVM_X2APIC_API_USE_32BIT_IDS feature of KVM_CAP_X2APIC_API is 1957enabled, then the format of APIC_ID register depends on the APIC mode 1958(reported by MSR_IA32_APICBASE) of its VCPU. x2APIC stores APIC ID in 1959the APIC_ID register (bytes 32-35). xAPIC only allows an 8-bit APIC ID 1960which is stored in bits 31-24 of the APIC register, or equivalently in 1961byte 35 of struct kvm_lapic_state's regs field. KVM_GET_LAPIC must then 1962be called after MSR_IA32_APICBASE has been set with KVM_SET_MSR. 1963 1964If KVM_X2APIC_API_USE_32BIT_IDS feature is disabled, struct kvm_lapic_state 1965always uses xAPIC format. 1966 1967 19684.58 KVM_SET_LAPIC 1969------------------ 1970 1971:Capability: KVM_CAP_IRQCHIP 1972:Architectures: x86 1973:Type: vcpu ioctl 1974:Parameters: struct kvm_lapic_state (in) 1975:Returns: 0 on success, -1 on error 1976 1977:: 1978 1979 #define KVM_APIC_REG_SIZE 0x400 1980 struct kvm_lapic_state { 1981 char regs[KVM_APIC_REG_SIZE]; 1982 }; 1983 1984Copies the input argument into the Local APIC registers. The data format 1985and layout are the same as documented in the architecture manual. 1986 1987The format of the APIC ID register (bytes 32-35 of struct kvm_lapic_state's 1988regs field) depends on the state of the KVM_CAP_X2APIC_API capability. 1989See the note in KVM_GET_LAPIC. 1990 1991 19924.59 KVM_IOEVENTFD 1993------------------ 1994 1995:Capability: KVM_CAP_IOEVENTFD 1996:Architectures: all 1997:Type: vm ioctl 1998:Parameters: struct kvm_ioeventfd (in) 1999:Returns: 0 on success, !0 on error 2000 2001This ioctl attaches or detaches an ioeventfd to a legal pio/mmio address 2002within the guest. A guest write in the registered address will signal the 2003provided event instead of triggering an exit. 2004 2005:: 2006 2007 struct kvm_ioeventfd { 2008 __u64 datamatch; 2009 __u64 addr; /* legal pio/mmio address */ 2010 __u32 len; /* 0, 1, 2, 4, or 8 bytes */ 2011 __s32 fd; 2012 __u32 flags; 2013 __u8 pad[36]; 2014 }; 2015 2016For the special case of virtio-ccw devices on s390, the ioevent is matched 2017to a subchannel/virtqueue tuple instead. 2018 2019The following flags are defined:: 2020 2021 #define KVM_IOEVENTFD_FLAG_DATAMATCH (1 << kvm_ioeventfd_flag_nr_datamatch) 2022 #define KVM_IOEVENTFD_FLAG_PIO (1 << kvm_ioeventfd_flag_nr_pio) 2023 #define KVM_IOEVENTFD_FLAG_DEASSIGN (1 << kvm_ioeventfd_flag_nr_deassign) 2024 #define KVM_IOEVENTFD_FLAG_VIRTIO_CCW_NOTIFY \ 2025 (1 << kvm_ioeventfd_flag_nr_virtio_ccw_notify) 2026 2027If datamatch flag is set, the event will be signaled only if the written value 2028to the registered address is equal to datamatch in struct kvm_ioeventfd. 2029 2030For virtio-ccw devices, addr contains the subchannel id and datamatch the 2031virtqueue index. 2032 2033With KVM_CAP_IOEVENTFD_ANY_LENGTH, a zero length ioeventfd is allowed, and 2034the kernel will ignore the length of guest write and may get a faster vmexit. 2035The speedup may only apply to specific architectures, but the ioeventfd will 2036work anyway. 2037 20384.60 KVM_DIRTY_TLB 2039------------------ 2040 2041:Capability: KVM_CAP_SW_TLB 2042:Architectures: ppc 2043:Type: vcpu ioctl 2044:Parameters: struct kvm_dirty_tlb (in) 2045:Returns: 0 on success, -1 on error 2046 2047:: 2048 2049 struct kvm_dirty_tlb { 2050 __u64 bitmap; 2051 __u32 num_dirty; 2052 }; 2053 2054This must be called whenever userspace has changed an entry in the shared 2055TLB, prior to calling KVM_RUN on the associated vcpu. 2056 2057The "bitmap" field is the userspace address of an array. This array 2058consists of a number of bits, equal to the total number of TLB entries as 2059determined by the last successful call to KVM_CONFIG_TLB, rounded up to the 2060nearest multiple of 64. 2061 2062Each bit corresponds to one TLB entry, ordered the same as in the shared TLB 2063array. 2064 2065The array is little-endian: the bit 0 is the least significant bit of the 2066first byte, bit 8 is the least significant bit of the second byte, etc. 2067This avoids any complications with differing word sizes. 2068 2069The "num_dirty" field is a performance hint for KVM to determine whether it 2070should skip processing the bitmap and just invalidate everything. It must 2071be set to the number of set bits in the bitmap. 2072 2073 20744.62 KVM_CREATE_SPAPR_TCE 2075------------------------- 2076 2077:Capability: KVM_CAP_SPAPR_TCE 2078:Architectures: powerpc 2079:Type: vm ioctl 2080:Parameters: struct kvm_create_spapr_tce (in) 2081:Returns: file descriptor for manipulating the created TCE table 2082 2083This creates a virtual TCE (translation control entry) table, which 2084is an IOMMU for PAPR-style virtual I/O. It is used to translate 2085logical addresses used in virtual I/O into guest physical addresses, 2086and provides a scatter/gather capability for PAPR virtual I/O. 2087 2088:: 2089 2090 /* for KVM_CAP_SPAPR_TCE */ 2091 struct kvm_create_spapr_tce { 2092 __u64 liobn; 2093 __u32 window_size; 2094 }; 2095 2096The liobn field gives the logical IO bus number for which to create a 2097TCE table. The window_size field specifies the size of the DMA window 2098which this TCE table will translate - the table will contain one 64 2099bit TCE entry for every 4kiB of the DMA window. 2100 2101When the guest issues an H_PUT_TCE hcall on a liobn for which a TCE 2102table has been created using this ioctl(), the kernel will handle it 2103in real mode, updating the TCE table. H_PUT_TCE calls for other 2104liobns will cause a vm exit and must be handled by userspace. 2105 2106The return value is a file descriptor which can be passed to mmap(2) 2107to map the created TCE table into userspace. This lets userspace read 2108the entries written by kernel-handled H_PUT_TCE calls, and also lets 2109userspace update the TCE table directly which is useful in some 2110circumstances. 2111 2112 21134.63 KVM_ALLOCATE_RMA 2114--------------------- 2115 2116:Capability: KVM_CAP_PPC_RMA 2117:Architectures: powerpc 2118:Type: vm ioctl 2119:Parameters: struct kvm_allocate_rma (out) 2120:Returns: file descriptor for mapping the allocated RMA 2121 2122This allocates a Real Mode Area (RMA) from the pool allocated at boot 2123time by the kernel. An RMA is a physically-contiguous, aligned region 2124of memory used on older POWER processors to provide the memory which 2125will be accessed by real-mode (MMU off) accesses in a KVM guest. 2126POWER processors support a set of sizes for the RMA that usually 2127includes 64MB, 128MB, 256MB and some larger powers of two. 2128 2129:: 2130 2131 /* for KVM_ALLOCATE_RMA */ 2132 struct kvm_allocate_rma { 2133 __u64 rma_size; 2134 }; 2135 2136The return value is a file descriptor which can be passed to mmap(2) 2137to map the allocated RMA into userspace. The mapped area can then be 2138passed to the KVM_SET_USER_MEMORY_REGION ioctl to establish it as the 2139RMA for a virtual machine. The size of the RMA in bytes (which is 2140fixed at host kernel boot time) is returned in the rma_size field of 2141the argument structure. 2142 2143The KVM_CAP_PPC_RMA capability is 1 or 2 if the KVM_ALLOCATE_RMA ioctl 2144is supported; 2 if the processor requires all virtual machines to have 2145an RMA, or 1 if the processor can use an RMA but doesn't require it, 2146because it supports the Virtual RMA (VRMA) facility. 2147 2148 21494.64 KVM_NMI 2150------------ 2151 2152:Capability: KVM_CAP_USER_NMI 2153:Architectures: x86 2154:Type: vcpu ioctl 2155:Parameters: none 2156:Returns: 0 on success, -1 on error 2157 2158Queues an NMI on the thread's vcpu. Note this is well defined only 2159when KVM_CREATE_IRQCHIP has not been called, since this is an interface 2160between the virtual cpu core and virtual local APIC. After KVM_CREATE_IRQCHIP 2161has been called, this interface is completely emulated within the kernel. 2162 2163To use this to emulate the LINT1 input with KVM_CREATE_IRQCHIP, use the 2164following algorithm: 2165 2166 - pause the vcpu 2167 - read the local APIC's state (KVM_GET_LAPIC) 2168 - check whether changing LINT1 will queue an NMI (see the LVT entry for LINT1) 2169 - if so, issue KVM_NMI 2170 - resume the vcpu 2171 2172Some guests configure the LINT1 NMI input to cause a panic, aiding in 2173debugging. 2174 2175 21764.65 KVM_S390_UCAS_MAP 2177---------------------- 2178 2179:Capability: KVM_CAP_S390_UCONTROL 2180:Architectures: s390 2181:Type: vcpu ioctl 2182:Parameters: struct kvm_s390_ucas_mapping (in) 2183:Returns: 0 in case of success 2184 2185The parameter is defined like this:: 2186 2187 struct kvm_s390_ucas_mapping { 2188 __u64 user_addr; 2189 __u64 vcpu_addr; 2190 __u64 length; 2191 }; 2192 2193This ioctl maps the memory at "user_addr" with the length "length" to 2194the vcpu's address space starting at "vcpu_addr". All parameters need to 2195be aligned by 1 megabyte. 2196 2197 21984.66 KVM_S390_UCAS_UNMAP 2199------------------------ 2200 2201:Capability: KVM_CAP_S390_UCONTROL 2202:Architectures: s390 2203:Type: vcpu ioctl 2204:Parameters: struct kvm_s390_ucas_mapping (in) 2205:Returns: 0 in case of success 2206 2207The parameter is defined like this:: 2208 2209 struct kvm_s390_ucas_mapping { 2210 __u64 user_addr; 2211 __u64 vcpu_addr; 2212 __u64 length; 2213 }; 2214 2215This ioctl unmaps the memory in the vcpu's address space starting at 2216"vcpu_addr" with the length "length". The field "user_addr" is ignored. 2217All parameters need to be aligned by 1 megabyte. 2218 2219 22204.67 KVM_S390_VCPU_FAULT 2221------------------------ 2222 2223:Capability: KVM_CAP_S390_UCONTROL 2224:Architectures: s390 2225:Type: vcpu ioctl 2226:Parameters: vcpu absolute address (in) 2227:Returns: 0 in case of success 2228 2229This call creates a page table entry on the virtual cpu's address space 2230(for user controlled virtual machines) or the virtual machine's address 2231space (for regular virtual machines). This only works for minor faults, 2232thus it's recommended to access subject memory page via the user page 2233table upfront. This is useful to handle validity intercepts for user 2234controlled virtual machines to fault in the virtual cpu's lowcore pages 2235prior to calling the KVM_RUN ioctl. 2236 2237 22384.68 KVM_SET_ONE_REG 2239-------------------- 2240 2241:Capability: KVM_CAP_ONE_REG 2242:Architectures: all 2243:Type: vcpu ioctl 2244:Parameters: struct kvm_one_reg (in) 2245:Returns: 0 on success, negative value on failure 2246 2247Errors: 2248 2249 ====== ============================================================ 2250 ENOENT no such register 2251 EINVAL invalid register ID, or no such register or used with VMs in 2252 protected virtualization mode on s390 2253 EPERM (arm64) register access not allowed before vcpu finalization 2254 ====== ============================================================ 2255 2256(These error codes are indicative only: do not rely on a specific error 2257code being returned in a specific situation.) 2258 2259:: 2260 2261 struct kvm_one_reg { 2262 __u64 id; 2263 __u64 addr; 2264 }; 2265 2266Using this ioctl, a single vcpu register can be set to a specific value 2267defined by user space with the passed in struct kvm_one_reg, where id 2268refers to the register identifier as described below and addr is a pointer 2269to a variable with the respective size. There can be architecture agnostic 2270and architecture specific registers. Each have their own range of operation 2271and their own constants and width. To keep track of the implemented 2272registers, find a list below: 2273 2274 ======= =============================== ============ 2275 Arch Register Width (bits) 2276 ======= =============================== ============ 2277 PPC KVM_REG_PPC_HIOR 64 2278 PPC KVM_REG_PPC_IAC1 64 2279 PPC KVM_REG_PPC_IAC2 64 2280 PPC KVM_REG_PPC_IAC3 64 2281 PPC KVM_REG_PPC_IAC4 64 2282 PPC KVM_REG_PPC_DAC1 64 2283 PPC KVM_REG_PPC_DAC2 64 2284 PPC KVM_REG_PPC_DABR 64 2285 PPC KVM_REG_PPC_DSCR 64 2286 PPC KVM_REG_PPC_PURR 64 2287 PPC KVM_REG_PPC_SPURR 64 2288 PPC KVM_REG_PPC_DAR 64 2289 PPC KVM_REG_PPC_DSISR 32 2290 PPC KVM_REG_PPC_AMR 64 2291 PPC KVM_REG_PPC_UAMOR 64 2292 PPC KVM_REG_PPC_MMCR0 64 2293 PPC KVM_REG_PPC_MMCR1 64 2294 PPC KVM_REG_PPC_MMCRA 64 2295 PPC KVM_REG_PPC_MMCR2 64 2296 PPC KVM_REG_PPC_MMCRS 64 2297 PPC KVM_REG_PPC_MMCR3 64 2298 PPC KVM_REG_PPC_SIAR 64 2299 PPC KVM_REG_PPC_SDAR 64 2300 PPC KVM_REG_PPC_SIER 64 2301 PPC KVM_REG_PPC_SIER2 64 2302 PPC KVM_REG_PPC_SIER3 64 2303 PPC KVM_REG_PPC_PMC1 32 2304 PPC KVM_REG_PPC_PMC2 32 2305 PPC KVM_REG_PPC_PMC3 32 2306 PPC KVM_REG_PPC_PMC4 32 2307 PPC KVM_REG_PPC_PMC5 32 2308 PPC KVM_REG_PPC_PMC6 32 2309 PPC KVM_REG_PPC_PMC7 32 2310 PPC KVM_REG_PPC_PMC8 32 2311 PPC KVM_REG_PPC_FPR0 64 2312 ... 2313 PPC KVM_REG_PPC_FPR31 64 2314 PPC KVM_REG_PPC_VR0 128 2315 ... 2316 PPC KVM_REG_PPC_VR31 128 2317 PPC KVM_REG_PPC_VSR0 128 2318 ... 2319 PPC KVM_REG_PPC_VSR31 128 2320 PPC KVM_REG_PPC_FPSCR 64 2321 PPC KVM_REG_PPC_VSCR 32 2322 PPC KVM_REG_PPC_VPA_ADDR 64 2323 PPC KVM_REG_PPC_VPA_SLB 128 2324 PPC KVM_REG_PPC_VPA_DTL 128 2325 PPC KVM_REG_PPC_EPCR 32 2326 PPC KVM_REG_PPC_EPR 32 2327 PPC KVM_REG_PPC_TCR 32 2328 PPC KVM_REG_PPC_TSR 32 2329 PPC KVM_REG_PPC_OR_TSR 32 2330 PPC KVM_REG_PPC_CLEAR_TSR 32 2331 PPC KVM_REG_PPC_MAS0 32 2332 PPC KVM_REG_PPC_MAS1 32 2333 PPC KVM_REG_PPC_MAS2 64 2334 PPC KVM_REG_PPC_MAS7_3 64 2335 PPC KVM_REG_PPC_MAS4 32 2336 PPC KVM_REG_PPC_MAS6 32 2337 PPC KVM_REG_PPC_MMUCFG 32 2338 PPC KVM_REG_PPC_TLB0CFG 32 2339 PPC KVM_REG_PPC_TLB1CFG 32 2340 PPC KVM_REG_PPC_TLB2CFG 32 2341 PPC KVM_REG_PPC_TLB3CFG 32 2342 PPC KVM_REG_PPC_TLB0PS 32 2343 PPC KVM_REG_PPC_TLB1PS 32 2344 PPC KVM_REG_PPC_TLB2PS 32 2345 PPC KVM_REG_PPC_TLB3PS 32 2346 PPC KVM_REG_PPC_EPTCFG 32 2347 PPC KVM_REG_PPC_ICP_STATE 64 2348 PPC KVM_REG_PPC_VP_STATE 128 2349 PPC KVM_REG_PPC_TB_OFFSET 64 2350 PPC KVM_REG_PPC_SPMC1 32 2351 PPC KVM_REG_PPC_SPMC2 32 2352 PPC KVM_REG_PPC_IAMR 64 2353 PPC KVM_REG_PPC_TFHAR 64 2354 PPC KVM_REG_PPC_TFIAR 64 2355 PPC KVM_REG_PPC_TEXASR 64 2356 PPC KVM_REG_PPC_FSCR 64 2357 PPC KVM_REG_PPC_PSPB 32 2358 PPC KVM_REG_PPC_EBBHR 64 2359 PPC KVM_REG_PPC_EBBRR 64 2360 PPC KVM_REG_PPC_BESCR 64 2361 PPC KVM_REG_PPC_TAR 64 2362 PPC KVM_REG_PPC_DPDES 64 2363 PPC KVM_REG_PPC_DAWR 64 2364 PPC KVM_REG_PPC_DAWRX 64 2365 PPC KVM_REG_PPC_CIABR 64 2366 PPC KVM_REG_PPC_IC 64 2367 PPC KVM_REG_PPC_VTB 64 2368 PPC KVM_REG_PPC_CSIGR 64 2369 PPC KVM_REG_PPC_TACR 64 2370 PPC KVM_REG_PPC_TCSCR 64 2371 PPC KVM_REG_PPC_PID 64 2372 PPC KVM_REG_PPC_ACOP 64 2373 PPC KVM_REG_PPC_VRSAVE 32 2374 PPC KVM_REG_PPC_LPCR 32 2375 PPC KVM_REG_PPC_LPCR_64 64 2376 PPC KVM_REG_PPC_PPR 64 2377 PPC KVM_REG_PPC_ARCH_COMPAT 32 2378 PPC KVM_REG_PPC_DABRX 32 2379 PPC KVM_REG_PPC_WORT 64 2380 PPC KVM_REG_PPC_SPRG9 64 2381 PPC KVM_REG_PPC_DBSR 32 2382 PPC KVM_REG_PPC_TIDR 64 2383 PPC KVM_REG_PPC_PSSCR 64 2384 PPC KVM_REG_PPC_DEC_EXPIRY 64 2385 PPC KVM_REG_PPC_PTCR 64 2386 PPC KVM_REG_PPC_DAWR1 64 2387 PPC KVM_REG_PPC_DAWRX1 64 2388 PPC KVM_REG_PPC_TM_GPR0 64 2389 ... 2390 PPC KVM_REG_PPC_TM_GPR31 64 2391 PPC KVM_REG_PPC_TM_VSR0 128 2392 ... 2393 PPC KVM_REG_PPC_TM_VSR63 128 2394 PPC KVM_REG_PPC_TM_CR 64 2395 PPC KVM_REG_PPC_TM_LR 64 2396 PPC KVM_REG_PPC_TM_CTR 64 2397 PPC KVM_REG_PPC_TM_FPSCR 64 2398 PPC KVM_REG_PPC_TM_AMR 64 2399 PPC KVM_REG_PPC_TM_PPR 64 2400 PPC KVM_REG_PPC_TM_VRSAVE 64 2401 PPC KVM_REG_PPC_TM_VSCR 32 2402 PPC KVM_REG_PPC_TM_DSCR 64 2403 PPC KVM_REG_PPC_TM_TAR 64 2404 PPC KVM_REG_PPC_TM_XER 64 2405 2406 MIPS KVM_REG_MIPS_R0 64 2407 ... 2408 MIPS KVM_REG_MIPS_R31 64 2409 MIPS KVM_REG_MIPS_HI 64 2410 MIPS KVM_REG_MIPS_LO 64 2411 MIPS KVM_REG_MIPS_PC 64 2412 MIPS KVM_REG_MIPS_CP0_INDEX 32 2413 MIPS KVM_REG_MIPS_CP0_ENTRYLO0 64 2414 MIPS KVM_REG_MIPS_CP0_ENTRYLO1 64 2415 MIPS KVM_REG_MIPS_CP0_CONTEXT 64 2416 MIPS KVM_REG_MIPS_CP0_CONTEXTCONFIG 32 2417 MIPS KVM_REG_MIPS_CP0_USERLOCAL 64 2418 MIPS KVM_REG_MIPS_CP0_XCONTEXTCONFIG 64 2419 MIPS KVM_REG_MIPS_CP0_PAGEMASK 32 2420 MIPS KVM_REG_MIPS_CP0_PAGEGRAIN 32 2421 MIPS KVM_REG_MIPS_CP0_SEGCTL0 64 2422 MIPS KVM_REG_MIPS_CP0_SEGCTL1 64 2423 MIPS KVM_REG_MIPS_CP0_SEGCTL2 64 2424 MIPS KVM_REG_MIPS_CP0_PWBASE 64 2425 MIPS KVM_REG_MIPS_CP0_PWFIELD 64 2426 MIPS KVM_REG_MIPS_CP0_PWSIZE 64 2427 MIPS KVM_REG_MIPS_CP0_WIRED 32 2428 MIPS KVM_REG_MIPS_CP0_PWCTL 32 2429 MIPS KVM_REG_MIPS_CP0_HWRENA 32 2430 MIPS KVM_REG_MIPS_CP0_BADVADDR 64 2431 MIPS KVM_REG_MIPS_CP0_BADINSTR 32 2432 MIPS KVM_REG_MIPS_CP0_BADINSTRP 32 2433 MIPS KVM_REG_MIPS_CP0_COUNT 32 2434 MIPS KVM_REG_MIPS_CP0_ENTRYHI 64 2435 MIPS KVM_REG_MIPS_CP0_COMPARE 32 2436 MIPS KVM_REG_MIPS_CP0_STATUS 32 2437 MIPS KVM_REG_MIPS_CP0_INTCTL 32 2438 MIPS KVM_REG_MIPS_CP0_CAUSE 32 2439 MIPS KVM_REG_MIPS_CP0_EPC 64 2440 MIPS KVM_REG_MIPS_CP0_PRID 32 2441 MIPS KVM_REG_MIPS_CP0_EBASE 64 2442 MIPS KVM_REG_MIPS_CP0_CONFIG 32 2443 MIPS KVM_REG_MIPS_CP0_CONFIG1 32 2444 MIPS KVM_REG_MIPS_CP0_CONFIG2 32 2445 MIPS KVM_REG_MIPS_CP0_CONFIG3 32 2446 MIPS KVM_REG_MIPS_CP0_CONFIG4 32 2447 MIPS KVM_REG_MIPS_CP0_CONFIG5 32 2448 MIPS KVM_REG_MIPS_CP0_CONFIG7 32 2449 MIPS KVM_REG_MIPS_CP0_XCONTEXT 64 2450 MIPS KVM_REG_MIPS_CP0_ERROREPC 64 2451 MIPS KVM_REG_MIPS_CP0_KSCRATCH1 64 2452 MIPS KVM_REG_MIPS_CP0_KSCRATCH2 64 2453 MIPS KVM_REG_MIPS_CP0_KSCRATCH3 64 2454 MIPS KVM_REG_MIPS_CP0_KSCRATCH4 64 2455 MIPS KVM_REG_MIPS_CP0_KSCRATCH5 64 2456 MIPS KVM_REG_MIPS_CP0_KSCRATCH6 64 2457 MIPS KVM_REG_MIPS_CP0_MAAR(0..63) 64 2458 MIPS KVM_REG_MIPS_COUNT_CTL 64 2459 MIPS KVM_REG_MIPS_COUNT_RESUME 64 2460 MIPS KVM_REG_MIPS_COUNT_HZ 64 2461 MIPS KVM_REG_MIPS_FPR_32(0..31) 32 2462 MIPS KVM_REG_MIPS_FPR_64(0..31) 64 2463 MIPS KVM_REG_MIPS_VEC_128(0..31) 128 2464 MIPS KVM_REG_MIPS_FCR_IR 32 2465 MIPS KVM_REG_MIPS_FCR_CSR 32 2466 MIPS KVM_REG_MIPS_MSA_IR 32 2467 MIPS KVM_REG_MIPS_MSA_CSR 32 2468 ======= =============================== ============ 2469 2470ARM registers are mapped using the lower 32 bits. The upper 16 of that 2471is the register group type, or coprocessor number: 2472 2473ARM core registers have the following id bit patterns:: 2474 2475 0x4020 0000 0010 <index into the kvm_regs struct:16> 2476 2477ARM 32-bit CP15 registers have the following id bit patterns:: 2478 2479 0x4020 0000 000F <zero:1> <crn:4> <crm:4> <opc1:4> <opc2:3> 2480 2481ARM 64-bit CP15 registers have the following id bit patterns:: 2482 2483 0x4030 0000 000F <zero:1> <zero:4> <crm:4> <opc1:4> <zero:3> 2484 2485ARM CCSIDR registers are demultiplexed by CSSELR value:: 2486 2487 0x4020 0000 0011 00 <csselr:8> 2488 2489ARM 32-bit VFP control registers have the following id bit patterns:: 2490 2491 0x4020 0000 0012 1 <regno:12> 2492 2493ARM 64-bit FP registers have the following id bit patterns:: 2494 2495 0x4030 0000 0012 0 <regno:12> 2496 2497ARM firmware pseudo-registers have the following bit pattern:: 2498 2499 0x4030 0000 0014 <regno:16> 2500 2501 2502arm64 registers are mapped using the lower 32 bits. The upper 16 of 2503that is the register group type, or coprocessor number: 2504 2505arm64 core/FP-SIMD registers have the following id bit patterns. Note 2506that the size of the access is variable, as the kvm_regs structure 2507contains elements ranging from 32 to 128 bits. The index is a 32bit 2508value in the kvm_regs structure seen as a 32bit array:: 2509 2510 0x60x0 0000 0010 <index into the kvm_regs struct:16> 2511 2512Specifically: 2513 2514======================= ========= ===== ======================================= 2515 Encoding Register Bits kvm_regs member 2516======================= ========= ===== ======================================= 2517 0x6030 0000 0010 0000 X0 64 regs.regs[0] 2518 0x6030 0000 0010 0002 X1 64 regs.regs[1] 2519 ... 2520 0x6030 0000 0010 003c X30 64 regs.regs[30] 2521 0x6030 0000 0010 003e SP 64 regs.sp 2522 0x6030 0000 0010 0040 PC 64 regs.pc 2523 0x6030 0000 0010 0042 PSTATE 64 regs.pstate 2524 0x6030 0000 0010 0044 SP_EL1 64 sp_el1 2525 0x6030 0000 0010 0046 ELR_EL1 64 elr_el1 2526 0x6030 0000 0010 0048 SPSR_EL1 64 spsr[KVM_SPSR_EL1] (alias SPSR_SVC) 2527 0x6030 0000 0010 004a SPSR_ABT 64 spsr[KVM_SPSR_ABT] 2528 0x6030 0000 0010 004c SPSR_UND 64 spsr[KVM_SPSR_UND] 2529 0x6030 0000 0010 004e SPSR_IRQ 64 spsr[KVM_SPSR_IRQ] 2530 0x6060 0000 0010 0050 SPSR_FIQ 64 spsr[KVM_SPSR_FIQ] 2531 0x6040 0000 0010 0054 V0 128 fp_regs.vregs[0] [1]_ 2532 0x6040 0000 0010 0058 V1 128 fp_regs.vregs[1] [1]_ 2533 ... 2534 0x6040 0000 0010 00d0 V31 128 fp_regs.vregs[31] [1]_ 2535 0x6020 0000 0010 00d4 FPSR 32 fp_regs.fpsr 2536 0x6020 0000 0010 00d5 FPCR 32 fp_regs.fpcr 2537======================= ========= ===== ======================================= 2538 2539.. [1] These encodings are not accepted for SVE-enabled vcpus. See 2540 KVM_ARM_VCPU_INIT. 2541 2542 The equivalent register content can be accessed via bits [127:0] of 2543 the corresponding SVE Zn registers instead for vcpus that have SVE 2544 enabled (see below). 2545 2546arm64 CCSIDR registers are demultiplexed by CSSELR value:: 2547 2548 0x6020 0000 0011 00 <csselr:8> 2549 2550arm64 system registers have the following id bit patterns:: 2551 2552 0x6030 0000 0013 <op0:2> <op1:3> <crn:4> <crm:4> <op2:3> 2553 2554.. warning:: 2555 2556 Two system register IDs do not follow the specified pattern. These 2557 are KVM_REG_ARM_TIMER_CVAL and KVM_REG_ARM_TIMER_CNT, which map to 2558 system registers CNTV_CVAL_EL0 and CNTVCT_EL0 respectively. These 2559 two had their values accidentally swapped, which means TIMER_CVAL is 2560 derived from the register encoding for CNTVCT_EL0 and TIMER_CNT is 2561 derived from the register encoding for CNTV_CVAL_EL0. As this is 2562 API, it must remain this way. 2563 2564arm64 firmware pseudo-registers have the following bit pattern:: 2565 2566 0x6030 0000 0014 <regno:16> 2567 2568arm64 SVE registers have the following bit patterns:: 2569 2570 0x6080 0000 0015 00 <n:5> <slice:5> Zn bits[2048*slice + 2047 : 2048*slice] 2571 0x6050 0000 0015 04 <n:4> <slice:5> Pn bits[256*slice + 255 : 256*slice] 2572 0x6050 0000 0015 060 <slice:5> FFR bits[256*slice + 255 : 256*slice] 2573 0x6060 0000 0015 ffff KVM_REG_ARM64_SVE_VLS pseudo-register 2574 2575Access to register IDs where 2048 * slice >= 128 * max_vq will fail with 2576ENOENT. max_vq is the vcpu's maximum supported vector length in 128-bit 2577quadwords: see [2]_ below. 2578 2579These registers are only accessible on vcpus for which SVE is enabled. 2580See KVM_ARM_VCPU_INIT for details. 2581 2582In addition, except for KVM_REG_ARM64_SVE_VLS, these registers are not 2583accessible until the vcpu's SVE configuration has been finalized 2584using KVM_ARM_VCPU_FINALIZE(KVM_ARM_VCPU_SVE). See KVM_ARM_VCPU_INIT 2585and KVM_ARM_VCPU_FINALIZE for more information about this procedure. 2586 2587KVM_REG_ARM64_SVE_VLS is a pseudo-register that allows the set of vector 2588lengths supported by the vcpu to be discovered and configured by 2589userspace. When transferred to or from user memory via KVM_GET_ONE_REG 2590or KVM_SET_ONE_REG, the value of this register is of type 2591__u64[KVM_ARM64_SVE_VLS_WORDS], and encodes the set of vector lengths as 2592follows:: 2593 2594 __u64 vector_lengths[KVM_ARM64_SVE_VLS_WORDS]; 2595 2596 if (vq >= SVE_VQ_MIN && vq <= SVE_VQ_MAX && 2597 ((vector_lengths[(vq - KVM_ARM64_SVE_VQ_MIN) / 64] >> 2598 ((vq - KVM_ARM64_SVE_VQ_MIN) % 64)) & 1)) 2599 /* Vector length vq * 16 bytes supported */ 2600 else 2601 /* Vector length vq * 16 bytes not supported */ 2602 2603.. [2] The maximum value vq for which the above condition is true is 2604 max_vq. This is the maximum vector length available to the guest on 2605 this vcpu, and determines which register slices are visible through 2606 this ioctl interface. 2607 2608(See Documentation/arm64/sve.rst for an explanation of the "vq" 2609nomenclature.) 2610 2611KVM_REG_ARM64_SVE_VLS is only accessible after KVM_ARM_VCPU_INIT. 2612KVM_ARM_VCPU_INIT initialises it to the best set of vector lengths that 2613the host supports. 2614 2615Userspace may subsequently modify it if desired until the vcpu's SVE 2616configuration is finalized using KVM_ARM_VCPU_FINALIZE(KVM_ARM_VCPU_SVE). 2617 2618Apart from simply removing all vector lengths from the host set that 2619exceed some value, support for arbitrarily chosen sets of vector lengths 2620is hardware-dependent and may not be available. Attempting to configure 2621an invalid set of vector lengths via KVM_SET_ONE_REG will fail with 2622EINVAL. 2623 2624After the vcpu's SVE configuration is finalized, further attempts to 2625write this register will fail with EPERM. 2626 2627arm64 bitmap feature firmware pseudo-registers have the following bit pattern:: 2628 2629 0x6030 0000 0016 <regno:16> 2630 2631The bitmap feature firmware registers exposes the hypercall services that 2632are available for userspace to configure. The set bits corresponds to the 2633services that are available for the guests to access. By default, KVM 2634sets all the supported bits during VM initialization. The userspace can 2635discover the available services via KVM_GET_ONE_REG, and write back the 2636bitmap corresponding to the features that it wishes guests to see via 2637KVM_SET_ONE_REG. 2638 2639Note: These registers are immutable once any of the vCPUs of the VM has 2640run at least once. A KVM_SET_ONE_REG in such a scenario will return 2641a -EBUSY to userspace. 2642 2643(See Documentation/virt/kvm/arm/hypercalls.rst for more details.) 2644 2645 2646MIPS registers are mapped using the lower 32 bits. The upper 16 of that is 2647the register group type: 2648 2649MIPS core registers (see above) have the following id bit patterns:: 2650 2651 0x7030 0000 0000 <reg:16> 2652 2653MIPS CP0 registers (see KVM_REG_MIPS_CP0_* above) have the following id bit 2654patterns depending on whether they're 32-bit or 64-bit registers:: 2655 2656 0x7020 0000 0001 00 <reg:5> <sel:3> (32-bit) 2657 0x7030 0000 0001 00 <reg:5> <sel:3> (64-bit) 2658 2659Note: KVM_REG_MIPS_CP0_ENTRYLO0 and KVM_REG_MIPS_CP0_ENTRYLO1 are the MIPS64 2660versions of the EntryLo registers regardless of the word size of the host 2661hardware, host kernel, guest, and whether XPA is present in the guest, i.e. 2662with the RI and XI bits (if they exist) in bits 63 and 62 respectively, and 2663the PFNX field starting at bit 30. 2664 2665MIPS MAARs (see KVM_REG_MIPS_CP0_MAAR(*) above) have the following id bit 2666patterns:: 2667 2668 0x7030 0000 0001 01 <reg:8> 2669 2670MIPS KVM control registers (see above) have the following id bit patterns:: 2671 2672 0x7030 0000 0002 <reg:16> 2673 2674MIPS FPU registers (see KVM_REG_MIPS_FPR_{32,64}() above) have the following 2675id bit patterns depending on the size of the register being accessed. They are 2676always accessed according to the current guest FPU mode (Status.FR and 2677Config5.FRE), i.e. as the guest would see them, and they become unpredictable 2678if the guest FPU mode is changed. MIPS SIMD Architecture (MSA) vector 2679registers (see KVM_REG_MIPS_VEC_128() above) have similar patterns as they 2680overlap the FPU registers:: 2681 2682 0x7020 0000 0003 00 <0:3> <reg:5> (32-bit FPU registers) 2683 0x7030 0000 0003 00 <0:3> <reg:5> (64-bit FPU registers) 2684 0x7040 0000 0003 00 <0:3> <reg:5> (128-bit MSA vector registers) 2685 2686MIPS FPU control registers (see KVM_REG_MIPS_FCR_{IR,CSR} above) have the 2687following id bit patterns:: 2688 2689 0x7020 0000 0003 01 <0:3> <reg:5> 2690 2691MIPS MSA control registers (see KVM_REG_MIPS_MSA_{IR,CSR} above) have the 2692following id bit patterns:: 2693 2694 0x7020 0000 0003 02 <0:3> <reg:5> 2695 2696RISC-V registers are mapped using the lower 32 bits. The upper 8 bits of 2697that is the register group type. 2698 2699RISC-V config registers are meant for configuring a Guest VCPU and it has 2700the following id bit patterns:: 2701 2702 0x8020 0000 01 <index into the kvm_riscv_config struct:24> (32bit Host) 2703 0x8030 0000 01 <index into the kvm_riscv_config struct:24> (64bit Host) 2704 2705Following are the RISC-V config registers: 2706 2707======================= ========= ============================================= 2708 Encoding Register Description 2709======================= ========= ============================================= 2710 0x80x0 0000 0100 0000 isa ISA feature bitmap of Guest VCPU 2711======================= ========= ============================================= 2712 2713The isa config register can be read anytime but can only be written before 2714a Guest VCPU runs. It will have ISA feature bits matching underlying host 2715set by default. 2716 2717RISC-V core registers represent the general excution state of a Guest VCPU 2718and it has the following id bit patterns:: 2719 2720 0x8020 0000 02 <index into the kvm_riscv_core struct:24> (32bit Host) 2721 0x8030 0000 02 <index into the kvm_riscv_core struct:24> (64bit Host) 2722 2723Following are the RISC-V core registers: 2724 2725======================= ========= ============================================= 2726 Encoding Register Description 2727======================= ========= ============================================= 2728 0x80x0 0000 0200 0000 regs.pc Program counter 2729 0x80x0 0000 0200 0001 regs.ra Return address 2730 0x80x0 0000 0200 0002 regs.sp Stack pointer 2731 0x80x0 0000 0200 0003 regs.gp Global pointer 2732 0x80x0 0000 0200 0004 regs.tp Task pointer 2733 0x80x0 0000 0200 0005 regs.t0 Caller saved register 0 2734 0x80x0 0000 0200 0006 regs.t1 Caller saved register 1 2735 0x80x0 0000 0200 0007 regs.t2 Caller saved register 2 2736 0x80x0 0000 0200 0008 regs.s0 Callee saved register 0 2737 0x80x0 0000 0200 0009 regs.s1 Callee saved register 1 2738 0x80x0 0000 0200 000a regs.a0 Function argument (or return value) 0 2739 0x80x0 0000 0200 000b regs.a1 Function argument (or return value) 1 2740 0x80x0 0000 0200 000c regs.a2 Function argument 2 2741 0x80x0 0000 0200 000d regs.a3 Function argument 3 2742 0x80x0 0000 0200 000e regs.a4 Function argument 4 2743 0x80x0 0000 0200 000f regs.a5 Function argument 5 2744 0x80x0 0000 0200 0010 regs.a6 Function argument 6 2745 0x80x0 0000 0200 0011 regs.a7 Function argument 7 2746 0x80x0 0000 0200 0012 regs.s2 Callee saved register 2 2747 0x80x0 0000 0200 0013 regs.s3 Callee saved register 3 2748 0x80x0 0000 0200 0014 regs.s4 Callee saved register 4 2749 0x80x0 0000 0200 0015 regs.s5 Callee saved register 5 2750 0x80x0 0000 0200 0016 regs.s6 Callee saved register 6 2751 0x80x0 0000 0200 0017 regs.s7 Callee saved register 7 2752 0x80x0 0000 0200 0018 regs.s8 Callee saved register 8 2753 0x80x0 0000 0200 0019 regs.s9 Callee saved register 9 2754 0x80x0 0000 0200 001a regs.s10 Callee saved register 10 2755 0x80x0 0000 0200 001b regs.s11 Callee saved register 11 2756 0x80x0 0000 0200 001c regs.t3 Caller saved register 3 2757 0x80x0 0000 0200 001d regs.t4 Caller saved register 4 2758 0x80x0 0000 0200 001e regs.t5 Caller saved register 5 2759 0x80x0 0000 0200 001f regs.t6 Caller saved register 6 2760 0x80x0 0000 0200 0020 mode Privilege mode (1 = S-mode or 0 = U-mode) 2761======================= ========= ============================================= 2762 2763RISC-V csr registers represent the supervisor mode control/status registers 2764of a Guest VCPU and it has the following id bit patterns:: 2765 2766 0x8020 0000 03 <index into the kvm_riscv_csr struct:24> (32bit Host) 2767 0x8030 0000 03 <index into the kvm_riscv_csr struct:24> (64bit Host) 2768 2769Following are the RISC-V csr registers: 2770 2771======================= ========= ============================================= 2772 Encoding Register Description 2773======================= ========= ============================================= 2774 0x80x0 0000 0300 0000 sstatus Supervisor status 2775 0x80x0 0000 0300 0001 sie Supervisor interrupt enable 2776 0x80x0 0000 0300 0002 stvec Supervisor trap vector base 2777 0x80x0 0000 0300 0003 sscratch Supervisor scratch register 2778 0x80x0 0000 0300 0004 sepc Supervisor exception program counter 2779 0x80x0 0000 0300 0005 scause Supervisor trap cause 2780 0x80x0 0000 0300 0006 stval Supervisor bad address or instruction 2781 0x80x0 0000 0300 0007 sip Supervisor interrupt pending 2782 0x80x0 0000 0300 0008 satp Supervisor address translation and protection 2783======================= ========= ============================================= 2784 2785RISC-V timer registers represent the timer state of a Guest VCPU and it has 2786the following id bit patterns:: 2787 2788 0x8030 0000 04 <index into the kvm_riscv_timer struct:24> 2789 2790Following are the RISC-V timer registers: 2791 2792======================= ========= ============================================= 2793 Encoding Register Description 2794======================= ========= ============================================= 2795 0x8030 0000 0400 0000 frequency Time base frequency (read-only) 2796 0x8030 0000 0400 0001 time Time value visible to Guest 2797 0x8030 0000 0400 0002 compare Time compare programmed by Guest 2798 0x8030 0000 0400 0003 state Time compare state (1 = ON or 0 = OFF) 2799======================= ========= ============================================= 2800 2801RISC-V F-extension registers represent the single precision floating point 2802state of a Guest VCPU and it has the following id bit patterns:: 2803 2804 0x8020 0000 05 <index into the __riscv_f_ext_state struct:24> 2805 2806Following are the RISC-V F-extension registers: 2807 2808======================= ========= ============================================= 2809 Encoding Register Description 2810======================= ========= ============================================= 2811 0x8020 0000 0500 0000 f[0] Floating point register 0 2812 ... 