brintos / linux-shallow public Read only

1.. SPDX-License-Identifier: GPL-2.02 3===================================================================4The Definitive KVM (Kernel-based Virtual Machine) API Documentation5===================================================================6 71. General description8======================9 10The kvm API is a set of ioctls that are issued to control various aspects11of a virtual machine.  The ioctls belong to the following classes:12 13 - System ioctls: These query and set global attributes which affect the14   whole kvm subsystem.  In addition a system ioctl is used to create15   virtual machines.16 17 - VM ioctls: These query and set attributes that affect an entire virtual18   machine, for example memory layout.  In addition a VM ioctl is used to19   create virtual cpus (vcpus) and devices.20 21   VM ioctls must be issued from the same process (address space) that was22   used to create the VM.23 24 - vcpu ioctls: These query and set attributes that control the operation25   of a single virtual cpu.26 27   vcpu ioctls should be issued from the same thread that was used to create28   the vcpu, except for asynchronous vcpu ioctl that are marked as such in29   the documentation.  Otherwise, the first ioctl after switching threads30   could see a performance impact.31 32 - device ioctls: These query and set attributes that control the operation33   of a single device.34 35   device ioctls must be issued from the same process (address space) that36   was used to create the VM.37 382. File descriptors39===================40 41The kvm API is centered around file descriptors.  An initial42open("/dev/kvm") obtains a handle to the kvm subsystem; this handle43can be used to issue system ioctls.  A KVM_CREATE_VM ioctl on this44handle will create a VM file descriptor which can be used to issue VM45ioctls.  A KVM_CREATE_VCPU or KVM_CREATE_DEVICE ioctl on a VM fd will46create a virtual cpu or device and return a file descriptor pointing to47the new resource.  Finally, ioctls on a vcpu or device fd can be used48to control the vcpu or device.  For vcpus, this includes the important49task of actually running guest code.50 51In general file descriptors can be migrated among processes by means52of fork() and the SCM_RIGHTS facility of unix domain socket.  These53kinds of tricks are explicitly not supported by kvm.  While they will54not cause harm to the host, their actual behavior is not guaranteed by55the API.  See "General description" for details on the ioctl usage56model that is supported by KVM.57 58It is important to note that although VM ioctls may only be issued from59the process that created the VM, a VM's lifecycle is associated with its60file descriptor, not its creator (process).  In other words, the VM and61its resources, *including the associated address space*, are not freed62until 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 will64not be freed until both the parent (original) process and its child have65put their references to the VM's file descriptor.66 67Because a VM's resources are not freed until the last reference to its68file descriptor is released, creating additional references to a VM69via fork(), dup(), etc... without careful consideration is strongly70discouraged and may have unwanted side effects, e.g. memory allocated71by and on behalf of the VM's process may not be freed/unaccounted when72the VM is shut down.73 74 753. Extensions76=============77 78As of Linux 2.6.22, the KVM ABI has been stabilized: no backward79incompatible change are allowed.  However, there is an extension80facility that allows backward-compatible extensions to the API to be81queried and used.82 83The extension mechanism is not based on the Linux version number.84Instead, kvm defines extension identifiers and a facility to query85whether a particular extension identifier is available.  If it is, a86set of ioctls is available for application use.87 88 894. API description90==================91 92This section describes ioctls that can be used to control kvm guests.93For each ioctl, the following information is provided along with a94description: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 supports99      API version 12 (see section 4.1), a KVM_CAP_xyz constant, which100      means availability needs to be checked with KVM_CHECK_EXTENSION101      (see section 4.4), or 'none' which means that while not all kernels102      support this ioctl, there's no capability bit to check its103      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_VERSION122-----------------------123 124:Capability: basic125:Architectures: all126:Type: system ioctl127:Parameters: none128:Returns: the constant KVM_API_VERSION (=12)129 130This identifies the API version as the stable kvm API. It is not131expected that this number will change.  However, Linux 2.6.20 and1322.6.21 report earlier versions; these are not documented and not133supported.  Applications should refuse to run if KVM_GET_API_VERSION134returns a value other than 12.  If this check passes, all ioctls135described as 'basic' will be available.136 137 1384.2 KVM_CREATE_VM139-----------------140 141:Capability: basic142:Architectures: all143:Type: system ioctl144: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 150X86:151^^^^152 153Supported X86 VM types can be queried via KVM_CAP_VM_TYPES.154 155S390:156^^^^^157 158In order to create user controlled virtual machines on S390, check159KVM_CAP_S390_UCONTROL and use the flag KVM_VM_S390_UCONTROL as160privileged user (CAP_SYS_ADMIN).161 162MIPS:163^^^^^164 165To use hardware assisted virtualization on MIPS (VZ ASE) rather than166the default trap & emulate implementation (which changes the virtual167memory layout to fit in user mode), check KVM_CAP_MIPS_VZ and use the168flag KVM_VM_MIPS_VZ.169 170ARM64:171^^^^^^172 173On arm64, the physical address size for a VM (IPA Size limit) is limited174to 40bits by default. The limit can be configured if the host supports the175extension KVM_CAP_ARM_VM_IPA_SIZE. When supported, use176KVM_VM_TYPE_ARM_IPA_SIZE(IPA_Bits) to set the size in the machine type177identifier, where IPA_Bits is the maximum width of any physical178address used by the VM. The IPA_Bits is encoded in bits[7-0] of the179machine type identifier.180 181e.g, to configure a guest to use 48bit physical address size::182 183    vm_fd = ioctl(dev_fd, KVM_CREATE_VM, KVM_VM_TYPE_ARM_IPA_SIZE(48));184 185The requested size (IPA_Bits) must be:186 187 ==   =========================================================188  0   Implies default size, 40bits (for backward compatibility)189  N   Implies N bits, where N is a positive integer such that,190      32 <= N <= Host_IPA_Limit191 ==   =========================================================192 193Host_IPA_Limit is the maximum possible value for IPA_Bits on the host and194is dependent on the CPU capability and the kernel configuration. The limit can195be retrieved using KVM_CAP_ARM_VM_IPA_SIZE of the KVM_CHECK_EXTENSION196ioctl() at run-time.197 198Creation of the VM will fail if the requested IPA size (whether it is199implicit or explicit) is unsupported on the host.200 201Please note that configuring the IPA size does not affect the capability202exposed by the guest CPUs in ID_AA64MMFR0_EL1[PARange]. It only affects203size of the address translated by the stage2 level (guest physical to204host physical address translations).205 206 2074.3 KVM_GET_MSR_INDEX_LIST, KVM_GET_MSR_FEATURE_INDEX_LIST208----------------------------------------------------------209 210:Capability: basic, KVM_CAP_GET_MSR_FEATURES for KVM_GET_MSR_FEATURE_INDEX_LIST211:Architectures: x86212:Type: system ioctl213:Parameters: struct kvm_msr_list (in/out)214:Returns: 0 on success; -1 on error215 216Errors:217 218  ======     ============================================================219  EFAULT     the msr index list cannot be read from or written to220  E2BIG      the msr index list is too big to fit in the array specified by221             the user.222  ======     ============================================================223 224::225 226  struct kvm_msr_list {227	__u32 nmsrs; /* number of msrs in entries */228	__u32 indices[0];229  };230 231The user fills in the size of the indices array in nmsrs, and in return232kvm adjusts nmsrs to reflect the actual number of msrs and fills in the233indices array with their numbers.234 235KVM_GET_MSR_INDEX_LIST returns the guest msrs that are supported.  The list236varies by kvm version and host processor, but does not change otherwise.237 238Note: if kvm indicates supports MCE (KVM_CAP_MCE), then the MCE bank MSRs are239not returned in the MSR list, as different vcpus can have a different number240of banks, as set via the KVM_X86_SETUP_MCE ioctl.241 242KVM_GET_MSR_FEATURE_INDEX_LIST returns the list of MSRs that can be passed243to the KVM_GET_MSRS system ioctl.  This lets userspace probe host capabilities244and processor features that are exposed via MSRs (e.g., VMX capabilities).245This list also varies by kvm version and host processor, but does not change246otherwise.247 248 2494.4 KVM_CHECK_EXTENSION250-----------------------251 252:Capability: basic, KVM_CAP_CHECK_EXTENSION_VM for vm ioctl253:Architectures: all254:Type: system ioctl, vm ioctl255:Parameters: extension identifier (KVM_CAP_*)256:Returns: 0 if unsupported; 1 (or some other positive integer) if supported257 258The API allows the application to query about extensions to the core259kvm API.  Userspace passes an extension identifier (an integer) and260receives an integer that describes the extension availability.261Generally 0 means no and 1 means yes, but some extensions may report262additional information in the integer return value.263 264Based on their initialization different VMs may have different capabilities.265It is thus encouraged to use the vm ioctl to query for capabilities (available266with KVM_CAP_CHECK_EXTENSION_VM on the vm fd)267 2684.5 KVM_GET_VCPU_MMAP_SIZE269--------------------------270 271:Capability: basic272:Architectures: all273:Type: system ioctl274:Parameters: none275:Returns: size of vcpu mmap area, in bytes276 277The KVM_RUN ioctl (cf.) communicates with userspace via a shared278memory region.  This ioctl returns the size of that region.  See the279KVM_RUN documentation for details.280 281Besides the size of the KVM_RUN communication region, other areas of282the VCPU file descriptor can be mmap-ed, including:283 284- if KVM_CAP_COALESCED_MMIO is available, a page at285  KVM_COALESCED_MMIO_PAGE_OFFSET * PAGE_SIZE; for historical reasons,286  this page is included in the result of KVM_GET_VCPU_MMAP_SIZE.287  KVM_CAP_COALESCED_MMIO is not documented yet.288 289- if KVM_CAP_DIRTY_LOG_RING is available, a number of pages at290  KVM_DIRTY_LOG_PAGE_OFFSET * PAGE_SIZE.  For more information on291  KVM_CAP_DIRTY_LOG_RING, see section 8.3.292 293 2944.7 KVM_CREATE_VCPU295-------------------296 297:Capability: basic298:Architectures: all299:Type: vm ioctl300:Parameters: vcpu id (apic id on x86)301:Returns: vcpu fd on success, -1 on error302 303This API adds a vcpu to a virtual machine. No more than max_vcpus may be added.304The vcpu id is an integer in the range [0, max_vcpu_id).305 306The recommended max_vcpus value can be retrieved using the KVM_CAP_NR_VCPUS of307the KVM_CHECK_EXTENSION ioctl() at run-time.308The maximum possible value for max_vcpus can be retrieved using the309KVM_CAP_MAX_VCPUS of the KVM_CHECK_EXTENSION ioctl() at run-time.310 311If the KVM_CAP_NR_VCPUS does not exist, you should assume that max_vcpus is 4312cpus max.313If the KVM_CAP_MAX_VCPUS does not exist, you should assume that max_vcpus is314same as the value returned from KVM_CAP_NR_VCPUS.315 316The maximum possible value for max_vcpu_id can be retrieved using the317KVM_CAP_MAX_VCPU_ID of the KVM_CHECK_EXTENSION ioctl() at run-time.318 319If the KVM_CAP_MAX_VCPU_ID does not exist, you should assume that max_vcpu_id320is the same as the value returned from KVM_CAP_MAX_VCPUS.321 322On powerpc using book3s_hv mode, the vcpus are mapped onto virtual323threads in one or more virtual CPU cores.  (This is because the324hardware requires all the hardware threads in a CPU core to be in the325same partition.)  The KVM_CAP_PPC_SMT capability indicates the number326of vcpus per virtual core (vcore).  The vcore id is obtained by327dividing the vcpu id by the number of vcpus per vcore.  The vcpus in a328given vcore will always be in the same physical core as each other329(though that might be a different physical core from time to time).330Userspace can control the threading (SMT) mode of the guest by its331allocation of vcpu ids.  