2813 0x8020 0000 0500 001f f[31] Floating point register 31 2814 0x8020 0000 0500 0020 fcsr Floating point control and status register 2815======================= ========= ============================================= 2816 2817RISC-V D-extension registers represent the double precision floating point 2818state of a Guest VCPU and it has the following id bit patterns:: 2819 2820 0x8020 0000 06 <index into the __riscv_d_ext_state struct:24> (fcsr) 2821 0x8030 0000 06 <index into the __riscv_d_ext_state struct:24> (non-fcsr) 2822 2823Following are the RISC-V D-extension registers: 2824 2825======================= ========= ============================================= 2826 Encoding Register Description 2827======================= ========= ============================================= 2828 0x8030 0000 0600 0000 f[0] Floating point register 0 2829 ... 2830 0x8030 0000 0600 001f f[31] Floating point register 31 2831 0x8020 0000 0600 0020 fcsr Floating point control and status register 2832======================= ========= ============================================= 2833 2834 28354.69 KVM_GET_ONE_REG 2836-------------------- 2837 2838:Capability: KVM_CAP_ONE_REG 2839:Architectures: all 2840:Type: vcpu ioctl 2841:Parameters: struct kvm_one_reg (in and out) 2842:Returns: 0 on success, negative value on failure 2843 2844Errors include: 2845 2846 ======== ============================================================ 2847 ENOENT no such register 2848 EINVAL invalid register ID, or no such register or used with VMs in 2849 protected virtualization mode on s390 2850 EPERM (arm64) register access not allowed before vcpu finalization 2851 ======== ============================================================ 2852 2853(These error codes are indicative only: do not rely on a specific error 2854code being returned in a specific situation.) 2855 2856This ioctl allows to receive the value of a single register implemented 2857in a vcpu. The register to read is indicated by the "id" field of the 2858kvm_one_reg struct passed in. On success, the register value can be found 2859at the memory location pointed to by "addr". 2860 2861The list of registers accessible using this interface is identical to the 2862list in 4.68. 2863 2864 28654.70 KVM_KVMCLOCK_CTRL 2866---------------------- 2867 2868:Capability: KVM_CAP_KVMCLOCK_CTRL 2869:Architectures: Any that implement pvclocks (currently x86 only) 2870:Type: vcpu ioctl 2871:Parameters: None 2872:Returns: 0 on success, -1 on error 2873 2874This ioctl sets a flag accessible to the guest indicating that the specified 2875vCPU has been paused by the host userspace. 2876 2877The host will set a flag in the pvclock structure that is checked from the 2878soft lockup watchdog. The flag is part of the pvclock structure that is 2879shared between guest and host, specifically the second bit of the flags 2880field of the pvclock_vcpu_time_info structure. It will be set exclusively by 2881the host and read/cleared exclusively by the guest. The guest operation of 2882checking and clearing the flag must be an atomic operation so 2883load-link/store-conditional, or equivalent must be used. There are two cases 2884where the guest will clear the flag: when the soft lockup watchdog timer resets 2885itself or when a soft lockup is detected. This ioctl can be called any time 2886after pausing the vcpu, but before it is resumed. 2887 2888 28894.71 KVM_SIGNAL_MSI 2890------------------- 2891 2892:Capability: KVM_CAP_SIGNAL_MSI 2893:Architectures: x86 arm64 2894:Type: vm ioctl 2895:Parameters: struct kvm_msi (in) 2896:Returns: >0 on delivery, 0 if guest blocked the MSI, and -1 on error 2897 2898Directly inject a MSI message. Only valid with in-kernel irqchip that handles 2899MSI messages. 2900 2901:: 2902 2903 struct kvm_msi { 2904 __u32 address_lo; 2905 __u32 address_hi; 2906 __u32 data; 2907 __u32 flags; 2908 __u32 devid; 2909 __u8 pad[12]; 2910 }; 2911 2912flags: 2913 KVM_MSI_VALID_DEVID: devid contains a valid value. The per-VM 2914 KVM_CAP_MSI_DEVID capability advertises the requirement to provide 2915 the device ID. If this capability is not available, userspace 2916 should never set the KVM_MSI_VALID_DEVID flag as the ioctl might fail. 2917 2918If KVM_MSI_VALID_DEVID is set, devid contains a unique device identifier 2919for the device that wrote the MSI message. For PCI, this is usually a 2920BFD identifier in the lower 16 bits. 2921 2922On x86, address_hi is ignored unless the KVM_X2APIC_API_USE_32BIT_IDS 2923feature of KVM_CAP_X2APIC_API capability is enabled. If it is enabled, 2924address_hi bits 31-8 provide bits 31-8 of the destination id. Bits 7-0 of 2925address_hi must be zero. 2926 2927 29284.71 KVM_CREATE_PIT2 2929-------------------- 2930 2931:Capability: KVM_CAP_PIT2 2932:Architectures: x86 2933:Type: vm ioctl 2934:Parameters: struct kvm_pit_config (in) 2935:Returns: 0 on success, -1 on error 2936 2937Creates an in-kernel device model for the i8254 PIT. This call is only valid 2938after enabling in-kernel irqchip support via KVM_CREATE_IRQCHIP. The following 2939parameters have to be passed:: 2940 2941 struct kvm_pit_config { 2942 __u32 flags; 2943 __u32 pad[15]; 2944 }; 2945 2946Valid flags are:: 2947 2948 #define KVM_PIT_SPEAKER_DUMMY 1 /* emulate speaker port stub */ 2949 2950PIT timer interrupts may use a per-VM kernel thread for injection. If it 2951exists, this thread will have a name of the following pattern:: 2952 2953 kvm-pit/<owner-process-pid> 2954 2955When running a guest with elevated priorities, the scheduling parameters of 2956this thread may have to be adjusted accordingly. 2957 2958This IOCTL replaces the obsolete KVM_CREATE_PIT. 2959 2960 29614.72 KVM_GET_PIT2 2962----------------- 2963 2964:Capability: KVM_CAP_PIT_STATE2 2965:Architectures: x86 2966:Type: vm ioctl 2967:Parameters: struct kvm_pit_state2 (out) 2968:Returns: 0 on success, -1 on error 2969 2970Retrieves the state of the in-kernel PIT model. Only valid after 2971KVM_CREATE_PIT2. The state is returned in the following structure:: 2972 2973 struct kvm_pit_state2 { 2974 struct kvm_pit_channel_state channels[3]; 2975 __u32 flags; 2976 __u32 reserved[9]; 2977 }; 2978 2979Valid flags are:: 2980 2981 /* disable PIT in HPET legacy mode */ 2982 #define KVM_PIT_FLAGS_HPET_LEGACY 0x00000001 2983 /* speaker port data bit enabled */ 2984 #define KVM_PIT_FLAGS_SPEAKER_DATA_ON 0x00000002 2985 2986This IOCTL replaces the obsolete KVM_GET_PIT. 2987 2988 29894.73 KVM_SET_PIT2 2990----------------- 2991 2992:Capability: KVM_CAP_PIT_STATE2 2993:Architectures: x86 2994:Type: vm ioctl 2995:Parameters: struct kvm_pit_state2 (in) 2996:Returns: 0 on success, -1 on error 2997 2998Sets the state of the in-kernel PIT model. Only valid after KVM_CREATE_PIT2. 2999See KVM_GET_PIT2 for details on struct kvm_pit_state2. 3000 3001This IOCTL replaces the obsolete KVM_SET_PIT. 3002 3003 30044.74 KVM_PPC_GET_SMMU_INFO 3005-------------------------- 3006 3007:Capability: KVM_CAP_PPC_GET_SMMU_INFO 3008:Architectures: powerpc 3009:Type: vm ioctl 3010:Parameters: None 3011:Returns: 0 on success, -1 on error 3012 3013This populates and returns a structure describing the features of 3014the "Server" class MMU emulation supported by KVM. 3015This can in turn be used by userspace to generate the appropriate 3016device-tree properties for the guest operating system. 3017 3018The structure contains some global information, followed by an 3019array of supported segment page sizes:: 3020 3021 struct kvm_ppc_smmu_info { 3022 __u64 flags; 3023 __u32 slb_size; 3024 __u32 pad; 3025 struct kvm_ppc_one_seg_page_size sps[KVM_PPC_PAGE_SIZES_MAX_SZ]; 3026 }; 3027 3028The supported flags are: 3029 3030 - KVM_PPC_PAGE_SIZES_REAL: 3031 When that flag is set, guest page sizes must "fit" the backing 3032 store page sizes. When not set, any page size in the list can 3033 be used regardless of how they are backed by userspace. 3034 3035 - KVM_PPC_1T_SEGMENTS 3036 The emulated MMU supports 1T segments in addition to the 3037 standard 256M ones. 3038 3039 - KVM_PPC_NO_HASH 3040 This flag indicates that HPT guests are not supported by KVM, 3041 thus all guests must use radix MMU mode. 3042 3043The "slb_size" field indicates how many SLB entries are supported 3044 3045The "sps" array contains 8 entries indicating the supported base 3046page sizes for a segment in increasing order. Each entry is defined 3047as follow:: 3048 3049 struct kvm_ppc_one_seg_page_size { 3050 __u32 page_shift; /* Base page shift of segment (or 0) */ 3051 __u32 slb_enc; /* SLB encoding for BookS */ 3052 struct kvm_ppc_one_page_size enc[KVM_PPC_PAGE_SIZES_MAX_SZ]; 3053 }; 3054 3055An entry with a "page_shift" of 0 is unused. Because the array is 3056organized in increasing order, a lookup can stop when encoutering 3057such an entry. 3058 3059The "slb_enc" field provides the encoding to use in the SLB for the 3060page size. The bits are in positions such as the value can directly 3061be OR'ed into the "vsid" argument of the slbmte instruction. 3062 3063The "enc" array is a list which for each of those segment base page 3064size provides the list of supported actual page sizes (which can be 3065only larger or equal to the base page size), along with the 3066corresponding encoding in the hash PTE. Similarly, the array is 30678 entries sorted by increasing sizes and an entry with a "0" shift 3068is an empty entry and a terminator:: 3069 3070 struct kvm_ppc_one_page_size { 3071 __u32 page_shift; /* Page shift (or 0) */ 3072 __u32 pte_enc; /* Encoding in the HPTE (>>12) */ 3073 }; 3074 3075The "pte_enc" field provides a value that can OR'ed into the hash 3076PTE's RPN field (ie, it needs to be shifted left by 12 to OR it 3077into the hash PTE second double word). 3078 30794.75 KVM_IRQFD 3080-------------- 3081 3082:Capability: KVM_CAP_IRQFD 3083:Architectures: x86 s390 arm64 3084:Type: vm ioctl 3085:Parameters: struct kvm_irqfd (in) 3086:Returns: 0 on success, -1 on error 3087 3088Allows setting an eventfd to directly trigger a guest interrupt. 3089kvm_irqfd.fd specifies the file descriptor to use as the eventfd and 3090kvm_irqfd.gsi specifies the irqchip pin toggled by this event. When 3091an event is triggered on the eventfd, an interrupt is injected into 3092the guest using the specified gsi pin. The irqfd is removed using 3093the KVM_IRQFD_FLAG_DEASSIGN flag, specifying both kvm_irqfd.fd 3094and kvm_irqfd.gsi. 3095 3096With KVM_CAP_IRQFD_RESAMPLE, KVM_IRQFD supports a de-assert and notify 3097mechanism allowing emulation of level-triggered, irqfd-based 3098interrupts. When KVM_IRQFD_FLAG_RESAMPLE is set the user must pass an 3099additional eventfd in the kvm_irqfd.resamplefd field. When operating 3100in resample mode, posting of an interrupt through kvm_irq.fd asserts 3101the specified gsi in the irqchip. When the irqchip is resampled, such 3102as from an EOI, the gsi is de-asserted and the user is notified via 3103kvm_irqfd.resamplefd. It is the user's responsibility to re-queue 3104the interrupt if the device making use of it still requires service. 3105Note that closing the resamplefd is not sufficient to disable the 3106irqfd. The KVM_IRQFD_FLAG_RESAMPLE is only necessary on assignment 3107and need not be specified with KVM_IRQFD_FLAG_DEASSIGN. 3108 3109On arm64, gsi routing being supported, the following can happen: 3110 3111- in case no routing entry is associated to this gsi, injection fails 3112- in case the gsi is associated to an irqchip routing entry, 3113 irqchip.pin + 32 corresponds to the injected SPI ID. 3114- in case the gsi is associated to an MSI routing entry, the MSI 3115 message and device ID are translated into an LPI (support restricted 3116 to GICv3 ITS in-kernel emulation). 3117 31184.76 KVM_PPC_ALLOCATE_HTAB 3119-------------------------- 3120 3121:Capability: KVM_CAP_PPC_ALLOC_HTAB 3122:Architectures: powerpc 3123:Type: vm ioctl 3124:Parameters: Pointer to u32 containing hash table order (in/out) 3125:Returns: 0 on success, -1 on error 3126 3127This requests the host kernel to allocate an MMU hash table for a 3128guest using the PAPR paravirtualization interface. This only does 3129anything if the kernel is configured to use the Book 3S HV style of 3130virtualization. Otherwise the capability doesn't exist and the ioctl 3131returns an ENOTTY error. The rest of this description assumes Book 3S 3132HV. 3133 3134There must be no vcpus running when this ioctl is called; if there 3135are, it will do nothing and return an EBUSY error. 3136 3137The parameter is a pointer to a 32-bit unsigned integer variable 3138containing the order (log base 2) of the desired size of the hash 3139table, which must be between 18 and 46. On successful return from the 3140ioctl, the value will not be changed by the kernel. 3141 3142If no hash table has been allocated when any vcpu is asked to run 3143(with the KVM_RUN ioctl), the host kernel will allocate a 3144default-sized hash table (16 MB). 3145 3146If this ioctl is called when a hash table has already been allocated, 3147with a different order from the existing hash table, the existing hash 3148table will be freed and a new one allocated. If this is ioctl is 3149called when a hash table has already been allocated of the same order 3150as specified, the kernel will clear out the existing hash table (zero 3151all HPTEs). In either case, if the guest is using the virtualized 3152real-mode area (VRMA) facility, the kernel will re-create the VMRA 3153HPTEs on the next KVM_RUN of any vcpu. 3154 31554.77 KVM_S390_INTERRUPT 3156----------------------- 3157 3158:Capability: basic 3159:Architectures: s390 3160:Type: vm ioctl, vcpu ioctl 3161:Parameters: struct kvm_s390_interrupt (in) 3162:Returns: 0 on success, -1 on error 3163 3164Allows to inject an interrupt to the guest. Interrupts can be floating 3165(vm ioctl) or per cpu (vcpu ioctl), depending on the interrupt type. 3166 3167Interrupt parameters are passed via kvm_s390_interrupt:: 3168 3169 struct kvm_s390_interrupt { 3170 __u32 type; 3171 __u32 parm; 3172 __u64 parm64; 3173 }; 3174 3175type can be one of the following: 3176 3177KVM_S390_SIGP_STOP (vcpu) 3178 - sigp stop; optional flags in parm 3179KVM_S390_PROGRAM_INT (vcpu) 3180 - program check; code in parm 3181KVM_S390_SIGP_SET_PREFIX (vcpu) 3182 - sigp set prefix; prefix address in parm 3183KVM_S390_RESTART (vcpu) 3184 - restart 3185KVM_S390_INT_CLOCK_COMP (vcpu) 3186 - clock comparator interrupt 3187KVM_S390_INT_CPU_TIMER (vcpu) 3188 - CPU timer interrupt 3189KVM_S390_INT_VIRTIO (vm) 3190 - virtio external interrupt; external interrupt 3191 parameters in parm and parm64 3192KVM_S390_INT_SERVICE (vm) 3193 - sclp external interrupt; sclp parameter in parm 3194KVM_S390_INT_EMERGENCY (vcpu) 3195 - sigp emergency; source cpu in parm 3196KVM_S390_INT_EXTERNAL_CALL (vcpu) 3197 - sigp external call; source cpu in parm 3198KVM_S390_INT_IO(ai,cssid,ssid,schid) (vm) 3199 - compound value to indicate an 3200 I/O interrupt (ai - adapter interrupt; cssid,ssid,schid - subchannel); 3201 I/O interruption parameters in parm (subchannel) and parm64 (intparm, 3202 interruption subclass) 3203KVM_S390_MCHK (vm, vcpu) 3204 - machine check interrupt; cr 14 bits in parm, machine check interrupt 3205 code in parm64 (note that machine checks needing further payload are not 3206 supported by this ioctl) 3207 3208This is an asynchronous vcpu ioctl and can be invoked from any thread. 3209 32104.78 KVM_PPC_GET_HTAB_FD 3211------------------------ 3212 3213:Capability: KVM_CAP_PPC_HTAB_FD 3214:Architectures: powerpc 3215:Type: vm ioctl 3216:Parameters: Pointer to struct kvm_get_htab_fd (in) 3217:Returns: file descriptor number (>= 0) on success, -1 on error 3218 3219This returns a file descriptor that can be used either to read out the 3220entries in the guest's hashed page table (HPT), or to write entries to 3221initialize the HPT. The returned fd can only be written to if the 3222KVM_GET_HTAB_WRITE bit is set in the flags field of the argument, and 3223can only be read if that bit is clear. The argument struct looks like 3224this:: 3225 3226 /* For KVM_PPC_GET_HTAB_FD */ 3227 struct kvm_get_htab_fd { 3228 __u64 flags; 3229 __u64 start_index; 3230 __u64 reserved[2]; 3231 }; 3232 3233 /* Values for kvm_get_htab_fd.flags */ 3234 #define KVM_GET_HTAB_BOLTED_ONLY ((__u64)0x1) 3235 #define KVM_GET_HTAB_WRITE ((__u64)0x2) 3236 3237The 'start_index' field gives the index in the HPT of the entry at 3238which to start reading. It is ignored when writing. 3239 3240Reads on the fd will initially supply information about all 3241"interesting" HPT entries. Interesting entries are those with the 3242bolted bit set, if the KVM_GET_HTAB_BOLTED_ONLY bit is set, otherwise 3243all entries. When the end of the HPT is reached, the read() will 3244return. If read() is called again on the fd, it will start again from 3245the beginning of the HPT, but will only return HPT entries that have 3246changed since they were last read. 3247 3248Data read or written is structured as a header (8 bytes) followed by a 3249series of valid HPT entries (16 bytes) each. The header indicates how 3250many valid HPT entries there are and how many invalid entries follow 3251the valid entries. The invalid entries are not represented explicitly 3252in the stream. The header format is:: 3253 3254 struct kvm_get_htab_header { 3255 __u32 index; 3256 __u16 n_valid; 3257 __u16 n_invalid; 3258 }; 3259 3260Writes to the fd create HPT entries starting at the index given in the 3261header; first 'n_valid' valid entries with contents from the data 3262written, then 'n_invalid' invalid entries, invalidating any previously 3263valid entries found. 3264 32654.79 KVM_CREATE_DEVICE 3266---------------------- 3267 3268:Capability: KVM_CAP_DEVICE_CTRL 3269:Architectures: all 3270:Type: vm ioctl 3271:Parameters: struct kvm_create_device (in/out) 3272:Returns: 0 on success, -1 on error 3273 3274Errors: 3275 3276 ====== ======================================================= 3277 ENODEV The device type is unknown or unsupported 3278 EEXIST Device already created, and this type of device may not 3279 be instantiated multiple times 3280 ====== ======================================================= 3281 3282 Other error conditions may be defined by individual device types or 3283 have their standard meanings. 3284 3285Creates an emulated device in the kernel. The file descriptor returned 3286in fd can be used with KVM_SET/GET/HAS_DEVICE_ATTR. 3287 3288If the KVM_CREATE_DEVICE_TEST flag is set, only test whether the 3289device type is supported (not necessarily whether it can be created 3290in the current vm). 3291 3292Individual devices should not define flags. Attributes should be used 3293for specifying any behavior that is not implied by the device type 3294number. 3295 3296:: 3297 3298 struct kvm_create_device { 3299 __u32 type; /* in: KVM_DEV_TYPE_xxx */ 3300 __u32 fd; /* out: device handle */ 3301 __u32 flags; /* in: KVM_CREATE_DEVICE_xxx */ 3302 }; 3303 33044.80 KVM_SET_DEVICE_ATTR/KVM_GET_DEVICE_ATTR 3305-------------------------------------------- 3306 3307:Capability: KVM_CAP_DEVICE_CTRL, KVM_CAP_VM_ATTRIBUTES for vm device, 3308 KVM_CAP_VCPU_ATTRIBUTES for vcpu device 3309 KVM_CAP_SYS_ATTRIBUTES for system (/dev/kvm) device (no set) 3310:Architectures: x86, arm64, s390 3311:Type: device ioctl, vm ioctl, vcpu ioctl 3312:Parameters: struct kvm_device_attr 3313:Returns: 0 on success, -1 on error 3314 3315Errors: 3316 3317 ===== ============================================================= 3318 ENXIO The group or attribute is unknown/unsupported for this device 3319 or hardware support is missing. 3320 EPERM The attribute cannot (currently) be accessed this way 3321 (e.g. read-only attribute, or attribute that only makes 3322 sense when the device is in a different state) 3323 ===== ============================================================= 3324 3325 Other error conditions may be defined by individual device types. 3326 3327Gets/sets a specified piece of device configuration and/or state. The 3328semantics are device-specific. See individual device documentation in 3329the "devices" directory. As with ONE_REG, the size of the data 3330transferred is defined by the particular attribute. 3331 3332:: 3333 3334 struct kvm_device_attr { 3335 __u32 flags; /* no flags currently defined */ 3336 __u32 group; /* device-defined */ 3337 __u64 attr; /* group-defined */ 3338 __u64 addr; /* userspace address of attr data */ 3339 }; 3340 33414.81 KVM_HAS_DEVICE_ATTR 3342------------------------ 3343 3344:Capability: KVM_CAP_DEVICE_CTRL, KVM_CAP_VM_ATTRIBUTES for vm device, 3345 KVM_CAP_VCPU_ATTRIBUTES for vcpu device 3346 KVM_CAP_SYS_ATTRIBUTES for system (/dev/kvm) device 3347:Type: device ioctl, vm ioctl, vcpu ioctl 3348:Parameters: struct kvm_device_attr 3349:Returns: 0 on success, -1 on error 3350 3351Errors: 3352 3353 ===== ============================================================= 3354 ENXIO The group or attribute is unknown/unsupported for this device 3355 or hardware support is missing. 3356 ===== ============================================================= 3357 3358Tests whether a device supports a particular attribute. A successful 3359return indicates the attribute is implemented. It does not necessarily 3360indicate that the attribute can be read or written in the device's 3361current state. "addr" is ignored. 3362 33634.82 KVM_ARM_VCPU_INIT 3364---------------------- 3365 3366:Capability: basic 3367:Architectures: arm64 3368:Type: vcpu ioctl 3369:Parameters: struct kvm_vcpu_init (in) 3370:Returns: 0 on success; -1 on error 3371 3372Errors: 3373 3374 ====== ================================================================= 3375 EINVAL the target is unknown, or the combination of features is invalid. 3376 ENOENT a features bit specified is unknown. 3377 ====== ================================================================= 3378 3379This tells KVM what type of CPU to present to the guest, and what 3380optional features it should have. This will cause a reset of the cpu 3381registers to their initial values. If this is not called, KVM_RUN will 3382return ENOEXEC for that vcpu. 3383 3384The initial values are defined as: 3385 - Processor state: 3386 * AArch64: EL1h, D, A, I and F bits set. All other bits 3387 are cleared. 3388 * AArch32: SVC, A, I and F bits set. All other bits are 3389 cleared. 3390 - General Purpose registers, including PC and SP: set to 0 3391 - FPSIMD/NEON registers: set to 0 3392 - SVE registers: set to 0 3393 - System registers: Reset to their architecturally defined 3394 values as for a warm reset to EL1 (resp. SVC) 3395 3396Note that because some registers reflect machine topology, all vcpus 3397should be created before this ioctl is invoked. 3398 3399Userspace can call this function multiple times for a given vcpu, including 3400after the vcpu has been run. This will reset the vcpu to its initial 3401state. All calls to this function after the initial call must use the same 3402target and same set of feature flags, otherwise EINVAL will be returned. 3403 3404Possible features: 3405 3406 - KVM_ARM_VCPU_POWER_OFF: Starts the CPU in a power-off state. 3407 Depends on KVM_CAP_ARM_PSCI. If not set, the CPU will be powered on 3408 and execute guest code when KVM_RUN is called. 3409 - KVM_ARM_VCPU_EL1_32BIT: Starts the CPU in a 32bit mode. 3410 Depends on KVM_CAP_ARM_EL1_32BIT (arm64 only). 3411 - KVM_ARM_VCPU_PSCI_0_2: Emulate PSCI v0.2 (or a future revision 3412 backward compatible with v0.2) for the CPU. 3413 Depends on KVM_CAP_ARM_PSCI_0_2. 3414 - KVM_ARM_VCPU_PMU_V3: Emulate PMUv3 for the CPU. 3415 Depends on KVM_CAP_ARM_PMU_V3. 3416 3417 - KVM_ARM_VCPU_PTRAUTH_ADDRESS: Enables Address Pointer authentication 3418 for arm64 only. 3419 Depends on KVM_CAP_ARM_PTRAUTH_ADDRESS. 3420 If KVM_CAP_ARM_PTRAUTH_ADDRESS and KVM_CAP_ARM_PTRAUTH_GENERIC are 3421 both present, then both KVM_ARM_VCPU_PTRAUTH_ADDRESS and 3422 KVM_ARM_VCPU_PTRAUTH_GENERIC must be requested or neither must be 3423 requested. 3424 3425 - KVM_ARM_VCPU_PTRAUTH_GENERIC: Enables Generic Pointer authentication 3426 for arm64 only. 3427 Depends on KVM_CAP_ARM_PTRAUTH_GENERIC. 3428 If KVM_CAP_ARM_PTRAUTH_ADDRESS and KVM_CAP_ARM_PTRAUTH_GENERIC are 3429 both present, then both KVM_ARM_VCPU_PTRAUTH_ADDRESS and 3430 KVM_ARM_VCPU_PTRAUTH_GENERIC must be requested or neither must be 3431 requested. 3432 3433 - KVM_ARM_VCPU_SVE: Enables SVE for the CPU (arm64 only). 3434 Depends on KVM_CAP_ARM_SVE. 3435 Requires KVM_ARM_VCPU_FINALIZE(KVM_ARM_VCPU_SVE): 3436 3437 * After KVM_ARM_VCPU_INIT: 3438 3439 - KVM_REG_ARM64_SVE_VLS may be read using KVM_GET_ONE_REG: the 3440 initial value of this pseudo-register indicates the best set of 3441 vector lengths possible for a vcpu on this host. 3442 3443 * Before KVM_ARM_VCPU_FINALIZE(KVM_ARM_VCPU_SVE): 3444 3445 - KVM_RUN and KVM_GET_REG_LIST are not available; 3446 3447 - KVM_GET_ONE_REG and KVM_SET_ONE_REG cannot be used to access 3448 the scalable archietctural SVE registers 3449 KVM_REG_ARM64_SVE_ZREG(), KVM_REG_ARM64_SVE_PREG() or 3450 KVM_REG_ARM64_SVE_FFR; 3451 3452 - KVM_REG_ARM64_SVE_VLS may optionally be written using 3453 KVM_SET_ONE_REG, to modify the set of vector lengths available 3454 for the vcpu. 3455 3456 * After KVM_ARM_VCPU_FINALIZE(KVM_ARM_VCPU_SVE): 3457 3458 - the KVM_REG_ARM64_SVE_VLS pseudo-register is immutable, and can 3459 no longer be written using KVM_SET_ONE_REG. 3460 34614.83 KVM_ARM_PREFERRED_TARGET 3462----------------------------- 3463 3464:Capability: basic 3465:Architectures: arm64 3466:Type: vm ioctl 3467:Parameters: struct kvm_vcpu_init (out) 3468:Returns: 0 on success; -1 on error 3469 3470Errors: 3471 3472 ====== ========================================== 3473 ENODEV no preferred target available for the host 3474 ====== ========================================== 3475 3476This queries KVM for preferred CPU target type which can be emulated 3477by KVM on underlying host. 3478 3479The ioctl returns struct kvm_vcpu_init instance containing information 3480about preferred CPU target type and recommended features for it. The 3481kvm_vcpu_init->features bitmap returned will have feature bits set if 3482the preferred target recommends setting these features, but this is 3483not mandatory. 3484 3485The information returned by this ioctl can be used to prepare an instance 3486of struct kvm_vcpu_init for KVM_ARM_VCPU_INIT ioctl which will result in 3487VCPU matching underlying host. 3488 3489 34904.84 KVM_GET_REG_LIST 3491--------------------- 3492 3493:Capability: basic 3494:Architectures: arm64, mips 3495:Type: vcpu ioctl 3496:Parameters: struct kvm_reg_list (in/out) 3497:Returns: 0 on success; -1 on error 3498 3499Errors: 3500 3501 ===== ============================================================== 3502 E2BIG the reg index list is too big to fit in the array specified by 3503 the user (the number required will be written into n). 3504 ===== ============================================================== 3505 3506:: 3507 3508 struct kvm_reg_list { 3509 __u64 n; /* number of registers in reg[] */ 3510 __u64 reg[0]; 3511 }; 3512 3513This ioctl returns the guest registers that are supported for the 3514KVM_GET_ONE_REG/KVM_SET_ONE_REG calls. 3515 3516 35174.85 KVM_ARM_SET_DEVICE_ADDR (deprecated) 3518----------------------------------------- 3519 3520:Capability: KVM_CAP_ARM_SET_DEVICE_ADDR 3521:Architectures: arm64 3522:Type: vm ioctl 3523:Parameters: struct kvm_arm_device_address (in) 3524:Returns: 0 on success, -1 on error 3525 3526Errors: 3527 3528 ====== ============================================ 3529 ENODEV The device id is unknown 3530 ENXIO Device not supported on current system 3531 EEXIST Address already set 3532 E2BIG Address outside guest physical address space 3533 EBUSY Address overlaps with other device range 3534 ====== ============================================ 3535 3536:: 3537 3538 struct kvm_arm_device_addr { 3539 __u64 id; 3540 __u64 addr; 3541 }; 3542 3543Specify a device address in the guest's physical address space where guests 3544can access emulated or directly exposed devices, which the host kernel needs 3545to know about. The id field is an architecture specific identifier for a 3546specific device. 3547 3548arm64 divides the id field into two parts, a device id and an 3549address type id specific to the individual device:: 3550 3551 bits: | 63 ... 32 | 31 ... 16 | 15 ... 0 | 3552 field: | 0x00000000 | device id | addr type id | 3553 3554arm64 currently only require this when using the in-kernel GIC 3555support for the hardware VGIC features, using KVM_ARM_DEVICE_VGIC_V2 3556as the device id. When setting the base address for the guest's 3557mapping of the VGIC virtual CPU and distributor interface, the ioctl 3558must be called after calling KVM_CREATE_IRQCHIP, but before calling 3559KVM_RUN on any of the VCPUs. Calling this ioctl twice for any of the 3560base addresses will return -EEXIST. 3561 3562Note, this IOCTL is deprecated and the more flexible SET/GET_DEVICE_ATTR API 3563should be used instead. 3564 3565 35664.86 KVM_PPC_RTAS_DEFINE_TOKEN 3567------------------------------ 3568 3569:Capability: KVM_CAP_PPC_RTAS 3570:Architectures: ppc 3571:Type: vm ioctl 3572:Parameters: struct kvm_rtas_token_args 3573:Returns: 0 on success, -1 on error 3574 3575Defines a token value for a RTAS (Run Time Abstraction Services) 3576service in order to allow it to be handled in the kernel. The 3577argument struct gives the name of the service, which must be the name 3578of a service that has a kernel-side implementation. If the token 3579value is non-zero, it will be associated with that service, and 3580subsequent RTAS calls by the guest specifying that token will be 3581handled by the kernel. If the token value is 0, then any token 3582associated with the service will be forgotten, and subsequent RTAS 3583calls by the guest for that service will be passed to userspace to be 3584handled. 3585 35864.87 KVM_SET_GUEST_DEBUG 3587------------------------ 3588 3589:Capability: KVM_CAP_SET_GUEST_DEBUG 3590:Architectures: x86, s390, ppc, arm64 3591:Type: vcpu ioctl 3592:Parameters: struct kvm_guest_debug (in) 3593:Returns: 0 on success; -1 on error 3594 3595:: 3596 3597 struct kvm_guest_debug { 3598 __u32 control; 3599 __u32 pad; 3600 struct kvm_guest_debug_arch arch; 3601 }; 3602 3603Set up the processor specific debug registers and configure vcpu for 3604handling guest debug events. There are two parts to the structure, the 3605first a control bitfield indicates the type of debug events to handle 3606when running. Common control bits are: 3607 3608 - KVM_GUESTDBG_ENABLE: guest debugging is enabled 3609 - KVM_GUESTDBG_SINGLESTEP: the next run should single-step 3610 3611The top 16 bits of the control field are architecture specific control 3612flags which can include the following: 3613 3614 - KVM_GUESTDBG_USE_SW_BP: using software breakpoints [x86, arm64] 3615 - KVM_GUESTDBG_USE_HW_BP: using hardware breakpoints [x86, s390] 3616 - KVM_GUESTDBG_USE_HW: using hardware debug events [arm64] 3617 - KVM_GUESTDBG_INJECT_DB: inject DB type exception [x86] 3618 - KVM_GUESTDBG_INJECT_BP: inject BP type exception [x86] 3619 - KVM_GUESTDBG_EXIT_PENDING: trigger an immediate guest exit [s390] 3620 - KVM_GUESTDBG_BLOCKIRQ: avoid injecting interrupts/NMI/SMI [x86] 3621 3622For example KVM_GUESTDBG_USE_SW_BP indicates that software breakpoints 3623are enabled in memory so we need to ensure breakpoint exceptions are 3624correctly trapped and the KVM run loop exits at the breakpoint and not 3625running off into the normal guest vector. For KVM_GUESTDBG_USE_HW_BP 3626we need to ensure the guest vCPUs architecture specific registers are 3627updated to the correct (supplied) values. 3628 3629The second part of the structure is architecture specific and 3630typically contains a set of debug registers. 3631 3632For arm64 the number of debug registers is implementation defined and 3633can be determined by querying the KVM_CAP_GUEST_DEBUG_HW_BPS and 3634KVM_CAP_GUEST_DEBUG_HW_WPS capabilities which return a positive number 3635indicating the number of supported registers. 3636 3637For ppc, the KVM_CAP_PPC_GUEST_DEBUG_SSTEP capability indicates whether 3638the single-step debug event (KVM_GUESTDBG_SINGLESTEP) is supported. 3639 3640Also when supported, KVM_CAP_SET_GUEST_DEBUG2 capability indicates the 3641supported KVM_GUESTDBG_* bits in the control field. 3642 3643When debug events exit the main run loop with the reason 3644KVM_EXIT_DEBUG with the kvm_debug_exit_arch part of the kvm_run 3645structure containing architecture specific debug information. 3646 36474.88 KVM_GET_EMULATED_CPUID 3648--------------------------- 3649 3650:Capability: KVM_CAP_EXT_EMUL_CPUID 3651:Architectures: x86 3652:Type: system ioctl 3653:Parameters: struct kvm_cpuid2 (in/out) 3654:Returns: 0 on success, -1 on error 3655 3656:: 3657 3658 struct kvm_cpuid2 { 3659 __u32 nent; 3660 __u32 flags; 3661 struct kvm_cpuid_entry2 entries[0]; 3662 }; 3663 3664The member 'flags' is used for passing flags from userspace. 3665 3666:: 3667 3668 #define KVM_CPUID_FLAG_SIGNIFCANT_INDEX BIT(0) 3669 #define KVM_CPUID_FLAG_STATEFUL_FUNC BIT(1) /* deprecated */ 3670 #define KVM_CPUID_FLAG_STATE_READ_NEXT BIT(2) /* deprecated */ 3671 3672 struct kvm_cpuid_entry2 { 3673 __u32 function; 3674 __u32 index; 3675 __u32 flags; 3676 __u32 eax; 3677 __u32 ebx; 3678 __u32 ecx; 3679 __u32 edx; 3680 __u32 padding[3]; 3681 }; 3682 3683This ioctl returns x86 cpuid features which are emulated by 3684kvm.Userspace can use the information returned by this ioctl to query 3685which features are emulated by kvm instead of being present natively. 3686 3687Userspace invokes KVM_GET_EMULATED_CPUID by passing a kvm_cpuid2 3688structure with the 'nent' field indicating the number of entries in 3689the variable-size array 'entries'. If the number of entries is too low 3690to describe the cpu capabilities, an error (E2BIG) is returned. If the 3691number is too high, the 'nent' field is adjusted and an error (ENOMEM) 3692is returned. If the number is just right, the 'nent' field is adjusted 3693to the number of valid entries in the 'entries' array, which is then 3694filled. 3695 3696The entries returned are the set CPUID bits of the respective features 3697which kvm emulates, as returned by the CPUID instruction, with unknown 3698or unsupported feature bits cleared. 3699 3700Features like x2apic, for example, may not be present in the host cpu 3701but are exposed by kvm in KVM_GET_SUPPORTED_CPUID because they can be 3702emulated efficiently and thus not included here. 3703 3704The fields in each entry are defined as follows: 3705 3706 function: 3707 the eax value used to obtain the entry 3708 index: 3709 the ecx value used to obtain the entry (for entries that are 3710 affected by ecx) 3711 flags: 3712 an OR of zero or more of the following: 3713 3714 KVM_CPUID_FLAG_SIGNIFCANT_INDEX: 3715 if the index field is valid 3716 3717 eax, ebx, ecx, edx: 3718 3719 the values returned by the cpuid instruction for 3720 this function/index combination 3721 37224.89 KVM_S390_MEM_OP 3723-------------------- 3724 3725:Capability: KVM_CAP_S390_MEM_OP, KVM_CAP_S390_PROTECTED, KVM_CAP_S390_MEM_OP_EXTENSION 3726:Architectures: s390 3727:Type: vm ioctl, vcpu ioctl 3728:Parameters: struct kvm_s390_mem_op (in) 3729:Returns: = 0 on success, 3730 < 0 on generic error (e.g. -EFAULT or -ENOMEM), 3731 > 0 if an exception occurred while walking the page tables 3732 3733Read or write data from/to the VM's memory. 3734The KVM_CAP_S390_MEM_OP_EXTENSION capability specifies what functionality is 3735supported. 3736 3737Parameters are specified via the following structure:: 3738 3739 struct kvm_s390_mem_op { 3740 __u64 gaddr; /* the guest address */ 3741 __u64 flags; /* flags */ 3742 __u32 size; /* amount of bytes */ 3743 __u32 op; /* type of operation */ 3744 __u64 buf; /* buffer in userspace */ 3745 union { 3746 struct { 3747 __u8 ar; /* the access register number */ 3748 __u8 key; /* access key, ignored if flag unset */ 3749 }; 3750 __u32 sida_offset; /* offset into the sida */ 3751 __u8 reserved[32]; /* ignored */ 3752 }; 3753 }; 3754 3755The start address of the memory region has to be specified in the "gaddr" 3756field, and the length of the region in the "size" field (which must not 3757be 0). The maximum value for "size" can be obtained by checking the 3758KVM_CAP_S390_MEM_OP capability. "buf" is the buffer supplied by the 3759userspace application where the read data should be written to for 3760a read access, or where the data that should be written is stored for 3761a write access. The "reserved" field is meant for future extensions. 3762Reserved and unused values are ignored. Future extension that add members must 3763introduce new flags. 3764 3765The type of operation is specified in the "op" field. Flags modifying 3766their behavior can be set in the "flags" field. Undefined flag bits must 3767be set to 0. 3768 3769Possible operations are: 3770 * ``KVM_S390_MEMOP_LOGICAL_READ`` 3771 * ``KVM_S390_MEMOP_LOGICAL_WRITE`` 3772 * ``KVM_S390_MEMOP_ABSOLUTE_READ`` 3773 * ``KVM_S390_MEMOP_ABSOLUTE_WRITE`` 3774 * ``KVM_S390_MEMOP_SIDA_READ`` 3775 * ``KVM_S390_MEMOP_SIDA_WRITE`` 3776 3777Logical read/write: 3778^^^^^^^^^^^^^^^^^^^ 3779 3780Access logical memory, i.e. translate the given guest address to an absolute 3781address given the state of the VCPU and use the absolute address as target of 3782the access. "ar" designates the access register number to be used; the valid 3783range is 0..15. 3784Logical accesses are permitted for the VCPU ioctl only. 3785Logical accesses are permitted for non-protected guests only. 3786 3787Supported flags: 3788 * ``KVM_S390_MEMOP_F_CHECK_ONLY`` 3789 * ``KVM_S390_MEMOP_F_INJECT_EXCEPTION`` 3790 * ``KVM_S390_MEMOP_F_SKEY_PROTECTION`` 3791 3792The KVM_S390_MEMOP_F_CHECK_ONLY flag can be set to check whether the 3793corresponding memory access would cause an access exception; however, 3794no actual access to the data in memory at the destination is performed. 3795In this case, "buf" is unused and can be NULL. 3796 3797In case an access exception occurred during the access (or would occur 3798in case of KVM_S390_MEMOP_F_CHECK_ONLY), the ioctl returns a positive 3799error number indicating the type of exception. This exception is also 3800raised directly at the corresponding VCPU if the flag 3801KVM_S390_MEMOP_F_INJECT_EXCEPTION is set. 3802On protection exceptions, unless specified otherwise, the injected 3803translation-exception identifier (TEID) indicates suppression. 3804 3805If the KVM_S390_MEMOP_F_SKEY_PROTECTION flag is set, storage key 3806protection is also in effect and may cause exceptions if accesses are 3807prohibited given the access key designated by "key"; the valid range is 0..15. 3808KVM_S390_MEMOP_F_SKEY_PROTECTION is available if KVM_CAP_S390_MEM_OP_EXTENSION 3809is > 0. 3810Since the accessed memory may span multiple pages and those pages might have 3811different storage keys, it is possible that a protection exception occurs 3812after memory has been modified. In this case, if the exception is injected, 3813the TEID does not indicate suppression. 3814 3815Absolute read/write: 3816^^^^^^^^^^^^^^^^^^^^ 3817 3818Access absolute memory. This operation is intended to be used with the 3819KVM_S390_MEMOP_F_SKEY_PROTECTION flag, to allow accessing memory and performing 3820the checks required for storage key protection as one operation (as opposed to 3821user space getting the storage keys, performing the checks, and accessing 3822memory thereafter, which could lead to a delay between check and access). 3823Absolute accesses are permitted for the VM ioctl if KVM_CAP_S390_MEM_OP_EXTENSION 3824is > 0. 3825Currently absolute accesses are not permitted for VCPU ioctls. 3826Absolute accesses are permitted for non-protected guests only. 3827 3828Supported flags: 3829 * ``KVM_S390_MEMOP_F_CHECK_ONLY`` 3830 * ``KVM_S390_MEMOP_F_SKEY_PROTECTION`` 3831 3832The semantics of the flags are as for logical accesses. 3833 3834SIDA read/write: 3835^^^^^^^^^^^^^^^^ 3836 3837Access the secure instruction data area which contains memory operands necessary 3838for instruction emulation for protected guests. 3839SIDA accesses are available if the KVM_CAP_S390_PROTECTED capability is available. 3840SIDA accesses are permitted for the VCPU ioctl only. 3841SIDA accesses are permitted for protected guests only. 3842 3843No flags are supported. 3844 38454.90 KVM_S390_GET_SKEYS 3846----------------------- 3847 3848:Capability: KVM_CAP_S390_SKEYS 3849:Architectures: s390 3850:Type: vm ioctl 3851:Parameters: struct kvm_s390_skeys 3852:Returns: 0 on success, KVM_S390_GET_SKEYS_NONE if guest is not using storage 3853 keys, negative value on error 3854 3855This ioctl is used to get guest storage key values on the s390 3856architecture. The ioctl takes parameters via the kvm_s390_skeys struct:: 3857 3858 struct kvm_s390_skeys { 3859 __u64 start_gfn; 3860 __u64 count; 3861 __u64 skeydata_addr; 3862 __u32 flags; 3863 __u32 reserved[9]; 3864 }; 3865 3866The start_gfn field is the number of the first guest frame whose storage keys 3867you want to get. 3868 3869The count field is the number of consecutive frames (starting from start_gfn) 3870whose storage keys to get. The count field must be at least 1 and the maximum 3871allowed value is defined as KVM_S390_SKEYS_MAX. Values outside this range 3872will cause the ioctl to return -EINVAL. 3873 3874The skeydata_addr field is the address to a buffer large enough to hold count 3875bytes. This buffer will be filled with storage key data by the ioctl. 3876 38774.91 KVM_S390_SET_SKEYS 3878----------------------- 3879 3880:Capability: KVM_CAP_S390_SKEYS 3881:Architectures: s390 3882:Type: vm ioctl 3883:Parameters: struct kvm_s390_skeys 3884:Returns: 0 on success, negative value on error 3885 3886This ioctl is used to set guest storage key values on the s390 3887architecture. The ioctl takes parameters via the kvm_s390_skeys struct. 3888See section on KVM_S390_GET_SKEYS for struct definition. 3889 3890The start_gfn field is the number of the first guest frame whose storage keys 3891you want to set. 3892 3893The count field is the number of consecutive frames (starting from start_gfn) 3894whose storage keys to get. The count field must be at least 1 and the maximum 3895allowed value is defined as KVM_S390_SKEYS_MAX. Values outside this range 3896will cause the ioctl to return -EINVAL. 3897 3898The skeydata_addr field is the address to a buffer containing count bytes of 3899storage keys. Each byte in the buffer will be set as the storage key for a 3900single frame starting at start_gfn for count frames. 3901 3902Note: If any architecturally invalid key value is found in the given data then 3903the ioctl will return -EINVAL. 3904 39054.92 KVM_S390_IRQ 3906----------------- 3907 3908:Capability: KVM_CAP_S390_INJECT_IRQ 3909:Architectures: s390 3910:Type: vcpu ioctl 3911:Parameters: struct kvm_s390_irq (in) 3912:Returns: 0 on success, -1 on error 3913 3914Errors: 3915 3916 3917 ====== ================================================================= 3918 EINVAL interrupt type is invalid 3919 type is KVM_S390_SIGP_STOP and flag parameter is invalid value, 3920 type is KVM_S390_INT_EXTERNAL_CALL and code is bigger 3921 than the maximum of VCPUs 3922 EBUSY type is KVM_S390_SIGP_SET_PREFIX and vcpu is not stopped, 3923 type is KVM_S390_SIGP_STOP and a stop irq is already pending, 3924 type is KVM_S390_INT_EXTERNAL_CALL and an external call interrupt 3925 is already pending 3926 ====== ================================================================= 3927 3928Allows to inject an interrupt to the guest. 3929 3930Using struct kvm_s390_irq as a parameter allows 3931to inject additional payload which is not 3932possible via KVM_S390_INTERRUPT. 3933 3934Interrupt parameters are passed via kvm_s390_irq:: 3935 3936 struct kvm_s390_irq { 3937 __u64 type; 3938 union { 3939 struct kvm_s390_io_info io; 3940 struct kvm_s390_ext_info ext; 3941 struct kvm_s390_pgm_info pgm; 3942 struct kvm_s390_emerg_info emerg; 3943 struct kvm_s390_extcall_info extcall; 3944 struct kvm_s390_prefix_info prefix; 3945 struct kvm_s390_stop_info stop; 3946 struct kvm_s390_mchk_info mchk; 3947 char reserved[64]; 3948 } u; 3949 }; 3950 3951type can be one of the following: 3952 3953- KVM_S390_SIGP_STOP - sigp stop; parameter in .stop 3954- KVM_S390_PROGRAM_INT - program check; parameters in .pgm 3955- KVM_S390_SIGP_SET_PREFIX - sigp set prefix; parameters in .prefix 3956- KVM_S390_RESTART - restart; no parameters 3957- KVM_S390_INT_CLOCK_COMP - clock comparator interrupt; no parameters 3958- KVM_S390_INT_CPU_TIMER - CPU timer interrupt; no parameters 3959- KVM_S390_INT_EMERGENCY - sigp emergency; parameters in .emerg 3960- KVM_S390_INT_EXTERNAL_CALL - sigp external call; parameters in .extcall 3961- KVM_S390_MCHK - machine check interrupt; parameters in .mchk 3962 3963This is an asynchronous vcpu ioctl and can be invoked from any thread. 3964 39654.94 KVM_S390_GET_IRQ_STATE 3966--------------------------- 3967 3968:Capability: KVM_CAP_S390_IRQ_STATE 3969:Architectures: s390 3970:Type: vcpu ioctl 3971:Parameters: struct kvm_s390_irq_state (out) 3972:Returns: >= number of bytes copied into buffer, 3973 -EINVAL if buffer size is 0, 3974 -ENOBUFS if buffer size is too small to fit all pending interrupts, 3975 -EFAULT if the buffer address was invalid 3976 3977This ioctl allows userspace to retrieve the complete state of all currently 3978pending interrupts in a single buffer. Use cases include migration 3979and introspection. The parameter structure contains the address of a 3980userspace buffer and its length:: 3981 3982 struct kvm_s390_irq_state { 3983 __u64 buf; 3984 __u32 flags; /* will stay unused for compatibility reasons */ 3985 __u32 len; 3986 __u32 reserved[4]; /* will stay unused for compatibility reasons */ 3987 }; 3988 3989Userspace passes in the above struct and for each pending interrupt a 3990struct kvm_s390_irq is copied to the provided buffer. 3991 3992The structure contains a flags and a reserved field for future extensions. As 3993the kernel never checked for flags == 0 and QEMU never pre-zeroed flags and 3994reserved, these fields can not be used in the future without breaking 3995compatibility. 3996 3997If -ENOBUFS is returned the buffer provided was too small and userspace 3998may retry with a bigger buffer. 3999 40004.95 KVM_S390_SET_IRQ_STATE 4001--------------------------- 4002 4003:Capability: KVM_CAP_S390_IRQ_STATE 4004:Architectures: s390 4005:Type: vcpu ioctl 4006:Parameters: struct kvm_s390_irq_state (in) 4007:Returns: 0 on success, 4008 -EFAULT if the buffer address was invalid, 4009 -EINVAL for an invalid buffer length (see below), 4010 -EBUSY if there were already interrupts pending, 4011 errors occurring when actually injecting the 4012 interrupt. See KVM_S390_IRQ. 4013 4014This ioctl allows userspace to set the complete state of all cpu-local 4015interrupts currently pending for the vcpu. It is intended for restoring 4016interrupt state after a migration. The input parameter is a userspace buffer 4017containing a struct kvm_s390_irq_state:: 4018 4019 struct kvm_s390_irq_state { 4020 __u64 buf; 4021 __u32 flags; /* will stay unused for compatibility reasons */ 4022 __u32 len; 4023 __u32 reserved[4]; /* will stay unused for compatibility reasons */ 4024 }; 4025 4026The restrictions for flags and reserved apply as well. 4027(see KVM_S390_GET_IRQ_STATE) 4028 4029The userspace memory referenced by buf contains a struct kvm_s390_irq 4030for each interrupt to be injected into the guest. 4031If one of the interrupts could not be injected for some reason the 4032ioctl aborts. 4033 4034len must be a multiple of sizeof(struct kvm_s390_irq). It must be > 0 4035and it must not exceed (max_vcpus + 32) * sizeof(struct kvm_s390_irq), 4036which is the maximum number of possibly pending cpu-local interrupts. 4037 40384.96 KVM_SMI 4039------------ 4040 4041:Capability: KVM_CAP_X86_SMM 4042:Architectures: x86 4043:Type: vcpu ioctl 4044:Parameters: none 4045:Returns: 0 on success, -1 on error 4046 4047Queues an SMI on the thread's vcpu. 4048 40494.97 KVM_X86_SET_MSR_FILTER 4050---------------------------- 4051 4052:Capability: KVM_CAP_X86_MSR_FILTER 4053:Architectures: x86 4054:Type: vm ioctl 4055:Parameters: struct kvm_msr_filter 4056:Returns: 0 on success, < 0 on error 4057 4058:: 4059 4060 struct kvm_msr_filter_range { 4061 #define KVM_MSR_FILTER_READ (1 << 0) 4062 #define KVM_MSR_FILTER_WRITE (1 << 1) 4063 __u32 flags; 4064 __u32 nmsrs; /* number of msrs in bitmap */ 4065 __u32 base; /* MSR index the bitmap starts at */ 4066 __u8 *bitmap; /* a 1 bit allows the operations in flags, 0 denies */ 4067 }; 4068 4069 #define KVM_MSR_FILTER_MAX_RANGES 16 4070 struct kvm_msr_filter { 4071 #define KVM_MSR_FILTER_DEFAULT_ALLOW (0 << 0) 4072 #define KVM_MSR_FILTER_DEFAULT_DENY (1 << 0) 4073 __u32 flags; 4074 struct kvm_msr_filter_range ranges[KVM_MSR_FILTER_MAX_RANGES]; 4075 }; 4076 4077flags values for ``struct kvm_msr_filter_range``: 4078 4079``KVM_MSR_FILTER_READ`` 4080 4081 Filter read accesses to MSRs using the given bitmap. A 0 in the bitmap 4082 indicates that read accesses should be denied, while a 1 indicates that 4083 a read for a particular MSR should be allowed regardless of the default 4084 filter action. 4085 4086``KVM_MSR_FILTER_WRITE`` 4087 4088 Filter write accesses to MSRs using the given bitmap. A 0 in the bitmap 4089 indicates that write accesses should be denied, while a 1 indicates that 4090 a write for a particular MSR should be allowed regardless of the default 4091 filter action. 4092 4093flags values for ``struct kvm_msr_filter``: 4094 4095``KVM_MSR_FILTER_DEFAULT_ALLOW`` 4096 4097 If no filter range matches an MSR index that is getting accessed, KVM will 4098 allow accesses to all MSRs by default. 4099 4100``KVM_MSR_FILTER_DEFAULT_DENY`` 4101 4102 If no filter range matches an MSR index that is getting accessed, KVM will 4103 deny accesses to all MSRs by default. 4104 4105This ioctl allows userspace to define up to 16 bitmaps of MSR ranges to deny 4106guest MSR accesses that would normally be allowed by KVM. If an MSR is not 4107covered by a specific range, the "default" filtering behavior applies. Each 4108bitmap range covers MSRs from [base .. base+nmsrs). 4109 4110If an MSR access is denied by userspace, the resulting KVM behavior depends on 4111whether or not KVM_CAP_X86_USER_SPACE_MSR's KVM_MSR_EXIT_REASON_FILTER is 4112enabled. If KVM_MSR_EXIT_REASON_FILTER is enabled, KVM will exit to userspace 4113on denied accesses, i.e. userspace effectively intercepts the MSR access. If 4114KVM_MSR_EXIT_REASON_FILTER is not enabled, KVM will inject a #GP into the guest 4115on denied accesses. 4116 4117If an MSR access is allowed by userspace, KVM will emulate and/or virtualize 4118the access in accordance with the vCPU model. Note, KVM may still ultimately 4119inject a #GP if an access is allowed by userspace, e.g. if KVM doesn't support 4120the MSR, or to follow architectural behavior for the MSR. 4121 4122By default, KVM operates in KVM_MSR_FILTER_DEFAULT_ALLOW mode with no MSR range 4123filters. 4124 4125Calling this ioctl with an empty set of ranges (all nmsrs == 0) disables MSR 4126filtering. In that mode, ``KVM_MSR_FILTER_DEFAULT_DENY`` is invalid and causes 4127an error. 4128 4129.. warning:: 4130 MSR accesses as part of nested VM-Enter/VM-Exit are not filtered. 4131 This includes both writes to individual VMCS fields and reads/writes 4132 through the MSR lists pointed to by the VMCS. 4133 4134 x2APIC MSR accesses cannot be filtered (KVM silently ignores filters that 4135 cover any x2APIC MSRs). 4136 4137Note, invoking this ioctl while a vCPU is running is inherently racy. However, 4138KVM does guarantee that vCPUs will see either the previous filter or the new 4139filter, e.g. MSRs with identical settings in both the old and new filter will 4140have deterministic behavior. 4141 4142Similarly, if userspace wishes to intercept on denied accesses, 4143KVM_MSR_EXIT_REASON_FILTER must be enabled before activating any filters, and 4144left enabled until after all filters are deactivated. Failure to do so may 4145result in KVM injecting a #GP instead of exiting to userspace. 4146 41474.98 KVM_CREATE_SPAPR_TCE_64 4148---------------------------- 4149 4150:Capability: KVM_CAP_SPAPR_TCE_64 4151:Architectures: powerpc 4152:Type: vm ioctl 4153:Parameters: struct kvm_create_spapr_tce_64 (in) 4154:Returns: file descriptor for manipulating the created TCE table 4155 4156This is an extension for KVM_CAP_SPAPR_TCE which only supports 32bit 4157windows, described in 4.62 KVM_CREATE_SPAPR_TCE 4158 4159This capability uses extended struct in ioctl interface:: 4160 4161 /* for KVM_CAP_SPAPR_TCE_64 */ 4162 struct kvm_create_spapr_tce_64 { 4163 __u64 liobn; 4164 __u32 page_shift; 4165 __u32 flags; 4166 __u64 offset; /* in pages */ 4167 __u64 size; /* in pages */ 4168 }; 4169 4170The aim of extension is to support an additional bigger DMA window with 4171a variable page size. 4172KVM_CREATE_SPAPR_TCE_64 receives a 64bit window size, an IOMMU page shift and 4173a bus offset of the corresponding DMA window, @size and @offset are numbers 4174of IOMMU pages. 4175 4176@flags are not used at the moment. 4177 4178The rest of functionality is identical to KVM_CREATE_SPAPR_TCE. 4179 41804.99 KVM_REINJECT_CONTROL 4181------------------------- 4182 4183:Capability: KVM_CAP_REINJECT_CONTROL 4184:Architectures: x86 4185:Type: vm ioctl 4186:Parameters: struct kvm_reinject_control (in) 4187:Returns: 0 on success, 4188 -EFAULT if struct kvm_reinject_control cannot be read, 4189 -ENXIO if KVM_CREATE_PIT or KVM_CREATE_PIT2 didn't succeed earlier. 4190 4191i8254 (PIT) has two modes, reinject and !reinject. The default is reinject, 4192where KVM queues elapsed i8254 ticks and monitors completion of interrupt from 4193vector(s) that i8254 injects. Reinject mode dequeues a tick and injects its 4194interrupt whenever there isn't a pending interrupt from i8254. 4195!reinject mode injects an interrupt as soon as a tick arrives. 4196 4197:: 4198 4199 struct kvm_reinject_control { 4200 __u8 pit_reinject; 4201 __u8 reserved[31]; 4202 }; 4203 4204pit_reinject = 0 (!reinject mode) is recommended, unless running an old 4205operating system that uses the PIT for timing (e.g. Linux 2.4.x). 4206 42074.100 KVM_PPC_CONFIGURE_V3_MMU 4208------------------------------ 4209 4210:Capability: KVM_CAP_PPC_RADIX_MMU or KVM_CAP_PPC_HASH_MMU_V3 4211:Architectures: ppc 4212:Type: vm ioctl 4213:Parameters: struct kvm_ppc_mmuv3_cfg (in) 4214:Returns: 0 on success, 4215 -EFAULT if struct kvm_ppc_mmuv3_cfg cannot be read, 4216 -EINVAL if the configuration is invalid 4217 4218This ioctl controls whether the guest will use radix or HPT (hashed 4219page table) translation, and sets the pointer to the process table for 4220the guest. 4221 4222:: 4223 4224 struct kvm_ppc_mmuv3_cfg { 4225 __u64 flags; 4226 __u64 process_table; 4227 }; 4228 4229There are two bits that can be set in flags; KVM_PPC_MMUV3_RADIX and 4230KVM_PPC_MMUV3_GTSE. KVM_PPC_MMUV3_RADIX, if set, configures the guest 4231to use radix tree translation, and if clear, to use HPT translation. 4232KVM_PPC_MMUV3_GTSE, if set and if KVM permits it, configures the guest 4233to be able to use the global TLB and SLB invalidation instructions; 4234if clear, the guest may not use these instructions. 4235 4236The process_table field specifies the address and size of the guest 4237process table, which is in the guest's space. This field is formatted 4238as the second doubleword of the partition table entry, as defined in 4239the Power ISA V3.00, Book III section 5.7.6.1. 4240 42414.101 KVM_PPC_GET_RMMU_INFO 4242--------------------------- 4243 4244:Capability: KVM_CAP_PPC_RADIX_MMU 4245:Architectures: ppc 4246:Type: vm ioctl 4247:Parameters: struct kvm_ppc_rmmu_info (out) 4248:Returns: 0 on success, 4249 -EFAULT if struct kvm_ppc_rmmu_info cannot be written, 4250 -EINVAL if no useful information can be returned 4251 4252This ioctl returns a structure containing two things: (a) a list 4253containing supported radix tree geometries, and (b) a list that maps 4254page sizes to put in the "AP" (actual page size) field for the tlbie 4255(TLB invalidate entry) instruction. 4256 4257:: 4258 4259 struct kvm_ppc_rmmu_info { 4260 struct kvm_ppc_radix_geom { 4261 __u8 page_shift; 4262 __u8 level_bits[4]; 4263 __u8 pad[3]; 4264 } geometries[8]; 4265 __u32 ap_encodings[8]; 4266 }; 4267 4268The geometries[] field gives up to 8 supported geometries for the 4269radix page table, in terms of the log base 2 of the smallest page 4270size, and the number of bits indexed at each level of the tree, from 4271the PTE level up to the PGD level in that order. Any unused entries 4272will have 0 in the page_shift field. 4273 4274The ap_encodings gives the supported page sizes and their AP field 4275encodings, encoded with the AP value in the top 3 bits and the log 4276base 2 of the page size in the bottom 6 bits. 4277 42784.102 KVM_PPC_RESIZE_HPT_PREPARE 4279-------------------------------- 4280 4281:Capability: KVM_CAP_SPAPR_RESIZE_HPT 4282:Architectures: powerpc 4283:Type: vm ioctl 4284:Parameters: struct kvm_ppc_resize_hpt (in) 4285:Returns: 0 on successful completion, 4286 >0 if a new HPT is being prepared, the value is an estimated 4287 number of milliseconds until preparation is complete, 4288 -EFAULT if struct kvm_reinject_control cannot be read, 4289 -EINVAL if the supplied shift or flags are invalid, 4290 -ENOMEM if unable to allocate the new HPT, 4291 4292Used to implement the PAPR extension for runtime resizing of a guest's 4293Hashed Page Table (HPT). Specifically this starts, stops or monitors 4294the preparation of a new potential HPT for the guest, essentially 4295implementing the H_RESIZE_HPT_PREPARE hypercall. 4296 4297:: 4298 4299 struct kvm_ppc_resize_hpt { 4300 __u64 flags; 4301 __u32 shift; 4302 __u32 pad; 4303 }; 4304 4305If called with shift > 0 when there is no pending HPT for the guest, 4306this begins preparation of a new pending HPT of size 2^(shift) bytes. 4307It then returns a positive integer with the estimated number of 4308milliseconds until preparation is complete. 4309 4310If called when there is a pending HPT whose size does not match that 4311requested in the parameters, discards the existing pending HPT and 4312creates a new one as above. 4313 4314If called when there is a pending HPT of the size requested, will: 4315 4316 * If preparation of the pending HPT is already complete, return 0 4317 * If preparation of the pending HPT has failed, return an error 4318 code, then discard the pending HPT. 4319 * If preparation of the pending HPT is still in progress, return an 4320 estimated number of milliseconds until preparation is complete. 4321 4322If called with shift == 0, discards any currently pending HPT and 4323returns 0 (i.e. cancels any in-progress preparation). 4324 4325flags is reserved for future expansion, currently setting any bits in 4326flags will result in an -EINVAL. 4327 4328Normally this will be called repeatedly with the same parameters until 4329it returns <= 0. The first call will initiate preparation, subsequent 4330ones will monitor preparation until it completes or fails. 4331 43324.103 KVM_PPC_RESIZE_HPT_COMMIT 4333------------------------------- 4334 4335:Capability: KVM_CAP_SPAPR_RESIZE_HPT 4336:Architectures: powerpc 4337:Type: vm ioctl 4338:Parameters: struct kvm_ppc_resize_hpt (in) 4339:Returns: 0 on successful completion, 4340 -EFAULT if struct kvm_reinject_control cannot be read, 4341 -EINVAL if the supplied shift or flags are invalid, 4342 -ENXIO is there is no pending HPT, or the pending HPT doesn't 4343 have the requested size, 4344 -EBUSY if the pending HPT is not fully prepared, 4345 -ENOSPC if there was a hash collision when moving existing 4346 HPT entries to the new HPT, 4347 -EIO on other error conditions 4348 4349Used to implement the PAPR extension for runtime resizing of a guest's 4350Hashed Page Table (HPT). Specifically this requests that the guest be 4351transferred to working with the new HPT, essentially implementing the 4352H_RESIZE_HPT_COMMIT hypercall. 