For example, if userspace wants332single-threaded guest vcpus, it should make all vcpu ids be a multiple333of the number of vcpus per vcore.334 335For virtual cpus that have been created with S390 user controlled virtual336machines, the resulting vcpu fd can be memory mapped at page offset337KVM_S390_SIE_PAGE_OFFSET in order to obtain a memory map of the virtual338cpu's hardware control block.339 340 3414.8 KVM_GET_DIRTY_LOG (vm ioctl)342--------------------------------343 344:Capability: basic345:Architectures: all346:Type: vm ioctl347:Parameters: struct kvm_dirty_log (in/out)348:Returns: 0 on success, -1 on error349 350::351 352  /* for KVM_GET_DIRTY_LOG */353  struct kvm_dirty_log {354	__u32 slot;355	__u32 padding;356	union {357		void __user *dirty_bitmap; /* one bit per page */358		__u64 padding;359	};360  };361 362Given a memory slot, return a bitmap containing any pages dirtied363since the last call to this ioctl.  Bit 0 is the first page in the364memory slot.  Ensure the entire structure is cleared to avoid padding365issues.366 367If KVM_CAP_MULTI_ADDRESS_SPACE is available, bits 16-31 of slot field specifies368the address space for which you want to return the dirty bitmap.  See369KVM_SET_USER_MEMORY_REGION for details on the usage of slot field.370 371The bits in the dirty bitmap are cleared before the ioctl returns, unless372KVM_CAP_MANUAL_DIRTY_LOG_PROTECT2 is enabled.  For more information,373see the description of the capability.374 375Note that the Xen shared_info page, if configured, shall always be assumed376to be dirty. KVM will not explicitly mark it such.377 378 3794.10 KVM_RUN380------------381 382:Capability: basic383:Architectures: all384:Type: vcpu ioctl385:Parameters: none386:Returns: 0 on success, -1 on error387 388Errors:389 390  =======    ==============================================================391  EINTR      an unmasked signal is pending392  ENOEXEC    the vcpu hasn't been initialized or the guest tried to execute393             instructions from device memory (arm64)394  ENOSYS     data abort outside memslots with no syndrome info and395             KVM_CAP_ARM_NISV_TO_USER not enabled (arm64)396  EPERM      SVE feature set but not finalized (arm64)397  =======    ==============================================================398 399This ioctl is used to run a guest virtual cpu.  While there are no400explicit parameters, there is an implicit parameter block that can be401obtained by mmap()ing the vcpu fd at offset 0, with the size given by402KVM_GET_VCPU_MMAP_SIZE.  The parameter block is formatted as a 'struct403kvm_run' (see below).404 405 4064.11 KVM_GET_REGS407-----------------408 409:Capability: basic410:Architectures: all except arm64411:Type: vcpu ioctl412:Parameters: struct kvm_regs (out)413:Returns: 0 on success, -1 on error414 415Reads the general purpose registers from the vcpu.416 417::418 419  /* x86 */420  struct kvm_regs {421	/* out (KVM_GET_REGS) / in (KVM_SET_REGS) */422	__u64 rax, rbx, rcx, rdx;423	__u64 rsi, rdi, rsp, rbp;424	__u64 r8,  r9,  r10, r11;425	__u64 r12, r13, r14, r15;426	__u64 rip, rflags;427  };428 429  /* mips */430  struct kvm_regs {431	/* out (KVM_GET_REGS) / in (KVM_SET_REGS) */432	__u64 gpr[32];433	__u64 hi;434	__u64 lo;435	__u64 pc;436  };437 438  /* LoongArch */439  struct kvm_regs {440	/* out (KVM_GET_REGS) / in (KVM_SET_REGS) */441	unsigned long gpr[32];442	unsigned long pc;443  };444 445 4464.12 KVM_SET_REGS447-----------------448 449:Capability: basic450:Architectures: all except arm64451:Type: vcpu ioctl452:Parameters: struct kvm_regs (in)453:Returns: 0 on success, -1 on error454 455Writes the general purpose registers into the vcpu.456 457See KVM_GET_REGS for the data structure.458 459 4604.13 KVM_GET_SREGS461------------------462 463:Capability: basic464:Architectures: x86, ppc465:Type: vcpu ioctl466:Parameters: struct kvm_sregs (out)467:Returns: 0 on success, -1 on error468 469Reads special registers from the vcpu.470 471::472 473  /* x86 */474  struct kvm_sregs {475	struct kvm_segment cs, ds, es, fs, gs, ss;476	struct kvm_segment tr, ldt;477	struct kvm_dtable gdt, idt;478	__u64 cr0, cr2, cr3, cr4, cr8;479	__u64 efer;480	__u64 apic_base;481	__u64 interrupt_bitmap[(KVM_NR_INTERRUPTS + 63) / 64];482  };483 484  /* ppc -- see arch/powerpc/include/uapi/asm/kvm.h */485 486interrupt_bitmap is a bitmap of pending external interrupts.  At most487one bit may be set.  This interrupt has been acknowledged by the APIC488but not yet injected into the cpu core.489 490 4914.14 KVM_SET_SREGS492------------------493 494:Capability: basic495:Architectures: x86, ppc496:Type: vcpu ioctl497:Parameters: struct kvm_sregs (in)498:Returns: 0 on success, -1 on error499 500Writes special registers into the vcpu.  See KVM_GET_SREGS for the501data structures.502 503 5044.15 KVM_TRANSLATE505------------------506 507:Capability: basic508:Architectures: x86509:Type: vcpu ioctl510:Parameters: struct kvm_translation (in/out)511:Returns: 0 on success, -1 on error512 513Translates a virtual address according to the vcpu's current address514translation mode.515 516::517 518  struct kvm_translation {519	/* in */520	__u64 linear_address;521 522	/* out */523	__u64 physical_address;524	__u8  valid;525	__u8  writeable;526	__u8  usermode;527	__u8  pad[5];528  };529 530 5314.16 KVM_INTERRUPT532------------------533 534:Capability: basic535:Architectures: x86, ppc, mips, riscv, loongarch536:Type: vcpu ioctl537:Parameters: struct kvm_interrupt (in)538:Returns: 0 on success, negative on failure.539 540Queues a hardware interrupt vector to be injected.541 542::543 544  /* for KVM_INTERRUPT */545  struct kvm_interrupt {546	/* in */547	__u32 irq;548  };549 550X86:551^^^^552 553:Returns:554 555	========= ===================================556	  0       on success,557	 -EEXIST  if an interrupt is already enqueued558	 -EINVAL  the irq number is invalid559	 -ENXIO   if the PIC is in the kernel560	 -EFAULT  if the pointer is invalid561	========= ===================================562 563Note 'irq' is an interrupt vector, not an interrupt pin or line. This564ioctl is useful if the in-kernel PIC is not used.565 566PPC:567^^^^568 569Queues an external interrupt to be injected. This ioctl is overloaded570with 3 different irq values:571 572a) KVM_INTERRUPT_SET573 574   This injects an edge type external interrupt into the guest once it's ready575   to receive interrupts. When injected, the interrupt is done.576 577b) KVM_INTERRUPT_UNSET578 579   This unsets any pending interrupt.580 581   Only available with KVM_CAP_PPC_UNSET_IRQ.582 583c) KVM_INTERRUPT_SET_LEVEL584 585   This injects a level type external interrupt into the guest context. The586   interrupt stays pending until a specific ioctl with KVM_INTERRUPT_UNSET587   is triggered.588 589   Only available with KVM_CAP_PPC_IRQ_LEVEL.590 591Note that any value for 'irq' other than the ones stated above is invalid592and incurs unexpected behavior.593 594This is an asynchronous vcpu ioctl and can be invoked from any thread.595 596MIPS:597^^^^^598 599Queues an external interrupt to be injected into the virtual CPU. A negative600interrupt number dequeues the interrupt.601 602This is an asynchronous vcpu ioctl and can be invoked from any thread.603 604RISC-V:605^^^^^^^606 607Queues an external interrupt to be injected into the virtual CPU. This ioctl608is overloaded with 2 different irq values:609 610a) KVM_INTERRUPT_SET611 612   This sets external interrupt for a virtual CPU and it will receive613   once it is ready.614 615b) KVM_INTERRUPT_UNSET616 617   This clears pending external interrupt for a virtual CPU.618 619This is an asynchronous vcpu ioctl and can be invoked from any thread.620 621LOONGARCH:622^^^^^^^^^^623 624Queues an external interrupt to be injected into the virtual CPU. A negative625interrupt number dequeues the interrupt.626 627This is an asynchronous vcpu ioctl and can be invoked from any thread.628 629 6304.18 KVM_GET_MSRS631-----------------632 633:Capability: basic (vcpu), KVM_CAP_GET_MSR_FEATURES (system)634:Architectures: x86635:Type: system ioctl, vcpu ioctl636:Parameters: struct kvm_msrs (in/out)637:Returns: number of msrs successfully returned;638          -1 on error639 640When used as a system ioctl:641Reads the values of MSR-based features that are available for the VM.  This642is similar to KVM_GET_SUPPORTED_CPUID, but it returns MSR indices and values.643The list of msr-based features can be obtained using KVM_GET_MSR_FEATURE_INDEX_LIST644in a system ioctl.645 646When used as a vcpu ioctl:647Reads model-specific registers from the vcpu.  Supported msr indices can648be obtained using KVM_GET_MSR_INDEX_LIST in a system ioctl.649 650::651 652  struct kvm_msrs {653	__u32 nmsrs; /* number of msrs in entries */654	__u32 pad;655 656	struct kvm_msr_entry entries[0];657  };658 659  struct kvm_msr_entry {660	__u32 index;661	__u32 reserved;662	__u64 data;663  };664 665Application code should set the 'nmsrs' member (which indicates the666size of the entries array) and the 'index' member of each array entry.667kvm will fill in the 'data' member.668 669 6704.19 KVM_SET_MSRS671-----------------672 673:Capability: basic674:Architectures: x86675:Type: vcpu ioctl676:Parameters: struct kvm_msrs (in)677:Returns: number of msrs successfully set (see below), -1 on error678 679Writes model-specific registers to the vcpu.  See KVM_GET_MSRS for the680data structures.681 682Application code should set the 'nmsrs' member (which indicates the683size of the entries array), and the 'index' and 'data' members of each684array entry.685 686It tries to set the MSRs in array entries[] one by one. If setting an MSR687fails, e.g., due to setting reserved bits, the MSR isn't supported/emulated688by KVM, etc..., it stops processing the MSR list and returns the number of689MSRs that have been set successfully.690 691 6924.20 KVM_SET_CPUID693------------------694 695:Capability: basic696:Architectures: x86697:Type: vcpu ioctl698:Parameters: struct kvm_cpuid (in)699:Returns: 0 on success, -1 on error700 701Defines the vcpu responses to the cpuid instruction.  Applications702should use the KVM_SET_CPUID2 ioctl if available.703 704Caveat emptor:705  - If this IOCTL fails, KVM gives no guarantees that previous valid CPUID706    configuration (if there is) is not corrupted. Userspace can get a copy707    of the resulting CPUID configuration through KVM_GET_CPUID2 in case.708  - Using KVM_SET_CPUID{,2} after KVM_RUN, i.e. changing the guest vCPU model709    after running the guest, may cause guest instability.710  - Using heterogeneous CPUID configurations, modulo APIC IDs, topology, etc...711    may cause guest instability.712 713::714 715  struct kvm_cpuid_entry {716	__u32 function;717	__u32 eax;718	__u32 ebx;719	__u32 ecx;720	__u32 edx;721	__u32 padding;722  };723 724  /* for KVM_SET_CPUID */725  struct kvm_cpuid {726	__u32 nent;727	__u32 padding;728	struct kvm_cpuid_entry entries[0];729  };730 731 7324.21 KVM_SET_SIGNAL_MASK733------------------------734 735:Capability: basic736:Architectures: all737:Type: vcpu ioctl738:Parameters: struct kvm_signal_mask (in)739:Returns: 0 on success, -1 on error740 741Defines which signals are blocked during execution of KVM_RUN.  This742signal mask temporarily overrides the threads signal mask.  Any743unblocked signal received (except SIGKILL and SIGSTOP, which retain744their traditional behaviour) will cause KVM_RUN to return with -EINTR.745 746Note the signal will only be delivered if not blocked by the original747signal mask.748 749::750 751  /* for KVM_SET_SIGNAL_MASK */752  struct kvm_signal_mask {753	__u32 len;754	__u8  sigset[0];755  };756 757 7584.22 KVM_GET_FPU759----------------760 761:Capability: basic762:Architectures: x86, loongarch763:Type: vcpu ioctl764:Parameters: struct kvm_fpu (out)765:Returns: 0 on success, -1 on error766 767Reads the floating point state from the vcpu.768 769::770 771  /* x86: for KVM_GET_FPU and KVM_SET_FPU */772  struct kvm_fpu {773	__u8  fpr[8][16];774	__u16 fcw;775	__u16 fsw;776	__u8  ftwx;  /* in fxsave format */777	__u8  pad1;778	__u16 last_opcode;779	__u64 last_ip;780	__u64 last_dp;781	__u8  xmm[16][16];782	__u32 mxcsr;783	__u32 pad2;784  };785 786  /* LoongArch: for KVM_GET_FPU and KVM_SET_FPU */787  struct kvm_fpu {788	__u32 fcsr;789	__u64 fcc;790	struct kvm_fpureg {791		__u64 val64[4];792	}fpr[32];793  };794 795 7964.23 KVM_SET_FPU797----------------798 799:Capability: basic800:Architectures: x86, loongarch801:Type: vcpu ioctl802:Parameters: struct kvm_fpu (in)803:Returns: 0 on success, -1 on error804 805Writes the floating point state to the vcpu.806 807::808 809  /* x86: for KVM_GET_FPU and KVM_SET_FPU */810  struct kvm_fpu {811	__u8  fpr[8][16];812	__u16 fcw;813	__u16 fsw;814	__u8  ftwx;  /* in fxsave format */815	__u8  pad1;816	__u16 last_opcode;817	__u64 last_ip;818	__u64 last_dp;819	__u8  xmm[16][16];820	__u32 mxcsr;821	__u32 pad2;822  };823 824  /* LoongArch: for KVM_GET_FPU and KVM_SET_FPU */825  struct kvm_fpu {826	__u32 fcsr;827	__u64 fcc;828	struct kvm_fpureg {829		__u64 val64[4];830	}fpr[32];831  };832 833 8344.24 KVM_CREATE_IRQCHIP835-----------------------836 837:Capability: KVM_CAP_IRQCHIP, KVM_CAP_S390_IRQCHIP (s390)838:Architectures: x86, arm64, s390839:Type: vm ioctl840:Parameters: none841:Returns: 0 on success, -1 on error842 843Creates an interrupt controller model in the kernel.844On x86, creates a virtual ioapic, a virtual PIC (two PICs, nested), and sets up845future vcpus to have a local APIC.  