4353 4354:: 4355 4356 struct kvm_ppc_resize_hpt { 4357 __u64 flags; 4358 __u32 shift; 4359 __u32 pad; 4360 }; 4361 4362This should only be called after KVM_PPC_RESIZE_HPT_PREPARE has 4363returned 0 with the same parameters. In other cases 4364KVM_PPC_RESIZE_HPT_COMMIT will return an error (usually -ENXIO or 4365-EBUSY, though others may be possible if the preparation was started, 4366but failed). 4367 4368This will have undefined effects on the guest if it has not already 4369placed itself in a quiescent state where no vcpu will make MMU enabled 4370memory accesses. 4371 4372On succsful completion, the pending HPT will become the guest's active 4373HPT and the previous HPT will be discarded. 4374 4375On failure, the guest will still be operating on its previous HPT. 4376 43774.104 KVM_X86_GET_MCE_CAP_SUPPORTED 4378----------------------------------- 4379 4380:Capability: KVM_CAP_MCE 4381:Architectures: x86 4382:Type: system ioctl 4383:Parameters: u64 mce_cap (out) 4384:Returns: 0 on success, -1 on error 4385 4386Returns supported MCE capabilities. The u64 mce_cap parameter 4387has the same format as the MSR_IA32_MCG_CAP register. Supported 4388capabilities will have the corresponding bits set. 4389 43904.105 KVM_X86_SETUP_MCE 4391----------------------- 4392 4393:Capability: KVM_CAP_MCE 4394:Architectures: x86 4395:Type: vcpu ioctl 4396:Parameters: u64 mcg_cap (in) 4397:Returns: 0 on success, 4398 -EFAULT if u64 mcg_cap cannot be read, 4399 -EINVAL if the requested number of banks is invalid, 4400 -EINVAL if requested MCE capability is not supported. 4401 4402Initializes MCE support for use. The u64 mcg_cap parameter 4403has the same format as the MSR_IA32_MCG_CAP register and 4404specifies which capabilities should be enabled. The maximum 4405supported number of error-reporting banks can be retrieved when 4406checking for KVM_CAP_MCE. The supported capabilities can be 4407retrieved with KVM_X86_GET_MCE_CAP_SUPPORTED. 4408 44094.106 KVM_X86_SET_MCE 4410--------------------- 4411 4412:Capability: KVM_CAP_MCE 4413:Architectures: x86 4414:Type: vcpu ioctl 4415:Parameters: struct kvm_x86_mce (in) 4416:Returns: 0 on success, 4417 -EFAULT if struct kvm_x86_mce cannot be read, 4418 -EINVAL if the bank number is invalid, 4419 -EINVAL if VAL bit is not set in status field. 4420 4421Inject a machine check error (MCE) into the guest. The input 4422parameter is:: 4423 4424 struct kvm_x86_mce { 4425 __u64 status; 4426 __u64 addr; 4427 __u64 misc; 4428 __u64 mcg_status; 4429 __u8 bank; 4430 __u8 pad1[7]; 4431 __u64 pad2[3]; 4432 }; 4433 4434If the MCE being reported is an uncorrected error, KVM will 4435inject it as an MCE exception into the guest. If the guest 4436MCG_STATUS register reports that an MCE is in progress, KVM 4437causes an KVM_EXIT_SHUTDOWN vmexit. 4438 4439Otherwise, if the MCE is a corrected error, KVM will just 4440store it in the corresponding bank (provided this bank is 4441not holding a previously reported uncorrected error). 4442 44434.107 KVM_S390_GET_CMMA_BITS 4444---------------------------- 4445 4446:Capability: KVM_CAP_S390_CMMA_MIGRATION 4447:Architectures: s390 4448:Type: vm ioctl 4449:Parameters: struct kvm_s390_cmma_log (in, out) 4450:Returns: 0 on success, a negative value on error 4451 4452This ioctl is used to get the values of the CMMA bits on the s390 4453architecture. It is meant to be used in two scenarios: 4454 4455- During live migration to save the CMMA values. Live migration needs 4456 to be enabled via the KVM_REQ_START_MIGRATION VM property. 4457- To non-destructively peek at the CMMA values, with the flag 4458 KVM_S390_CMMA_PEEK set. 4459 4460The ioctl takes parameters via the kvm_s390_cmma_log struct. The desired 4461values are written to a buffer whose location is indicated via the "values" 4462member in the kvm_s390_cmma_log struct. The values in the input struct are 4463also updated as needed. 4464 4465Each CMMA value takes up one byte. 4466 4467:: 4468 4469 struct kvm_s390_cmma_log { 4470 __u64 start_gfn; 4471 __u32 count; 4472 __u32 flags; 4473 union { 4474 __u64 remaining; 4475 __u64 mask; 4476 }; 4477 __u64 values; 4478 }; 4479 4480start_gfn is the number of the first guest frame whose CMMA values are 4481to be retrieved, 4482 4483count is the length of the buffer in bytes, 4484 4485values points to the buffer where the result will be written to. 4486 4487If count is greater than KVM_S390_SKEYS_MAX, then it is considered to be 4488KVM_S390_SKEYS_MAX. KVM_S390_SKEYS_MAX is re-used for consistency with 4489other ioctls. 4490 4491The result is written in the buffer pointed to by the field values, and 4492the values of the input parameter are updated as follows. 4493 4494Depending on the flags, different actions are performed. The only 4495supported flag so far is KVM_S390_CMMA_PEEK. 4496 4497The default behaviour if KVM_S390_CMMA_PEEK is not set is: 4498start_gfn will indicate the first page frame whose CMMA bits were dirty. 4499It is not necessarily the same as the one passed as input, as clean pages 4500are skipped. 4501 4502count will indicate the number of bytes actually written in the buffer. 4503It can (and very often will) be smaller than the input value, since the 4504buffer is only filled until 16 bytes of clean values are found (which 4505are then not copied in the buffer). Since a CMMA migration block needs 4506the base address and the length, for a total of 16 bytes, we will send 4507back some clean data if there is some dirty data afterwards, as long as 4508the size of the clean data does not exceed the size of the header. This 4509allows to minimize the amount of data to be saved or transferred over 4510the network at the expense of more roundtrips to userspace. The next 4511invocation of the ioctl will skip over all the clean values, saving 4512potentially more than just the 16 bytes we found. 4513 4514If KVM_S390_CMMA_PEEK is set: 4515the existing storage attributes are read even when not in migration 4516mode, and no other action is performed; 4517 4518the output start_gfn will be equal to the input start_gfn, 4519 4520the output count will be equal to the input count, except if the end of 4521memory has been reached. 4522 4523In both cases: 4524the field "remaining" will indicate the total number of dirty CMMA values 4525still remaining, or 0 if KVM_S390_CMMA_PEEK is set and migration mode is 4526not enabled. 4527 4528mask is unused. 4529 4530values points to the userspace buffer where the result will be stored. 4531 4532This ioctl can fail with -ENOMEM if not enough memory can be allocated to 4533complete the task, with -ENXIO if CMMA is not enabled, with -EINVAL if 4534KVM_S390_CMMA_PEEK is not set but migration mode was not enabled, with 4535-EFAULT if the userspace address is invalid or if no page table is 4536present for the addresses (e.g. when using hugepages). 4537 45384.108 KVM_S390_SET_CMMA_BITS 4539---------------------------- 4540 4541:Capability: KVM_CAP_S390_CMMA_MIGRATION 4542:Architectures: s390 4543:Type: vm ioctl 4544:Parameters: struct kvm_s390_cmma_log (in) 4545:Returns: 0 on success, a negative value on error 4546 4547This ioctl is used to set the values of the CMMA bits on the s390 4548architecture. It is meant to be used during live migration to restore 4549the CMMA values, but there are no restrictions on its use. 4550The ioctl takes parameters via the kvm_s390_cmma_values struct. 4551Each CMMA value takes up one byte. 4552 4553:: 4554 4555 struct kvm_s390_cmma_log { 4556 __u64 start_gfn; 4557 __u32 count; 4558 __u32 flags; 4559 union { 4560 __u64 remaining; 4561 __u64 mask; 4562 }; 4563 __u64 values; 4564 }; 4565 4566start_gfn indicates the starting guest frame number, 4567 4568count indicates how many values are to be considered in the buffer, 4569 4570flags is not used and must be 0. 4571 4572mask indicates which PGSTE bits are to be considered. 4573 4574remaining is not used. 4575 4576values points to the buffer in userspace where to store the values. 4577 4578This ioctl can fail with -ENOMEM if not enough memory can be allocated to 4579complete the task, with -ENXIO if CMMA is not enabled, with -EINVAL if 4580the count field is too large (e.g. more than KVM_S390_CMMA_SIZE_MAX) or 4581if the flags field was not 0, with -EFAULT if the userspace address is 4582invalid, if invalid pages are written to (e.g. after the end of memory) 4583or if no page table is present for the addresses (e.g. when using 4584hugepages). 4585 45864.109 KVM_PPC_GET_CPU_CHAR 4587-------------------------- 4588 4589:Capability: KVM_CAP_PPC_GET_CPU_CHAR 4590:Architectures: powerpc 4591:Type: vm ioctl 4592:Parameters: struct kvm_ppc_cpu_char (out) 4593:Returns: 0 on successful completion, 4594 -EFAULT if struct kvm_ppc_cpu_char cannot be written 4595 4596This ioctl gives userspace information about certain characteristics 4597of the CPU relating to speculative execution of instructions and 4598possible information leakage resulting from speculative execution (see 4599CVE-2017-5715, CVE-2017-5753 and CVE-2017-5754). The information is 4600returned in struct kvm_ppc_cpu_char, which looks like this:: 4601 4602 struct kvm_ppc_cpu_char { 4603 __u64 character; /* characteristics of the CPU */ 4604 __u64 behaviour; /* recommended software behaviour */ 4605 __u64 character_mask; /* valid bits in character */ 4606 __u64 behaviour_mask; /* valid bits in behaviour */ 4607 }; 4608 4609For extensibility, the character_mask and behaviour_mask fields 4610indicate which bits of character and behaviour have been filled in by 4611the kernel. If the set of defined bits is extended in future then 4612userspace will be able to tell whether it is running on a kernel that 4613knows about the new bits. 4614 4615The character field describes attributes of the CPU which can help 4616with preventing inadvertent information disclosure - specifically, 4617whether there is an instruction to flash-invalidate the L1 data cache 4618(ori 30,30,0 or mtspr SPRN_TRIG2,rN), whether the L1 data cache is set 4619to a mode where entries can only be used by the thread that created 4620them, whether the bcctr[l] instruction prevents speculation, and 4621whether a speculation barrier instruction (ori 31,31,0) is provided. 4622 4623The behaviour field describes actions that software should take to 4624prevent inadvertent information disclosure, and thus describes which 4625vulnerabilities the hardware is subject to; specifically whether the 4626L1 data cache should be flushed when returning to user mode from the 4627kernel, and whether a speculation barrier should be placed between an 4628array bounds check and the array access. 4629 4630These fields use the same bit definitions as the new 4631H_GET_CPU_CHARACTERISTICS hypercall. 4632 46334.110 KVM_MEMORY_ENCRYPT_OP 4634--------------------------- 4635 4636:Capability: basic 4637:Architectures: x86 4638:Type: vm 4639:Parameters: an opaque platform specific structure (in/out) 4640:Returns: 0 on success; -1 on error 4641 4642If the platform supports creating encrypted VMs then this ioctl can be used 4643for issuing platform-specific memory encryption commands to manage those 4644encrypted VMs. 4645 4646Currently, this ioctl is used for issuing Secure Encrypted Virtualization 4647(SEV) commands on AMD Processors. The SEV commands are defined in 4648Documentation/virt/kvm/x86/amd-memory-encryption.rst. 4649 46504.111 KVM_MEMORY_ENCRYPT_REG_REGION 4651----------------------------------- 4652 4653:Capability: basic 4654:Architectures: x86 4655:Type: system 4656:Parameters: struct kvm_enc_region (in) 4657:Returns: 0 on success; -1 on error 4658 4659This ioctl can be used to register a guest memory region which may 4660contain encrypted data (e.g. guest RAM, SMRAM etc). 4661 4662It is used in the SEV-enabled guest. When encryption is enabled, a guest 4663memory region may contain encrypted data. The SEV memory encryption 4664engine uses a tweak such that two identical plaintext pages, each at 4665different locations will have differing ciphertexts. So swapping or 4666moving ciphertext of those pages will not result in plaintext being 4667swapped. So relocating (or migrating) physical backing pages for the SEV 4668guest will require some additional steps. 4669 4670Note: The current SEV key management spec does not provide commands to 4671swap or migrate (move) ciphertext pages. Hence, for now we pin the guest 4672memory region registered with the ioctl. 4673 46744.112 KVM_MEMORY_ENCRYPT_UNREG_REGION 4675------------------------------------- 4676 4677:Capability: basic 4678:Architectures: x86 4679:Type: system 4680:Parameters: struct kvm_enc_region (in) 4681:Returns: 0 on success; -1 on error 4682 4683This ioctl can be used to unregister the guest memory region registered 4684with KVM_MEMORY_ENCRYPT_REG_REGION ioctl above. 4685 46864.113 KVM_HYPERV_EVENTFD 4687------------------------ 4688 4689:Capability: KVM_CAP_HYPERV_EVENTFD 4690:Architectures: x86 4691:Type: vm ioctl 4692:Parameters: struct kvm_hyperv_eventfd (in) 4693 4694This ioctl (un)registers an eventfd to receive notifications from the guest on 4695the specified Hyper-V connection id through the SIGNAL_EVENT hypercall, without 4696causing a user exit. SIGNAL_EVENT hypercall with non-zero event flag number 4697(bits 24-31) still triggers a KVM_EXIT_HYPERV_HCALL user exit. 4698 4699:: 4700 4701 struct kvm_hyperv_eventfd { 4702 __u32 conn_id; 4703 __s32 fd; 4704 __u32 flags; 4705 __u32 padding[3]; 4706 }; 4707 4708The conn_id field should fit within 24 bits:: 4709 4710 #define KVM_HYPERV_CONN_ID_MASK 0x00ffffff 4711 4712The acceptable values for the flags field are:: 4713 4714 #define KVM_HYPERV_EVENTFD_DEASSIGN (1 << 0) 4715 4716:Returns: 0 on success, 4717 -EINVAL if conn_id or flags is outside the allowed range, 4718 -ENOENT on deassign if the conn_id isn't registered, 4719 -EEXIST on assign if the conn_id is already registered 4720 47214.114 KVM_GET_NESTED_STATE 4722-------------------------- 4723 4724:Capability: KVM_CAP_NESTED_STATE 4725:Architectures: x86 4726:Type: vcpu ioctl 4727:Parameters: struct kvm_nested_state (in/out) 4728:Returns: 0 on success, -1 on error 4729 4730Errors: 4731 4732 ===== ============================================================= 4733 E2BIG the total state size exceeds the value of 'size' specified by 4734 the user; the size required will be written into size. 4735 ===== ============================================================= 4736 4737:: 4738 4739 struct kvm_nested_state { 4740 __u16 flags; 4741 __u16 format; 4742 __u32 size; 4743 4744 union { 4745 struct kvm_vmx_nested_state_hdr vmx; 4746 struct kvm_svm_nested_state_hdr svm; 4747 4748 /* Pad the header to 128 bytes. */ 4749 __u8 pad[120]; 4750 } hdr; 4751 4752 union { 4753 struct kvm_vmx_nested_state_data vmx[0]; 4754 struct kvm_svm_nested_state_data svm[0]; 4755 } data; 4756 }; 4757 4758 #define KVM_STATE_NESTED_GUEST_MODE 0x00000001 4759 #define KVM_STATE_NESTED_RUN_PENDING 0x00000002 4760 #define KVM_STATE_NESTED_EVMCS 0x00000004 4761 4762 #define KVM_STATE_NESTED_FORMAT_VMX 0 4763 #define KVM_STATE_NESTED_FORMAT_SVM 1 4764 4765 #define KVM_STATE_NESTED_VMX_VMCS_SIZE 0x1000 4766 4767 #define KVM_STATE_NESTED_VMX_SMM_GUEST_MODE 0x00000001 4768 #define KVM_STATE_NESTED_VMX_SMM_VMXON 0x00000002 4769 4770 #define KVM_STATE_VMX_PREEMPTION_TIMER_DEADLINE 0x00000001 4771 4772 struct kvm_vmx_nested_state_hdr { 4773 __u64 vmxon_pa; 4774 __u64 vmcs12_pa; 4775 4776 struct { 4777 __u16 flags; 4778 } smm; 4779 4780 __u32 flags; 4781 __u64 preemption_timer_deadline; 4782 }; 4783 4784 struct kvm_vmx_nested_state_data { 4785 __u8 vmcs12[KVM_STATE_NESTED_VMX_VMCS_SIZE]; 4786 __u8 shadow_vmcs12[KVM_STATE_NESTED_VMX_VMCS_SIZE]; 4787 }; 4788 4789This ioctl copies the vcpu's nested virtualization state from the kernel to 4790userspace. 4791 4792The maximum size of the state can be retrieved by passing KVM_CAP_NESTED_STATE 4793to the KVM_CHECK_EXTENSION ioctl(). 4794 47954.115 KVM_SET_NESTED_STATE 4796-------------------------- 4797 4798:Capability: KVM_CAP_NESTED_STATE 4799:Architectures: x86 4800:Type: vcpu ioctl 4801:Parameters: struct kvm_nested_state (in) 4802:Returns: 0 on success, -1 on error 4803 4804This copies the vcpu's kvm_nested_state struct from userspace to the kernel. 4805For the definition of struct kvm_nested_state, see KVM_GET_NESTED_STATE. 4806 48074.116 KVM_(UN)REGISTER_COALESCED_MMIO 4808------------------------------------- 4809 4810:Capability: KVM_CAP_COALESCED_MMIO (for coalesced mmio) 4811 KVM_CAP_COALESCED_PIO (for coalesced pio) 4812:Architectures: all 4813:Type: vm ioctl 4814:Parameters: struct kvm_coalesced_mmio_zone 4815:Returns: 0 on success, < 0 on error 4816 4817Coalesced I/O is a performance optimization that defers hardware 4818register write emulation so that userspace exits are avoided. It is 4819typically used to reduce the overhead of emulating frequently accessed 4820hardware registers. 4821 4822When a hardware register is configured for coalesced I/O, write accesses 4823do not exit to userspace and their value is recorded in a ring buffer 4824that is shared between kernel and userspace. 4825 4826Coalesced I/O is used if one or more write accesses to a hardware 4827register can be deferred until a read or a write to another hardware 4828register on the same device. This last access will cause a vmexit and 4829userspace will process accesses from the ring buffer before emulating 4830it. That will avoid exiting to userspace on repeated writes. 4831 4832Coalesced pio is based on coalesced mmio. There is little difference 4833between coalesced mmio and pio except that coalesced pio records accesses 4834to I/O ports. 4835 48364.117 KVM_CLEAR_DIRTY_LOG (vm ioctl) 4837------------------------------------ 4838 4839:Capability: KVM_CAP_MANUAL_DIRTY_LOG_PROTECT2 4840:Architectures: x86, arm64, mips 4841:Type: vm ioctl 4842:Parameters: struct kvm_clear_dirty_log (in) 4843:Returns: 0 on success, -1 on error 4844 4845:: 4846 4847 /* for KVM_CLEAR_DIRTY_LOG */ 4848 struct kvm_clear_dirty_log { 4849 __u32 slot; 4850 __u32 num_pages; 4851 __u64 first_page; 4852 union { 4853 void __user *dirty_bitmap; /* one bit per page */ 4854 __u64 padding; 4855 }; 4856 }; 4857 4858The ioctl clears the dirty status of pages in a memory slot, according to 4859the bitmap that is passed in struct kvm_clear_dirty_log's dirty_bitmap 4860field. Bit 0 of the bitmap corresponds to page "first_page" in the 4861memory slot, and num_pages is the size in bits of the input bitmap. 4862first_page must be a multiple of 64; num_pages must also be a multiple of 486364 unless first_page + num_pages is the size of the memory slot. For each 4864bit that is set in the input bitmap, the corresponding page is marked "clean" 4865in KVM's dirty bitmap, and dirty tracking is re-enabled for that page 4866(for example via write-protection, or by clearing the dirty bit in 4867a page table entry). 4868 4869If KVM_CAP_MULTI_ADDRESS_SPACE is available, bits 16-31 of slot field specifies 4870the address space for which you want to clear the dirty status. See 4871KVM_SET_USER_MEMORY_REGION for details on the usage of slot field. 4872 4873This ioctl is mostly useful when KVM_CAP_MANUAL_DIRTY_LOG_PROTECT2 4874is enabled; for more information, see the description of the capability. 4875However, it can always be used as long as KVM_CHECK_EXTENSION confirms 4876that KVM_CAP_MANUAL_DIRTY_LOG_PROTECT2 is present. 4877 48784.118 KVM_GET_SUPPORTED_HV_CPUID 4879-------------------------------- 4880 4881:Capability: KVM_CAP_HYPERV_CPUID (vcpu), KVM_CAP_SYS_HYPERV_CPUID (system) 4882:Architectures: x86 4883:Type: system ioctl, vcpu ioctl 4884:Parameters: struct kvm_cpuid2 (in/out) 4885:Returns: 0 on success, -1 on error 4886 4887:: 4888 4889 struct kvm_cpuid2 { 4890 __u32 nent; 4891 __u32 padding; 4892 struct kvm_cpuid_entry2 entries[0]; 4893 }; 4894 4895 struct kvm_cpuid_entry2 { 4896 __u32 function; 4897 __u32 index; 4898 __u32 flags; 4899 __u32 eax; 4900 __u32 ebx; 4901 __u32 ecx; 4902 __u32 edx; 4903 __u32 padding[3]; 4904 }; 4905 4906This ioctl returns x86 cpuid features leaves related to Hyper-V emulation in 4907KVM. Userspace can use the information returned by this ioctl to construct 4908cpuid information presented to guests consuming Hyper-V enlightenments (e.g. 4909Windows or Hyper-V guests). 4910 4911CPUID feature leaves returned by this ioctl are defined by Hyper-V Top Level 4912Functional Specification (TLFS). These leaves can't be obtained with 4913KVM_GET_SUPPORTED_CPUID ioctl because some of them intersect with KVM feature 4914leaves (0x40000000, 0x40000001). 4915 4916Currently, the following list of CPUID leaves are returned: 4917 4918 - HYPERV_CPUID_VENDOR_AND_MAX_FUNCTIONS 4919 - HYPERV_CPUID_INTERFACE 4920 - HYPERV_CPUID_VERSION 4921 - HYPERV_CPUID_FEATURES 4922 - HYPERV_CPUID_ENLIGHTMENT_INFO 4923 - HYPERV_CPUID_IMPLEMENT_LIMITS 4924 - HYPERV_CPUID_NESTED_FEATURES 4925 - HYPERV_CPUID_SYNDBG_VENDOR_AND_MAX_FUNCTIONS 4926 - HYPERV_CPUID_SYNDBG_INTERFACE 4927 - HYPERV_CPUID_SYNDBG_PLATFORM_CAPABILITIES 4928 4929Userspace invokes KVM_GET_SUPPORTED_HV_CPUID by passing a kvm_cpuid2 structure 4930with the 'nent' field indicating the number of entries in the variable-size 4931array 'entries'. If the number of entries is too low to describe all Hyper-V 4932feature leaves, an error (E2BIG) is returned. If the number is more or equal 4933to the number of Hyper-V feature leaves, the 'nent' field is adjusted to the 4934number of valid entries in the 'entries' array, which is then filled. 4935 4936'index' and 'flags' fields in 'struct kvm_cpuid_entry2' are currently reserved, 4937userspace should not expect to get any particular value there. 4938 4939Note, vcpu version of KVM_GET_SUPPORTED_HV_CPUID is currently deprecated. Unlike 4940system ioctl which exposes all supported feature bits unconditionally, vcpu 4941version has the following quirks: 4942 4943- HYPERV_CPUID_NESTED_FEATURES leaf and HV_X64_ENLIGHTENED_VMCS_RECOMMENDED 4944 feature bit are only exposed when Enlightened VMCS was previously enabled 4945 on the corresponding vCPU (KVM_CAP_HYPERV_ENLIGHTENED_VMCS). 4946- HV_STIMER_DIRECT_MODE_AVAILABLE bit is only exposed with in-kernel LAPIC. 4947 (presumes KVM_CREATE_IRQCHIP has already been called). 4948 49494.119 KVM_ARM_VCPU_FINALIZE 4950--------------------------- 4951 4952:Architectures: arm64 4953:Type: vcpu ioctl 4954:Parameters: int feature (in) 4955:Returns: 0 on success, -1 on error 4956 4957Errors: 4958 4959 ====== ============================================================== 4960 EPERM feature not enabled, needs configuration, or already finalized 4961 EINVAL feature unknown or not present 4962 ====== ============================================================== 4963 4964Recognised values for feature: 4965 4966 ===== =========================================== 4967 arm64 KVM_ARM_VCPU_SVE (requires KVM_CAP_ARM_SVE) 4968 ===== =========================================== 4969 4970Finalizes the configuration of the specified vcpu feature. 4971 4972The vcpu must already have been initialised, enabling the affected feature, by 4973means of a successful KVM_ARM_VCPU_INIT call with the appropriate flag set in 4974features[]. 4975 4976For affected vcpu features, this is a mandatory step that must be performed 4977before the vcpu is fully usable. 4978 4979Between KVM_ARM_VCPU_INIT and KVM_ARM_VCPU_FINALIZE, the feature may be 4980configured by use of ioctls such as KVM_SET_ONE_REG. The exact configuration 4981that should be performaned and how to do it are feature-dependent. 4982 4983Other calls that depend on a particular feature being finalized, such as 4984KVM_RUN, KVM_GET_REG_LIST, KVM_GET_ONE_REG and KVM_SET_ONE_REG, will fail with 4985-EPERM unless the feature has already been finalized by means of a 4986KVM_ARM_VCPU_FINALIZE call. 4987 4988See KVM_ARM_VCPU_INIT for details of vcpu features that require finalization 4989using this ioctl. 4990 49914.120 KVM_SET_PMU_EVENT_FILTER 4992------------------------------ 4993 4994:Capability: KVM_CAP_PMU_EVENT_FILTER 4995:Architectures: x86 4996:Type: vm ioctl 4997:Parameters: struct kvm_pmu_event_filter (in) 4998:Returns: 0 on success, -1 on error 4999 5000:: 5001 5002 struct kvm_pmu_event_filter { 5003 __u32 action; 5004 __u32 nevents; 5005 __u32 fixed_counter_bitmap; 5006 __u32 flags; 5007 __u32 pad[4]; 5008 __u64 events[0]; 5009 }; 5010 5011This ioctl restricts the set of PMU events that the guest can program. 5012The argument holds a list of events which will be allowed or denied. 5013The eventsel+umask of each event the guest attempts to program is compared 5014against the events field to determine whether the guest should have access. 5015The events field only controls general purpose counters; fixed purpose 5016counters are controlled by the fixed_counter_bitmap. 5017 5018No flags are defined yet, the field must be zero. 5019 5020Valid values for 'action':: 5021 5022 #define KVM_PMU_EVENT_ALLOW 0 5023 #define KVM_PMU_EVENT_DENY 1 5024 50254.121 KVM_PPC_SVM_OFF 5026--------------------- 5027 5028:Capability: basic 5029:Architectures: powerpc 5030:Type: vm ioctl 5031:Parameters: none 5032:Returns: 0 on successful completion, 5033 5034Errors: 5035 5036 ====== ================================================================ 5037 EINVAL if ultravisor failed to terminate the secure guest 5038 ENOMEM if hypervisor failed to allocate new radix page tables for guest 5039 ====== ================================================================ 5040 5041This ioctl is used to turn off the secure mode of the guest or transition 5042the guest from secure mode to normal mode. This is invoked when the guest 5043is reset. This has no effect if called for a normal guest. 5044 5045This ioctl issues an ultravisor call to terminate the secure guest, 5046unpins the VPA pages and releases all the device pages that are used to 5047track the secure pages by hypervisor. 5048 50494.122 KVM_S390_NORMAL_RESET 5050--------------------------- 5051 5052:Capability: KVM_CAP_S390_VCPU_RESETS 5053:Architectures: s390 5054:Type: vcpu ioctl 5055:Parameters: none 5056:Returns: 0 5057 5058This ioctl resets VCPU registers and control structures according to 5059the cpu reset definition in the POP (Principles Of Operation). 5060 50614.123 KVM_S390_INITIAL_RESET 5062---------------------------- 5063 5064:Capability: none 5065:Architectures: s390 5066:Type: vcpu ioctl 5067:Parameters: none 5068:Returns: 0 5069 5070This ioctl resets VCPU registers and control structures according to 5071the initial cpu reset definition in the POP. However, the cpu is not 5072put into ESA mode. This reset is a superset of the normal reset. 5073 50744.124 KVM_S390_CLEAR_RESET 5075-------------------------- 5076 5077:Capability: KVM_CAP_S390_VCPU_RESETS 5078:Architectures: s390 5079:Type: vcpu ioctl 5080:Parameters: none 5081:Returns: 0 5082 5083This ioctl resets VCPU registers and control structures according to 5084the clear cpu reset definition in the POP. However, the cpu is not put 5085into ESA mode. This reset is a superset of the initial reset. 5086 5087 50884.125 KVM_S390_PV_COMMAND 5089------------------------- 5090 5091:Capability: KVM_CAP_S390_PROTECTED 5092:Architectures: s390 5093:Type: vm ioctl 5094:Parameters: struct kvm_pv_cmd 5095:Returns: 0 on success, < 0 on error 5096 5097:: 5098 5099 struct kvm_pv_cmd { 5100 __u32 cmd; /* Command to be executed */ 5101 __u16 rc; /* Ultravisor return code */ 5102 __u16 rrc; /* Ultravisor return reason code */ 5103 __u64 data; /* Data or address */ 5104 __u32 flags; /* flags for future extensions. Must be 0 for now */ 5105 __u32 reserved[3]; 5106 }; 5107 5108**Ultravisor return codes** 5109The Ultravisor return (reason) codes are provided by the kernel if a 5110Ultravisor call has been executed to achieve the results expected by 5111the command. Therefore they are independent of the IOCTL return 5112code. If KVM changes `rc`, its value will always be greater than 0 5113hence setting it to 0 before issuing a PV command is advised to be 5114able to detect a change of `rc`. 5115 5116**cmd values:** 5117 5118KVM_PV_ENABLE 5119 Allocate memory and register the VM with the Ultravisor, thereby 5120 donating memory to the Ultravisor that will become inaccessible to 5121 KVM. All existing CPUs are converted to protected ones. After this 5122 command has succeeded, any CPU added via hotplug will become 5123 protected during its creation as well. 5124 5125 Errors: 5126 5127 ===== ============================= 5128 EINTR an unmasked signal is pending 5129 ===== ============================= 5130 5131KVM_PV_DISABLE 5132 Deregister the VM from the Ultravisor and reclaim the memory that had 5133 been donated to the Ultravisor, making it usable by the kernel again. 5134 All registered VCPUs are converted back to non-protected ones. If a 5135 previous protected VM had been prepared for asynchonous teardown with 5136 KVM_PV_ASYNC_CLEANUP_PREPARE and not subsequently torn down with 5137 KVM_PV_ASYNC_CLEANUP_PERFORM, it will be torn down in this call 5138 together with the current protected VM. 5139 5140KVM_PV_VM_SET_SEC_PARMS 5141 Pass the image header from VM memory to the Ultravisor in 5142 preparation of image unpacking and verification. 5143 5144KVM_PV_VM_UNPACK 5145 Unpack (protect and decrypt) a page of the encrypted boot image. 5146 5147KVM_PV_VM_VERIFY 5148 Verify the integrity of the unpacked image. Only if this succeeds, 5149 KVM is allowed to start protected VCPUs. 5150 5151KVM_PV_INFO 5152 :Capability: KVM_CAP_S390_PROTECTED_DUMP 5153 5154 Presents an API that provides Ultravisor related data to userspace 5155 via subcommands. len_max is the size of the user space buffer, 5156 len_written is KVM's indication of how much bytes of that buffer 5157 were actually written to. len_written can be used to determine the 5158 valid fields if more response fields are added in the future. 5159 5160 :: 5161 5162 enum pv_cmd_info_id { 5163 KVM_PV_INFO_VM, 5164 KVM_PV_INFO_DUMP, 5165 }; 5166 5167 struct kvm_s390_pv_info_header { 5168 __u32 id; 5169 __u32 len_max; 5170 __u32 len_written; 5171 __u32 reserved; 5172 }; 5173 5174 struct kvm_s390_pv_info { 5175 struct kvm_s390_pv_info_header header; 5176 struct kvm_s390_pv_info_dump dump; 5177 struct kvm_s390_pv_info_vm vm; 5178 }; 5179 5180**subcommands:** 5181 5182 KVM_PV_INFO_VM 5183 This subcommand provides basic Ultravisor information for PV 5184 hosts. These values are likely also exported as files in the sysfs 5185 firmware UV query interface but they are more easily available to 5186 programs in this API. 5187 5188 The installed calls and feature_indication members provide the 5189 installed UV calls and the UV's other feature indications. 5190 5191 The max_* members provide information about the maximum number of PV 5192 vcpus, PV guests and PV guest memory size. 5193 5194 :: 5195 5196 struct kvm_s390_pv_info_vm { 5197 __u64 inst_calls_list[4]; 5198 __u64 max_cpus; 5199 __u64 max_guests; 5200 __u64 max_guest_addr; 5201 __u64 feature_indication; 5202 }; 5203 5204 5205 KVM_PV_INFO_DUMP 5206 This subcommand provides information related to dumping PV guests. 