IRQ routing for GSIs 0-15 is set to both846PIC and IOAPIC; GSI 16-23 only go to the IOAPIC.847On arm64, a GICv2 is created. Any other GIC versions require the usage of848KVM_CREATE_DEVICE, which also supports creating a GICv2.  Using849KVM_CREATE_DEVICE is preferred over KVM_CREATE_IRQCHIP for GICv2.850On s390, a dummy irq routing table is created.851 852Note that on s390 the KVM_CAP_S390_IRQCHIP vm capability needs to be enabled853before KVM_CREATE_IRQCHIP can be used.854 855 8564.25 KVM_IRQ_LINE857-----------------858 859:Capability: KVM_CAP_IRQCHIP860:Architectures: x86, arm64861:Type: vm ioctl862:Parameters: struct kvm_irq_level863:Returns: 0 on success, -1 on error864 865Sets the level of a GSI input to the interrupt controller model in the kernel.866On some architectures it is required that an interrupt controller model has867been previously created with KVM_CREATE_IRQCHIP.  Note that edge-triggered868interrupts require the level to be set to 1 and then back to 0.869 870On real hardware, interrupt pins can be active-low or active-high.  This871does not matter for the level field of struct kvm_irq_level: 1 always872means active (asserted), 0 means inactive (deasserted).873 874x86 allows the operating system to program the interrupt polarity875(active-low/active-high) for level-triggered interrupts, and KVM used876to consider the polarity.  However, due to bitrot in the handling of877active-low interrupts, the above convention is now valid on x86 too.878This is signaled by KVM_CAP_X86_IOAPIC_POLARITY_IGNORED.  Userspace879should not present interrupts to the guest as active-low unless this880capability is present (or unless it is not using the in-kernel irqchip,881of course).882 883 884arm64 can signal an interrupt either at the CPU level, or at the885in-kernel irqchip (GIC), and for in-kernel irqchip can tell the GIC to886use PPIs designated for specific cpus.  The irq field is interpreted887like this::888 889  bits:  |  31 ... 28  | 27 ... 24 | 23  ... 16 | 15 ... 0 |890  field: | vcpu2_index | irq_type  | vcpu_index |  irq_id  |891 892The irq_type field has the following values:893 894- KVM_ARM_IRQ_TYPE_CPU:895	       out-of-kernel GIC: irq_id 0 is IRQ, irq_id 1 is FIQ896- KVM_ARM_IRQ_TYPE_SPI:897	       in-kernel GIC: SPI, irq_id between 32 and 1019 (incl.)898               (the vcpu_index field is ignored)899- KVM_ARM_IRQ_TYPE_PPI:900	       in-kernel GIC: PPI, irq_id between 16 and 31 (incl.)901 902(The irq_id field thus corresponds nicely to the IRQ ID in the ARM GIC specs)903 904In both cases, level is used to assert/deassert the line.905 906When KVM_CAP_ARM_IRQ_LINE_LAYOUT_2 is supported, the target vcpu is907identified as (256 * vcpu2_index + vcpu_index). Otherwise, vcpu2_index908must be zero.909 910Note that on arm64, the KVM_CAP_IRQCHIP capability only conditions911injection of interrupts for the in-kernel irqchip. KVM_IRQ_LINE can always912be used for a userspace interrupt controller.913 914::915 916  struct kvm_irq_level {917	union {918		__u32 irq;     /* GSI */919		__s32 status;  /* not used for KVM_IRQ_LEVEL */920	};921	__u32 level;           /* 0 or 1 */922  };923 924 9254.26 KVM_GET_IRQCHIP926--------------------927 928:Capability: KVM_CAP_IRQCHIP929:Architectures: x86930:Type: vm ioctl931:Parameters: struct kvm_irqchip (in/out)932:Returns: 0 on success, -1 on error933 934Reads the state of a kernel interrupt controller created with935KVM_CREATE_IRQCHIP into a buffer provided by the caller.936 937::938 939  struct kvm_irqchip {940	__u32 chip_id;  /* 0 = PIC1, 1 = PIC2, 2 = IOAPIC */941	__u32 pad;942        union {943		char dummy[512];  /* reserving space */944		struct kvm_pic_state pic;945		struct kvm_ioapic_state ioapic;946	} chip;947  };948 949 9504.27 KVM_SET_IRQCHIP951--------------------952 953:Capability: KVM_CAP_IRQCHIP954:Architectures: x86955:Type: vm ioctl956:Parameters: struct kvm_irqchip (in)957:Returns: 0 on success, -1 on error958 959Sets the state of a kernel interrupt controller created with960KVM_CREATE_IRQCHIP from a buffer provided by the caller.961 962::963 964  struct kvm_irqchip {965	__u32 chip_id;  /* 0 = PIC1, 1 = PIC2, 2 = IOAPIC */966	__u32 pad;967        union {968		char dummy[512];  /* reserving space */969		struct kvm_pic_state pic;970		struct kvm_ioapic_state ioapic;971	} chip;972  };973 974 9754.28 KVM_XEN_HVM_CONFIG976-----------------------977 978:Capability: KVM_CAP_XEN_HVM979:Architectures: x86980:Type: vm ioctl981:Parameters: struct kvm_xen_hvm_config (in)982:Returns: 0 on success, -1 on error983 984Sets the MSR that the Xen HVM guest uses to initialize its hypercall985page, and provides the starting address and size of the hypercall986blobs in userspace.  When the guest writes the MSR, kvm copies one987page of a blob (32- or 64-bit, depending on the vcpu mode) to guest988memory.989 990::991 992  struct kvm_xen_hvm_config {993	__u32 flags;994	__u32 msr;995	__u64 blob_addr_32;996	__u64 blob_addr_64;997	__u8 blob_size_32;998	__u8 blob_size_64;999	__u8 pad2[30];1000  };1001 1002If certain flags are returned from the KVM_CAP_XEN_HVM check, they may1003be set in the flags field of this ioctl:1004 1005The KVM_XEN_HVM_CONFIG_INTERCEPT_HCALL flag requests KVM to generate1006the contents of the hypercall page automatically; hypercalls will be1007intercepted and passed to userspace through KVM_EXIT_XEN.  In this1008case, all of the blob size and address fields must be zero.1009 1010The KVM_XEN_HVM_CONFIG_EVTCHN_SEND flag indicates to KVM that userspace1011will always use the KVM_XEN_HVM_EVTCHN_SEND ioctl to deliver event1012channel interrupts rather than manipulating the guest's shared_info1013structures directly. This, in turn, may allow KVM to enable features1014such as intercepting the SCHEDOP_poll hypercall to accelerate PV1015spinlock operation for the guest. Userspace may still use the ioctl1016to deliver events if it was advertised, even if userspace does not1017send this indication that it will always do so1018 1019No other flags are currently valid in the struct kvm_xen_hvm_config.1020 10214.29 KVM_GET_CLOCK1022------------------1023 1024:Capability: KVM_CAP_ADJUST_CLOCK1025:Architectures: x861026:Type: vm ioctl1027:Parameters: struct kvm_clock_data (out)1028:Returns: 0 on success, -1 on error1029 1030Gets the current timestamp of kvmclock as seen by the current guest. In1031conjunction with KVM_SET_CLOCK, it is used to ensure monotonicity on scenarios1032such as migration.1033 1034When KVM_CAP_ADJUST_CLOCK is passed to KVM_CHECK_EXTENSION, it returns the1035set of bits that KVM can return in struct kvm_clock_data's flag member.1036 1037The following flags are defined:1038 1039KVM_CLOCK_TSC_STABLE1040  If set, the returned value is the exact kvmclock1041  value seen by all VCPUs at the instant when KVM_GET_CLOCK was called.1042  If clear, the returned value is simply CLOCK_MONOTONIC plus a constant1043  offset; the offset can be modified with KVM_SET_CLOCK.  KVM will try1044  to make all VCPUs follow this clock, but the exact value read by each1045  VCPU could differ, because the host TSC is not stable.1046 1047KVM_CLOCK_REALTIME1048  If set, the `realtime` field in the kvm_clock_data1049  structure is populated with the value of the host's real time1050  clocksource at the instant when KVM_GET_CLOCK was called. If clear,1051  the `realtime` field does not contain a value.1052 1053KVM_CLOCK_HOST_TSC1054  If set, the `host_tsc` field in the kvm_clock_data1055  structure is populated with the value of the host's timestamp counter (TSC)1056  at the instant when KVM_GET_CLOCK was called. If clear, the `host_tsc` field1057  does not contain a value.1058 1059::1060 1061  struct kvm_clock_data {1062	__u64 clock;  /* kvmclock current value */1063	__u32 flags;1064	__u32 pad0;1065	__u64 realtime;1066	__u64 host_tsc;1067	__u32 pad[4];1068  };1069 1070 10714.30 KVM_SET_CLOCK1072------------------1073 1074:Capability: KVM_CAP_ADJUST_CLOCK1075:Architectures: x861076:Type: vm ioctl1077:Parameters: struct kvm_clock_data (in)1078:Returns: 0 on success, -1 on error1079 1080Sets the current timestamp of kvmclock to the value specified in its parameter.1081In conjunction with KVM_GET_CLOCK, it is used to ensure monotonicity on scenarios1082such as migration.1083 1084The following flags can be passed:1085 1086KVM_CLOCK_REALTIME1087  If set, KVM will compare the value of the `realtime` field1088  with the value of the host's real time clocksource at the instant when1089  KVM_SET_CLOCK was called. The difference in elapsed time is added to the final1090  kvmclock value that will be provided to guests.1091 1092Other flags returned by ``KVM_GET_CLOCK`` are accepted but ignored.1093 1094::1095 1096  struct kvm_clock_data {1097	__u64 clock;  /* kvmclock current value */1098	__u32 flags;1099	__u32 pad0;1100	__u64 realtime;1101	__u64 host_tsc;1102	__u32 pad[4];1103  };1104 1105 11064.31 KVM_GET_VCPU_EVENTS1107------------------------1108 1109:Capability: KVM_CAP_VCPU_EVENTS1110:Extended by: KVM_CAP_INTR_SHADOW1111:Architectures: x86, arm641112:Type: vcpu ioctl1113:Parameters: struct kvm_vcpu_events (out)1114:Returns: 0 on success, -1 on error1115 1116X86:1117^^^^1118 1119Gets currently pending exceptions, interrupts, and NMIs as well as related1120states of the vcpu.1121 1122::1123 1124  struct kvm_vcpu_events {1125	struct {1126		__u8 injected;1127		__u8 nr;1128		__u8 has_error_code;1129		__u8 pending;1130		__u32 error_code;1131	} exception;1132	struct {1133		__u8 injected;1134		__u8 nr;1135		__u8 soft;1136		__u8 shadow;1137	} interrupt;1138	struct {1139		__u8 injected;1140		__u8 pending;1141		__u8 masked;1142		__u8 pad;1143	} nmi;1144	__u32 sipi_vector;1145	__u32 flags;1146	struct {1147		__u8 smm;1148		__u8 pending;1149		__u8 smm_inside_nmi;1150		__u8 latched_init;1151	} smi;1152	__u8 reserved[27];1153	__u8 exception_has_payload;1154	__u64 exception_payload;1155  };1156 1157The following bits are defined in the flags field:1158 1159- KVM_VCPUEVENT_VALID_SHADOW may be set to signal that1160  interrupt.shadow contains a valid state.1161 1162- KVM_VCPUEVENT_VALID_SMM may be set to signal that smi contains a1163  valid state.1164 1165- KVM_VCPUEVENT_VALID_PAYLOAD may be set to signal that the1166  exception_has_payload, exception_payload, and exception.pending1167  fields contain a valid state. This bit will be set whenever1168  KVM_CAP_EXCEPTION_PAYLOAD is enabled.1169 1170- KVM_VCPUEVENT_VALID_TRIPLE_FAULT may be set to signal that the1171  triple_fault_pending field contains a valid state. This bit will1172  be set whenever KVM_CAP_X86_TRIPLE_FAULT_EVENT is enabled.1173 1174ARM64:1175^^^^^^1176 1177If the guest accesses a device that is being emulated by the host kernel in1178such a way that a real device would generate a physical SError, KVM may make1179a virtual SError pending for that VCPU. This system error interrupt remains1180pending until the guest takes the exception by unmasking PSTATE.A.1181 1182Running the VCPU may cause it to take a pending SError, or make an access that1183causes an SError to become pending. The event's description is only valid while1184the VPCU is not running.1185 1186This API provides a way to read and write the pending 'event' state that is not1187visible to the guest. To save, restore or migrate a VCPU the struct representing1188the state can be read then written using this GET/SET API, along with the other1189guest-visible registers. It is not possible to 'cancel' an SError that has been1190made pending.1191 1192A device being emulated in user-space may also wish to generate an SError. To do1193this the events structure can be populated by user-space. The current state1194should be read first, to ensure no existing SError is pending. If an existing1195SError is pending, the architecture's 'Multiple SError interrupts' rules should1196be followed. (2.5.3 of DDI0587.a "ARM Reliability, Availability, and1197Serviceability (RAS) Specification").1198 1199SError exceptions always have an ESR value. Some CPUs have the ability to1200specify what the