5207 5208 :: 5209 5210 struct kvm_s390_pv_info_dump { 5211 __u64 dump_cpu_buffer_len; 5212 __u64 dump_config_mem_buffer_per_1m; 5213 __u64 dump_config_finalize_len; 5214 }; 5215 5216KVM_PV_DUMP 5217 :Capability: KVM_CAP_S390_PROTECTED_DUMP 5218 5219 Presents an API that provides calls which facilitate dumping a 5220 protected VM. 5221 5222 :: 5223 5224 struct kvm_s390_pv_dmp { 5225 __u64 subcmd; 5226 __u64 buff_addr; 5227 __u64 buff_len; 5228 __u64 gaddr; /* For dump storage state */ 5229 }; 5230 5231 **subcommands:** 5232 5233 KVM_PV_DUMP_INIT 5234 Initializes the dump process of a protected VM. If this call does 5235 not succeed all other subcommands will fail with -EINVAL. This 5236 subcommand will return -EINVAL if a dump process has not yet been 5237 completed. 5238 5239 Not all PV vms can be dumped, the owner needs to set `dump 5240 allowed` PCF bit 34 in the SE header to allow dumping. 5241 5242 KVM_PV_DUMP_CONFIG_STOR_STATE 5243 Stores `buff_len` bytes of tweak component values starting with 5244 the 1MB block specified by the absolute guest address 5245 (`gaddr`). `buff_len` needs to be `conf_dump_storage_state_len` 5246 aligned and at least >= the `conf_dump_storage_state_len` value 5247 provided by the dump uv_info data. buff_user might be written to 5248 even if an error rc is returned. For instance if we encounter a 5249 fault after writing the first page of data. 5250 5251 KVM_PV_DUMP_COMPLETE 5252 If the subcommand succeeds it completes the dump process and lets 5253 KVM_PV_DUMP_INIT be called again. 5254 5255 On success `conf_dump_finalize_len` bytes of completion data will be 5256 stored to the `buff_addr`. The completion data contains a key 5257 derivation seed, IV, tweak nonce and encryption keys as well as an 5258 authentication tag all of which are needed to decrypt the dump at a 5259 later time. 5260 5261KVM_PV_ASYNC_CLEANUP_PREPARE 5262 :Capability: KVM_CAP_S390_PROTECTED_ASYNC_DISABLE 5263 5264 Prepare the current protected VM for asynchronous teardown. Most 5265 resources used by the current protected VM will be set aside for a 5266 subsequent asynchronous teardown. The current protected VM will then 5267 resume execution immediately as non-protected. There can be at most 5268 one protected VM prepared for asynchronous teardown at any time. If 5269 a protected VM had already been prepared for teardown without 5270 subsequently calling KVM_PV_ASYNC_CLEANUP_PERFORM, this call will 5271 fail. In that case, the userspace process should issue a normal 5272 KVM_PV_DISABLE. The resources set aside with this call will need to 5273 be cleaned up with a subsequent call to KVM_PV_ASYNC_CLEANUP_PERFORM 5274 or KVM_PV_DISABLE, otherwise they will be cleaned up when KVM 5275 terminates. KVM_PV_ASYNC_CLEANUP_PREPARE can be called again as soon 5276 as cleanup starts, i.e. before KVM_PV_ASYNC_CLEANUP_PERFORM finishes. 5277 5278KVM_PV_ASYNC_CLEANUP_PERFORM 5279 :Capability: KVM_CAP_S390_PROTECTED_ASYNC_DISABLE 5280 5281 Tear down the protected VM previously prepared for teardown with 5282 KVM_PV_ASYNC_CLEANUP_PREPARE. The resources that had been set aside 5283 will be freed during the execution of this command. This PV command 5284 should ideally be issued by userspace from a separate thread. If a 5285 fatal signal is received (or the process terminates naturally), the 5286 command will terminate immediately without completing, and the normal 5287 KVM shutdown procedure will take care of cleaning up all remaining 5288 protected VMs, including the ones whose teardown was interrupted by 5289 process termination. 5290 52914.126 KVM_XEN_HVM_SET_ATTR 5292-------------------------- 5293 5294:Capability: KVM_CAP_XEN_HVM / KVM_XEN_HVM_CONFIG_SHARED_INFO 5295:Architectures: x86 5296:Type: vm ioctl 5297:Parameters: struct kvm_xen_hvm_attr 5298:Returns: 0 on success, < 0 on error 5299 5300:: 5301 5302 struct kvm_xen_hvm_attr { 5303 __u16 type; 5304 __u16 pad[3]; 5305 union { 5306 __u8 long_mode; 5307 __u8 vector; 5308 __u8 runstate_update_flag; 5309 struct { 5310 __u64 gfn; 5311 } shared_info; 5312 struct { 5313 __u32 send_port; 5314 __u32 type; /* EVTCHNSTAT_ipi / EVTCHNSTAT_interdomain */ 5315 __u32 flags; 5316 union { 5317 struct { 5318 __u32 port; 5319 __u32 vcpu; 5320 __u32 priority; 5321 } port; 5322 struct { 5323 __u32 port; /* Zero for eventfd */ 5324 __s32 fd; 5325 } eventfd; 5326 __u32 padding[4]; 5327 } deliver; 5328 } evtchn; 5329 __u32 xen_version; 5330 __u64 pad[8]; 5331 } u; 5332 }; 5333 5334type values: 5335 5336KVM_XEN_ATTR_TYPE_LONG_MODE 5337 Sets the ABI mode of the VM to 32-bit or 64-bit (long mode). This 5338 determines the layout of the shared info pages exposed to the VM. 5339 5340KVM_XEN_ATTR_TYPE_SHARED_INFO 5341 Sets the guest physical frame number at which the Xen "shared info" 5342 page resides. Note that although Xen places vcpu_info for the first 5343 32 vCPUs in the shared_info page, KVM does not automatically do so 5344 and instead requires that KVM_XEN_VCPU_ATTR_TYPE_VCPU_INFO be used 5345 explicitly even when the vcpu_info for a given vCPU resides at the 5346 "default" location in the shared_info page. This is because KVM may 5347 not be aware of the Xen CPU id which is used as the index into the 5348 vcpu_info[] array, so may know the correct default location. 5349 5350 Note that the shared info page may be constantly written to by KVM; 5351 it contains the event channel bitmap used to deliver interrupts to 5352 a Xen guest, amongst other things. It is exempt from dirty tracking 5353 mechanisms — KVM will not explicitly mark the page as dirty each 5354 time an event channel interrupt is delivered to the guest! Thus, 5355 userspace should always assume that the designated GFN is dirty if 5356 any vCPU has been running or any event channel interrupts can be 5357 routed to the guest. 5358 5359 Setting the gfn to KVM_XEN_INVALID_GFN will disable the shared info 5360 page. 5361 5362KVM_XEN_ATTR_TYPE_UPCALL_VECTOR 5363 Sets the exception vector used to deliver Xen event channel upcalls. 5364 This is the HVM-wide vector injected directly by the hypervisor 5365 (not through the local APIC), typically configured by a guest via 5366 HVM_PARAM_CALLBACK_IRQ. This can be disabled again (e.g. for guest 5367 SHUTDOWN_soft_reset) by setting it to zero. 5368 5369KVM_XEN_ATTR_TYPE_EVTCHN 5370 This attribute is available when the KVM_CAP_XEN_HVM ioctl indicates 5371 support for KVM_XEN_HVM_CONFIG_EVTCHN_SEND features. It configures 5372 an outbound port number for interception of EVTCHNOP_send requests 5373 from the guest. A given sending port number may be directed back to 5374 a specified vCPU (by APIC ID) / port / priority on the guest, or to 5375 trigger events on an eventfd. The vCPU and priority can be changed 5376 by setting KVM_XEN_EVTCHN_UPDATE in a subsequent call, but but other 5377 fields cannot change for a given sending port. A port mapping is 5378 removed by using KVM_XEN_EVTCHN_DEASSIGN in the flags field. Passing 5379 KVM_XEN_EVTCHN_RESET in the flags field removes all interception of 5380 outbound event channels. The values of the flags field are mutually 5381 exclusive and cannot be combined as a bitmask. 5382 5383KVM_XEN_ATTR_TYPE_XEN_VERSION 5384 This attribute is available when the KVM_CAP_XEN_HVM ioctl indicates 5385 support for KVM_XEN_HVM_CONFIG_EVTCHN_SEND features. It configures 5386 the 32-bit version code returned to the guest when it invokes the 5387 XENVER_version call; typically (XEN_MAJOR << 16 | XEN_MINOR). PV 5388 Xen guests will often use this to as a dummy hypercall to trigger 5389 event channel delivery, so responding within the kernel without 5390 exiting to userspace is beneficial. 5391 5392KVM_XEN_ATTR_TYPE_RUNSTATE_UPDATE_FLAG 5393 This attribute is available when the KVM_CAP_XEN_HVM ioctl indicates 5394 support for KVM_XEN_HVM_CONFIG_RUNSTATE_UPDATE_FLAG. It enables the 5395 XEN_RUNSTATE_UPDATE flag which allows guest vCPUs to safely read 5396 other vCPUs' vcpu_runstate_info. Xen guests enable this feature via 5397 the VMASST_TYPE_runstate_update_flag of the HYPERVISOR_vm_assist 5398 hypercall. 5399 54004.127 KVM_XEN_HVM_GET_ATTR 5401-------------------------- 5402 5403:Capability: KVM_CAP_XEN_HVM / KVM_XEN_HVM_CONFIG_SHARED_INFO 5404:Architectures: x86 5405:Type: vm ioctl 5406:Parameters: struct kvm_xen_hvm_attr 5407:Returns: 0 on success, < 0 on error 5408 5409Allows Xen VM attributes to be read. For the structure and types, 5410see KVM_XEN_HVM_SET_ATTR above. The KVM_XEN_ATTR_TYPE_EVTCHN 5411attribute cannot be read. 5412 54134.128 KVM_XEN_VCPU_SET_ATTR 5414--------------------------- 5415 5416:Capability: KVM_CAP_XEN_HVM / KVM_XEN_HVM_CONFIG_SHARED_INFO 5417:Architectures: x86 5418:Type: vcpu ioctl 5419:Parameters: struct kvm_xen_vcpu_attr 5420:Returns: 0 on success, < 0 on error 5421 5422:: 5423 5424 struct kvm_xen_vcpu_attr { 5425 __u16 type; 5426 __u16 pad[3]; 5427 union { 5428 __u64 gpa; 5429 __u64 pad[4]; 5430 struct { 5431 __u64 state; 5432 __u64 state_entry_time; 5433 __u64 time_running; 5434 __u64 time_runnable; 5435 __u64 time_blocked; 5436 __u64 time_offline; 5437 } runstate; 5438 __u32 vcpu_id; 5439 struct { 5440 __u32 port; 5441 __u32 priority; 5442 __u64 expires_ns; 5443 } timer; 5444 __u8 vector; 5445 } u; 5446 }; 5447 5448type values: 5449 5450KVM_XEN_VCPU_ATTR_TYPE_VCPU_INFO 5451 Sets the guest physical address of the vcpu_info for a given vCPU. 5452 As with the shared_info page for the VM, the corresponding page may be 5453 dirtied at any time if event channel interrupt delivery is enabled, so 5454 userspace should always assume that the page is dirty without relying 5455 on dirty logging. Setting the gpa to KVM_XEN_INVALID_GPA will disable 5456 the vcpu_info. 5457 5458KVM_XEN_VCPU_ATTR_TYPE_VCPU_TIME_INFO 5459 Sets the guest physical address of an additional pvclock structure 5460 for a given vCPU. This is typically used for guest vsyscall support. 5461 Setting the gpa to KVM_XEN_INVALID_GPA will disable the structure. 5462 5463KVM_XEN_VCPU_ATTR_TYPE_RUNSTATE_ADDR 5464 Sets the guest physical address of the vcpu_runstate_info for a given 5465 vCPU. This is how a Xen guest tracks CPU state such as steal time. 5466 Setting the gpa to KVM_XEN_INVALID_GPA will disable the runstate area. 5467 5468KVM_XEN_VCPU_ATTR_TYPE_RUNSTATE_CURRENT 5469 Sets the runstate (RUNSTATE_running/_runnable/_blocked/_offline) of 5470 the given vCPU from the .u.runstate.state member of the structure. 5471 KVM automatically accounts running and runnable time but blocked 5472 and offline states are only entered explicitly. 5473 5474KVM_XEN_VCPU_ATTR_TYPE_RUNSTATE_DATA 5475 Sets all fields of the vCPU runstate data from the .u.runstate member 5476 of the structure, including the current runstate. The state_entry_time 5477 must equal the sum of the other four times. 5478 5479KVM_XEN_VCPU_ATTR_TYPE_RUNSTATE_ADJUST 5480 This *adds* the contents of the .u.runstate members of the structure 5481 to the corresponding members of the given vCPU's runstate data, thus 5482 permitting atomic adjustments to the runstate times. The adjustment 5483 to the state_entry_time must equal the sum of the adjustments to the 5484 other four times. The state field must be set to -1, or to a valid 5485 runstate value (RUNSTATE_running, RUNSTATE_runnable, RUNSTATE_blocked 5486 or RUNSTATE_offline) to set the current accounted state as of the 5487 adjusted state_entry_time. 5488 5489KVM_XEN_VCPU_ATTR_TYPE_VCPU_ID 5490 This attribute is available when the KVM_CAP_XEN_HVM ioctl indicates 5491 support for KVM_XEN_HVM_CONFIG_EVTCHN_SEND features. It sets the Xen 5492 vCPU ID of the given vCPU, to allow timer-related VCPU operations to 5493 be intercepted by KVM. 5494 5495KVM_XEN_VCPU_ATTR_TYPE_TIMER 5496 This attribute is available when the KVM_CAP_XEN_HVM ioctl indicates 5497 support for KVM_XEN_HVM_CONFIG_EVTCHN_SEND features. It sets the 5498 event channel port/priority for the VIRQ_TIMER of the vCPU, as well 5499 as allowing a pending timer to be saved/restored. Setting the timer 5500 port to zero disables kernel handling of the singleshot timer. 5501 5502KVM_XEN_VCPU_ATTR_TYPE_UPCALL_VECTOR 5503 This attribute is available when the KVM_CAP_XEN_HVM ioctl indicates 5504 support for KVM_XEN_HVM_CONFIG_EVTCHN_SEND features. It sets the 5505 per-vCPU local APIC upcall vector, configured by a Xen guest with 5506 the HVMOP_set_evtchn_upcall_vector hypercall. This is typically 5507 used by Windows guests, and is distinct from the HVM-wide upcall 5508 vector configured with HVM_PARAM_CALLBACK_IRQ. It is disabled by 5509 setting the vector to zero. 5510 5511 55124.129 KVM_XEN_VCPU_GET_ATTR 5513--------------------------- 5514 5515:Capability: KVM_CAP_XEN_HVM / KVM_XEN_HVM_CONFIG_SHARED_INFO 5516:Architectures: x86 5517:Type: vcpu ioctl 5518:Parameters: struct kvm_xen_vcpu_attr 5519:Returns: 0 on success, < 0 on error 5520 5521Allows Xen vCPU attributes to be read. For the structure and types, 5522see KVM_XEN_VCPU_SET_ATTR above. 5523 5524The KVM_XEN_VCPU_ATTR_TYPE_RUNSTATE_ADJUST type may not be used 5525with the KVM_XEN_VCPU_GET_ATTR ioctl. 5526 55274.130 KVM_ARM_MTE_COPY_TAGS 5528--------------------------- 5529 5530:Capability: KVM_CAP_ARM_MTE 5531:Architectures: arm64 5532:Type: vm ioctl 5533:Parameters: struct kvm_arm_copy_mte_tags 5534:Returns: number of bytes copied, < 0 on error (-EINVAL for incorrect 5535 arguments, -EFAULT if memory cannot be accessed). 5536 5537:: 5538 5539 struct kvm_arm_copy_mte_tags { 5540 __u64 guest_ipa; 5541 __u64 length; 5542 void __user *addr; 5543 __u64 flags; 5544 __u64 reserved[2]; 5545 }; 5546 5547Copies Memory Tagging Extension (MTE) tags to/from guest tag memory. The 5548``guest_ipa`` and ``length`` fields must be ``PAGE_SIZE`` aligned. The ``addr`` 5549field must point to a buffer which the tags will be copied to or from. 5550 5551``flags`` specifies the direction of copy, either ``KVM_ARM_TAGS_TO_GUEST`` or 5552``KVM_ARM_TAGS_FROM_GUEST``. 5553 5554The size of the buffer to store the tags is ``(length / 16)`` bytes 5555(granules in MTE are 16 bytes long). Each byte contains a single tag 5556value. This matches the format of ``PTRACE_PEEKMTETAGS`` and 5557``PTRACE_POKEMTETAGS``. 5558 5559If an error occurs before any data is copied then a negative error code is 5560returned. If some tags have been copied before an error occurs then the number 5561of bytes successfully copied is returned. If the call completes successfully 5562then ``length`` is returned. 5563 55644.131 KVM_GET_SREGS2 5565-------------------- 5566 5567:Capability: KVM_CAP_SREGS2 5568:Architectures: x86 5569:Type: vcpu ioctl 5570:Parameters: struct kvm_sregs2 (out) 5571:Returns: 0 on success, -1 on error 5572 5573Reads special registers from the vcpu. 5574This ioctl (when supported) replaces the KVM_GET_SREGS. 5575 5576:: 5577 5578 struct kvm_sregs2 { 5579 /* out (KVM_GET_SREGS2) / in (KVM_SET_SREGS2) */ 5580 struct kvm_segment cs, ds, es, fs, gs, ss; 5581 struct kvm_segment tr, ldt; 5582 struct kvm_dtable gdt, idt; 5583 __u64 cr0, cr2, cr3, cr4, cr8; 5584 __u64 efer; 5585 __u64 apic_base; 5586 __u64 flags; 5587 __u64 pdptrs[4]; 5588 }; 5589 5590flags values for ``kvm_sregs2``: 5591 5592``KVM_SREGS2_FLAGS_PDPTRS_VALID`` 5593 5594 Indicates thats the struct contain valid PDPTR values. 5595 5596 55974.132 KVM_SET_SREGS2 5598-------------------- 5599 5600:Capability: KVM_CAP_SREGS2 5601:Architectures: x86 5602:Type: vcpu ioctl 5603:Parameters: struct kvm_sregs2 (in) 5604:Returns: 0 on success, -1 on error 5605 5606Writes special registers into the vcpu. 5607See KVM_GET_SREGS2 for the data structures. 5608This ioctl (when supported) replaces the KVM_SET_SREGS. 5609 56104.133 KVM_GET_STATS_FD 5611---------------------- 5612 5613:Capability: KVM_CAP_STATS_BINARY_FD 5614:Architectures: all 5615:Type: vm ioctl, vcpu ioctl 5616:Parameters: none 5617:Returns: statistics file descriptor on success, < 0 on error 5618 5619Errors: 5620 5621 ====== ====================================================== 5622 ENOMEM if the fd could not be created due to lack of memory 5623 EMFILE if the number of opened files exceeds the limit 5624 ====== ====================================================== 5625 5626The returned file descriptor can be used to read VM/vCPU statistics data in 5627binary format. The data in the file descriptor consists of four blocks 5628organized as follows: 5629 5630+-------------+ 5631| Header | 5632+-------------+ 5633| id string | 5634+-------------+ 5635| Descriptors | 5636+-------------+ 5637| Stats Data | 5638+-------------+ 5639 5640Apart from the header starting at offset 0, please be aware that it is 5641not guaranteed that the four blocks are adjacent or in the above order; 5642the offsets of the id, descriptors and data blocks are found in the 5643header. However, all four blocks are aligned to 64 bit offsets in the 5644file and they do not overlap. 5645 5646All blocks except the data block are immutable. Userspace can read them 5647only one time after retrieving the file descriptor, and then use ``pread`` or 5648``lseek`` to read the statistics repeatedly. 5649 5650All data is in system endianness. 5651 5652The format of the header is as follows:: 5653 5654 struct kvm_stats_header { 5655 __u32 flags; 5656 __u32 name_size; 5657 __u32 num_desc; 5658 __u32 id_offset; 5659 __u32 desc_offset; 5660 __u32 data_offset; 5661 }; 5662 5663The ``flags`` field is not used at the moment. It is always read as 0. 5664 5665The ``name_size`` field is the size (in byte) of the statistics name string 5666(including trailing '\0') which is contained in the "id string" block and 5667appended at the end of every descriptor. 5668 5669The ``num_desc`` field is the number of descriptors that are included in the 5670descriptor block. (The actual number of values in the data block may be 5671larger, since each descriptor may comprise more than one value). 5672 5673The ``id_offset`` field is the offset of the id string from the start of the 5674file indicated by the file descriptor. It is a multiple of 8. 5675 5676The ``desc_offset`` field is the offset of the Descriptors block from the start 5677of the file indicated by the file descriptor. It is a multiple of 8. 5678 5679The ``data_offset`` field is the offset of the Stats Data block from the start 5680of the file indicated by the file descriptor. It is a multiple of 8. 5681 5682The id string block contains a string which identifies the file descriptor on 5683which KVM_GET_STATS_FD was invoked. The size of the block, including the 5684trailing ``'\0'``, is indicated by the ``name_size`` field in the header. 5685 5686The descriptors block is only needed to be read once for the lifetime of the 5687file descriptor contains a sequence of ``struct kvm_stats_desc``, each followed 5688by a string of size ``name_size``. 5689:: 5690 5691 #define KVM_STATS_TYPE_SHIFT 0 5692 #define KVM_STATS_TYPE_MASK (0xF << KVM_STATS_TYPE_SHIFT) 5693 #define KVM_STATS_TYPE_CUMULATIVE (0x0 << KVM_STATS_TYPE_SHIFT) 5694 #define KVM_STATS_TYPE_INSTANT (0x1 << KVM_STATS_TYPE_SHIFT) 5695 #define KVM_STATS_TYPE_PEAK (0x2 << KVM_STATS_TYPE_SHIFT) 5696 #define KVM_STATS_TYPE_LINEAR_HIST (0x3 << KVM_STATS_TYPE_SHIFT) 5697 #define KVM_STATS_TYPE_LOG_HIST (0x4 << KVM_STATS_TYPE_SHIFT) 5698 #define KVM_STATS_TYPE_MAX KVM_STATS_TYPE_LOG_HIST 5699 5700 #define KVM_STATS_UNIT_SHIFT 4 5701 #define KVM_STATS_UNIT_MASK (0xF << KVM_STATS_UNIT_SHIFT) 5702 #define KVM_STATS_UNIT_NONE (0x0 << KVM_STATS_UNIT_SHIFT) 5703 #define KVM_STATS_UNIT_BYTES (0x1 << KVM_STATS_UNIT_SHIFT) 5704 #define KVM_STATS_UNIT_SECONDS (0x2 << KVM_STATS_UNIT_SHIFT) 5705 #define KVM_STATS_UNIT_CYCLES (0x3 << KVM_STATS_UNIT_SHIFT) 5706 #define KVM_STATS_UNIT_BOOLEAN (0x4 << KVM_STATS_UNIT_SHIFT) 5707 #define KVM_STATS_UNIT_MAX KVM_STATS_UNIT_BOOLEAN 5708 5709 #define KVM_STATS_BASE_SHIFT 8 5710 #define KVM_STATS_BASE_MASK (0xF << KVM_STATS_BASE_SHIFT) 5711 #define KVM_STATS_BASE_POW10 (0x0 << KVM_STATS_BASE_SHIFT) 5712 #define KVM_STATS_BASE_POW2 (0x1 << KVM_STATS_BASE_SHIFT) 5713 #define KVM_STATS_BASE_MAX KVM_STATS_BASE_POW2 5714 5715 struct kvm_stats_desc { 5716 __u32 flags; 5717 __s16 exponent; 5718 __u16 size; 5719 __u32 offset; 5720 __u32 bucket_size; 5721 char name[]; 5722 }; 5723 5724The ``flags`` field contains the type and unit of the statistics data described 5725by this descriptor. Its endianness is CPU native. 5726The following flags are supported: 5727 5728Bits 0-3 of ``flags`` encode the type: 5729 5730 * ``KVM_STATS_TYPE_CUMULATIVE`` 5731 The statistics reports a cumulative count. The value of data can only be increased. 5732 Most of the counters used in KVM are of this type. 5733 The corresponding ``size`` field for this type is always 1. 5734 All cumulative statistics data are read/write. 5735 * ``KVM_STATS_TYPE_INSTANT`` 5736 The statistics reports an instantaneous value. Its value can be increased or 5737 decreased. This type is usually used as a measurement of some resources, 5738 like the number of dirty pages, the number of large pages, etc. 5739 All instant statistics are read only. 5740 The corresponding ``size`` field for this type is always 1. 5741 * ``KVM_STATS_TYPE_PEAK`` 5742 The statistics data reports a peak value, for example the maximum number 5743 of items in a hash table bucket, the longest time waited and so on. 5744 The value of data can only be increased. 5745 The corresponding ``size`` field for this type is always 1. 5746 * ``KVM_STATS_TYPE_LINEAR_HIST`` 5747 The statistic is reported as a linear histogram. The number of 5748 buckets is specified by the ``size`` field. The size of buckets is specified 5749 by the ``hist_param`` field. The range of the Nth bucket (1 <= N < ``size``) 5750 is [``hist_param``*(N-1), ``hist_param``*N), while the range of the last 5751 bucket is [``hist_param``*(``size``-1), +INF). (+INF means positive infinity 5752 value.) 5753 * ``KVM_STATS_TYPE_LOG_HIST`` 5754 The statistic is reported as a logarithmic histogram. The number of 5755 buckets is specified by the ``size`` field. The range of the first bucket is 5756 [0, 1), while the range of the last bucket is [pow(2, ``size``-2), +INF). 5757 Otherwise, The Nth bucket (1 < N < ``size``) covers 5758 [pow(2, N-2), pow(2, N-1)). 5759 5760Bits 4-7 of ``flags`` encode the unit: 5761 5762 * ``KVM_STATS_UNIT_NONE`` 5763 There is no unit for the value of statistics data. This usually means that 5764 the value is a simple counter of an event. 5765 * ``KVM_STATS_UNIT_BYTES`` 5766 It indicates that the statistics data is used to measure memory size, in the 5767 unit of Byte, KiByte, MiByte, GiByte, etc. The unit of the data is 5768 determined by the ``exponent`` field in the descriptor. 5769 * ``KVM_STATS_UNIT_SECONDS`` 5770 It indicates that the statistics data is used to measure time or latency. 5771 * ``KVM_STATS_UNIT_CYCLES`` 5772 It indicates that the statistics data is used to measure CPU clock cycles. 5773 * ``KVM_STATS_UNIT_BOOLEAN`` 5774 It indicates that the statistic will always be either 0 or 1. Boolean 5775 statistics of "peak" type will never go back from 1 to 0. Boolean 5776 statistics can be linear histograms (with two buckets) but not logarithmic 5777 histograms. 5778 5779Note that, in the case of histograms, the unit applies to the bucket 5780ranges, while the bucket value indicates how many samples fell in the 5781bucket's range. 5782 5783Bits 8-11 of ``flags``, together with ``exponent``, encode the scale of the 5784unit: 5785 5786 * ``KVM_STATS_BASE_POW10`` 5787 The scale is based on power of 10. It is used for measurement of time and 5788 CPU clock cycles. For example, an exponent of -9 can be used with 5789 ``KVM_STATS_UNIT_SECONDS`` to express that the unit is nanoseconds. 5790 * ``KVM_STATS_BASE_POW2`` 5791 The scale is based on power of 2. It is used for measurement of memory size. 5792 For example, an exponent of 20 can be used with ``KVM_STATS_UNIT_BYTES`` to 5793 express that the unit is MiB. 5794 5795The ``size`` field is the number of values of this statistics data. Its 5796value is usually 1 for most of simple statistics. 1 means it contains an 5797unsigned 64bit data. 5798 5799The ``offset`` field is the offset from the start of Data Block to the start of 5800the corresponding statistics data. 5801 5802The ``bucket_size`` field is used as a parameter for histogram statistics data. 5803It is only used by linear histogram statistics data, specifying the size of a 5804bucket in the unit expressed by bits 4-11 of ``flags`` together with ``exponent``. 5805 5806The ``name`` field is the name string of the statistics data. The name string 5807starts at the end of ``struct kvm_stats_desc``. The maximum length including 5808the trailing ``'\0'``, is indicated by ``name_size`` in the header. 5809 5810The Stats Data block contains an array of 64-bit values in the same order 5811as the descriptors in Descriptors block. 5812 58134.134 KVM_GET_XSAVE2 5814-------------------- 5815 5816:Capability: KVM_CAP_XSAVE2 5817:Architectures: x86 5818:Type: vcpu ioctl 5819:Parameters: struct kvm_xsave (out) 5820:Returns: 0 on success, -1 on error 5821 5822 5823:: 5824 5825 struct kvm_xsave { 5826 __u32 region[1024]; 5827 __u32 extra[0]; 5828 }; 5829 5830This ioctl would copy current vcpu's xsave struct to the userspace. It 5831copies as many bytes as are returned by KVM_CHECK_EXTENSION(KVM_CAP_XSAVE2) 5832when invoked on the vm file descriptor. The size value returned by 5833KVM_CHECK_EXTENSION(KVM_CAP_XSAVE2) will always be at least 4096. 5834Currently, it is only greater than 4096 if a dynamic feature has been 5835enabled with ``arch_prctl()``, but this may change in the future. 5836 5837The offsets of the state save areas in struct kvm_xsave follow the contents 5838of CPUID leaf 0xD on the host. 5839 58404.135 KVM_XEN_HVM_EVTCHN_SEND 5841----------------------------- 5842 5843:Capability: KVM_CAP_XEN_HVM / KVM_XEN_HVM_CONFIG_EVTCHN_SEND 5844:Architectures: x86 5845:Type: vm ioctl 5846:Parameters: struct kvm_irq_routing_xen_evtchn 5847:Returns: 0 on success, < 0 on error 5848 5849 5850:: 5851 5852 struct kvm_irq_routing_xen_evtchn { 5853 __u32 port; 5854 __u32 vcpu; 5855 __u32 priority; 5856 }; 5857 5858This ioctl injects an event channel interrupt directly to the guest vCPU. 5859 58604.136 KVM_S390_PV_CPU_COMMAND 5861----------------------------- 5862 5863:Capability: KVM_CAP_S390_PROTECTED_DUMP 5864:Architectures: s390 5865:Type: vcpu ioctl 5866:Parameters: none 5867:Returns: 0 on success, < 0 on error 5868 5869This ioctl closely mirrors `KVM_S390_PV_COMMAND` but handles requests 5870for vcpus. It re-uses the kvm_s390_pv_dmp struct and hence also shares 5871the command ids. 5872 5873**command:** 5874 5875KVM_PV_DUMP 5876 Presents an API that provides calls which facilitate dumping a vcpu 5877 of a protected VM. 5878 5879**subcommand:** 5880 5881KVM_PV_DUMP_CPU 5882 Provides encrypted dump data like register values. 5883 The length of the returned data is provided by uv_info.guest_cpu_stor_len. 5884 58854.137 KVM_S390_ZPCI_OP 5886---------------------- 5887 5888:Capability: KVM_CAP_S390_ZPCI_OP 5889:Architectures: s390 5890:Type: vm ioctl 5891:Parameters: struct kvm_s390_zpci_op (in) 5892:Returns: 0 on success, <0 on error 5893 5894Used to manage hardware-assisted virtualization features for zPCI devices. 5895 5896Parameters are specified via the following structure:: 5897 5898 struct kvm_s390_zpci_op { 5899 /* in */ 5900 __u32 fh; /* target device */ 5901 __u8 op; /* operation to perform */ 5902 __u8 pad[3]; 5903 union { 5904 /* for KVM_S390_ZPCIOP_REG_AEN */ 5905 struct { 5906 __u64 ibv; /* Guest addr of interrupt bit vector */ 5907 __u64 sb; /* Guest addr of summary bit */ 5908 __u32 flags; 5909 __u32 noi; /* Number of interrupts */ 5910 __u8 isc; /* Guest interrupt subclass */ 5911 __u8 sbo; /* Offset of guest summary bit vector */ 5912 __u16 pad; 5913 } reg_aen; 5914 __u64 reserved[8]; 5915 } u; 5916 }; 5917 5918The type of operation is specified in the "op" field. 5919KVM_S390_ZPCIOP_REG_AEN is used to register the VM for adapter event 5920notification interpretation, which will allow firmware delivery of adapter 5921events directly to the vm, with KVM providing a backup delivery mechanism; 5922KVM_S390_ZPCIOP_DEREG_AEN is used to subsequently disable interpretation of 5923adapter event notifications. 5924 5925The target zPCI function must also be specified via the "fh" field. For the 5926KVM_S390_ZPCIOP_REG_AEN operation, additional information to establish firmware 5927delivery must be provided via the "reg_aen" struct. 5928 5929The "pad" and "reserved" fields may be used for future extensions and should be 5930set to 0s by userspace. 5931 59325. The kvm_run structure 5933======================== 5934 5935Application code obtains a pointer to the kvm_run structure by 5936mmap()ing a vcpu fd. From that point, application code can control 5937execution by changing fields in kvm_run prior to calling the KVM_RUN 5938ioctl, and obtain information about the reason KVM_RUN returned by 5939looking up structure members. 5940 5941:: 5942 5943 struct kvm_run { 5944 /* in */ 5945 __u8 request_interrupt_window; 5946 5947Request that KVM_RUN return when it becomes possible to inject external 5948interrupts into the guest. Useful in conjunction with KVM_INTERRUPT. 5949 5950:: 5951 5952 __u8 immediate_exit; 5953 5954This field is polled once when KVM_RUN starts; if non-zero, KVM_RUN 5955exits immediately, returning -EINTR. In the common scenario where a 5956signal is used to "kick" a VCPU out of KVM_RUN, this field can be used 5957to avoid usage of KVM_SET_SIGNAL_MASK, which has worse scalability. 5958Rather than blocking the signal outside KVM_RUN, userspace can set up 5959a signal handler that sets run->immediate_exit to a non-zero value. 5960 5961This field is ignored if KVM_CAP_IMMEDIATE_EXIT is not available. 5962 5963:: 5964 5965 __u8 padding1[6]; 5966 5967 /* out */ 5968 __u32 exit_reason; 5969 5970When KVM_RUN has returned successfully (return value 0), this informs 5971application code why KVM_RUN has returned. Allowable values for this 5972field are detailed below. 5973 5974:: 5975 5976 __u8 ready_for_interrupt_injection; 5977 5978If request_interrupt_window has been specified, this field indicates 5979an interrupt can be injected now with KVM_INTERRUPT. 5980 5981:: 5982 5983 __u8 if_flag; 5984 5985The value of the current interrupt flag. Only valid if in-kernel 5986local APIC is not used. 5987 5988:: 5989 5990 __u16 flags; 5991 5992More architecture-specific flags detailing state of the VCPU that may 5993affect the device's behavior. Current defined flags:: 5994 5995 /* x86, set if the VCPU is in system management mode */ 5996 #define KVM_RUN_X86_SMM (1 << 0) 5997 /* x86, set if bus lock detected in VM */ 5998 #define KVM_RUN_BUS_LOCK (1 << 1) 5999 /* arm64, set for KVM_EXIT_DEBUG */ 6000 #define KVM_DEBUG_ARCH_HSR_HIGH_VALID (1 << 0) 6001 6002:: 6003 6004 /* in (pre_kvm_run), out (post_kvm_run) */ 6005 __u64 cr8; 6006 6007The value of the cr8 register. Only valid if in-kernel local APIC is 6008not used. Both input and output. 