virtual SError's ESR value should be. These systems will1201advertise KVM_CAP_ARM_INJECT_SERROR_ESR. In this case exception.has_esr will1202always have a non-zero value when read, and the agent making an SError pending1203should specify the ISS field in the lower 24 bits of exception.serror_esr. If1204the system supports KVM_CAP_ARM_INJECT_SERROR_ESR, but user-space sets the events1205with exception.has_esr as zero, KVM will choose an ESR.1206 1207Specifying exception.has_esr on a system that does not support it will return1208-EINVAL. Setting anything other than the lower 24bits of exception.serror_esr1209will return -EINVAL.1210 1211It is not possible to read back a pending external abort (injected via1212KVM_SET_VCPU_EVENTS or otherwise) because such an exception is always delivered1213directly to the virtual CPU).1214 1215::1216 1217  struct kvm_vcpu_events {1218	struct {1219		__u8 serror_pending;1220		__u8 serror_has_esr;1221		__u8 ext_dabt_pending;1222		/* Align it to 8 bytes */1223		__u8 pad[5];1224		__u64 serror_esr;1225	} exception;1226	__u32 reserved[12];1227  };1228 12294.32 KVM_SET_VCPU_EVENTS1230------------------------1231 1232:Capability: KVM_CAP_VCPU_EVENTS1233:Extended by: KVM_CAP_INTR_SHADOW1234:Architectures: x86, arm641235:Type: vcpu ioctl1236:Parameters: struct kvm_vcpu_events (in)1237:Returns: 0 on success, -1 on error1238 1239X86:1240^^^^1241 1242Set pending exceptions, interrupts, and NMIs as well as related states of the1243vcpu.1244 1245See KVM_GET_VCPU_EVENTS for the data structure.1246 1247Fields that may be modified asynchronously by running VCPUs can be excluded1248from the update. These fields are nmi.pending, sipi_vector, smi.smm,1249smi.pending. Keep the corresponding bits in the flags field cleared to1250suppress overwriting the current in-kernel state. The bits are:1251 1252===============================  ==================================1253KVM_VCPUEVENT_VALID_NMI_PENDING  transfer nmi.pending to the kernel1254KVM_VCPUEVENT_VALID_SIPI_VECTOR  transfer sipi_vector1255KVM_VCPUEVENT_VALID_SMM          transfer the smi sub-struct.1256===============================  ==================================1257 1258If KVM_CAP_INTR_SHADOW is available, KVM_VCPUEVENT_VALID_SHADOW can be set in1259the flags field to signal that interrupt.shadow contains a valid state and1260shall be written into the VCPU.1261 1262KVM_VCPUEVENT_VALID_SMM can only be set if KVM_CAP_X86_SMM is available.1263 1264If KVM_CAP_EXCEPTION_PAYLOAD is enabled, KVM_VCPUEVENT_VALID_PAYLOAD1265can be set in the flags field to signal that the1266exception_has_payload, exception_payload, and exception.pending fields1267contain a valid state and shall be written into the VCPU.1268 1269If KVM_CAP_X86_TRIPLE_FAULT_EVENT is enabled, KVM_VCPUEVENT_VALID_TRIPLE_FAULT1270can be set in flags field to signal that the triple_fault field contains1271a valid state and shall be written into the VCPU.1272 1273ARM64:1274^^^^^^1275 1276User space may need to inject several types of events to the guest.1277 1278Set the pending SError exception state for this VCPU. It is not possible to1279'cancel' an Serror that has been made pending.1280 1281If the guest performed an access to I/O memory which could not be handled by1282userspace, for example because of missing instruction syndrome decode1283information or because there is no device mapped at the accessed IPA, then1284userspace can ask the kernel to inject an external abort using the address1285from the exiting fault on the VCPU. It is a programming error to set1286ext_dabt_pending after an exit which was not either KVM_EXIT_MMIO or1287KVM_EXIT_ARM_NISV. This feature is only available if the system supports1288KVM_CAP_ARM_INJECT_EXT_DABT. This is a helper which provides commonality in1289how userspace reports accesses for the above cases to guests, across different1290userspace implementations. Nevertheless, userspace can still emulate all Arm1291exceptions by manipulating individual registers using the KVM_SET_ONE_REG API.1292 1293See KVM_GET_VCPU_EVENTS for the data structure.1294 1295 12964.33 KVM_GET_DEBUGREGS1297----------------------1298 1299:Capability: KVM_CAP_DEBUGREGS1300:Architectures: x861301:Type: vm ioctl1302:Parameters: struct kvm_debugregs (out)1303:Returns: 0 on success, -1 on error1304 1305Reads debug registers from the vcpu.1306 1307::1308 1309  struct kvm_debugregs {1310	__u64 db[4];1311	__u64 dr6;1312	__u64 dr7;1313	__u64 flags;1314	__u64 reserved[9];1315  };1316 1317 13184.34 KVM_SET_DEBUGREGS1319----------------------1320 1321:Capability: KVM_CAP_DEBUGREGS1322:Architectures: x861323:Type: vm ioctl1324:Parameters: struct kvm_debugregs (in)1325:Returns: 0 on success, -1 on error1326 1327Writes debug registers into the vcpu.1328 1329See KVM_GET_DEBUGREGS for the data structure. The flags field is unused1330yet and must be cleared on entry.1331 1332 13334.35 KVM_SET_USER_MEMORY_REGION1334-------------------------------1335 1336:Capability: KVM_CAP_USER_MEMORY1337:Architectures: all1338:Type: vm ioctl1339:Parameters: struct kvm_userspace_memory_region (in)1340:Returns: 0 on success, -1 on error1341 1342::1343 1344  struct kvm_userspace_memory_region {1345	__u32 slot;1346	__u32 flags;1347	__u64 guest_phys_addr;1348	__u64 memory_size; /* bytes */1349	__u64 userspace_addr; /* start of the userspace allocated memory */1350  };1351 1352  /* for kvm_userspace_memory_region::flags */1353  #define KVM_MEM_LOG_DIRTY_PAGES	(1UL << 0)1354  #define KVM_MEM_READONLY	(1UL << 1)1355 1356This ioctl allows the user to create, modify or delete a guest physical1357memory slot.  Bits 0-15 of "slot" specify the slot id and this value1358should be less than the maximum number of user memory slots supported per1359VM.  The maximum allowed slots can be queried using KVM_CAP_NR_MEMSLOTS.1360Slots may not overlap in guest physical address space.1361 1362If KVM_CAP_MULTI_ADDRESS_SPACE is available, bits 16-31 of "slot"1363specifies the address space which is being modified.  They must be1364less than the value that KVM_CHECK_EXTENSION returns for the1365KVM_CAP_MULTI_ADDRESS_SPACE capability.  Slots in separate address spaces1366are unrelated; the restriction on overlapping slots only applies within1367each address space.1368 1369Deleting a slot is done by passing zero for memory_size.  When changing1370an existing slot, it may be moved in the guest physical memory space,1371or its flags may be modified, but it may not be resized.1372 1373Memory for the region is taken starting at the address denoted by the1374field userspace_addr, which must point at user addressable memory for1375the entire memory slot size.  Any object may back this memory, including1376anonymous memory, ordinary files, and hugetlbfs.1377 1378On architectures that support a form of address tagging, userspace_addr must1379be an untagged address.1380 1381It is recommended that the lower 21 bits of guest_phys_addr and userspace_addr1382be identical.  This allows large pages in the guest to be backed by large1383pages in the host.1384 1385The flags field supports two flags: KVM_MEM_LOG_DIRTY_PAGES and1386KVM_MEM_READONLY.  The former can be set to instruct KVM to keep track of1387writes to memory within the slot.  See KVM_GET_DIRTY_LOG ioctl to know how to1388use it.  The latter can be set, if KVM_CAP_READONLY_MEM capability allows it,1389to make a new slot read-only.  In this case, writes to this memory will be1390posted to userspace as KVM_EXIT_MMIO exits.1391 1392When the KVM_CAP_SYNC_MMU capability is available, changes in the backing of1393the memory region are automatically reflected into the guest.  For example, an1394mmap() that affects the region will be made visible immediately.  Another1395example is madvise(MADV_DROP).1396 1397Note: On arm64, a write generated by the page-table walker (to update1398the Access and Dirty flags, for example) never results in a1399KVM_EXIT_MMIO exit when the slot has the KVM_MEM_READONLY flag. This1400is because KVM cannot provide the data that would be written by the1401page-table walker, making it impossible to emulate the access.1402Instead, an abort (data abort if the cause of the page-table update1403was a load or a store, instruction abort if it was an instruction1404fetch) is injected in the guest.1405 1406S390:1407^^^^^1408 1409Returns -EINVAL if the VM has the KVM_VM_S390_UCONTROL flag set.1410Returns -EINVAL if called on a protected VM.1411 14124.36 KVM_SET_TSS_ADDR1413---------------------1414 1415:Capability: KVM_CAP_SET_TSS_ADDR1416:Architectures: x861417:Type: vm ioctl1418:Parameters: unsigned long tss_address (in)1419:Returns: 0 on success, -1 on error1420 1421This ioctl defines the physical address of a three-page region in the guest1422physical address space.  The region must be within the first 4GB of the1423guest physical address space and must not conflict with any memory slot1424or any mmio address.  The guest may malfunction if it accesses this memory1425region.1426 1427This ioctl is required on Intel-based hosts.  This is needed on Intel hardware1428because of a quirk in the virtualization implementation (see the internals1429documentation when it pops into existence).1430 1431 14324.37 KVM_ENABLE_CAP1433-------------------1434 1435:Capability: KVM_CAP_ENABLE_CAP1436:Architectures: mips, ppc, s390, x86, loongarch1437:Type: vcpu ioctl1438:Parameters: struct kvm_enable_cap (in)1439:Returns: 0 on success; -1 on error1440 1441:Capability: KVM_CAP_ENABLE_CAP_VM1442:Architectures: all1443:Type: vm ioctl1444:Parameters: struct kvm_enable_cap (in)1445:Returns: 0 on success; -1 on error1446 1447.. note::1448 1449   Not all extensions are enabled by default. Using this ioctl the application1450   can enable an extension, making it available to the guest.1451 1452On systems that do not support this ioctl, it always fails. On systems that1453do support it, it only works for extensions that are supported for enablement.1454 1455To check if a capability can be enabled, the KVM_CHECK_EXTENSION ioctl should1456be used.1457 1458::1459 1460  struct kvm_enable_cap {1461       /* in */1462       __u32 cap;1463 1464The capability that is supposed to get enabled.1465 1466::1467 1468       __u32 flags;1469 1470A bitfield indicating future enhancements. Has to be 0 for now.1471 1472::1473 1474       __u64 args[4];1475 1476Arguments for enabling a feature. If a feature needs initial values to1477function properly, this is the place to put them.1478 1479::1480 1481       __u8  pad[64];1482  };1483 1484The vcpu ioctl should be used for vcpu-specific capabilities, the vm ioctl1485for vm-wide capabilities.1486 14874.38 KVM_GET_MP_STATE1488---------------------1489 1490:Capability: KVM_CAP_MP_STATE1491:Architectures: x86, s390, arm64, riscv, loongarch1492:Type: vcpu ioctl1493:Parameters: struct kvm_mp_state (out)1494:Returns: 0 on success; -1 on error1495 1496::1497 1498  struct kvm_mp_state {1499	__u32 mp_state;1500  };1501 1502Returns the vcpu's current "multiprocessing state" (though also valid on1503uniprocessor guests).1504 1505Possible values are:1506 1507   ==========================    ===============================================1508   KVM_MP_STATE_RUNNABLE         the vcpu is currently running1509                                 [x86,arm64,riscv,loongarch]1510   KVM_MP_STATE_UNINITIALIZED    the vcpu is an application processor (AP)1511                                 which has not yet received an INIT signal [x86]1512   KVM_MP_STATE_INIT_RECEIVED    the vcpu has received an INIT signal, and is1513                                 now ready for a SIPI [x86]1514   KVM_MP_STATE_HALTED           the vcpu has executed a HLT instruction and1515                                 is waiting for an interrupt [x86]1516   KVM_MP_STATE_SIPI_RECEIVED    the vcpu has just received a SIPI (vector1517                                 accessible via KVM_GET_VCPU_EVENTS) [x86]1518   KVM_MP_STATE_STOPPED          the vcpu is stopped [s390,arm64,riscv]1519   KVM_MP_STATE_CHECK_STOP       the vcpu is in a special error state [s390]1520   KVM_MP_STATE_OPERATING        the vcpu is operating (running or halted)1521                                 [s390]1522   KVM_MP_STATE_LOAD             the vcpu is in a special load/startup state1523                                 [s390]1524   KVM_MP_STATE_SUSPENDED        the vcpu is in a suspend state and is waiting1525                                 for a wakeup event [arm64]1526   ==========================    ===============================================1527 1528On x86, this ioctl is only useful after KVM_CREATE_IRQCHIP. Without an1529in-kernel irqchip, the multiprocessing state must be maintained by userspace on1530these architectures.1531 1532For arm64:1533^^^^^^^^^^1534 