6009 6010:: 6011 6012 __u64 apic_base; 6013 6014The value of the APIC BASE msr. Only valid if in-kernel local 6015APIC is not used. Both input and output. 6016 6017:: 6018 6019 union { 6020 /* KVM_EXIT_UNKNOWN */ 6021 struct { 6022 __u64 hardware_exit_reason; 6023 } hw; 6024 6025If exit_reason is KVM_EXIT_UNKNOWN, the vcpu has exited due to unknown 6026reasons. Further architecture-specific information is available in 6027hardware_exit_reason. 6028 6029:: 6030 6031 /* KVM_EXIT_FAIL_ENTRY */ 6032 struct { 6033 __u64 hardware_entry_failure_reason; 6034 __u32 cpu; /* if KVM_LAST_CPU */ 6035 } fail_entry; 6036 6037If exit_reason is KVM_EXIT_FAIL_ENTRY, the vcpu could not be run due 6038to unknown reasons. Further architecture-specific information is 6039available in hardware_entry_failure_reason. 6040 6041:: 6042 6043 /* KVM_EXIT_EXCEPTION */ 6044 struct { 6045 __u32 exception; 6046 __u32 error_code; 6047 } ex; 6048 6049Unused. 6050 6051:: 6052 6053 /* KVM_EXIT_IO */ 6054 struct { 6055 #define KVM_EXIT_IO_IN 0 6056 #define KVM_EXIT_IO_OUT 1 6057 __u8 direction; 6058 __u8 size; /* bytes */ 6059 __u16 port; 6060 __u32 count; 6061 __u64 data_offset; /* relative to kvm_run start */ 6062 } io; 6063 6064If exit_reason is KVM_EXIT_IO, then the vcpu has 6065executed a port I/O instruction which could not be satisfied by kvm. 6066data_offset describes where the data is located (KVM_EXIT_IO_OUT) or 6067where kvm expects application code to place the data for the next 6068KVM_RUN invocation (KVM_EXIT_IO_IN). Data format is a packed array. 6069 6070:: 6071 6072 /* KVM_EXIT_DEBUG */ 6073 struct { 6074 struct kvm_debug_exit_arch arch; 6075 } debug; 6076 6077If the exit_reason is KVM_EXIT_DEBUG, then a vcpu is processing a debug event 6078for which architecture specific information is returned. 6079 6080:: 6081 6082 /* KVM_EXIT_MMIO */ 6083 struct { 6084 __u64 phys_addr; 6085 __u8 data[8]; 6086 __u32 len; 6087 __u8 is_write; 6088 } mmio; 6089 6090If exit_reason is KVM_EXIT_MMIO, then the vcpu has 6091executed a memory-mapped I/O instruction which could not be satisfied 6092by kvm. The 'data' member contains the written data if 'is_write' is 6093true, and should be filled by application code otherwise. 6094 6095The 'data' member contains, in its first 'len' bytes, the value as it would 6096appear if the VCPU performed a load or store of the appropriate width directly 6097to the byte array. 6098 6099.. note:: 6100 6101 For KVM_EXIT_IO, KVM_EXIT_MMIO, KVM_EXIT_OSI, KVM_EXIT_PAPR, KVM_EXIT_XEN, 6102 KVM_EXIT_EPR, KVM_EXIT_X86_RDMSR and KVM_EXIT_X86_WRMSR the corresponding 6103 operations are complete (and guest state is consistent) only after userspace 6104 has re-entered the kernel with KVM_RUN. The kernel side will first finish 6105 incomplete operations and then check for pending signals. 6106 6107 The pending state of the operation is not preserved in state which is 6108 visible to userspace, thus userspace should ensure that the operation is 6109 completed before performing a live migration. Userspace can re-enter the 6110 guest with an unmasked signal pending or with the immediate_exit field set 6111 to complete pending operations without allowing any further instructions 6112 to be executed. 6113 6114:: 6115 6116 /* KVM_EXIT_HYPERCALL */ 6117 struct { 6118 __u64 nr; 6119 __u64 args[6]; 6120 __u64 ret; 6121 __u32 longmode; 6122 __u32 pad; 6123 } hypercall; 6124 6125Unused. This was once used for 'hypercall to userspace'. To implement 6126such functionality, use KVM_EXIT_IO (x86) or KVM_EXIT_MMIO (all except s390). 6127 6128.. note:: KVM_EXIT_IO is significantly faster than KVM_EXIT_MMIO. 6129 6130:: 6131 6132 /* KVM_EXIT_TPR_ACCESS */ 6133 struct { 6134 __u64 rip; 6135 __u32 is_write; 6136 __u32 pad; 6137 } tpr_access; 6138 6139To be documented (KVM_TPR_ACCESS_REPORTING). 6140 6141:: 6142 6143 /* KVM_EXIT_S390_SIEIC */ 6144 struct { 6145 __u8 icptcode; 6146 __u64 mask; /* psw upper half */ 6147 __u64 addr; /* psw lower half */ 6148 __u16 ipa; 6149 __u32 ipb; 6150 } s390_sieic; 6151 6152s390 specific. 6153 6154:: 6155 6156 /* KVM_EXIT_S390_RESET */ 6157 #define KVM_S390_RESET_POR 1 6158 #define KVM_S390_RESET_CLEAR 2 6159 #define KVM_S390_RESET_SUBSYSTEM 4 6160 #define KVM_S390_RESET_CPU_INIT 8 6161 #define KVM_S390_RESET_IPL 16 6162 __u64 s390_reset_flags; 6163 6164s390 specific. 6165 6166:: 6167 6168 /* KVM_EXIT_S390_UCONTROL */ 6169 struct { 6170 __u64 trans_exc_code; 6171 __u32 pgm_code; 6172 } s390_ucontrol; 6173 6174s390 specific. A page fault has occurred for a user controlled virtual 6175machine (KVM_VM_S390_UNCONTROL) on it's host page table that cannot be 6176resolved by the kernel. 6177The program code and the translation exception code that were placed 6178in the cpu's lowcore are presented here as defined by the z Architecture 6179Principles of Operation Book in the Chapter for Dynamic Address Translation 6180(DAT) 6181 6182:: 6183 6184 /* KVM_EXIT_DCR */ 6185 struct { 6186 __u32 dcrn; 6187 __u32 data; 6188 __u8 is_write; 6189 } dcr; 6190 6191Deprecated - was used for 440 KVM. 6192 6193:: 6194 6195 /* KVM_EXIT_OSI */ 6196 struct { 6197 __u64 gprs[32]; 6198 } osi; 6199 6200MOL uses a special hypercall interface it calls 'OSI'. To enable it, we catch 6201hypercalls and exit with this exit struct that contains all the guest gprs. 6202 6203If exit_reason is KVM_EXIT_OSI, then the vcpu has triggered such a hypercall. 6204Userspace can now handle the hypercall and when it's done modify the gprs as 6205necessary. Upon guest entry all guest GPRs will then be replaced by the values 6206in this struct. 6207 6208:: 6209 6210 /* KVM_EXIT_PAPR_HCALL */ 6211 struct { 6212 __u64 nr; 6213 __u64 ret; 6214 __u64 args[9]; 6215 } papr_hcall; 6216 6217This is used on 64-bit PowerPC when emulating a pSeries partition, 6218e.g. with the 'pseries' machine type in qemu. It occurs when the 6219guest does a hypercall using the 'sc 1' instruction. The 'nr' field 6220contains the hypercall number (from the guest R3), and 'args' contains 6221the arguments (from the guest R4 - R12). Userspace should put the 6222return code in 'ret' and any extra returned values in args[]. 6223The possible hypercalls are defined in the Power Architecture Platform 6224Requirements (PAPR) document available from www.power.org (free 6225developer registration required to access it). 6226 6227:: 6228 6229 /* KVM_EXIT_S390_TSCH */ 6230 struct { 6231 __u16 subchannel_id; 6232 __u16 subchannel_nr; 6233 __u32 io_int_parm; 6234 __u32 io_int_word; 6235 __u32 ipb; 6236 __u8 dequeued; 6237 } s390_tsch; 6238 6239s390 specific. This exit occurs when KVM_CAP_S390_CSS_SUPPORT has been enabled 6240and TEST SUBCHANNEL was intercepted. If dequeued is set, a pending I/O 6241interrupt for the target subchannel has been dequeued and subchannel_id, 6242subchannel_nr, io_int_parm and io_int_word contain the parameters for that 6243interrupt. ipb is needed for instruction parameter decoding. 6244 6245:: 6246 6247 /* KVM_EXIT_EPR */ 6248 struct { 6249 __u32 epr; 6250 } epr; 6251 6252On FSL BookE PowerPC chips, the interrupt controller has a fast patch 6253interrupt acknowledge path to the core. When the core successfully 6254delivers an interrupt, it automatically populates the EPR register with 6255the interrupt vector number and acknowledges the interrupt inside 6256the interrupt controller. 6257 6258In case the interrupt controller lives in user space, we need to do 6259the interrupt acknowledge cycle through it to fetch the next to be 6260delivered interrupt vector using this exit. 6261 6262It gets triggered whenever both KVM_CAP_PPC_EPR are enabled and an 6263external interrupt has just been delivered into the guest. User space 6264should put the acknowledged interrupt vector into the 'epr' field. 6265 6266:: 6267 6268 /* KVM_EXIT_SYSTEM_EVENT */ 6269 struct { 6270 #define KVM_SYSTEM_EVENT_SHUTDOWN 1 6271 #define KVM_SYSTEM_EVENT_RESET 2 6272 #define KVM_SYSTEM_EVENT_CRASH 3 6273 #define KVM_SYSTEM_EVENT_WAKEUP 4 6274 #define KVM_SYSTEM_EVENT_SUSPEND 5 6275 #define KVM_SYSTEM_EVENT_SEV_TERM 6 6276 __u32 type; 6277 __u32 ndata; 6278 __u64 data[16]; 6279 } system_event; 6280 6281If exit_reason is KVM_EXIT_SYSTEM_EVENT then the vcpu has triggered 6282a system-level event using some architecture specific mechanism (hypercall 6283or some special instruction). In case of ARM64, this is triggered using 6284HVC instruction based PSCI call from the vcpu. 6285 6286The 'type' field describes the system-level event type. 6287Valid values for 'type' are: 6288 6289 - KVM_SYSTEM_EVENT_SHUTDOWN -- the guest has requested a shutdown of the 6290 VM. Userspace is not obliged to honour this, and if it does honour 6291 this does not need to destroy the VM synchronously (ie it may call 6292 KVM_RUN again before shutdown finally occurs). 6293 - KVM_SYSTEM_EVENT_RESET -- the guest has requested a reset of the VM. 6294 As with SHUTDOWN, userspace can choose to ignore the request, or 6295 to schedule the reset to occur in the future and may call KVM_RUN again. 6296 - KVM_SYSTEM_EVENT_CRASH -- the guest crash occurred and the guest 6297 has requested a crash condition maintenance. Userspace can choose 6298 to ignore the request, or to gather VM memory core dump and/or 6299 reset/shutdown of the VM. 6300 - KVM_SYSTEM_EVENT_SEV_TERM -- an AMD SEV guest requested termination. 6301 The guest physical address of the guest's GHCB is stored in `data[0]`. 6302 - KVM_SYSTEM_EVENT_WAKEUP -- the exiting vCPU is in a suspended state and 6303 KVM has recognized a wakeup event. Userspace may honor this event by 6304 marking the exiting vCPU as runnable, or deny it and call KVM_RUN again. 6305 - KVM_SYSTEM_EVENT_SUSPEND -- the guest has requested a suspension of 6306 the VM. 6307 6308If KVM_CAP_SYSTEM_EVENT_DATA is present, the 'data' field can contain 6309architecture specific information for the system-level event. Only 6310the first `ndata` items (possibly zero) of the data array are valid. 6311 6312 - for arm64, data[0] is set to KVM_SYSTEM_EVENT_RESET_FLAG_PSCI_RESET2 if 6313 the guest issued a SYSTEM_RESET2 call according to v1.1 of the PSCI 6314 specification. 6315 6316 - for RISC-V, data[0] is set to the value of the second argument of the 6317 ``sbi_system_reset`` call. 6318 6319Previous versions of Linux defined a `flags` member in this struct. The 6320field is now aliased to `data[0]`. Userspace can assume that it is only 6321written if ndata is greater than 0. 6322 6323For arm/arm64: 6324-------------- 6325 6326KVM_SYSTEM_EVENT_SUSPEND exits are enabled with the 6327KVM_CAP_ARM_SYSTEM_SUSPEND VM capability. If a guest invokes the PSCI 6328SYSTEM_SUSPEND function, KVM will exit to userspace with this event 6329type. 6330 6331It is the sole responsibility of userspace to implement the PSCI 6332SYSTEM_SUSPEND call according to ARM DEN0022D.b 5.19 "SYSTEM_SUSPEND". 6333KVM does not change the vCPU's state before exiting to userspace, so 6334the call parameters are left in-place in the vCPU registers. 6335 6336Userspace is _required_ to take action for such an exit. It must 6337either: 6338 6339 - Honor the guest request to suspend the VM. Userspace can request 6340 in-kernel emulation of suspension by setting the calling vCPU's 6341 state to KVM_MP_STATE_SUSPENDED. Userspace must configure the vCPU's 6342 state according to the parameters passed to the PSCI function when 6343 the calling vCPU is resumed. See ARM DEN0022D.b 5.19.1 "Intended use" 6344 for details on the function parameters. 6345 6346 - Deny the guest request to suspend the VM. See ARM DEN0022D.b 5.19.2 6347 "Caller responsibilities" for possible return values. 6348 6349:: 6350 6351 /* KVM_EXIT_IOAPIC_EOI */ 6352 struct { 6353 __u8 vector; 6354 } eoi; 6355 6356Indicates that the VCPU's in-kernel local APIC received an EOI for a 6357level-triggered IOAPIC interrupt. This exit only triggers when the 6358IOAPIC is implemented in userspace (i.e. KVM_CAP_SPLIT_IRQCHIP is enabled); 6359the userspace IOAPIC should process the EOI and retrigger the interrupt if 6360it is still asserted. Vector is the LAPIC interrupt vector for which the 6361EOI was received. 6362 6363:: 6364 6365 struct kvm_hyperv_exit { 6366 #define KVM_EXIT_HYPERV_SYNIC 1 6367 #define KVM_EXIT_HYPERV_HCALL 2 6368 #define KVM_EXIT_HYPERV_SYNDBG 3 6369 __u32 type; 6370 __u32 pad1; 6371 union { 6372 struct { 6373 __u32 msr; 6374 __u32 pad2; 6375 __u64 control; 6376 __u64 evt_page; 6377 __u64 msg_page; 6378 } synic; 6379 struct { 6380 __u64 input; 6381 __u64 result; 6382 __u64 params[2]; 6383 } hcall; 6384 struct { 6385 __u32 msr; 6386 __u32 pad2; 6387 __u64 control; 6388 __u64 status; 6389 __u64 send_page; 6390 __u64 recv_page; 6391 __u64 pending_page; 6392 } syndbg; 6393 } u; 6394 }; 6395 /* KVM_EXIT_HYPERV */ 6396 struct kvm_hyperv_exit hyperv; 6397 6398Indicates that the VCPU exits into userspace to process some tasks 6399related to Hyper-V emulation. 6400 6401Valid values for 'type' are: 6402 6403 - KVM_EXIT_HYPERV_SYNIC -- synchronously notify user-space about 6404 6405Hyper-V SynIC state change. Notification is used to remap SynIC 6406event/message pages and to enable/disable SynIC messages/events processing 6407in userspace. 6408 6409 - KVM_EXIT_HYPERV_SYNDBG -- synchronously notify user-space about 6410 6411Hyper-V Synthetic debugger state change. Notification is used to either update 6412the pending_page location or to send a control command (send the buffer located 6413in send_page or recv a buffer to recv_page). 6414 6415:: 6416 6417 /* KVM_EXIT_ARM_NISV */ 6418 struct { 6419 __u64 esr_iss; 6420 __u64 fault_ipa; 6421 } arm_nisv; 6422 6423Used on arm64 systems. If a guest accesses memory not in a memslot, 6424KVM will typically return to userspace and ask it to do MMIO emulation on its 6425behalf. However, for certain classes of instructions, no instruction decode 6426(direction, length of memory access) is provided, and fetching and decoding 6427the instruction from the VM is overly complicated to live in the kernel. 6428 6429Historically, when this situation occurred, KVM would print a warning and kill 6430the VM. KVM assumed that if the guest accessed non-memslot memory, it was 6431trying to do I/O, which just couldn't be emulated, and the warning message was 6432phrased accordingly. However, what happened more often was that a guest bug 6433caused access outside the guest memory areas which should lead to a more 6434meaningful warning message and an external abort in the guest, if the access 6435did not fall within an I/O window. 6436 6437Userspace implementations can query for KVM_CAP_ARM_NISV_TO_USER, and enable 6438this capability at VM creation. Once this is done, these types of errors will 6439instead return to userspace with KVM_EXIT_ARM_NISV, with the valid bits from 6440the ESR_EL2 in the esr_iss field, and the faulting IPA in the fault_ipa field. 6441Userspace can either fix up the access if it's actually an I/O access by 6442decoding the instruction from guest memory (if it's very brave) and continue 6443executing the guest, or it can decide to suspend, dump, or restart the guest. 6444 6445Note that KVM does not skip the faulting instruction as it does for 6446KVM_EXIT_MMIO, but userspace has to emulate any change to the processing state 6447if it decides to decode and emulate the instruction. 6448 6449:: 6450 6451 /* KVM_EXIT_X86_RDMSR / KVM_EXIT_X86_WRMSR */ 6452 struct { 6453 __u8 error; /* user -> kernel */ 6454 __u8 pad[7]; 6455 __u32 reason; /* kernel -> user */ 6456 __u32 index; /* kernel -> user */ 6457 __u64 data; /* kernel <-> user */ 6458 } msr; 6459 6460Used on x86 systems. When the VM capability KVM_CAP_X86_USER_SPACE_MSR is 6461enabled, MSR accesses to registers that would invoke a #GP by KVM kernel code 6462may instead trigger a KVM_EXIT_X86_RDMSR exit for reads and KVM_EXIT_X86_WRMSR 6463exit for writes. 6464 6465The "reason" field specifies why the MSR interception occurred. Userspace will 6466only receive MSR exits when a particular reason was requested during through 6467ENABLE_CAP. Currently valid exit reasons are: 6468 6469============================ ======================================== 6470 KVM_MSR_EXIT_REASON_UNKNOWN access to MSR that is unknown to KVM 6471 KVM_MSR_EXIT_REASON_INVAL access to invalid MSRs or reserved bits 6472 KVM_MSR_EXIT_REASON_FILTER access blocked by KVM_X86_SET_MSR_FILTER 6473============================ ======================================== 6474 6475For KVM_EXIT_X86_RDMSR, the "index" field tells userspace which MSR the guest 6476wants to read. To respond to this request with a successful read, userspace 6477writes the respective data into the "data" field and must continue guest 6478execution to ensure the read data is transferred into guest register state. 6479 6480If the RDMSR request was unsuccessful, userspace indicates that with a "1" in 6481the "error" field. This will inject a #GP into the guest when the VCPU is 6482executed again. 6483 6484For KVM_EXIT_X86_WRMSR, the "index" field tells userspace which MSR the guest 6485wants to write. Once finished processing the event, userspace must continue 6486vCPU execution. If the MSR write was unsuccessful, userspace also sets the 6487"error" field to "1". 6488 6489See KVM_X86_SET_MSR_FILTER for details on the interaction with MSR filtering. 6490 6491:: 6492 6493 6494 struct kvm_xen_exit { 6495 #define KVM_EXIT_XEN_HCALL 1 6496 __u32 type; 6497 union { 6498 struct { 6499 __u32 longmode; 6500 __u32 cpl; 6501 __u64 input; 6502 __u64 result; 6503 __u64 params[6]; 6504 } hcall; 6505 } u; 6506 }; 6507 /* KVM_EXIT_XEN */ 6508 struct kvm_hyperv_exit xen; 6509 6510Indicates that the VCPU exits into userspace to process some tasks 6511related to Xen emulation. 6512 6513Valid values for 'type' are: 6514 6515 - KVM_EXIT_XEN_HCALL -- synchronously notify user-space about Xen hypercall. 6516 Userspace is expected to place the hypercall result into the appropriate 6517 field before invoking KVM_RUN again. 6518 6519:: 6520 6521 /* KVM_EXIT_RISCV_SBI */ 6522 struct { 6523 unsigned long extension_id; 6524 unsigned long function_id; 6525 unsigned long args[6]; 6526 unsigned long ret[2]; 6527 } riscv_sbi; 6528 6529If exit reason is KVM_EXIT_RISCV_SBI then it indicates that the VCPU has 6530done a SBI call which is not handled by KVM RISC-V kernel module. The details 6531of the SBI call are available in 'riscv_sbi' member of kvm_run structure. The 6532'extension_id' field of 'riscv_sbi' represents SBI extension ID whereas the 6533'function_id' field represents function ID of given SBI extension. The 'args' 6534array field of 'riscv_sbi' represents parameters for the SBI call and 'ret' 6535array field represents return values. The userspace should update the return 6536values of SBI call before resuming the VCPU. For more details on RISC-V SBI 6537spec refer, https://github.com/riscv/riscv-sbi-doc. 6538 6539:: 6540 6541 /* KVM_EXIT_NOTIFY */ 6542 struct { 6543 #define KVM_NOTIFY_CONTEXT_INVALID (1 << 0) 6544 __u32 flags; 6545 } notify; 6546 6547Used on x86 systems. When the VM capability KVM_CAP_X86_NOTIFY_VMEXIT is 6548enabled, a VM exit generated if no event window occurs in VM non-root mode 6549for a specified amount of time. Once KVM_X86_NOTIFY_VMEXIT_USER is set when 6550enabling the cap, it would exit to userspace with the exit reason 6551KVM_EXIT_NOTIFY for further handling. The "flags" field contains more 6552detailed info. 6553 6554The valid value for 'flags' is: 6555 6556 - KVM_NOTIFY_CONTEXT_INVALID -- the VM context is corrupted and not valid 6557 in VMCS. It would run into unknown result if resume the target VM. 6558 6559:: 6560 6561 /* Fix the size of the union. */ 6562 char padding[256]; 6563 }; 6564 6565 /* 6566 * shared registers between kvm and userspace. 6567 * kvm_valid_regs specifies the register classes set by the host 6568 * kvm_dirty_regs specified the register classes dirtied by userspace 6569 * struct kvm_sync_regs is architecture specific, as well as the 6570 * bits for kvm_valid_regs and kvm_dirty_regs 6571 */ 6572 __u64 kvm_valid_regs; 6573 __u64 kvm_dirty_regs; 6574 union { 6575 struct kvm_sync_regs regs; 6576 char padding[SYNC_REGS_SIZE_BYTES]; 6577 } s; 6578 6579If KVM_CAP_SYNC_REGS is defined, these fields allow userspace to access 6580certain guest registers without having to call SET/GET_*REGS. Thus we can 6581avoid some system call overhead if userspace has to handle the exit. 6582Userspace can query the validity of the structure by checking 6583kvm_valid_regs for specific bits. These bits are architecture specific 6584and usually define the validity of a groups of registers. (e.g. one bit 6585for general purpose registers) 6586 6587Please note that the kernel is allowed to use the kvm_run structure as the 6588primary storage for certain register types. Therefore, the kernel may use the 6589values in kvm_run even if the corresponding bit in kvm_dirty_regs is not set. 6590 6591 65926. Capabilities that can be enabled on vCPUs 6593============================================ 6594 6595There are certain capabilities that change the behavior of the virtual CPU or 6596the virtual machine when enabled. To enable them, please see section 4.37. 6597Below you can find a list of capabilities and what their effect on the vCPU or 6598the virtual machine is when enabling them. 6599 6600The following information is provided along with the description: 6601 6602 Architectures: 6603 which instruction set architectures provide this ioctl. 6604 x86 includes both i386 and x86_64. 6605 6606 Target: 6607 whether this is a per-vcpu or per-vm capability. 6608 6609 Parameters: 6610 what parameters are accepted by the capability. 6611 6612 Returns: 6613 the return value. General error numbers (EBADF, ENOMEM, EINVAL) 6614 are not detailed, but errors with specific meanings are. 6615 6616 66176.1 KVM_CAP_PPC_OSI 6618------------------- 6619 6620:Architectures: ppc 6621:Target: vcpu 6622:Parameters: none 6623:Returns: 0 on success; -1 on error 6624 6625This capability enables interception of OSI hypercalls that otherwise would 6626be treated as normal system calls to be injected into the guest. OSI hypercalls 6627were invented by Mac-on-Linux to have a standardized communication mechanism 6628between the guest and the host. 6629 6630When this capability is enabled, KVM_EXIT_OSI can occur. 6631 6632 66336.2 KVM_CAP_PPC_PAPR 6634-------------------- 6635 6636:Architectures: ppc 6637:Target: vcpu 6638:Parameters: none 6639:Returns: 0 on success; -1 on error 6640 6641This capability enables interception of PAPR hypercalls. PAPR hypercalls are 6642done using the hypercall instruction "sc 1". 6643 6644It also sets the guest privilege level to "supervisor" mode. Usually the guest 6645runs in "hypervisor" privilege mode with a few missing features. 6646 6647In addition to the above, it changes the semantics of SDR1. In this mode, the 6648HTAB address part of SDR1 contains an HVA instead of a GPA, as PAPR keeps the 6649HTAB invisible to the guest. 6650 6651When this capability is enabled, KVM_EXIT_PAPR_HCALL can occur. 6652 6653 66546.3 KVM_CAP_SW_TLB 6655------------------ 6656 6657:Architectures: ppc 6658:Target: vcpu 6659:Parameters: args[0] is the address of a struct kvm_config_tlb 6660:Returns: 0 on success; -1 on error 6661 6662:: 6663 6664 struct kvm_config_tlb { 6665 __u64 params; 6666 __u64 array; 6667 __u32 mmu_type; 6668 __u32 array_len; 6669 }; 6670 6671Configures the virtual CPU's TLB array, establishing a shared memory area 6672between userspace and KVM. The "params" and "array" fields are userspace 6673addresses of mmu-type-specific data structures. The "array_len" field is an 6674safety mechanism, and should be set to the size in bytes of the memory that 6675userspace has reserved for the array. It must be at least the size dictated 6676by "mmu_type" and "params". 6677 6678While KVM_RUN is active, the shared region is under control of KVM. Its 6679contents are undefined, and any modification by userspace results in 6680boundedly undefined behavior. 6681 6682On return from KVM_RUN, the shared region will reflect the current state of 6683the guest's TLB. If userspace makes any changes, it must call KVM_DIRTY_TLB 6684to tell KVM which entries have been changed, prior to calling KVM_RUN again 6685on this vcpu. 6686 6687For mmu types KVM_MMU_FSL_BOOKE_NOHV and KVM_MMU_FSL_BOOKE_HV: 6688 6689 - The "params" field is of type "struct kvm_book3e_206_tlb_params". 6690 - The "array" field points to an array of type "struct 6691 kvm_book3e_206_tlb_entry". 6692 - The array consists of all entries in the first TLB, followed by all 6693 entries in the second TLB. 6694 - Within a TLB, entries are ordered first by increasing set number. Within a 6695 set, entries are ordered by way (increasing ESEL). 6696 - The hash for determining set number in TLB0 is: (MAS2 >> 12) & (num_sets - 1) 6697 where "num_sets" is the tlb_sizes[] value divided by the tlb_ways[] value. 6698 - The tsize field of mas1 shall be set to 4K on TLB0, even though the 6699 hardware ignores this value for TLB0. 6700 67016.4 KVM_CAP_S390_CSS_SUPPORT 6702---------------------------- 6703 6704:Architectures: s390 6705:Target: vcpu 6706:Parameters: none 6707:Returns: 0 on success; -1 on error 6708 6709This capability enables support for handling of channel I/O instructions. 6710 6711TEST PENDING INTERRUPTION and the interrupt portion of TEST SUBCHANNEL are 6712handled in-kernel, while the other I/O instructions are passed to userspace. 6713 6714When this capability is enabled, KVM_EXIT_S390_TSCH will occur on TEST 6715SUBCHANNEL intercepts. 6716 6717Note that even though this capability is enabled per-vcpu, the complete 6718virtual machine is affected. 6719 67206.5 KVM_CAP_PPC_EPR 6721------------------- 6722 6723:Architectures: ppc 6724:Target: vcpu 6725:Parameters: args[0] defines whether the proxy facility is active 6726:Returns: 0 on success; -1 on error 6727 6728This capability enables or disables the delivery of interrupts through the 6729external proxy facility. 6730 6731When enabled (args[0] != 0), every time the guest gets an external interrupt 6732delivered, it automatically exits into user space with a KVM_EXIT_EPR exit 6733to receive the topmost interrupt vector. 6734 6735When disabled (args[0] == 0), behavior is as if this facility is unsupported. 6736 6737When this capability is enabled, KVM_EXIT_EPR can occur. 6738 67396.6 KVM_CAP_IRQ_MPIC 6740-------------------- 6741 6742:Architectures: ppc 6743:Parameters: args[0] is the MPIC device fd; 6744 args[1] is the MPIC CPU number for this vcpu 6745 6746This capability connects the vcpu to an in-kernel MPIC device. 6747 67486.7 KVM_CAP_IRQ_XICS 6749-------------------- 6750 6751:Architectures: ppc 6752:Target: vcpu 6753:Parameters: args[0] is the XICS device fd; 6754 args[1] is the XICS CPU number (server ID) for this vcpu 6755 6756This capability connects the vcpu to an in-kernel XICS device. 6757 67586.8 KVM_CAP_S390_IRQCHIP 6759------------------------ 6760 6761:Architectures: s390 6762:Target: vm 6763:Parameters: none 6764 6765This capability enables the in-kernel irqchip for s390. Please refer to 6766"4.24 KVM_CREATE_IRQCHIP" for details. 6767 67686.9 KVM_CAP_MIPS_FPU 6769-------------------- 6770 6771:Architectures: mips 6772:Target: vcpu 6773:Parameters: args[0] is reserved for future use (should be 0). 6774 6775This capability allows the use of the host Floating Point Unit by the guest. It 6776allows the Config1.FP bit to be set to enable the FPU in the guest. Once this is 6777done the ``KVM_REG_MIPS_FPR_*`` and ``KVM_REG_MIPS_FCR_*`` registers can be 6778accessed (depending on the current guest FPU register mode), and the Status.FR, 6779Config5.FRE bits are accessible via the KVM API and also from the guest, 6780depending on them being supported by the FPU. 6781 67826.10 KVM_CAP_MIPS_MSA 6783--------------------- 6784 6785:Architectures: mips 6786:Target: vcpu 6787:Parameters: args[0] is reserved for future use (should be 0). 6788 6789This capability allows the use of the MIPS SIMD Architecture (MSA) by the guest. 6790It allows the Config3.MSAP bit to be set to enable the use of MSA by the guest. 6791Once this is done the ``KVM_REG_MIPS_VEC_*`` and ``KVM_REG_MIPS_MSA_*`` 6792registers can be accessed, and the Config5.MSAEn bit is accessible via the 6793KVM API and also from the guest. 6794 67956.74 KVM_CAP_SYNC_REGS 6796---------------------- 6797 6798:Architectures: s390, x86 6799:Target: s390: always enabled, x86: vcpu 6800:Parameters: none 6801:Returns: x86: KVM_CHECK_EXTENSION returns a bit-array indicating which register 6802 sets are supported 6803 (bitfields defined in arch/x86/include/uapi/asm/kvm.h). 6804 6805As described above in the kvm_sync_regs struct info in section 5 (kvm_run): 6806KVM_CAP_SYNC_REGS "allow[s] userspace to access certain guest registers 6807without having to call SET/GET_*REGS". This reduces overhead by eliminating 6808repeated ioctl calls for setting and/or getting register values. This is 6809particularly important when userspace is making synchronous guest state 6810modifications, e.g. when emulating and/or intercepting instructions in 6811userspace. 6812 6813For s390 specifics, please refer to the source code. 6814 6815For x86: 6816 6817- the register sets to be copied out to kvm_run are selectable 6818 by userspace (rather that all sets being copied out for every exit). 6819- vcpu_events are available in addition to regs and sregs. 6820 6821For x86, the 'kvm_valid_regs' field of struct kvm_run is overloaded to 6822function as an input bit-array field set by userspace to indicate the 6823specific register sets to be copied out on the next exit. 6824 6825To indicate when userspace has modified values that should be copied into 6826the vCPU, the all architecture bitarray field, 'kvm_dirty_regs' must be set. 6827This is done using the same bitflags as for the 'kvm_valid_regs' field. 6828If the dirty bit is not set, then the register set values will not be copied 6829into the vCPU even if they've been modified. 6830 6831Unused bitfields in the bitarrays must be set to zero. 6832 6833:: 6834 6835 struct kvm_sync_regs { 6836 struct kvm_regs regs; 6837 struct kvm_sregs sregs; 6838 struct kvm_vcpu_events events; 6839 }; 6840 68416.75 KVM_CAP_PPC_IRQ_XIVE 6842------------------------- 6843 6844:Architectures: ppc 6845:Target: vcpu 6846:Parameters: args[0] is the XIVE device fd; 6847 args[1] is the XIVE CPU number (server ID) for this vcpu 6848 6849This capability connects the vcpu to an in-kernel XIVE device. 6850 68517. Capabilities that can be enabled on VMs 6852========================================== 6853 6854There are certain capabilities that change the behavior of the virtual 6855machine when enabled. To enable them, please see section 4.37. Below 6856you can find a list of capabilities and what their effect on the VM 6857is when enabling them. 6858 6859The following information is provided along with the description: 6860 6861 Architectures: 6862 which instruction set architectures provide this ioctl. 6863 x86 includes both i386 and x86_64. 6864 6865 Parameters: 6866 what parameters are accepted by the capability. 6867 6868 Returns: 6869 the return value. General error numbers (EBADF, ENOMEM, EINVAL) 6870 are not detailed, but errors with specific meanings are. 6871 6872 68737.1 KVM_CAP_PPC_ENABLE_HCALL 6874---------------------------- 6875 6876:Architectures: ppc 6877:Parameters: args[0] is the sPAPR hcall number; 6878 args[1] is 0 to disable, 1 to enable in-kernel handling 6879 6880This capability controls whether individual sPAPR hypercalls (hcalls) 6881get handled by the kernel or not. Enabling or disabling in-kernel 6882handling of an hcall is effective across the VM. On creation, an 6883initial set of hcalls are enabled for in-kernel handling, which 6884consists of those hcalls for which in-kernel handlers were implemented 6885before this capability was implemented. If disabled, the kernel will 6886not to attempt to handle the hcall, but will always exit to userspace 6887to handle it. Note that it may not make sense to enable some and 6888disable others of a group of related hcalls, but KVM does not prevent 6889userspace from doing that. 6890 6891If the hcall number specified is not one that has an in-kernel 6892implementation, the KVM_ENABLE_CAP ioctl will fail with an EINVAL 6893error. 6894 68957.2 KVM_CAP_S390_USER_SIGP 6896-------------------------- 6897 6898:Architectures: s390 6899:Parameters: none 6900 6901This capability controls which SIGP orders will be handled completely in user 6902space. With this capability enabled, all fast orders will be handled completely 6903in the kernel: 6904 6905- SENSE 6906- SENSE RUNNING 6907- EXTERNAL CALL 6908- EMERGENCY SIGNAL 6909- CONDITIONAL EMERGENCY SIGNAL 6910 6911All other orders will be handled completely in user space. 6912 6913Only privileged operation exceptions will be checked for in the kernel (or even 6914in the hardware prior to interception). If this capability is not enabled, the 6915old way of handling SIGP orders is used (partially in kernel and user space). 6916 69177.3 KVM_CAP_S390_VECTOR_REGISTERS 6918--------------------------------- 6919 6920:Architectures: s390 6921:Parameters: none 6922:Returns: 0 on success, negative value on error 6923 6924Allows use of the vector registers introduced with z13 processor, and 6925provides for the synchronization between host and user space. Will 6926return -EINVAL if the machine does not support vectors. 6927 69287.4 KVM_CAP_S390_USER_STSI 6929-------------------------- 6930 6931:Architectures: s390 6932:Parameters: none 6933 6934This capability allows post-handlers for the STSI instruction. After 6935initial handling in the kernel, KVM exits to user space with 6936KVM_EXIT_S390_STSI to allow user space to insert further data. 6937 6938Before exiting to userspace, kvm handlers should fill in s390_stsi field of 6939vcpu->run:: 6940 6941 struct { 6942 __u64 addr; 6943 __u8 ar; 6944 __u8 reserved; 6945 __u8 fc; 6946 __u8 sel1; 6947 __u16 sel2; 6948 } s390_stsi; 6949 6950 @addr - guest address of STSI SYSIB 6951 @fc - function code 6952 @sel1 - selector 1 6953 @sel2 - selector 2 6954 @ar - access register number 6955 6956KVM handlers should exit to userspace with rc = -EREMOTE. 6957 69587.5 KVM_CAP_SPLIT_IRQCHIP 6959------------------------- 6960 6961:Architectures: x86 6962:Parameters: args[0] - number of routes reserved for userspace IOAPICs 6963:Returns: 0 on success, -1 on error 6964 6965Create a local apic for each processor in the kernel. This can be used 6966instead of KVM_CREATE_IRQCHIP if the userspace VMM wishes to emulate the 6967IOAPIC and PIC (and also the PIT, even though this has to be enabled 6968separately). 6969 6970This capability also enables in kernel routing of interrupt requests; 6971when KVM_CAP_SPLIT_IRQCHIP only routes of KVM_IRQ_ROUTING_MSI type are 6972used in the IRQ routing table. The first args[0] MSI routes are reserved 6973for the IOAPIC pins. Whenever the LAPIC receives an EOI for these routes, 6974a KVM_EXIT_IOAPIC_EOI vmexit will be reported to userspace. 6975 6976Fails if VCPU has already been created, or if the irqchip is already in the 6977kernel (i.e. KVM_CREATE_IRQCHIP has already been called). 6978 69797.6 KVM_CAP_S390_RI 6980------------------- 6981 6982:Architectures: s390 6983:Parameters: none 6984 6985Allows use of runtime-instrumentation introduced with zEC12 processor. 6986Will return -EINVAL if the machine does not support runtime-instrumentation. 6987Will return -EBUSY if a VCPU has already been created. 6988 69897.7 KVM_CAP_X2APIC_API 6990---------------------- 6991 6992:Architectures: x86 6993:Parameters: args[0] - features that should be enabled 6994:Returns: 0 on success, -EINVAL when args[0] contains invalid features 6995 6996Valid feature flags in args[0] are:: 6997 6998 #define KVM_X2APIC_API_USE_32BIT_IDS (1ULL << 0) 6999 #define KVM_X2APIC_API_DISABLE_BROADCAST_QUIRK (1ULL << 1) 7000 7001Enabling KVM_X2APIC_API_USE_32BIT_IDS changes the behavior of 7002KVM_SET_GSI_ROUTING, KVM_SIGNAL_MSI, KVM_SET_LAPIC, and KVM_GET_LAPIC, 7003allowing the use of 32-bit APIC IDs. See KVM_CAP_X2APIC_API in their 7004respective sections. 7005 7006KVM_X2APIC_API_DISABLE_BROADCAST_QUIRK must be enabled for x2APIC to work 7007in logical mode or with more than 255 VCPUs. Otherwise, KVM treats 0xff 7008as a broadcast even in x2APIC mode in order to support physical x2APIC 7009without interrupt remapping. This is undesirable in logical mode, 7010where 0xff represents CPUs 0-7 in cluster 0. 7011 70127.8 KVM_CAP_S390_USER_INSTR0 7013---------------------------- 7014 7015:Architectures: s390 7016:Parameters: none 7017 7018With this capability enabled, all illegal instructions 0x0000 (2 bytes) will 7019be intercepted and forwarded to user space. User space can use this 7020mechanism e.g. to realize 2-byte software breakpoints. The kernel will 7021not inject an operating exception for these instructions, user space has 7022to take care of that. 7023 7024This capability can be enabled dynamically even if VCPUs were already 7025created and are running. 7026 70277.9 KVM_CAP_S390_GS 7028------------------- 7029 7030:Architectures: s390 7031:Parameters: none 7032:Returns: 0 on success; -EINVAL if the machine does not support 7033 guarded storage; -EBUSY if a VCPU has already been created. 7034 7035Allows use of guarded storage for the KVM guest. 7036 70377.10 KVM_CAP_S390_AIS 7038--------------------- 7039 7040:Architectures: s390 7041:Parameters: none 7042 7043Allow use of adapter-interruption suppression. 7044:Returns: 0 on success; -EBUSY if a VCPU has already been created. 7045 70467.11 KVM_CAP_PPC_SMT 7047-------------------- 7048 7049:Architectures: ppc 7050:Parameters: vsmt_mode, flags 7051 7052Enabling this capability on a VM provides userspace with a way to set 7053the desired virtual SMT mode (i.e. the number of virtual CPUs per 7054virtual core). The virtual SMT mode, vsmt_mode, must be a power of 2 7055between 1 and 8. On POWER8, vsmt_mode must also be no greater than 7056the number of threads per subcore for the host. Currently flags must 7057be 0. A successful call to enable this capability will result in 7058vsmt_mode being returned when the KVM_CAP_PPC_SMT capability is 7059subsequently queried for the VM. This capability is only supported by 7060HV KVM, and can only be set before any VCPUs have been created. 7061The KVM_CAP_PPC_SMT_POSSIBLE capability indicates which virtual SMT 7062modes are available. 7063 70647.12 KVM_CAP_PPC_FWNMI 7065---------------------- 7066 7067:Architectures: ppc 7068:Parameters: none 7069 7070With this capability a machine check exception in the guest address 7071space will cause KVM to exit the guest with NMI exit reason. This 7072enables QEMU to build error log and branch to guest kernel registered 7073machine check handling routine. Without this capability KVM will 7074branch to guests' 0x200 interrupt vector. 7075 70767.13 KVM_CAP_X86_DISABLE_EXITS 7077------------------------------ 7078 7079:Architectures: x86 7080:Parameters: args[0] defines which exits are disabled 7081:Returns: 0 on success, -EINVAL when args[0] contains invalid exits 7082 7083Valid bits in args[0] are:: 7084 7085 #define KVM_X86_DISABLE_EXITS_MWAIT (1 << 0) 7086 #define KVM_X86_DISABLE_EXITS_HLT (1 << 1) 7087 #define KVM_X86_DISABLE_EXITS_PAUSE (1 << 2) 7088 #define KVM_X86_DISABLE_EXITS_CSTATE (1 << 3) 7089 7090Enabling this capability on a VM provides userspace with a way to no 7091longer intercept some instructions for improved latency in some 7092workloads, and is suggested when vCPUs are associated to dedicated 7093physical CPUs. More bits can be added in the future; userspace can 7094just pass the KVM_CHECK_EXTENSION result to KVM_ENABLE_CAP to disable 7095all such vmexits. 7096 7097Do not enable KVM_FEATURE_PV_UNHALT if you disable HLT exits. 7098 70997.14 KVM_CAP_S390_HPAGE_1M 7100-------------------------- 7101 7102:Architectures: s390 7103:Parameters: none 7104:Returns: 0 on success, -EINVAL if hpage module parameter was not set 7105 or cmma is enabled, or the VM has the KVM_VM_S390_UCONTROL 7106 flag set 7107 7108With this capability the KVM support for memory backing with 1m pages 7109through hugetlbfs can be enabled for a VM. After the capability is 7110enabled, cmma can't be enabled anymore and pfmfi and the storage key 7111interpretation are disabled. If cmma has already been enabled or the 7112hpage module parameter is not set to 1, -EINVAL is returned. 7113 7114While it is generally possible to create a huge page backed VM without 7115this capability, the VM will not be able to run. 7116 71177.15 KVM_CAP_MSR_PLATFORM_INFO 7118------------------------------ 7119 7120:Architectures: x86 7121:Parameters: args[0] whether feature should be enabled or not 7122 7123With this capability, a guest may read the MSR_PLATFORM_INFO MSR. Otherwise, 7124a #GP would be raised when the guest tries to access. Currently, this 7125capability does not enable write permissions of this MSR for the guest. 7126 71277.16 KVM_CAP_PPC_NESTED_HV 7128-------------------------- 7129 7130:Architectures: ppc 7131:Parameters: none 7132:Returns: 0 on success, -EINVAL when the implementation doesn't support 7133 nested-HV virtualization. 7134 7135HV-KVM on POWER9 and later systems allows for "nested-HV" 7136virtualization, which provides a way for a guest VM to run guests that 7137can run using the CPU's supervisor mode (privileged non-hypervisor 7138state). Enabling this capability on a VM depends on the CPU having 7139the necessary functionality and on the facility being enabled with a 7140kvm-hv module parameter. 7141 71427.17 KVM_CAP_EXCEPTION_PAYLOAD 7143------------------------------ 7144 7145:Architectures: x86 7146:Parameters: args[0] whether feature should be enabled or not 7147 7148With this capability enabled, CR2 will not be modified prior to the 7149emulated VM-exit when L1 intercepts a #PF exception that occurs in 7150L2. Similarly, for kvm-intel only, DR6 will not be modified prior to 7151the emulated VM-exit when L1 intercepts a #DB exception that occurs in 7152L2. As a result, when KVM_GET_VCPU_EVENTS reports a pending #PF (or 7153#DB) exception for L2, exception.has_payload will be set and the 7154faulting address (or the new DR6 bits*) will be reported in the 7155exception_payload field. Similarly, when userspace injects a #PF (or 7156#DB) into L2 using KVM_SET_VCPU_EVENTS, it is expected to set 7157exception.has_payload and to put the faulting address - or the new DR6 7158bits\ [#]_ - in the exception_payload field. 7159 7160This capability also enables exception.pending in struct 7161kvm_vcpu_events, which allows userspace to distinguish between pending 7162and injected exceptions. 7163 7164 7165.. [#] For the new DR6 bits, note that bit 16 is set iff the #DB exception 7166 will clear DR6.RTM. 7167 71687.18 KVM_CAP_MANUAL_DIRTY_LOG_PROTECT2 7169 7170:Architectures: x86, arm64, mips 7171:Parameters: args[0] whether feature should be enabled or not 7172 7173Valid flags are:: 7174 7175 #define KVM_DIRTY_LOG_MANUAL_PROTECT_ENABLE (1 << 0) 7176 #define KVM_DIRTY_LOG_INITIALLY_SET (1 << 1) 7177 7178With KVM_DIRTY_LOG_MANUAL_PROTECT_ENABLE is set, KVM_GET_DIRTY_LOG will not 7179automatically clear and write-protect all pages that are returned as dirty. 7180Rather, userspace will have to do this operation separately using 7181KVM_CLEAR_DIRTY_LOG. 7182 7183At the cost of a slightly more complicated operation, this provides better 7184scalability and responsiveness for two reasons. First, 7185KVM_CLEAR_DIRTY_LOG ioctl can operate on a 64-page granularity rather 7186than requiring to sync a full memslot; this ensures that KVM does not 7187take spinlocks for an extended period of time. Second, in some cases a 7188large amount of time can pass between a call to KVM_GET_DIRTY_LOG and 7189userspace actually using the data in the page. Pages can be modified 7190during this time, which is inefficient for both the guest and userspace: 7191the guest will incur a higher penalty due to write protection faults, 7192while userspace can see false reports of dirty pages. Manual reprotection 7193helps reducing this time, improving guest performance and reducing the 7194number of dirty log false positives. 7195 7196With KVM_DIRTY_LOG_INITIALLY_SET set, all the bits of the dirty bitmap 7197will be initialized to 1 when created. This also improves performance because 7198dirty logging can be enabled gradually in small chunks on the first call 7199to KVM_CLEAR_DIRTY_LOG. KVM_DIRTY_LOG_INITIALLY_SET depends on 7200KVM_DIRTY_LOG_MANUAL_PROTECT_ENABLE (it is also only available on 7201x86 and arm64 for now). 7202 7203KVM_CAP_MANUAL_DIRTY_LOG_PROTECT2 was previously available under the name 7204KVM_CAP_MANUAL_DIRTY_LOG_PROTECT, but the implementation had bugs that make 7205it hard or impossible to use it correctly. The availability of 7206KVM_CAP_MANUAL_DIRTY_LOG_PROTECT2 signals that those bugs are fixed. 7207Userspace should not try to use KVM_CAP_MANUAL_DIRTY_LOG_PROTECT. 7208 72097.19 KVM_CAP_PPC_SECURE_GUEST 7210------------------------------ 7211 7212:Architectures: ppc 7213 7214This capability indicates that KVM is running on a host that has 7215ultravisor firmware and thus can support a secure guest. On such a 7216system, a guest can ask the ultravisor to make it a secure guest, 7217one whose memory is inaccessible to the host except for pages which 7218are explicitly requested to be shared with the host. The ultravisor 7219notifies KVM when a guest requests to become a secure guest, and KVM 7220has the opportunity to veto the transition. 7221 7222If present, this capability can be enabled for a VM, meaning that KVM 7223will allow the transition to secure guest mode. Otherwise KVM will 7224veto the transition. 7225 72267.20 KVM_CAP_HALT_POLL 7227---------------------- 7228 7229:Architectures: all 7230:Target: VM 7231:Parameters: args[0] is the maximum poll time in nanoseconds 7232:Returns: 0 on success; -1 on error 7233 7234KVM_CAP_HALT_POLL overrides the kvm.halt_poll_ns module parameter to set the 7235maximum halt-polling time for all vCPUs in the target VM. This capability can 7236be invoked at any time and any number of times to dynamically change the 7237maximum halt-polling time. 7238 7239See Documentation/virt/kvm/halt-polling.rst for more information on halt 7240polling. 7241 72427.21 KVM_CAP_X86_USER_SPACE_MSR 7243------------------------------- 7244 7245:Architectures: x86 7246:Target: VM 7247:Parameters: args[0] contains the mask of KVM_MSR_EXIT_REASON_* events to report 7248:Returns: 0 on success; -1 on error 7249 7250This capability allows userspace to intercept RDMSR and WRMSR instructions if 7251access to an MSR is denied. By default, KVM injects #GP on denied accesses. 7252 7253When a guest requests to read or write an MSR, KVM may not implement all MSRs 7254that are relevant to a respective system. It also does not differentiate by 7255CPU type. 7256 7257To allow more fine grained control over MSR handling, userspace may enable 7258this capability. With it enabled, MSR accesses that match the mask specified in 7259args[0] and would trigger a #GP inside the guest will instead trigger 7260KVM_EXIT_X86_RDMSR and KVM_EXIT_X86_WRMSR exit notifications. Userspace 7261can then implement model specific MSR handling and/or user notifications 7262to inform a user that an MSR was not emulated/virtualized by KVM. 7263 7264The valid mask flags are: 7265 7266============================ =============================================== 7267 KVM_MSR_EXIT_REASON_UNKNOWN intercept accesses to unknown (to KVM) MSRs 7268 KVM_MSR_EXIT_REASON_INVAL intercept accesses that are architecturally 7269 invalid according to the vCPU model and/or mode 7270 KVM_MSR_EXIT_REASON_FILTER intercept accesses that are denied by userspace 7271 via KVM_X86_SET_MSR_FILTER 7272============================ =============================================== 7273 72747.22 KVM_CAP_X86_BUS_LOCK_EXIT 7275------------------------------- 7276 7277:Architectures: x86 7278:Target: VM 7279:Parameters: args[0] defines the policy used when bus locks detected in guest 7280:Returns: 0 on success, -EINVAL when args[0] contains invalid bits 7281 7282Valid bits in args[0] are:: 7283 7284 #define KVM_BUS_LOCK_DETECTION_OFF (1 << 0) 7285 #define KVM_BUS_LOCK_DETECTION_EXIT (1 << 1) 7286 7287Enabling this capability on a VM provides userspace with a way to select 7288a policy to handle the bus locks detected in guest. Userspace can obtain 7289the supported modes from the result of KVM_CHECK_EXTENSION and define it 7290through the KVM_ENABLE_CAP. 7291 7292KVM_BUS_LOCK_DETECTION_OFF and KVM_BUS_LOCK_DETECTION_EXIT are supported 7293currently and mutually exclusive with each other. More bits can be added in 7294the future. 7295 7296With KVM_BUS_LOCK_DETECTION_OFF set, bus locks in guest will not cause vm exits 7297so that no additional actions are needed. This is the default mode. 7298 7299With KVM_BUS_LOCK_DETECTION_EXIT set, vm exits happen when bus lock detected 7300in VM. KVM just exits to userspace when handling them. Userspace can enforce 7301its own throttling or other policy based mitigations. 7302 7303This capability is aimed to address the thread that VM can exploit bus locks to 7304degree the performance of the whole system. Once the userspace enable this 7305capability and select the KVM_BUS_LOCK_DETECTION_EXIT mode, KVM will set the 7306KVM_RUN_BUS_LOCK flag in vcpu-run->flags field and exit to userspace. Concerning 7307the bus lock vm exit can be preempted by a higher priority VM exit, the exit 7308notifications to userspace can be KVM_EXIT_BUS_LOCK or other reasons. 7309KVM_RUN_BUS_LOCK flag is used to distinguish between them. 7310 73117.23 KVM_CAP_PPC_DAWR1 7312---------------------- 7313 7314:Architectures: ppc 7315:Parameters: none 7316:Returns: 0 on success, -EINVAL when CPU doesn't support 2nd DAWR 7317 7318This capability can be used to check / enable 2nd DAWR feature provided 7319by POWER10 processor. 7320 7321 73227.24 KVM_CAP_VM_COPY_ENC_CONTEXT_FROM 7323------------------------------------- 7324 7325Architectures: x86 SEV enabled 7326Type: vm 7327Parameters: args[0] is the fd of the source vm 7328Returns: 0 on success; ENOTTY on error 7329 7330This capability enables userspace to copy encryption context from the vm 7331indicated by the fd to the vm this is called on. 7332 7333This is intended to support in-guest workloads scheduled by the host. This 7334allows the in-guest workload to maintain its own NPTs and keeps the two vms 7335from accidentally clobbering each other with interrupts and the like (separate 7336APIC/MSRs/etc). 7337 73387.25 KVM_CAP_SGX_ATTRIBUTE 7339-------------------------- 7340 7341:Architectures: x86 7342:Target: VM 7343:Parameters: args[0] is a file handle of a SGX attribute file in securityfs 7344:Returns: 0 on success, -EINVAL if the file handle is invalid or if a requested 7345 attribute is not supported by KVM. 7346 7347KVM_CAP_SGX_ATTRIBUTE enables a userspace VMM to grant a VM access to one or 7348more priveleged enclave attributes. args[0] must hold a file handle to a valid 7349SGX attribute file corresponding to an attribute that is supported/restricted 7350by KVM (currently only PROVISIONKEY). 7351 7352The SGX subsystem restricts access to a subset of enclave attributes to provide 7353additional security for an uncompromised kernel, e.g. use of the PROVISIONKEY 7354is restricted to deter malware from using the PROVISIONKEY to obtain a stable 7355system fingerprint. To prevent userspace from circumventing such restrictions 7356by running an enclave in a VM, KVM prevents access to privileged attributes by 7357default. 7358 7359See Documentation/x86/sgx.rst for more details. 7360 73617.26 KVM_CAP_PPC_RPT_INVALIDATE 7362------------------------------- 7363 7364:Capability: KVM_CAP_PPC_RPT_INVALIDATE 7365:Architectures: ppc 7366:Type: vm 7367 7368This capability indicates that the kernel is capable of handling 7369H_RPT_INVALIDATE hcall. 7370 7371In order to enable the use of H_RPT_INVALIDATE in the guest, 7372user space might have to advertise it for the guest. For example, 7373IBM pSeries (sPAPR) guest starts using it if "hcall-rpt-invalidate" is 7374present in the "ibm,hypertas-functions" device-tree property. 7375 7376This capability is enabled for hypervisors on platforms like POWER9 7377that support radix MMU. 7378 73797.27 KVM_CAP_EXIT_ON_EMULATION_FAILURE 7380-------------------------------------- 7381 7382:Architectures: x86 7383:Parameters: args[0] whether the feature should be enabled or not 7384 7385When this capability is enabled, an emulation failure will result in an exit 7386to userspace with KVM_INTERNAL_ERROR (except when the emulator was invoked 7387to handle a VMware backdoor instruction). Furthermore, KVM will now provide up 7388to 15 instruction bytes for any exit to userspace resulting from an emulation 7389failure. When these exits to userspace occur use the emulation_failure struct 7390instead of the internal struct. They both have the same layout, but the 7391emulation_failure struct matches the content better. It also explicitly 7392defines the 'flags' field which is used to describe the fields in the struct 7393that are valid (ie: if KVM_INTERNAL_ERROR_EMULATION_FLAG_INSTRUCTION_BYTES is 7394set in the 'flags' field then both 'insn_size' and 'insn_bytes' have valid data 7395in them.) 7396 73977.28 KVM_CAP_ARM_MTE 7398-------------------- 7399 7400:Architectures: arm64 7401:Parameters: none 7402 7403This capability indicates that KVM (and the hardware) supports exposing the 7404Memory Tagging Extensions (MTE) to the guest. It must also be enabled by the 7405VMM before creating any VCPUs to allow the guest access. Note that MTE is only 7406available to a guest running in AArch64 mode and enabling this capability will 7407cause attempts to create AArch32 VCPUs to fail. 7408 7409When enabled the guest is able to access tags associated with any memory given 7410to the guest. KVM will ensure that the tags are maintained during swap or 7411hibernation of the host; however the VMM needs to manually save/restore the 7412tags as appropriate if the VM is migrated. 7413 7414When this capability is enabled all memory in memslots must be mapped as 7415``MAP_ANONYMOUS`` or with a RAM-based file mapping (``tmpfs``, ``memfd``), 7416attempts to create a memslot with an invalid mmap will result in an 7417-EINVAL return. 7418 7419When enabled the VMM may make use of the ``KVM_ARM_MTE_COPY_TAGS`` ioctl to 7420perform a bulk copy of tags to/from the guest. 7421 74227.29 KVM_CAP_VM_MOVE_ENC_CONTEXT_FROM 7423------------------------------------- 7424 7425Architectures: x86 SEV enabled 7426Type: vm 7427Parameters: args[0] is the fd of the source vm 7428Returns: 0 on success 7429 7430This capability enables userspace to migrate the encryption context from the VM 7431indicated by the fd to the VM this is called on. 7432 7433This is intended to support intra-host migration of VMs between userspace VMMs, 7434upgrading the VMM process without interrupting the guest. 7435 74367.30 KVM_CAP_PPC_AIL_MODE_3 7437------------------------------- 7438 7439:Capability: KVM_CAP_PPC_AIL_MODE_3 7440:Architectures: ppc 7441:Type: vm 7442 7443This capability indicates that the kernel supports the mode 3 setting for the 7444"Address Translation Mode on Interrupt" aka "Alternate Interrupt Location" 7445resource that is controlled with the H_SET_MODE hypercall. 7446 7447This capability allows a guest kernel to use a better-performance mode for 7448handling interrupts and system calls. 7449 74507.31 KVM_CAP_DISABLE_QUIRKS2 7451---------------------------- 7452 7453:Capability: KVM_CAP_DISABLE_QUIRKS2 7454:Parameters: args[0] - set of KVM quirks to disable 7455:Architectures: x86 7456:Type: vm 7457 7458This capability, if enabled, will cause KVM to disable some behavior 7459quirks. 7460 7461Calling KVM_CHECK_EXTENSION for this capability returns a bitmask of 7462quirks that can be disabled in KVM. 7463 7464The argument to KVM_ENABLE_CAP for this capability is a bitmask of 7465quirks to disable, and must be a subset of the bitmask returned by 7466KVM_CHECK_EXTENSION. 7467 7468The valid bits in cap.args[0] are: 7469 7470=================================== ============================================ 7471 KVM_X86_QUIRK_LINT0_REENABLED By default, the reset value for the LVT 7472 LINT0 register is 0x700 (APIC_MODE_EXTINT). 7473 When this quirk is disabled, the reset value 7474 is 0x10000 (APIC_LVT_MASKED). 7475 7476 KVM_X86_QUIRK_CD_NW_CLEARED By default, KVM clears CR0.CD and CR0.NW. 7477 When this quirk is disabled, KVM does not 7478 change the value of CR0.CD and CR0.NW. 7479 7480 KVM_X86_QUIRK_LAPIC_MMIO_HOLE By default, the MMIO LAPIC interface is 7481 available even when configured for x2APIC 7482 mode. When this quirk is disabled, KVM 7483 disables the MMIO LAPIC interface if the 7484 LAPIC is in x2APIC mode. 7485 7486 KVM_X86_QUIRK_OUT_7E_INC_RIP By default, KVM pre-increments %rip before 7487 exiting to userspace for an OUT instruction 7488 to port 0x7e. When this quirk is disabled, 7489 KVM does not pre-increment %rip before 7490 exiting to userspace. 7491 7492 KVM_X86_QUIRK_MISC_ENABLE_NO_MWAIT When this quirk is disabled, KVM sets 7493 CPUID.01H:ECX[bit 3] (MONITOR/MWAIT) if 7494 IA32_MISC_ENABLE[bit 18] (MWAIT) is set. 7495 Additionally, when this quirk is disabled, 7496 KVM clears CPUID.01H:ECX[bit 3] if 7497 IA32_MISC_ENABLE[bit 18] is cleared. 7498 7499 KVM_X86_QUIRK_FIX_HYPERCALL_INSN By default, KVM rewrites guest 7500 VMMCALL/VMCALL instructions to match the 7501 vendor's hypercall instruction for the 7502 system. When this quirk is disabled, KVM 7503 will no longer rewrite invalid guest 7504 hypercall instructions. Executing the 7505 incorrect hypercall instruction will 7506 generate a #UD within the guest. 7507 7508KVM_X86_QUIRK_MWAIT_NEVER_UD_FAULTS By default, KVM emulates MONITOR/MWAIT (if 7509 they are intercepted) as NOPs regardless of 7510 whether or not MONITOR/MWAIT are supported 7511 according to guest CPUID. When this quirk 7512 is disabled and KVM_X86_DISABLE_EXITS_MWAIT 7513 is not set (MONITOR/MWAIT are intercepted), 7514 KVM will inject a #UD on MONITOR/MWAIT if 7515 they're unsupported per guest CPUID. Note, 7516 KVM will modify MONITOR/MWAIT support in 7517 guest CPUID on writes to MISC_ENABLE if 7518 KVM_X86_QUIRK_MISC_ENABLE_NO_MWAIT is 7519 disabled. 7520=================================== ============================================ 7521 75227.32 KVM_CAP_MAX_VCPU_ID 7523------------------------ 7524 7525:Architectures: x86 7526:Target: VM 7527:Parameters: args[0] - maximum APIC ID value set for current VM 7528:Returns: 0 on success, -EINVAL if args[0] is beyond KVM_MAX_VCPU_IDS 7529 supported in KVM or if it has been set. 7530 7531This capability allows userspace to specify maximum possible APIC ID 7532assigned for current VM session prior to the creation of vCPUs, saving 7533memory for data structures indexed by the APIC ID. Userspace is able 7534to calculate the limit to APIC ID values from designated 7535CPU topology. 7536 7537The value can be changed only until KVM_ENABLE_CAP is set to a nonzero 7538value or until a vCPU is created. Upon creation of the first vCPU, 7539if the value was set to zero or KVM_ENABLE_CAP was not invoked, KVM 7540uses the return value of KVM_CHECK_EXTENSION(KVM_CAP_MAX_VCPU_ID) as 7541the maximum APIC ID. 7542 75437.33 KVM_CAP_X86_NOTIFY_VMEXIT 7544------------------------------ 7545 7546:Architectures: x86 7547:Target: VM 7548:Parameters: args[0] is the value of notify window as well as some flags 7549:Returns: 0 on success, -EINVAL if args[0] contains invalid flags or notify 7550 VM exit is unsupported. 7551 7552Bits 63:32 of args[0] are used for notify window. 7553Bits 31:0 of args[0] are for some flags. Valid bits are:: 7554 7555 #define KVM_X86_NOTIFY_VMEXIT_ENABLED (1 << 0) 7556 #define KVM_X86_NOTIFY_VMEXIT_USER (1 << 1) 7557 7558This capability allows userspace to configure the notify VM exit on/off 7559in per-VM scope during VM creation. Notify VM exit is disabled by default. 7560When userspace sets KVM_X86_NOTIFY_VMEXIT_ENABLED bit in args[0], VMM will 7561enable this feature with the notify window provided, which will generate 7562a VM exit if no event window occurs in VM non-root mode for a specified of 7563time (notify window). 7564 7565If KVM_X86_NOTIFY_VMEXIT_USER is set in args[0], upon notify VM exits happen, 7566KVM would exit to userspace for handling. 7567 7568This capability is aimed to mitigate the threat that malicious VMs can 7569cause CPU stuck (due to event windows don't open up) and make the CPU 7570unavailable to host or other VMs. 7571 75728. Other capabilities. 7573====================== 7574 7575This section lists capabilities that give information about other 7576features of the KVM implementation. 7577 75788.1 KVM_CAP_PPC_HWRNG 7579--------------------- 7580 7581:Architectures: ppc 7582 7583This capability, if KVM_CHECK_EXTENSION indicates that it is 7584available, means that the kernel has an implementation of the 7585H_RANDOM hypercall backed by a hardware random-number generator. 7586If present, the kernel H_RANDOM handler can be enabled for guest use 7587with the KVM_CAP_PPC_ENABLE_HCALL capability. 7588 75898.2 KVM_CAP_HYPERV_SYNIC 7590------------------------ 7591 7592:Architectures: x86 7593 7594This capability, if KVM_CHECK_EXTENSION indicates that it is 7595available, means that the kernel has an implementation of the 7596Hyper-V Synthetic interrupt controller(SynIC). Hyper-V SynIC is 7597used to support Windows Hyper-V based guest paravirt drivers(VMBus). 7598 7599In order to use SynIC, it has to be activated by setting this 7600capability via KVM_ENABLE_CAP ioctl on the vcpu fd. Note that this 7601will disable the use of APIC hardware virtualization even if supported 7602by the CPU, as it's incompatible with SynIC auto-EOI behavior. 7603 76048.3 KVM_CAP_PPC_RADIX_MMU 7605------------------------- 7606 7607:Architectures: ppc 7608 7609This capability, if KVM_CHECK_EXTENSION indicates that it is 7610available, means that the kernel can support guests using the 7611radix MMU defined in Power ISA V3.00 (as implemented in the POWER9 7612processor). 