1535If a vCPU is in the KVM_MP_STATE_SUSPENDED state, KVM will emulate the1536architectural execution of a WFI instruction.1537 1538If a wakeup event is recognized, KVM will exit to userspace with a1539KVM_SYSTEM_EVENT exit, where the event type is KVM_SYSTEM_EVENT_WAKEUP. If1540userspace wants to honor the wakeup, it must set the vCPU's MP state to1541KVM_MP_STATE_RUNNABLE. If it does not, KVM will continue to await a wakeup1542event in subsequent calls to KVM_RUN.1543 1544.. warning::1545 1546     If userspace intends to keep the vCPU in a SUSPENDED state, it is1547     strongly recommended that userspace take action to suppress the1548     wakeup event (such as masking an interrupt). Otherwise, subsequent1549     calls to KVM_RUN will immediately exit with a KVM_SYSTEM_EVENT_WAKEUP1550     event and inadvertently waste CPU cycles.1551 1552     Additionally, if userspace takes action to suppress a wakeup event,1553     it is strongly recommended that it also restores the vCPU to its1554     original state when the vCPU is made RUNNABLE again. For example,1555     if userspace masked a pending interrupt to suppress the wakeup,1556     the interrupt should be unmasked before returning control to the1557     guest.1558 1559For riscv:1560^^^^^^^^^^1561 1562The only states that are valid are KVM_MP_STATE_STOPPED and1563KVM_MP_STATE_RUNNABLE which reflect if the vcpu is paused or not.1564 1565On LoongArch, only the KVM_MP_STATE_RUNNABLE state is used to reflect1566whether the vcpu is runnable.1567 15684.39 KVM_SET_MP_STATE1569---------------------1570 1571:Capability: KVM_CAP_MP_STATE1572:Architectures: x86, s390, arm64, riscv, loongarch1573:Type: vcpu ioctl1574:Parameters: struct kvm_mp_state (in)1575:Returns: 0 on success; -1 on error1576 1577Sets the vcpu's current "multiprocessing state"; see KVM_GET_MP_STATE for1578arguments.1579 1580On x86, this ioctl is only useful after KVM_CREATE_IRQCHIP. Without an1581in-kernel irqchip, the multiprocessing state must be maintained by userspace on1582these architectures.1583 1584For arm64/riscv:1585^^^^^^^^^^^^^^^^1586 1587The only states that are valid are KVM_MP_STATE_STOPPED and1588KVM_MP_STATE_RUNNABLE which reflect if the vcpu should be paused or not.1589 1590On LoongArch, only the KVM_MP_STATE_RUNNABLE state is used to reflect1591whether the vcpu is runnable.1592 15934.40 KVM_SET_IDENTITY_MAP_ADDR1594------------------------------1595 1596:Capability: KVM_CAP_SET_IDENTITY_MAP_ADDR1597:Architectures: x861598:Type: vm ioctl1599:Parameters: unsigned long identity (in)1600:Returns: 0 on success, -1 on error1601 1602This ioctl defines the physical address of a one-page region in the guest1603physical address space.  The region must be within the first 4GB of the1604guest physical address space and must not conflict with any memory slot1605or any mmio address.  The guest may malfunction if it accesses this memory1606region.1607 1608Setting the address to 0 will result in resetting the address to its default1609(0xfffbc000).1610 1611This ioctl is required on Intel-based hosts.  This is needed on Intel hardware1612because of a quirk in the virtualization implementation (see the internals1613documentation when it pops into existence).1614 1615Fails if any VCPU has already been created.1616 16174.41 KVM_SET_BOOT_CPU_ID1618------------------------1619 1620:Capability: KVM_CAP_SET_BOOT_CPU_ID1621:Architectures: x861622:Type: vm ioctl1623:Parameters: unsigned long vcpu_id1624:Returns: 0 on success, -1 on error1625 1626Define which vcpu is the Bootstrap Processor (BSP).  Values are the same1627as the vcpu id in KVM_CREATE_VCPU.  If this ioctl is not called, the default1628is vcpu 0. This ioctl has to be called before vcpu creation,1629otherwise it will return EBUSY error.1630 1631 16324.42 KVM_GET_XSAVE1633------------------1634 1635:Capability: KVM_CAP_XSAVE1636:Architectures: x861637:Type: vcpu ioctl1638:Parameters: struct kvm_xsave (out)1639:Returns: 0 on success, -1 on error1640 1641 1642::1643 1644  struct kvm_xsave {1645	__u32 region[1024];1646	__u32 extra[0];1647  };1648 1649This ioctl would copy current vcpu's xsave struct to the userspace.1650 1651 16524.43 KVM_SET_XSAVE1653------------------1654 1655:Capability: KVM_CAP_XSAVE and KVM_CAP_XSAVE21656:Architectures: x861657:Type: vcpu ioctl1658:Parameters: struct kvm_xsave (in)1659:Returns: 0 on success, -1 on error1660 1661::1662 1663 1664  struct kvm_xsave {1665	__u32 region[1024];1666	__u32 extra[0];1667  };1668 1669This ioctl would copy userspace's xsave struct to the kernel. It copies1670as many bytes as are returned by KVM_CHECK_EXTENSION(KVM_CAP_XSAVE2),1671when invoked on the vm file descriptor. The size value returned by1672KVM_CHECK_EXTENSION(KVM_CAP_XSAVE2) will always be at least 4096.1673Currently, it is only greater than 4096 if a dynamic feature has been1674enabled with ``arch_prctl()``, but this may change in the future.1675 1676The offsets of the state save areas in struct kvm_xsave follow the1677contents of CPUID leaf 0xD on the host.1678 1679 16804.44 KVM_GET_XCRS1681-----------------1682 1683:Capability: KVM_CAP_XCRS1684:Architectures: x861685:Type: vcpu ioctl1686:Parameters: struct kvm_xcrs (out)1687:Returns: 0 on success, -1 on error1688 1689::1690 1691  struct kvm_xcr {1692	__u32 xcr;1693	__u32 reserved;1694	__u64 value;1695  };1696 1697  struct kvm_xcrs {1698	__u32 nr_xcrs;1699	__u32 flags;1700	struct kvm_xcr xcrs[KVM_MAX_XCRS];1701	__u64 padding[16];1702  };1703 1704This ioctl would copy current vcpu's xcrs to the userspace.1705 1706 17074.45 KVM_SET_XCRS1708-----------------1709 1710:Capability: KVM_CAP_XCRS1711:Architectures: x861712:Type: vcpu ioctl1713:Parameters: struct kvm_xcrs (in)1714:Returns: 0 on success, -1 on error1715 1716::1717 1718  struct kvm_xcr {1719	__u32 xcr;1720	__u32 reserved;1721	__u64 value;1722  };1723 1724  struct kvm_xcrs {1725	__u32 nr_xcrs;1726	__u32 flags;1727	struct kvm_xcr xcrs[KVM_MAX_XCRS];1728	__u64 padding[16];1729  };1730 1731This ioctl would set vcpu's xcr to the value userspace specified.1732 1733 17344.46 KVM_GET_SUPPORTED_CPUID1735----------------------------1736 1737:Capability: KVM_CAP_EXT_CPUID1738:Architectures: x861739:Type: system ioctl1740:Parameters: struct kvm_cpuid2 (in/out)1741:Returns: 0 on success, -1 on error1742 1743::1744 1745  struct kvm_cpuid2 {1746	__u32 nent;1747	__u32 padding;1748	struct kvm_cpuid_entry2 entries[0];1749  };1750 1751  #define KVM_CPUID_FLAG_SIGNIFCANT_INDEX		BIT(0)1752  #define KVM_CPUID_FLAG_STATEFUL_FUNC		BIT(1) /* deprecated */1753  #define KVM_CPUID_FLAG_STATE_READ_NEXT		BIT(2) /* deprecated */1754 1755  struct kvm_cpuid_entry2 {1756	__u32 function;1757	__u32 index;1758	__u32 flags;1759	__u32 eax;1760	__u32 ebx;1761	__u32 ecx;1762	__u32 edx;1763	__u32 padding[3];1764  };1765 1766This ioctl returns x86 cpuid features which are supported by both the1767hardware and kvm in its default configuration.  Userspace can use the1768information returned by this ioctl to construct cpuid information (for1769KVM_SET_CPUID2) that is consistent with hardware, kernel, and1770userspace capabilities, and with user requirements (for example, the1771user may wish to constrain cpuid to emulate older hardware, or for1772feature consistency across a cluster).1773 1774Dynamically-enabled feature bits need to be requested with1775``arch_prctl()`` before calling this ioctl. Feature bits that have not1776been requested are excluded from the result.1777 1778Note that certain capabilities, such as KVM_CAP_X86_DISABLE_EXITS, may1779expose cpuid features (e.g. MONITOR) which are not supported by kvm in1780its default configuration. If userspace enables such capabilities, it1781is responsible for modifying the results of this ioctl appropriately.1782 1783Userspace invokes KVM_GET_SUPPORTED_CPUID by passing a kvm_cpuid2 structure1784with the 'nent' field indicating the number of entries in the variable-size1785array 'entries'.  If the number of entries is too low to describe the cpu1786capabilities, an error (E2BIG) is returned.  If the number is too high,1787the 'nent' field is adjusted and an error (ENOMEM) is returned.  If the1788number is just right, the 'nent' field is adjusted to the number of valid1789entries in the 'entries' array, which is then filled.1790 1791The entries returned are the host cpuid as returned by the cpuid instruction,1792with unknown or unsupported features masked out.  Some features (for example,1793x2apic), may not be present in the host cpu, but are exposed by kvm if it can1794emulate them efficiently. The fields in each entry are defined as follows:1795 1796  function:1797         the eax value used to obtain the entry1798 1799  index:1800         the ecx value used to obtain the entry (for entries that are1801         affected by ecx)1802 1803  flags:1804     an OR of zero or more of the following:1805 1806        KVM_CPUID_FLAG_SIGNIFCANT_INDEX:1807           if the index field is valid1808 1809   eax, ebx, ecx, edx:1810         the values returned by the cpuid instruction for1811         this function/index combination1812 1813The TSC deadline timer feature (CPUID leaf 1, ecx[24]) is always returned1814as false, since the feature depends on KVM_CREATE_IRQCHIP for local APIC1815support.  Instead it is reported via::1816 1817  ioctl(KVM_CHECK_EXTENSION, KVM_CAP_TSC_DEADLINE_TIMER)1818 1819if that returns true and you use KVM_CREATE_IRQCHIP, or if you emulate the1820feature in userspace, then you can enable the feature for KVM_SET_CPUID2.1821 1822 18234.47 KVM_PPC_GET_PVINFO1824-----------------------1825 1826:Capability: KVM_CAP_PPC_GET_PVINFO1827:Architectures: ppc1828:Type: vm ioctl1829:Parameters: struct kvm_ppc_pvinfo (out)1830:Returns: 0 on success, !0 on error1831 1832::1833 1834  struct kvm_ppc_pvinfo {1835	__u32 flags;1836	__u32 hcall[4];1837	__u8  pad[108];1838  };1839 1840This ioctl fetches PV specific information that need to be passed to the guest1841using the device tree or other means from vm context.1842 1843The hcall array defines 4 instructions that make up a hypercall.1844 1845If any additional field gets added to this structure later on, a bit for that1846additional piece of information will be set in the flags bitmap.1847 1848The flags bitmap is defined as::1849 1850   /* the host supports the ePAPR idle hcall1851   #define KVM_PPC_PVINFO_FLAGS_EV_IDLE   (1<<0)1852 18534.52 KVM_SET_GSI_ROUTING1854------------------------1855 1856:Capability: KVM_CAP_IRQ_ROUTING1857:Architectures: x86 s390 arm641858:Type: vm ioctl1859:Parameters: struct kvm_irq_routing (in)1860:Returns: 0 on success, -1 on error1861 1862Sets the GSI routing table entries, overwriting any previously set entries.1863 1864On arm64, GSI routing has the following limitation:1865 1866- GSI routing does not apply to KVM_IRQ_LINE but only to KVM_IRQFD.1867 1868::1869 1870  struct kvm_irq_routing {1871	__u32 nr;1872	__u32 flags;1873	struct kvm_irq_routing_entry entries[0];1874  };1875 1876No flags are specified so far, the corresponding field must be set to zero.1877 1878::1879 1880  struct kvm_irq_routing_entry {1881	__u32 gsi;1882	__u32 type;1883	__u32 flags;1884	__u32 pad;1885	union {1886		struct kvm_irq_routing_irqchip irqchip;1887		struct kvm_irq_routing_msi msi;1888		struct kvm_irq_routing_s390_adapter adapter;1889		struct kvm_irq_routing_hv_sint hv_sint;1890		struct kvm_irq_routing_xen_evtchn xen_evtchn;1891		__u32 pad[8];1892	} u;1893  };1894 1895  /* gsi routing entry types */1896  #define KVM_IRQ_ROUTING_IRQCHIP 11897  #define KVM_IRQ_ROUTING_MSI 21898  #define KVM_IRQ_ROUTING_S390_ADAPTER 31899  #define KVM_IRQ_ROUTING_HV_SINT 41900  #define KVM_IRQ_ROUTING_XEN_EVTCHN 51901 1902flags:1903 1904- KVM_MSI_VALID_DEVID: used along with KVM_IRQ_ROUTING_MSI routing entry1905  type, specifies that the devid field contains a valid value.  The per-VM1906  KVM_CAP_MSI_DEVID capability advertises the requirement to provide1907  the device ID.  If this capability is not available, userspace should1908  never set the KVM_MSI_VALID_DEVID flag as the ioctl might fail.1909- zero otherwise1910 1911::1912 1913  struct kvm_irq_routing_irqchip {1914	__u32 irqchip;1915	__u32 pin;1916  };1917 1918  struct kvm_irq_routing_msi {1919	__u32 address_lo;1920	__u32 address_hi;1921	__u32 data;1922	union {1923		__u32 pad;1924		__u32 devid;1925	};1926  };1927 1928If KVM_MSI_VALID_DEVID is set, devid contains a unique device identifier1929for the device that wrote the MSI message.  For PCI, this is usually a1930BDF identifier in the lower 16 bits.1931 1932On x86, address_hi is ignored unless the KVM_X2APIC_API_USE_32BIT_IDS1933feature of KVM_CAP_X2APIC_API capability is enabled.  If it is enabled,1934address_hi bits 31-8 provide bits 31-8 of the destination id.  Bits 7-0 of1935address_hi must be zero.1936 1937::1938 1939  struct kvm_irq_routing_s390_adapter {1940	__u64 ind_addr;1941	__u64 summary_addr;1942	__u64 ind_offset;1943	__u32 summary_offset;1944	__u32 adapter_id;1945  };1946 1947  struct kvm_irq_routing_hv_sint {1948	__u32 vcpu;1949	__u32 sint;1950  };1951 1952  struct kvm_irq_routing_xen_evtchn {1953	__u32 port;1954	__u32 vcpu;1955	__u32 priority;1956  };1957 1958 1959When KVM_CAP_XEN_HVM includes the KVM_XEN_HVM_CONFIG_EVTCHN_2LEVEL bit1960in its indication of supported features, routing to Xen event channels1961is supported. Although the priority field is present, only the value1962KVM_XEN_HVM_CONFIG_EVTCHN_2LEVEL is supported, which means delivery by19632 level event channels. FIFO event channel support may be added in1964the future.1965 1966 19674.55 KVM_SET_TSC_KHZ1968--------------------1969 1970:Capability: KVM_CAP_TSC_CONTROL / KVM_CAP_VM_TSC_CONTROL1971:Architectures: x861972:Type: vcpu ioctl / vm ioctl1973:Parameters: virtual tsc_khz1974:Returns: 0 on success, -1 on error1975 1976Specifies the tsc frequency for the virtual machine. The unit of the1977frequency is KHz.1978 1979If the KVM_CAP_VM_TSC_CONTROL capability is advertised, this can also1980be used as a vm ioctl to set the initial tsc frequency of subsequently1981created vCPUs.1982 19834.56 KVM_GET_TSC_KHZ1984--------------------1985 1986:Capability: KVM_CAP_GET_TSC_KHZ / KVM_CAP_VM_TSC_CONTROL1987:Architectures: x861988:Type: vcpu ioctl / vm ioctl1989:Parameters: none1990:Returns: virtual tsc-khz on success, negative value on error1991 1992Returns the tsc frequency of the guest. The unit of the return value is1993KHz. If the host has unstable tsc this ioctl returns -EIO instead as an1994error.1995 1996 19974.57 KVM_GET_LAPIC1998------------------1999 2000:Capability: KVM_CAP_IRQCHIP2001:Architectures: x862002:Type: vcpu ioctl2003:Parameters: struct kvm_lapic_state (out)2004:Returns: 0 on success, -1 on error2005 2006::2007 2008  #define KVM_APIC_REG_SIZE 0x4002009  struct kvm_lapic_state {2010	char regs[KVM_APIC_REG_SIZE];2011  };2012 2013Reads the Local APIC registers and copies them into the input argument.  The2014data format and layout are the same as documented in the architecture manual.2015 2016If KVM_X2APIC_API_USE_32BIT_IDS feature of KVM_CAP_X2APIC_API is2017enabled, then the format of APIC_ID register depends on the APIC mode2018(reported by MSR_IA32_APICBASE) of its VCPU.  x2APIC stores APIC ID in2019the APIC_ID register (bytes 32-35).  xAPIC only allows an 8-bit APIC ID2020which is stored in bits 31-24 of the APIC register, or equivalently in2021byte 35 of struct kvm_lapic_state's regs field.  KVM_GET_LAPIC must then2022be called after MSR_IA32_APICBASE has been set with KVM_SET_MSR.2023 2024If KVM_X2APIC_API_USE_32BIT_IDS feature is disabled, struct kvm_lapic_state2025always uses xAPIC format.2026 2027 20284.58 KVM_SET_LAPIC2029------------------2030 2031:Capability: KVM_CAP_IRQCHIP2032:Architectures: x862033:Type: vcpu ioctl2034:Parameters: struct kvm_lapic_state (in)2035:Returns: 0 on success, -1 on error2036 2037::2038 2039  #define KVM_APIC_REG_SIZE 0x4002040  struct kvm_lapic_state {2041	char regs[KVM_APIC_REG_SIZE];2042  };2043 2044Copies the input argument into the Local APIC registers.  The data format2045and layout are the same as documented in the architecture manual.2046 2047The format of the APIC ID register (bytes 32-35 of struct kvm_lapic_state's2048regs field) depends on the state of the KVM_CAP_X2APIC_API capability.2049See the note in KVM_GET_LAPIC.2050 2051 20524.59 KVM_IOEVENTFD2053------------------2054 2055:Capability: KVM_CAP_IOEVENTFD2056:Architectures: all2057:Type: vm ioctl2058:Parameters: struct kvm_ioeventfd (in)2059:Returns: 0 on success, !0 on error2060 2061This ioctl attaches or detaches an ioeventfd to a legal pio/mmio address2062within the guest.  A guest write in the registered address will signal the2063provided event instead of triggering an exit.2064 2065::2066 2067  struct kvm_ioeventfd {2068	__u64 datamatch;2069	__u64 addr;        /* legal pio/mmio address */2070	__u32 len;         /* 0, 1, 2, 4, or 8 bytes    */2071	__s32 fd;2072	__u32 flags;2073	__u8  pad[36];2074  };2075 2076For the special case of virtio-ccw devices on s390, the ioevent is matched2077to a subchannel/virtqueue tuple instead.2078 2079The following flags are defined::2080 2081  #define KVM_IOEVENTFD_FLAG_DATAMATCH (1 << kvm_ioeventfd_flag_nr_datamatch)2082  #define KVM_IOEVENTFD_FLAG_PIO       (1 << kvm_ioeventfd_flag_nr_pio)2083  #define KVM_IOEVENTFD_FLAG_DEASSIGN  (1 << kvm_ioeventfd_flag_nr_deassign)2084  #define KVM_IOEVENTFD_FLAG_VIRTIO_CCW_NOTIFY \2085	(1 << kvm_ioeventfd_flag_nr_virtio_ccw_notify)2086 2087If datamatch flag is set, the event will be signaled only if the written value2088to the registered address is equal to datamatch in struct kvm_ioeventfd.2089 2090For virtio-ccw devices, addr contains the subchannel id and datamatch the2091virtqueue index.2092 2093With KVM_CAP_IOEVENTFD_ANY_LENGTH, a zero length ioeventfd is allowed, and2094the kernel will ignore the length of guest write and may get a faster vmexit.2095The speedup may only apply to specific architectures, but the ioeventfd will2096work anyway.2097 20984.60 KVM_DIRTY_TLB2099------------------2100 2101:Capability: KVM_CAP_SW_TLB2102:Architectures: ppc2103:Type: vcpu ioctl2104:Parameters: struct kvm_dirty_tlb (in)2105:Returns: 0 on success, -1 on error2106 2107::2108 2109  struct kvm_dirty_tlb {2110	__u64 bitmap;2111	__u32 num_dirty;2112  };2113 2114This must be called whenever userspace has changed an entry in the shared2115TLB, prior to calling KVM_RUN on the associated vcpu.2116 2117The "bitmap" field is the userspace address of an array.  This array2118consists of a number of bits, equal to the total number of TLB entries as2119determined by the last successful call to KVM_CONFIG_TLB, rounded up to the2120nearest multiple of 64.2121 2122Each bit corresponds to one TLB entry, ordered the same as in the shared TLB2123array.2124 2125The array is little-endian: the bit 0 is the least significant bit of the2126first byte, bit 8 is the least significant bit of the second byte, etc.2127This avoids any complications with differing word sizes.2128 2129The "num_dirty" field is a performance hint for KVM to determine whether it2130should skip processing the bitmap and just invalidate everything.  It must2131be set to the number of set bits in the bitmap.2132 2133 21344.62 KVM_CREATE_SPAPR_TCE2135-------------------------2136 2137:Capability: KVM_CAP_SPAPR_TCE2138:Architectures: powerpc2139:Type: vm ioctl2140:Parameters: struct kvm_create_spapr_tce (in)2141:Returns: file descriptor for manipulating the created TCE table2142 2143This creates a virtual TCE (translation control entry) table, which2144is an IOMMU for PAPR-style virtual I/O.  It is used to translate2145logical addresses used in virtual I/O into guest physical addresses,2146and provides a scatter/gather capability for PAPR virtual I/O.2147 2148::2149 2150  /* for KVM_CAP_SPAPR_TCE */2151  struct kvm_create_spapr_tce {2152	__u64 liobn;2153	__u32 window_size;2154  };2155 2156The liobn field gives the logical IO bus number for which to create a2157TCE table.  The window_size field specifies the size of the DMA window2158which this TCE table will translate - the table will contain one 642159bit TCE entry for every 4kiB of the DMA window.2160 2161When the guest issues an H_PUT_TCE hcall on a liobn for which a TCE2162table has been created using this ioctl(), the kernel will handle it2163in real mode, updating the TCE table.  H_PUT_TCE calls for other2164liobns will cause a vm exit and must be handled by userspace.2165 2166The return value is a file descriptor which can be passed to mmap(2)2167to map the created TCE table into userspace.  This lets userspace read2168the entries written by kernel-handled H_PUT_TCE calls, and also lets2169userspace update the TCE table directly which is useful in some2170circumstances.2171 2172 21734.63 KVM_ALLOCATE_RMA2174---------------------2175 2176:Capability: KVM_CAP_PPC_RMA2177:Architectures: powerpc2178:Type: vm ioctl2179:Parameters: struct kvm_allocate_rma (out)2180:Returns: file descriptor for mapping the allocated RMA2181 2182This allocates a Real Mode Area (RMA) from the pool allocated at boot2183time by the kernel.  An RMA is a physically-contiguous, aligned region2184of memory used on older POWER processors to provide the memory which2185will be accessed by real-mode (MMU off) accesses in a KVM guest.2186POWER processors support a set of sizes for the RMA that usually2187includes 64MB, 128MB, 256MB and some larger powers of two.2188 2189::2190 2191  /* for KVM_ALLOCATE_RMA */2192  struct kvm_allocate_rma {2193	__u64 rma_size;2194  };2195 2196The return value is a file descriptor which can be passed to mmap(2)2197to map the allocated RMA into userspace.  The mapped area can then be2198passed to the KVM_SET_USER_MEMORY_REGION ioctl to establish it as the2199RMA for a virtual machine.  The size of the RMA in bytes (which is2200fixed at host kernel boot time) is returned in the rma_size field of2201the argument structure.2202 2203The KVM_CAP_PPC_RMA capability is 1 or 2 if the KVM_ALLOCATE_RMA ioctl2204is supported; 2 if the processor requires all virtual machines to have2205an RMA, or 1 if the processor can use an RMA but doesn't require it,2206because it supports the Virtual RMA (VRMA) facility.2207 2208 22094.64 KVM_NMI2210------------2211 2212:Capability: KVM_CAP_USER_NMI2213:Architectures: x862214:Type: vcpu ioctl2215:Parameters: none2216:Returns: 0 on success, -1 on error2217 2218Queues an NMI on the thread's vcpu.  Note this is well defined only2219when KVM_CREATE_IRQCHIP has not been called, since this is an interface2220between the virtual cpu core and virtual local APIC.  After KVM_CREATE_IRQCHIP2221has been called, this interface is completely emulated within the kernel.2222 2223To use this to emulate the LINT1 input with KVM_CREATE_IRQCHIP, use the2224following algorithm:2225 2226  - pause the vcpu2227  - read the local APIC's state (KVM_GET_LAPIC)2228  - check whether changing LINT1 will queue an NMI (see the LVT entry for LINT1)2229  - if so, issue KVM_NMI2230  - resume the vcpu2231 2232Some guests configure the LINT1 NMI input to cause a panic, aiding in2233debugging.2234 2235 22364.65 KVM_S390_UCAS_MAP2237----------------------2238 2239:Capability: KVM_CAP_S390_UCONTROL2240:Architectures: s3902241:Type: vcpu ioctl2242:Parameters: struct kvm_s390_ucas_mapping (in)2243:Returns: 0 in case of success2244 2245The parameter is defined like this::2246 2247	struct kvm_s390_ucas_mapping {2248		__u64 user_addr;2249		__u64 vcpu_addr;2250		__u64 length;2251	};2252 2253This ioctl maps the memory at "user_addr" with the length "length" to2254the vcpu's address space starting at "vcpu_addr". All parameters need to2255be aligned by 1 megabyte.2256 2257 22584.66 KVM_S390_UCAS_UNMAP2259------------------------2260 2261:Capability: KVM_CAP_S390_UCONTROL2262:Architectures: s3902263:Type: vcpu ioctl2264:Parameters: struct kvm_s390_ucas_mapping (in)2265:Returns: 0 in case of success2266 2267The parameter is defined like this::2268 2269	struct kvm_s390_ucas_mapping {2270		__u64 user_addr;2271		__u64 vcpu_addr;2272		__u64 length;2273	};2274 2275This ioctl unmaps the memory in the vcpu's address space starting at2276"vcpu_addr" with the length "length". The field "user_addr" is ignored.2277All parameters need to be aligned by 1 megabyte.2278 2279 22804.67 KVM_S390_VCPU_FAULT2281------------------------2282 2283:Capability: KVM_CAP_S390_UCONTROL2284:Architectures: s3902285:Type: vcpu ioctl2286:Parameters: vcpu absolute address (in)2287:Returns: 0 in case of success2288 2289This call creates a page table entry on the virtual cpu's address space2290(for user controlled virtual machines) or the virtual machine's address2291space (for regular virtual machines). This only works for minor faults,2292thus it's recommended to access subject memory page via the user page2293table upfront. This is useful to handle validity intercepts for user2294controlled virtual machines to fault in the virtual cpu's lowcore pages2295prior to calling the KVM_RUN ioctl.2296 2297 22984.68 KVM_SET_ONE_REG2299--------------------2300 2301:Capability: KVM_CAP_ONE_REG2302:Architectures: all2303:Type: vcpu ioctl2304:Parameters: struct kvm_one_reg (in)2305:Returns: 0 on success, negative value on failure2306 2307Errors:2308 2309  ======   ============================================================2310  ENOENT   no such register2311  EINVAL   invalid register ID, or no such register or used with VMs in2312           protected virtualization mode on