7613 76148.4 KVM_CAP_PPC_HASH_MMU_V3 7615--------------------------- 7616 7617:Architectures: ppc 7618 7619This capability, if KVM_CHECK_EXTENSION indicates that it is 7620available, means that the kernel can support guests using the 7621hashed page table MMU defined in Power ISA V3.00 (as implemented in 7622the POWER9 processor), including in-memory segment tables. 7623 76248.5 KVM_CAP_MIPS_VZ 7625------------------- 7626 7627:Architectures: mips 7628 7629This capability, if KVM_CHECK_EXTENSION on the main kvm handle indicates that 7630it is available, means that full hardware assisted virtualization capabilities 7631of the hardware are available for use through KVM. An appropriate 7632KVM_VM_MIPS_* type must be passed to KVM_CREATE_VM to create a VM which 7633utilises it. 7634 7635If KVM_CHECK_EXTENSION on a kvm VM handle indicates that this capability is 7636available, it means that the VM is using full hardware assisted virtualization 7637capabilities of the hardware. This is useful to check after creating a VM with 7638KVM_VM_MIPS_DEFAULT. 7639 7640The value returned by KVM_CHECK_EXTENSION should be compared against known 7641values (see below). All other values are reserved. This is to allow for the 7642possibility of other hardware assisted virtualization implementations which 7643may be incompatible with the MIPS VZ ASE. 7644 7645== ========================================================================== 7646 0 The trap & emulate implementation is in use to run guest code in user 7647 mode. Guest virtual memory segments are rearranged to fit the guest in the 7648 user mode address space. 7649 7650 1 The MIPS VZ ASE is in use, providing full hardware assisted 7651 virtualization, including standard guest virtual memory segments. 7652== ========================================================================== 7653 76548.6 KVM_CAP_MIPS_TE 7655------------------- 7656 7657:Architectures: mips 7658 7659This capability, if KVM_CHECK_EXTENSION on the main kvm handle indicates that 7660it is available, means that the trap & emulate implementation is available to 7661run guest code in user mode, even if KVM_CAP_MIPS_VZ indicates that hardware 7662assisted virtualisation is also available. KVM_VM_MIPS_TE (0) must be passed 7663to KVM_CREATE_VM to create a VM which utilises it. 7664 7665If KVM_CHECK_EXTENSION on a kvm VM handle indicates that this capability is 7666available, it means that the VM is using trap & emulate. 7667 76688.7 KVM_CAP_MIPS_64BIT 7669---------------------- 7670 7671:Architectures: mips 7672 7673This capability indicates the supported architecture type of the guest, i.e. the 7674supported register and address width. 7675 7676The values returned when this capability is checked by KVM_CHECK_EXTENSION on a 7677kvm VM handle correspond roughly to the CP0_Config.AT register field, and should 7678be checked specifically against known values (see below). All other values are 7679reserved. 7680 7681== ======================================================================== 7682 0 MIPS32 or microMIPS32. 7683 Both registers and addresses are 32-bits wide. 7684 It will only be possible to run 32-bit guest code. 7685 7686 1 MIPS64 or microMIPS64 with access only to 32-bit compatibility segments. 7687 Registers are 64-bits wide, but addresses are 32-bits wide. 7688 64-bit guest code may run but cannot access MIPS64 memory segments. 7689 It will also be possible to run 32-bit guest code. 7690 7691 2 MIPS64 or microMIPS64 with access to all address segments. 7692 Both registers and addresses are 64-bits wide. 7693 It will be possible to run 64-bit or 32-bit guest code. 7694== ======================================================================== 7695 76968.9 KVM_CAP_ARM_USER_IRQ 7697------------------------ 7698 7699:Architectures: arm64 7700 7701This capability, if KVM_CHECK_EXTENSION indicates that it is available, means 7702that if userspace creates a VM without an in-kernel interrupt controller, it 7703will be notified of changes to the output level of in-kernel emulated devices, 7704which can generate virtual interrupts, presented to the VM. 7705For such VMs, on every return to userspace, the kernel 7706updates the vcpu's run->s.regs.device_irq_level field to represent the actual 7707output level of the device. 7708 7709Whenever kvm detects a change in the device output level, kvm guarantees at 7710least one return to userspace before running the VM. This exit could either 7711be a KVM_EXIT_INTR or any other exit event, like KVM_EXIT_MMIO. This way, 7712userspace can always sample the device output level and re-compute the state of 7713the userspace interrupt controller. Userspace should always check the state 7714of run->s.regs.device_irq_level on every kvm exit. 7715The value in run->s.regs.device_irq_level can represent both level and edge 7716triggered interrupt signals, depending on the device. Edge triggered interrupt 7717signals will exit to userspace with the bit in run->s.regs.device_irq_level 7718set exactly once per edge signal. 7719 7720The field run->s.regs.device_irq_level is available independent of 7721run->kvm_valid_regs or run->kvm_dirty_regs bits. 7722 7723If KVM_CAP_ARM_USER_IRQ is supported, the KVM_CHECK_EXTENSION ioctl returns a 7724number larger than 0 indicating the version of this capability is implemented 7725and thereby which bits in run->s.regs.device_irq_level can signal values. 7726 7727Currently the following bits are defined for the device_irq_level bitmap:: 7728 7729 KVM_CAP_ARM_USER_IRQ >= 1: 7730 7731 KVM_ARM_DEV_EL1_VTIMER - EL1 virtual timer 7732 KVM_ARM_DEV_EL1_PTIMER - EL1 physical timer 7733 KVM_ARM_DEV_PMU - ARM PMU overflow interrupt signal 7734 7735Future versions of kvm may implement additional events. These will get 7736indicated by returning a higher number from KVM_CHECK_EXTENSION and will be 7737listed above. 7738 77398.10 KVM_CAP_PPC_SMT_POSSIBLE 7740----------------------------- 7741 7742:Architectures: ppc 7743 7744Querying this capability returns a bitmap indicating the possible 7745virtual SMT modes that can be set using KVM_CAP_PPC_SMT. If bit N 7746(counting from the right) is set, then a virtual SMT mode of 2^N is 7747available. 7748 77498.11 KVM_CAP_HYPERV_SYNIC2 7750-------------------------- 7751 7752:Architectures: x86 7753 7754This capability enables a newer version of Hyper-V Synthetic interrupt 7755controller (SynIC). The only difference with KVM_CAP_HYPERV_SYNIC is that KVM 7756doesn't clear SynIC message and event flags pages when they are enabled by 7757writing to the respective MSRs. 7758 77598.12 KVM_CAP_HYPERV_VP_INDEX 7760---------------------------- 7761 7762:Architectures: x86 7763 7764This capability indicates that userspace can load HV_X64_MSR_VP_INDEX msr. Its 7765value is used to denote the target vcpu for a SynIC interrupt. For 7766compatibilty, KVM initializes this msr to KVM's internal vcpu index. When this 7767capability is absent, userspace can still query this msr's value. 7768 77698.13 KVM_CAP_S390_AIS_MIGRATION 7770------------------------------- 7771 7772:Architectures: s390 7773:Parameters: none 7774 7775This capability indicates if the flic device will be able to get/set the 7776AIS states for migration via the KVM_DEV_FLIC_AISM_ALL attribute and allows 7777to discover this without having to create a flic device. 7778 77798.14 KVM_CAP_S390_PSW 7780--------------------- 7781 7782:Architectures: s390 7783 7784This capability indicates that the PSW is exposed via the kvm_run structure. 7785 77868.15 KVM_CAP_S390_GMAP 7787---------------------- 7788 7789:Architectures: s390 7790 7791This capability indicates that the user space memory used as guest mapping can 7792be anywhere in the user memory address space, as long as the memory slots are 7793aligned and sized to a segment (1MB) boundary. 7794 77958.16 KVM_CAP_S390_COW 7796--------------------- 7797 7798:Architectures: s390 7799 7800This capability indicates that the user space memory used as guest mapping can 7801use copy-on-write semantics as well as dirty pages tracking via read-only page 7802tables. 7803 78048.17 KVM_CAP_S390_BPB 7805--------------------- 7806 7807:Architectures: s390 7808 7809This capability indicates that kvm will implement the interfaces to handle 7810reset, migration and nested KVM for branch prediction blocking. The stfle 7811facility 82 should not be provided to the guest without this capability. 7812 78138.18 KVM_CAP_HYPERV_TLBFLUSH 7814---------------------------- 7815 7816:Architectures: x86 7817 7818This capability indicates that KVM supports paravirtualized Hyper-V TLB Flush 7819hypercalls: 7820HvFlushVirtualAddressSpace, HvFlushVirtualAddressSpaceEx, 7821HvFlushVirtualAddressList, HvFlushVirtualAddressListEx. 7822 78238.19 KVM_CAP_ARM_INJECT_SERROR_ESR 7824---------------------------------- 7825 7826:Architectures: arm64 7827 7828This capability indicates that userspace can specify (via the 7829KVM_SET_VCPU_EVENTS ioctl) the syndrome value reported to the guest when it 7830takes a virtual SError interrupt exception. 7831If KVM advertises this capability, userspace can only specify the ISS field for 7832the ESR syndrome. Other parts of the ESR, such as the EC are generated by the 7833CPU when the exception is taken. If this virtual SError is taken to EL1 using 7834AArch64, this value will be reported in the ISS field of ESR_ELx. 7835 7836See KVM_CAP_VCPU_EVENTS for more details. 7837 78388.20 KVM_CAP_HYPERV_SEND_IPI 7839---------------------------- 7840 7841:Architectures: x86 7842 7843This capability indicates that KVM supports paravirtualized Hyper-V IPI send 7844hypercalls: 7845HvCallSendSyntheticClusterIpi, HvCallSendSyntheticClusterIpiEx. 7846 78478.21 KVM_CAP_HYPERV_DIRECT_TLBFLUSH 7848----------------------------------- 7849 7850:Architectures: x86 7851 7852This capability indicates that KVM running on top of Hyper-V hypervisor 7853enables Direct TLB flush for its guests meaning that TLB flush 7854hypercalls are handled by Level 0 hypervisor (Hyper-V) bypassing KVM. 7855Due to the different ABI for hypercall parameters between Hyper-V and 7856KVM, enabling this capability effectively disables all hypercall 7857handling by KVM (as some KVM hypercall may be mistakenly treated as TLB 7858flush hypercalls by Hyper-V) so userspace should disable KVM identification 7859in CPUID and only exposes Hyper-V identification. In this case, guest 7860thinks it's running on Hyper-V and only use Hyper-V hypercalls. 7861 78628.22 KVM_CAP_S390_VCPU_RESETS 7863----------------------------- 7864 7865:Architectures: s390 7866 7867This capability indicates that the KVM_S390_NORMAL_RESET and 7868KVM_S390_CLEAR_RESET ioctls are available. 7869 78708.23 KVM_CAP_S390_PROTECTED 7871--------------------------- 7872 7873:Architectures: s390 7874 7875This capability indicates that the Ultravisor has been initialized and 7876KVM can therefore start protected VMs. 7877This capability governs the KVM_S390_PV_COMMAND ioctl and the 7878KVM_MP_STATE_LOAD MP_STATE. KVM_SET_MP_STATE can fail for protected 7879guests when the state change is invalid. 7880 78818.24 KVM_CAP_STEAL_TIME 7882----------------------- 7883 7884:Architectures: arm64, x86 7885 7886This capability indicates that KVM supports steal time accounting. 7887When steal time accounting is supported it may be enabled with 7888architecture-specific interfaces. This capability and the architecture- 7889specific interfaces must be consistent, i.e. if one says the feature 7890is supported, than the other should as well and vice versa. For arm64 7891see Documentation/virt/kvm/devices/vcpu.rst "KVM_ARM_VCPU_PVTIME_CTRL". 7892For x86 see Documentation/virt/kvm/x86/msr.rst "MSR_KVM_STEAL_TIME". 7893 78948.25 KVM_CAP_S390_DIAG318 7895------------------------- 7896 7897:Architectures: s390 7898 7899This capability enables a guest to set information about its control program 7900(i.e. guest kernel type and version). The information is helpful during 7901system/firmware service events, providing additional data about the guest 7902environments running on the machine. 7903 7904The information is associated with the DIAGNOSE 0x318 instruction, which sets 7905an 8-byte value consisting of a one-byte Control Program Name Code (CPNC) and 7906a 7-byte Control Program Version Code (CPVC). The CPNC determines what 7907environment the control program is running in (e.g. Linux, z/VM...), and the 7908CPVC is used for information specific to OS (e.g. Linux version, Linux 7909distribution...) 7910 7911If this capability is available, then the CPNC and CPVC can be synchronized 7912between KVM and userspace via the sync regs mechanism (KVM_SYNC_DIAG318). 7913 79148.26 KVM_CAP_X86_USER_SPACE_MSR 7915------------------------------- 7916 7917:Architectures: x86 7918 7919This capability indicates that KVM supports deflection of MSR reads and 7920writes to user space. It can be enabled on a VM level. If enabled, MSR 7921accesses that would usually trigger a #GP by KVM into the guest will 7922instead get bounced to user space through the KVM_EXIT_X86_RDMSR and 7923KVM_EXIT_X86_WRMSR exit notifications. 7924 79258.27 KVM_CAP_X86_MSR_FILTER 7926--------------------------- 7927 7928:Architectures: x86 7929 7930This capability indicates that KVM supports that accesses to user defined MSRs 7931may be rejected. With this capability exposed, KVM exports new VM ioctl 7932KVM_X86_SET_MSR_FILTER which user space can call to specify bitmaps of MSR 7933ranges that KVM should deny access to. 7934 7935In combination with KVM_CAP_X86_USER_SPACE_MSR, this allows user space to 7936trap and emulate MSRs that are outside of the scope of KVM as well as 7937limit the attack surface on KVM's MSR emulation code. 7938 79398.28 KVM_CAP_ENFORCE_PV_FEATURE_CPUID 7940------------------------------------- 7941 7942Architectures: x86 7943 7944When enabled, KVM will disable paravirtual features provided to the 7945guest according to the bits in the KVM_CPUID_FEATURES CPUID leaf 7946(0x40000001). Otherwise, a guest may use the paravirtual features 7947regardless of what has actually been exposed through the CPUID leaf. 7948 79498.29 KVM_CAP_DIRTY_LOG_RING/KVM_CAP_DIRTY_LOG_RING_ACQ_REL 7950---------------------------------------------------------- 7951 7952:Architectures: x86, arm64 7953:Parameters: args[0] - size of the dirty log ring 7954 7955KVM is capable of tracking dirty memory using ring buffers that are 7956mmaped into userspace; there is one dirty ring per vcpu. 7957 7958The dirty ring is available to userspace as an array of 7959``struct kvm_dirty_gfn``. Each dirty entry it's defined as:: 7960 7961 struct kvm_dirty_gfn { 7962 __u32 flags; 7963 __u32 slot; /* as_id | slot_id */ 7964 __u64 offset; 7965 }; 7966 7967The following values are defined for the flags field to define the 7968current state of the entry:: 7969 7970 #define KVM_DIRTY_GFN_F_DIRTY BIT(0) 7971 #define KVM_DIRTY_GFN_F_RESET BIT(1) 7972 #define KVM_DIRTY_GFN_F_MASK 0x3 7973 7974Userspace should call KVM_ENABLE_CAP ioctl right after KVM_CREATE_VM 7975ioctl to enable this capability for the new guest and set the size of 7976the rings. Enabling the capability is only allowed before creating any 7977vCPU, and the size of the ring must be a power of two. The larger the 7978ring buffer, the less likely the ring is full and the VM is forced to 7979exit to userspace. The optimal size depends on the workload, but it is 7980recommended that it be at least 64 KiB (4096 entries). 7981 7982Just like for dirty page bitmaps, the buffer tracks writes to 7983all user memory regions for which the KVM_MEM_LOG_DIRTY_PAGES flag was 7984set in KVM_SET_USER_MEMORY_REGION. Once a memory region is registered 7985with the flag set, userspace can start harvesting dirty pages from the 7986ring buffer. 7987 7988An entry in the ring buffer can be unused (flag bits ``00``), 7989dirty (flag bits ``01``) or harvested (flag bits ``1X``). The 7990state machine for the entry is as follows:: 7991 7992 dirtied harvested reset 7993 00 -----------> 01 -------------> 1X -------+ 7994 ^ | 7995 | | 7996 +------------------------------------------+ 7997 7998To harvest the dirty pages, userspace accesses the mmaped ring buffer 7999to read the dirty GFNs. If the flags has the DIRTY bit set (at this stage 8000the RESET bit must be cleared), then it means this GFN is a dirty GFN. 8001The userspace should harvest this GFN and mark the flags from state 8002``01b`` to ``1Xb`` (bit 0 will be ignored by KVM, but bit 1 must be set 8003to show that this GFN is harvested and waiting for a reset), and move 8004on to the next GFN. The userspace should continue to do this until the 8005flags of a GFN have the DIRTY bit cleared, meaning that it has harvested 8006all the dirty GFNs that were available. 8007 8008Note that on weakly ordered architectures, userspace accesses to the 8009ring buffer (and more specifically the 'flags' field) must be ordered, 8010using load-acquire/store-release accessors when available, or any 8011other memory barrier that will ensure this ordering. 8012 8013It's not necessary for userspace to harvest the all dirty GFNs at once. 8014However it must collect the dirty GFNs in sequence, i.e., the userspace 8015program cannot skip one dirty GFN to collect the one next to it. 8016 8017After processing one or more entries in the ring buffer, userspace 8018calls the VM ioctl KVM_RESET_DIRTY_RINGS to notify the kernel about 8019it, so that the kernel will reprotect those collected GFNs. 8020Therefore, the ioctl must be called *before* reading the content of 8021the dirty pages. 8022 8023The dirty ring can get full. When it happens, the KVM_RUN of the 8024vcpu will return with exit reason KVM_EXIT_DIRTY_LOG_FULL. 8025 8026The dirty ring interface has a major difference comparing to the 8027KVM_GET_DIRTY_LOG interface in that, when reading the dirty ring from 8028userspace, it's still possible that the kernel has not yet flushed the 8029processor's dirty page buffers into the kernel buffer (with dirty bitmaps, the 8030flushing is done by the KVM_GET_DIRTY_LOG ioctl). To achieve that, one 8031needs to kick the vcpu out of KVM_RUN using a signal. The resulting 8032vmexit ensures that all dirty GFNs are flushed to the dirty rings. 8033 8034NOTE: KVM_CAP_DIRTY_LOG_RING_ACQ_REL is the only capability that 8035should be exposed by weakly ordered architecture, in order to indicate 8036the additional memory ordering requirements imposed on userspace when 8037reading the state of an entry and mutating it from DIRTY to HARVESTED. 8038Architecture with TSO-like ordering (such as x86) are allowed to 8039expose both KVM_CAP_DIRTY_LOG_RING and KVM_CAP_DIRTY_LOG_RING_ACQ_REL 8040to userspace. 8041 8042After enabling the dirty rings, the userspace needs to detect the 8043capability of KVM_CAP_DIRTY_LOG_RING_WITH_BITMAP to see whether the 8044ring structures can be backed by per-slot bitmaps. With this capability 8045advertised, it means the architecture can dirty guest pages without 8046vcpu/ring context, so that some of the dirty information will still be 8047maintained in the bitmap structure. KVM_CAP_DIRTY_LOG_RING_WITH_BITMAP 8048can't be enabled if the capability of KVM_CAP_DIRTY_LOG_RING_ACQ_REL 8049hasn't been enabled, or any memslot has been existing. 8050 8051Note that the bitmap here is only a backup of the ring structure. The 8052use of the ring and bitmap combination is only beneficial if there is 8053only a very small amount of memory that is dirtied out of vcpu/ring 8054context. Otherwise, the stand-alone per-slot bitmap mechanism needs to 8055be considered. 8056 8057To collect dirty bits in the backup bitmap, userspace can use the same 8058KVM_GET_DIRTY_LOG ioctl. KVM_CLEAR_DIRTY_LOG isn't needed as long as all 8059the generation of the dirty bits is done in a single pass. Collecting 8060the dirty bitmap should be the very last thing that the VMM does before 8061considering the state as complete. VMM needs to ensure that the dirty 8062state is final and avoid missing dirty pages from another ioctl ordered 8063after the bitmap collection. 8064 8065NOTE: One example of using the backup bitmap is saving arm64 vgic/its 8066tables through KVM_DEV_ARM_{VGIC_GRP_CTRL, ITS_SAVE_TABLES} command on 8067KVM device "kvm-arm-vgic-its" when dirty ring is enabled. 8068 80698.30 KVM_CAP_XEN_HVM 8070-------------------- 8071 8072:Architectures: x86 8073 8074This capability indicates the features that Xen supports for hosting Xen 8075PVHVM guests. Valid flags are:: 8076 8077 #define KVM_XEN_HVM_CONFIG_HYPERCALL_MSR (1 << 0) 8078 #define KVM_XEN_HVM_CONFIG_INTERCEPT_HCALL (1 << 1) 8079 #define KVM_XEN_HVM_CONFIG_SHARED_INFO (1 << 2) 8080 #define KVM_XEN_HVM_CONFIG_RUNSTATE (1 << 3) 8081 #define KVM_XEN_HVM_CONFIG_EVTCHN_2LEVEL (1 << 4) 8082 #define KVM_XEN_HVM_CONFIG_EVTCHN_SEND (1 << 5) 8083 #define KVM_XEN_HVM_CONFIG_RUNSTATE_UPDATE_FLAG (1 << 6) 8084 8085The KVM_XEN_HVM_CONFIG_HYPERCALL_MSR flag indicates that the KVM_XEN_HVM_CONFIG 8086ioctl is available, for the guest to set its hypercall page. 8087 8088If KVM_XEN_HVM_CONFIG_INTERCEPT_HCALL is also set, the same flag may also be 8089provided in the flags to KVM_XEN_HVM_CONFIG, without providing hypercall page 8090contents, to request that KVM generate hypercall page content automatically 8091and also enable interception of guest hypercalls with KVM_EXIT_XEN. 8092 8093The KVM_XEN_HVM_CONFIG_SHARED_INFO flag indicates the availability of the 8094KVM_XEN_HVM_SET_ATTR, KVM_XEN_HVM_GET_ATTR, KVM_XEN_VCPU_SET_ATTR and 8095KVM_XEN_VCPU_GET_ATTR ioctls, as well as the delivery of exception vectors 8096for event channel upcalls when the evtchn_upcall_pending field of a vcpu's 8097vcpu_info is set. 8098 8099The KVM_XEN_HVM_CONFIG_RUNSTATE flag indicates that the runstate-related 8100features KVM_XEN_VCPU_ATTR_TYPE_RUNSTATE_ADDR/_CURRENT/_DATA/_ADJUST are 8101supported by the KVM_XEN_VCPU_SET_ATTR/KVM_XEN_VCPU_GET_ATTR ioctls. 8102 8103The KVM_XEN_HVM_CONFIG_EVTCHN_2LEVEL flag indicates that IRQ routing entries 8104of the type KVM_IRQ_ROUTING_XEN_EVTCHN are supported, with the priority 8105field set to indicate 2 level event channel delivery. 8106 8107The KVM_XEN_HVM_CONFIG_EVTCHN_SEND flag indicates that KVM supports 8108injecting event channel events directly into the guest with the 8109KVM_XEN_HVM_EVTCHN_SEND ioctl. It also indicates support for the 8110KVM_XEN_ATTR_TYPE_EVTCHN/XEN_VERSION HVM attributes and the 8111KVM_XEN_VCPU_ATTR_TYPE_VCPU_ID/TIMER/UPCALL_VECTOR vCPU attributes. 8112related to event channel delivery, timers, and the XENVER_version 8113interception. 8114 8115The KVM_XEN_HVM_CONFIG_RUNSTATE_UPDATE_FLAG flag indicates that KVM supports 8116the KVM_XEN_ATTR_TYPE_RUNSTATE_UPDATE_FLAG attribute in the KVM_XEN_SET_ATTR 8117and KVM_XEN_GET_ATTR ioctls. This controls whether KVM will set the 8118XEN_RUNSTATE_UPDATE flag in guest memory mapped vcpu_runstate_info during 8119updates of the runstate information. Note that versions of KVM which support 8120the RUNSTATE feature above, but not thie RUNSTATE_UPDATE_FLAG feature, will 8121always set the XEN_RUNSTATE_UPDATE flag when updating the guest structure, 8122which is perhaps counterintuitive. When this flag is advertised, KVM will 8123behave more correctly, not using the XEN_RUNSTATE_UPDATE flag until/unless 8124specifically enabled (by the guest making the hypercall, causing the VMM 8125to enable the KVM_XEN_ATTR_TYPE_RUNSTATE_UPDATE_FLAG attribute). 8126 81278.31 KVM_CAP_PPC_MULTITCE 8128------------------------- 8129 8130:Capability: KVM_CAP_PPC_MULTITCE 8131:Architectures: ppc 8132:Type: vm 8133 8134This capability means the kernel is capable of handling hypercalls 8135H_PUT_TCE_INDIRECT and H_STUFF_TCE without passing those into the user 8136space. This significantly accelerates DMA operations for PPC KVM guests. 8137User space should expect that its handlers for these hypercalls 8138are not going to be called if user space previously registered LIOBN 8139in KVM (via KVM_CREATE_SPAPR_TCE or similar calls). 8140 8141In order to enable H_PUT_TCE_INDIRECT and H_STUFF_TCE use in the guest, 8142user space might have to advertise it for the guest. For example, 8143IBM pSeries (sPAPR) guest starts using them if "hcall-multi-tce" is 8144present in the "ibm,hypertas-functions" device-tree property. 8145 8146The hypercalls mentioned above may or may not be processed successfully 8147in the kernel based fast path. If they can not be handled by the kernel, 8148they will get passed on to user space. So user space still has to have 8149an implementation for these despite the in kernel acceleration. 8150 8151This capability is always enabled. 8152 81538.32 KVM_CAP_PTP_KVM 8154-------------------- 8155 8156:Architectures: arm64 8157 8158This capability indicates that the KVM virtual PTP service is 8159supported in the host. A VMM can check whether the service is 8160available to the guest on migration. 8161 81628.33 KVM_CAP_HYPERV_ENFORCE_CPUID 8163--------------------------------- 8164 8165Architectures: x86 8166 8167When enabled, KVM will disable emulated Hyper-V features provided to the 8168guest according to the bits Hyper-V CPUID feature leaves. Otherwise, all 8169currently implmented Hyper-V features are provided unconditionally when 8170Hyper-V identification is set in the HYPERV_CPUID_INTERFACE (0x40000001) 8171leaf. 8172 81738.34 KVM_CAP_EXIT_HYPERCALL 8174--------------------------- 8175 8176:Capability: KVM_CAP_EXIT_HYPERCALL 8177:Architectures: x86 8178:Type: vm 8179 8180This capability, if enabled, will cause KVM to exit to userspace 8181with KVM_EXIT_HYPERCALL exit reason to process some hypercalls. 8182 8183Calling KVM_CHECK_EXTENSION for this capability will return a bitmask 8184of hypercalls that can be configured to exit to userspace. 8185Right now, the only such hypercall is KVM_HC_MAP_GPA_RANGE. 8186 8187The argument to KVM_ENABLE_CAP is also a bitmask, and must be a subset 8188of the result of KVM_CHECK_EXTENSION. KVM will forward to userspace 8189the hypercalls whose corresponding bit is in the argument, and return 8190ENOSYS for the others. 8191 81928.35 KVM_CAP_PMU_CAPABILITY 8193--------------------------- 8194 8195:Capability KVM_CAP_PMU_CAPABILITY 8196:Architectures: x86 8197:Type: vm 8198:Parameters: arg[0] is bitmask of PMU virtualization capabilities. 8199:Returns 0 on success, -EINVAL when arg[0] contains invalid bits 8200 8201This capability alters PMU virtualization in KVM. 8202 8203Calling KVM_CHECK_EXTENSION for this capability returns a bitmask of 8204PMU virtualization capabilities that can be adjusted on a VM. 8205 8206The argument to KVM_ENABLE_CAP is also a bitmask and selects specific 8207PMU virtualization capabilities to be applied to the VM. This can 8208only be invoked on a VM prior to the creation of VCPUs. 8209 8210At this time, KVM_PMU_CAP_DISABLE is the only capability. Setting 8211this capability will disable PMU virtualization for that VM. Usermode 8212should adjust CPUID leaf 0xA to reflect that the PMU is disabled. 8213 82148.36 KVM_CAP_ARM_SYSTEM_SUSPEND 8215------------------------------- 8216 8217:Capability: KVM_CAP_ARM_SYSTEM_SUSPEND 8218:Architectures: arm64 8219:Type: vm 8220 8221When enabled, KVM will exit to userspace with KVM_EXIT_SYSTEM_EVENT of 8222type KVM_SYSTEM_EVENT_SUSPEND to process the guest suspend request. 8223 82248.37 KVM_CAP_S390_PROTECTED_DUMP 8225-------------------------------- 8226 8227:Capability: KVM_CAP_S390_PROTECTED_DUMP 8228:Architectures: s390 8229:Type: vm 8230 8231This capability indicates that KVM and the Ultravisor support dumping 8232PV guests. The `KVM_PV_DUMP` command is available for the 8233`KVM_S390_PV_COMMAND` ioctl and the `KVM_PV_INFO` command provides 8234dump related UV data. Also the vcpu ioctl `KVM_S390_PV_CPU_COMMAND` is 8235available and supports the `KVM_PV_DUMP_CPU` subcommand. 8236 82378.38 KVM_CAP_VM_DISABLE_NX_HUGE_PAGES 8238------------------------------------- 8239 8240:Capability: KVM_CAP_VM_DISABLE_NX_HUGE_PAGES 8241:Architectures: x86 8242:Type: vm 8243:Parameters: arg[0] must be 0. 8244:Returns: 0 on success, -EPERM if the userspace process does not 8245 have CAP_SYS_BOOT, -EINVAL if args[0] is not 0 or any vCPUs have been 8246 created. 8247 8248This capability disables the NX huge pages mitigation for iTLB MULTIHIT. 8249 8250The capability has no effect if the nx_huge_pages module parameter is not set. 8251 8252This capability may only be set before any vCPUs are created. 8253 82548.39 KVM_CAP_S390_CPU_TOPOLOGY 8255------------------------------ 8256 8257:Capability: KVM_CAP_S390_CPU_TOPOLOGY 8258:Architectures: s390 8259:Type: vm 8260 8261This capability indicates that KVM will provide the S390 CPU Topology 8262facility which consist of the interpretation of the PTF instruction for 8263the function code 2 along with interception and forwarding of both the 8264PTF instruction with function codes 0 or 1 and the STSI(15,1,x) 8265instruction to the userland hypervisor. 8266 8267The stfle facility 11, CPU Topology facility, should not be indicated 8268to the guest without this capability. 8269 8270When this capability is present, KVM provides a new attribute group 8271on vm fd, KVM_S390_VM_CPU_TOPOLOGY. 8272This new attribute allows to get, set or clear the Modified Change 8273Topology Report (MTCR) bit of the SCA through the kvm_device_attr 8274structure. 8275 8276When getting the Modified Change Topology Report value, the attr->addr 8277must point to a byte where the value will be stored or retrieved from. 8278 82799. Known KVM API problems 8280========================= 8281 8282In some cases, KVM's API has some inconsistencies or common pitfalls 8283that userspace need to be aware of. This section details some of 8284these issues. 8285 8286Most of them are architecture specific, so the section is split by 8287architecture. 8288 82899.1. x86 8290-------- 8291 8292``KVM_GET_SUPPORTED_CPUID`` issues 8293^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ 8294 8295In general, ``KVM_GET_SUPPORTED_CPUID`` is designed so that it is possible 8296to take its result and pass it directly to ``KVM_SET_CPUID2``. This section 8297documents some cases in which that requires some care. 8298 8299Local APIC features 8300~~~~~~~~~~~~~~~~~~~ 8301 8302CPU[EAX=1]:ECX[21] (X2APIC) is reported by ``KVM_GET_SUPPORTED_CPUID``, 8303but it can only be enabled if ``KVM_CREATE_IRQCHIP`` or 8304``KVM_ENABLE_CAP(KVM_CAP_IRQCHIP_SPLIT)`` are used to enable in-kernel emulation of 8305the local APIC. 8306 8307The same is true for the ``KVM_FEATURE_PV_UNHALT`` paravirtualized feature. 8308 8309CPU[EAX=1]:ECX[24] (TSC_DEADLINE) is not reported by ``KVM_GET_SUPPORTED_CPUID``. 8310It can be enabled if ``KVM_CAP_TSC_DEADLINE_TIMER`` is present and the kernel 8311has enabled in-kernel emulation of the local APIC. 8312 8313Obsolete ioctls and capabilities 8314^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ 8315 8316KVM_CAP_DISABLE_QUIRKS does not let userspace know which quirks are actually 8317available. Use ``KVM_CHECK_EXTENSION(KVM_CAP_DISABLE_QUIRKS2)`` instead if 8318available. 8319 8320Ordering of KVM_GET_*/KVM_SET_* ioctls 8321^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ 8322 8323TBD 8324