s3902313  EPERM    (arm64) register access not allowed before vcpu finalization2314  EBUSY    (riscv) changing register value not allowed after the vcpu2315           has run at least once2316  ======   ============================================================2317 2318(These error codes are indicative only: do not rely on a specific error2319code being returned in a specific situation.)2320 2321::2322 2323  struct kvm_one_reg {2324       __u64 id;2325       __u64 addr;2326 };2327 2328Using this ioctl, a single vcpu register can be set to a specific value2329defined by user space with the passed in struct kvm_one_reg, where id2330refers to the register identifier as described below and addr is a pointer2331to a variable with the respective size. There can be architecture agnostic2332and architecture specific registers. Each have their own range of operation2333and their own constants and width. To keep track of the implemented2334registers, find a list below:2335 2336  ======= =============================== ============2337  Arch              Register              Width (bits)2338  ======= =============================== ============2339  PPC     KVM_REG_PPC_HIOR                642340  PPC     KVM_REG_PPC_IAC1                642341  PPC     KVM_REG_PPC_IAC2                642342  PPC     KVM_REG_PPC_IAC3                642343  PPC     KVM_REG_PPC_IAC4                642344  PPC     KVM_REG_PPC_DAC1                642345  PPC     KVM_REG_PPC_DAC2                642346  PPC     KVM_REG_PPC_DABR                642347  PPC     KVM_REG_PPC_DSCR                642348  PPC     KVM_REG_PPC_PURR                642349  PPC     KVM_REG_PPC_SPURR               642350  PPC     KVM_REG_PPC_DAR                 642351  PPC     KVM_REG_PPC_DSISR               322352  PPC     KVM_REG_PPC_AMR                 642353  PPC     KVM_REG_PPC_UAMOR               642354  PPC     KVM_REG_PPC_MMCR0               642355  PPC     KVM_REG_PPC_MMCR1               642356  PPC     KVM_REG_PPC_MMCRA               642357  PPC     KVM_REG_PPC_MMCR2               642358  PPC     KVM_REG_PPC_MMCRS               642359  PPC     KVM_REG_PPC_MMCR3               642360  PPC     KVM_REG_PPC_SIAR                642361  PPC     KVM_REG_PPC_SDAR                642362  PPC     KVM_REG_PPC_SIER                642363  PPC     KVM_REG_PPC_SIER2               642364  PPC     KVM_REG_PPC_SIER3               642365  PPC     KVM_REG_PPC_PMC1                322366  PPC     KVM_REG_PPC_PMC2                322367  PPC     KVM_REG_PPC_PMC3                322368  PPC     KVM_REG_PPC_PMC4                322369  PPC     KVM_REG_PPC_PMC5                322370  PPC     KVM_REG_PPC_PMC6                322371  PPC     KVM_REG_PPC_PMC7                322372  PPC     KVM_REG_PPC_PMC8                322373  PPC     KVM_REG_PPC_FPR0                642374  ...2375  PPC     KVM_REG_PPC_FPR31               642376  PPC     KVM_REG_PPC_VR0                 1282377  ...2378  PPC     KVM_REG_PPC_VR31                1282379  PPC     KVM_REG_PPC_VSR0                1282380  ...2381  PPC     KVM_REG_PPC_VSR31               1282382  PPC     KVM_REG_PPC_FPSCR               642383  PPC     KVM_REG_PPC_VSCR                322384  PPC     KVM_REG_PPC_VPA_ADDR            642385  PPC     KVM_REG_PPC_VPA_SLB             1282386  PPC     KVM_REG_PPC_VPA_DTL             1282387  PPC     KVM_REG_PPC_EPCR                322388  PPC     KVM_REG_PPC_EPR                 322389  PPC     KVM_REG_PPC_TCR                 322390  PPC     KVM_REG_PPC_TSR                 322391  PPC     KVM_REG_PPC_OR_TSR              322392  PPC     KVM_REG_PPC_CLEAR_TSR           322393  PPC     KVM_REG_PPC_MAS0                322394  PPC     KVM_REG_PPC_MAS1                322395  PPC     KVM_REG_PPC_MAS2                642396  PPC     KVM_REG_PPC_MAS7_3              642397  PPC     KVM_REG_PPC_MAS4                322398  PPC     KVM_REG_PPC_MAS6                322399  PPC     KVM_REG_PPC_MMUCFG              322400  PPC     KVM_REG_PPC_TLB0CFG             322401  PPC     KVM_REG_PPC_TLB1CFG             322402  PPC     KVM_REG_PPC_TLB2CFG             322403  PPC     KVM_REG_PPC_TLB3CFG             322404  PPC     KVM_REG_PPC_TLB0PS              322405  PPC     KVM_REG_PPC_TLB1PS              322406  PPC     KVM_REG_PPC_TLB2PS              322407  PPC     KVM_REG_PPC_TLB3PS              322408  PPC     KVM_REG_PPC_EPTCFG              322409  PPC     KVM_REG_PPC_ICP_STATE           642410  PPC     KVM_REG_PPC_VP_STATE            1282411  PPC     KVM_REG_PPC_TB_OFFSET           642412  PPC     KVM_REG_PPC_SPMC1               322413  PPC     KVM_REG_PPC_SPMC2               322414  PPC     KVM_REG_PPC_IAMR                642415  PPC     KVM_REG_PPC_TFHAR               642416  PPC     KVM_REG_PPC_TFIAR               642417  PPC     KVM_REG_PPC_TEXASR              642418  PPC     KVM_REG_PPC_FSCR                642419  PPC     KVM_REG_PPC_PSPB                322420  PPC     KVM_REG_PPC_EBBHR               642421  PPC     KVM_REG_PPC_EBBRR               642422  PPC     KVM_REG_PPC_BESCR               642423  PPC     KVM_REG_PPC_TAR                 642424  PPC     KVM_REG_PPC_DPDES               642425  PPC     KVM_REG_PPC_DAWR                642426  PPC     KVM_REG_PPC_DAWRX               642427  PPC     KVM_REG_PPC_CIABR               642428  PPC     KVM_REG_PPC_IC                  642429  PPC     KVM_REG_PPC_VTB                 642430  PPC     KVM_REG_PPC_CSIGR               642431  PPC     KVM_REG_PPC_TACR                642432  PPC     KVM_REG_PPC_TCSCR               642433  PPC     KVM_REG_PPC_PID                 642434  PPC     KVM_REG_PPC_ACOP                642435  PPC     KVM_REG_PPC_VRSAVE              322436  PPC     KVM_REG_PPC_LPCR                322437  PPC     KVM_REG_PPC_LPCR_64             642438  PPC     KVM_REG_PPC_PPR                 642439  PPC     KVM_REG_PPC_ARCH_COMPAT         322440  PPC     KVM_REG_PPC_DABRX               322441  PPC     KVM_REG_PPC_WORT                642442  PPC	  KVM_REG_PPC_SPRG9               642443  PPC	  KVM_REG_PPC_DBSR                322444  PPC     KVM_REG_PPC_TIDR                642445  PPC     KVM_REG_PPC_PSSCR               642446  PPC     KVM_REG_PPC_DEC_EXPIRY          642447  PPC     KVM_REG_PPC_PTCR                642448  PPC     KVM_REG_PPC_HASHKEYR            642449  PPC     KVM_REG_PPC_HASHPKEYR           642450  PPC     KVM_REG_PPC_DAWR1               642451  PPC     KVM_REG_PPC_DAWRX1              642452  PPC     KVM_REG_PPC_DEXCR               642453  PPC     KVM_REG_PPC_TM_GPR0             642454  ...2455  PPC     KVM_REG_PPC_TM_GPR31            642456  PPC     KVM_REG_PPC_TM_VSR0             1282457  ...2458  PPC     KVM_REG_PPC_TM_VSR63            1282459  PPC     KVM_REG_PPC_TM_CR               642460  PPC     KVM_REG_PPC_TM_LR               642461  PPC     KVM_REG_PPC_TM_CTR              642462  PPC     KVM_REG_PPC_TM_FPSCR            642463  PPC     KVM_REG_PPC_TM_AMR              642464  PPC     KVM_REG_PPC_TM_PPR              642465  PPC     KVM_REG_PPC_TM_VRSAVE           642466  PPC     KVM_REG_PPC_TM_VSCR             322467  PPC     KVM_REG_PPC_TM_DSCR             642468  PPC     KVM_REG_PPC_TM_TAR              642469  PPC     KVM_REG_PPC_TM_XER              642470 2471  MIPS    KVM_REG_MIPS_R0                 642472  ...2473  MIPS    KVM_REG_MIPS_R31                642474  MIPS    KVM_REG_MIPS_HI                 642475  MIPS    KVM_REG_MIPS_LO                 642476  MIPS    KVM_REG_MIPS_PC                 642477  MIPS    KVM_REG_MIPS_CP0_INDEX          322478  MIPS    KVM_REG_MIPS_CP0_ENTRYLO0       642479  MIPS    KVM_REG_MIPS_CP0_ENTRYLO1       642480  MIPS    KVM_REG_MIPS_CP0_CONTEXT        642481  MIPS    KVM_REG_MIPS_CP0_CONTEXTCONFIG  322482  MIPS    KVM_REG_MIPS_CP0_USERLOCAL      642483  MIPS    KVM_REG_MIPS_CP0_XCONTEXTCONFIG 642484  MIPS    KVM_REG_MIPS_CP0_PAGEMASK       322485  MIPS    KVM_REG_MIPS_CP0_PAGEGRAIN      322486  MIPS    KVM_REG_MIPS_CP0_SEGCTL0        642487  MIPS    KVM_REG_MIPS_CP0_SEGCTL1        642488  MIPS    KVM_REG_MIPS_CP0_SEGCTL2        642489  MIPS    KVM_REG_MIPS_CP0_PWBASE         642490  MIPS    KVM_REG_MIPS_CP0_PWFIELD        642491  MIPS    KVM_REG_MIPS_CP0_PWSIZE         642492  MIPS    KVM_REG_MIPS_CP0_WIRED          322493  MIPS    KVM_REG_MIPS_CP0_PWCTL          322494  MIPS    KVM_REG_MIPS_CP0_HWRENA         322495  MIPS    KVM_REG_MIPS_CP0_BADVADDR       642496  MIPS    KVM_REG_MIPS_CP0_BADINSTR       322497  MIPS    KVM_REG_MIPS_CP0_BADINSTRP      322498  MIPS    KVM_REG_MIPS_CP0_COUNT          322499  MIPS    KVM_REG_MIPS_CP0_ENTRYHI        642500  MIPS    KVM_REG_MIPS_CP0_COMPARE        322501  MIPS    KVM_REG_MIPS_CP0_STATUS         322502  MIPS    KVM_REG_MIPS_CP0_INTCTL         322503  MIPS    KVM_REG_MIPS_CP0_CAUSE          322504  MIPS    KVM_REG_MIPS_CP0_EPC            642505  MIPS    KVM_REG_MIPS_CP0_PRID           322506  MIPS    KVM_REG_MIPS_CP0_EBASE          642507  MIPS    KVM_REG_MIPS_CP0_CONFIG         322508  MIPS    KVM_REG_MIPS_CP0_CONFIG1        322509  MIPS    KVM_REG_MIPS_CP0_CONFIG2        322510  MIPS    KVM_REG_MIPS_CP0_CONFIG3        322511  MIPS    KVM_REG_MIPS_CP0_CONFIG4        322512  MIPS    KVM_REG_MIPS_CP0_CONFIG5        322513  MIPS    KVM_REG_MIPS_CP0_CONFIG7        322514  MIPS    KVM_REG_MIPS_CP0_XCONTEXT       642515  MIPS    KVM_REG_MIPS_CP0_ERROREPC       642516  MIPS    KVM_REG_MIPS_CP0_KSCRATCH1      642517  MIPS    KVM_REG_MIPS_CP0_KSCRATCH2      642518  MIPS    KVM_REG_MIPS_CP0_KSCRATCH3      642519  MIPS    KVM_REG_MIPS_CP0_KSCRATCH4      642520  MIPS    KVM_REG_MIPS_CP0_KSCRATCH5      642521  MIPS    KVM_REG_MIPS_CP0_KSCRATCH6      642522  MIPS    KVM_REG_MIPS_CP0_MAAR(0..63)    642523  MIPS    KVM_REG_MIPS_COUNT_CTL          642524  MIPS    KVM_REG_MIPS_COUNT_RESUME       642525  MIPS    KVM_REG_MIPS_COUNT_HZ           642526  MIPS    KVM_REG_MIPS_FPR_32(0..31)      322527  MIPS    KVM_REG_MIPS_FPR_64(0..31)      642528  MIPS    KVM_REG_MIPS_VEC_128(0..31)     1282529  MIPS    KVM_REG_MIPS_FCR_IR             322530  MIPS    KVM_REG_MIPS_FCR_CSR            322531  MIPS    KVM_REG_MIPS_MSA_IR             322532  MIPS    KVM_REG_MIPS_MSA_CSR            322533  ======= =============================== ============2534 2535ARM registers are mapped using the lower 32 bits.  The upper 16 of that2536is the register group type, or coprocessor number:2537 2538ARM core registers have the following id bit patterns::2539 2540  0x4020 0000 0010 <index into the kvm_regs struct:16>2541 2542ARM 32-bit CP15 registers have the following id bit patterns::2543 2544  0x4020 0000 000F <zero:1> <crn:4> <crm:4> <opc1:4> <opc2:3>2545 2546ARM 64-bit CP15 registers have the following id bit patterns::2547 2548  0x4030 0000 000F <zero:1> <zero:4> <crm:4> <opc1:4> <zero:3>2549 2550ARM CCSIDR registers are demultiplexed by CSSELR value::2551 2552  0x4020 0000 0011 00 <csselr:8>2553 2554ARM 32-bit VFP control registers have the following id bit patterns::2555 2556  0x4020 0000 0012 1 <regno:12>2557 2558ARM 64-bit FP registers have the following id bit patterns::2559 2560  0x4030 0000 0012 0 <regno:12>2561 2562ARM firmware pseudo-registers have the following bit pattern::2563 2564  0x4030 0000 0014 <regno:16>2565 2566 2567arm64 registers are mapped using the lower 32 bits. The upper 16 of2568that is the register group type, or coprocessor number:2569 2570arm64 core/FP-SIMD registers have the following id bit patterns. Note2571that the size of the access is variable, as the kvm_regs structure2572contains elements ranging from 32 to 128 bits. The index is a 32bit2573value in the kvm_regs structure seen as a 32bit array::2574 2575  0x60x0 0000 0010 <index into the kvm_regs struct:16>2576 2577Specifically:2578 2579======================= ========= ===== =======================================2580    Encoding            Register  Bits  kvm_regs member2581======================= ========= ===== =======================================2582  0x6030 0000 0010 0000 X0          64  regs.regs[0]2583  0x6030 0000 0010 0002 X1          64  regs.regs[1]2584  ...2585  0x6030 0000 0010 003c X30         64  regs.regs[30]2586  0x6030 0000 0010 003e SP          64  regs.sp2587  0x6030 0000 0010 0040 PC          64  regs.pc2588  0x6030 0000 0010 0042 PSTATE      64  regs.pstate2589  0x6030 0000 0010 0044 SP_EL1      64  sp_el12590  0x6030 0000 0010 0046 ELR_EL1     64  elr_el12591  0x6030 0000 0010 0048 SPSR_EL1    64  spsr[KVM_SPSR_EL1] (alias SPSR_SVC)2592  0x6030 0000 0010 004a SPSR_ABT    64  spsr[KVM_SPSR_ABT]2593  0x6030 0000 0010 004c SPSR_UND    64  spsr[KVM_SPSR_UND]2594  0x6030 0000 0010 004e SPSR_IRQ    64  spsr[KVM_SPSR_IRQ]2595  0x6030 0000 0010 0050 SPSR_FIQ    64  spsr[KVM_SPSR_FIQ]2596  0x6040 0000 0010 0054 V0         128  fp_regs.vregs[0]    [1]_2597  0x6040 0000 0010 0058 V1         128  fp_regs.vregs[1]    [1]_2598  ...2599  0x6040 0000 0010 00d0 V31        128  fp_regs.vregs[31]   [1]_2600  0x6020 0000 0010 00d4 FPSR        32  fp_regs.fpsr2601  0x6020 0000 0010 00d5 FPCR        32  fp_regs.fpcr2602======================= ========= ===== =======================================2603 2604.. [1] These encodings are not accepted for SVE-enabled vcpus.  See2605       KVM_ARM_VCPU_INIT.2606 2607       The equivalent register content can be accessed via bits [127:0] of2608       the corresponding SVE Zn registers instead for vcpus that have SVE2609       enabled (see below).2610 2611arm64 CCSIDR registers are demultiplexed by CSSELR value::2612 2613  0x6020 0000 0011 00 <csselr:8>2614 2615arm64 system registers have the following id bit patterns::2616 2617  0x6030 0000 0013 <op0:2> <op1:3> <crn:4> <crm:4> <op2:3>2618 2619.. warning::2620 2621     Two system register IDs do not follow the specified pattern.  These2622     are KVM_REG_ARM_TIMER_CVAL and KVM_REG_ARM_TIMER_CNT, which map to2623     system registers CNTV_CVAL_EL0 and CNTVCT_EL0 respectively.  These2624     two had their values accidentally swapped, which means TIMER_CVAL is2625     derived from the register encoding for CNTVCT_EL0 and TIMER_CNT is2626     derived from the register encoding for CNTV_CVAL_EL0.  As this is2627     API, it must remain this way.2628 2629arm64 firmware pseudo-registers have the following bit pattern::2630 2631  0x6030 0000 0014 <regno:16>2632 2633arm64 SVE registers have the following bit patterns::2634 2635  0x6080 0000 0015 00 <n:5> <slice:5>   Zn bits[2048*slice + 2047 : 2048*slice]2636  0x6050 0000 0015 04 <n:4> <slice:5>   Pn bits[256*slice + 255 : 256*slice]2637  0x6050 0000 0015 060 <slice:5>        FFR bits[256*slice + 255 : 256*slice]2638  0x6060 0000 0015 ffff                 KVM_REG_ARM64_SVE_VLS pseudo-register2639 2640Access to register IDs where 2048 * slice >= 128 * max_vq will fail with2641ENOENT.  max_vq is the vcpu's maximum supported vector length in 128-bit2642quadwords: see [2]_ below.2643 2644These registers are only accessible on vcpus for which SVE is enabled.2645See KVM_ARM_VCPU_INIT for details.2646 2647In addition, except for KVM_REG_ARM64_SVE_VLS, these registers are not2648accessible until the vcpu's SVE configuration has been finalized2649using KVM_ARM_VCPU_FINALIZE(KVM_ARM_VCPU_SVE).  See KVM_ARM_VCPU_INIT2650and KVM_ARM_VCPU_FINALIZE for more information about this procedure.2651 2652KVM_REG_ARM64_SVE_VLS is a pseudo-register that allows the set of vector2653lengths supported by the vcpu to be discovered and configured by2654userspace.  When transferred to or from user memory via KVM_GET_ONE_REG2655or KVM_SET_ONE_REG, the value of this register is of type2656__u64[KVM_ARM64_SVE_VLS_WORDS], and encodes the set of vector lengths as2657follows::2658 2659  __u64 vector_lengths[KVM_ARM64_SVE_VLS_WORDS];2660 2661  if (vq >= SVE_VQ_MIN && vq <= SVE_VQ_MAX &&2662      ((vector_lengths[(vq - KVM_ARM64_SVE_VQ_MIN) / 64] >>2663		((vq - KVM_ARM64_SVE_VQ_MIN) % 64)) & 1))2664	/* Vector length vq * 16 bytes supported */2665  else2666	/* Vector length vq * 16 bytes not supported */2667 2668.. [2] The maximum value vq for which the above condition is true is2669       max_vq.  This is the maximum vector length available to the guest on2670       this vcpu, and determines which register slices are visible through2671       this ioctl interface.2672 2673(See Documentation/arch/arm64/sve.rst for an explanation of the "vq"2674nomenclature.)2675 2676KVM_REG_ARM64_SVE_VLS is only accessible after KVM_ARM_VCPU_INIT.2677KVM_ARM_VCPU_INIT initialises it to the best set of vector lengths that2678the host supports.2679 2680Userspace may subsequently modify it if desired until the vcpu's SVE2681configuration is finalized using KVM_ARM_VCPU_FINALIZE(KVM_ARM_VCPU_SVE).2682 2683Apart from simply removing all vector lengths from the host set that2684exceed some value, support for arbitrarily chosen sets of vector lengths2685is hardware-dependent and may not be available.  Attempting to configure2686an invalid set of vector lengths via KVM_SET_ONE_REG will fail with2687EINVAL.2688 2689After the vcpu's SVE configuration is finalized, further attempts to2690write this register will fail with EPERM.2691 2692arm64 bitmap feature firmware pseudo-registers have the following bit pattern::2693 2694  0x6030 0000 0016 <regno:16>2695 2696The bitmap feature firmware registers exposes the hypercall services that2697are available for userspace to configure. The set bits corresponds to the2698services that are available for the guests to access. By default, KVM2699sets all the supported bits during VM initialization. The userspace can2700discover the available services via KVM_GET_ONE_REG, and write back the2701bitmap corresponding to the features that it wishes guests to see via2702KVM_SET_ONE_REG.2703 2704Note: These registers are immutable once any of the vCPUs of the VM has2705run at least once. A KVM_SET_ONE_REG in such a scenario will return2706a -EBUSY to userspace.2707 2708(See Documentation/virt/kvm/arm/hypercalls.rst for more details.)2709 2710 2711MIPS registers are mapped using the lower 32 bits.  The upper 16 of that is2712the register group type:2713 2714MIPS core registers (see above) have the following id bit patterns::2715 2716  0x7030 0000 0000 <reg:16>2717 2718MIPS CP0 registers (see KVM_REG_MIPS_CP0_* above) have the following id bit2719patterns depending on whether they're 32-bit or 64-bit registers::2720 2721  0x7020 0000 0001 00 <reg:5> <sel:3>   (32-bit)2722  0x7030 0000 0001 00 <reg:5> <sel:3>   (64-bit)2723 2724Note: KVM_REG_MIPS_CP0_ENTRYLO0 and KVM_REG_MIPS_CP0_ENTRYLO1 are the MIPS642725versions of the EntryLo registers regardless of the word size of the host2726hardware, host kernel, guest, and whether XPA is present in the guest, i.e.2727with the RI and XI bits (if they exist) in bits 63 and 62 respectively, and2728the PFNX field starting at bit 30.2729 2730MIPS MAARs (see KVM_REG_MIPS_CP0_MAAR(*) above) have the following id bit2731patterns::2732 2733  0x7030 0000 0001 01 <reg:8>2734 2735MIPS KVM control registers (see above) have the following id bit patterns::2736 2737  0x7030 0000 0002 <reg:16>2738 2739MIPS FPU registers (see KVM_REG_MIPS_FPR_{32,64}() above) have the following2740id bit patterns depending on the size of the register being accessed. They are2741always accessed according to the current guest FPU mode (Status.FR and2742Config5.FRE), i.e. as the guest would see them, and they become unpredictable2743if the guest FPU mode is changed. MIPS SIMD Architecture (MSA) vector2744registers (see KVM_REG_MIPS_VEC_128() above) have similar patterns as they2745overlap the FPU registers::2746 2747  0x7020 0000 0003 00 <0:3> <reg:5> (32-bit FPU registers)2748  0x7030 0000 0003 00 <0:3> <reg:5> (64-bit FPU registers)2749  0x7040 0000 0003 00 <0:3> <reg:5> (128-bit MSA vector registers)2750 2751MIPS FPU control registers (see KVM_REG_MIPS_FCR_{IR,CSR} above) have the2752following id bit patterns::2753 2754  0x7020 0000 0003 01 <0:3> <reg:5>2755 2756MIPS MSA control registers (see KVM_REG_MIPS_MSA_{IR,CSR} above) have the2757following id bit patterns::2758 2759  0x7020 0000 0003 02 <0:3> <reg:5>2760 2761RISC-V registers are mapped using the lower 32 bits. The upper 8 bits of2762that is the register group type.2763 2764RISC-V config registers are meant for configuring a Guest VCPU and it has2765the following id bit patterns::2766 2767  0x8020 0000 01 <index into the kvm_riscv_config struct:24> (32bit Host)2768  0x8030 0000 01 <index into the kvm_riscv_config struct:24> (64bit Host)2769 2770Following are the RISC-V config registers:2771 2772======================= ========= =============================================2773    Encoding            Register  Description2774======================= ========= =============================================2775  0x80x0 0000 0100 0000 isa       ISA feature bitmap of Guest VCPU2776======================= ========= =============================================2777 2778The isa config register can be read anytime but can only be written before2779a Guest VCPU runs. It will have ISA feature bits matching underlying host2780set by default.2781 2782RISC-V core registers represent the general execution state of a Guest VCPU2783and it has the following id bit patterns::2784 2785  0x8020 0000 02 <index into the kvm_riscv_core struct:24> (32bit Host)2786  0x8030 0000 02 <index into the kvm_riscv_core struct:24> (64bit Host)2787 2788Following are the RISC-V core registers:2789 2790======================= ========= =============================================2791    Encoding            Register  Description2792======================= ========= =============================================2793  0x80x0 0000 0200 0000 regs.pc   Program counter2794  0x80x0 0000 0200 0001 regs.ra   Return address2795  0x80x0 0000 0200 0002 regs.sp   Stack pointer2796  0x80x0 0000 0200 0003 regs.gp   Global pointer2797  0x80x0 0000 0200 0004 regs.tp   Task pointer2798  0x80x0 0000 0200 0005 regs.t0   Caller saved register 02799  0x80x0 0000 0200 0006 regs.t1   Caller saved register 12800  0x80x0 0000 0200 0007 regs.t2   Caller saved register 22801  0x80x0 0000 0200 0008 regs.s0   Callee saved register 02802  0x80x0 0000 0200 0009 regs.s1   Callee saved register 12803  0x80x0 0000 0200 000a regs.a0   Function argument (or return value) 02804  0x80x0 0000 0200 000b regs.a1   Function argument (or return value) 12805  0x80x0 0000 0200 000c regs.a2   Function argument 22806  0x80x0 0000 0200 000d regs.a3   Function argument 32807  0x80x0 0000 0200 000e regs.a4   Function argument 42808  0x80x0 0000 0200 000f regs.a5   Function argument 52809  0x80x0 0000 0200 0010 regs.a6   Function argument 62810  0x80x0 0000 0200 0011 regs.a7   Function argument 72811  0x80x0 0000 0200 0012 regs.s2   Callee saved register 22812  0x80x0 0000 0200 0013 regs.s3   Callee saved register 32813  0x80x0 0000 0200 0014 regs.s4   Callee saved register 42814  0x80x0 0000 0200 0015 regs.s5   Callee saved register 52815  0x80x0 0000 0200 0016 regs.s6   Callee saved register 62816  0x80x0 0000 0200 0017 regs.s7   Callee saved register 72817  0x80x0 0000 0200 0018 regs.s8   Callee saved register 82818  0x80x0 0000 0200 0019 regs.s9   Callee saved register 92819  0x80x0 0000 0200 001a regs.s10  Callee saved register 102820  0x80x0 0000 0200 001b regs.s11  Callee saved register 112821  0x80x0 0000 0200 001c regs.t3   Caller saved register 32822  0x80x0 0000 0200 001d regs.t4   Caller saved register 42823  0x80x0 0000 0200 001e regs.t5   Caller saved register 52824  0x80x0 0000 0200 001f regs.t6   Caller saved register 62825  0x80x0 0000 0200 0020 mode      Privilege mode (1 = S-mode or 0 = U-mode)2826======================= ========= =============================================2827 2828RISC-V csr registers represent the supervisor mode control/status registers2829of a Guest VCPU and it has the following id bit patterns::2830 2831  0x8020 0000 03 <index into the kvm_riscv_csr struct:24> (32bit Host)2832  0x8030 0000 03 <index into the kvm_riscv_csr struct:24> (64bit Host)2833 2834Following are the RISC-V csr registers:2835 2836======================= ========= =============================================2837    Encoding            Register  Description2838======================= ========= =============================================2839  0x80x0 0000 0300 0000 sstatus   Supervisor status2840  0x80x0 0000 0300 0001 sie       Supervisor interrupt enable2841  0x80x0 0000 0300 0002 stvec     Supervisor trap vector base2842  0x80x0 0000 0300 0003 sscratch  Supervisor scratch register2843  0x80x0 0000 0300 0004 sepc      Supervisor exception program counter2844  0x80x0 0000 0300 0005 scause    Supervisor trap cause2845  0x80x0 0000 0300 0006 stval     Supervisor bad address or instruction2846  0x80x0 0000 0300 0007 sip       Supervisor interrupt pending2847  0x80x0 0000 0300 0008 satp      Supervisor address translation and protection2848======================= ========= =============================================2849 2850RISC-V timer registers represent the timer state of a Guest VCPU and it has2851the following id bit patterns::2852 2853  0x8030 0000 04 <index into the kvm_riscv_timer struct:24>2854 2855Following are the RISC-V timer registers:2856 2857======================= ========= =============================================2858    Encoding            Register  Description2859======================= ========= =============================================2860  0x8030 0000 0400 0000 frequency Time base frequency (read-only)2861  0x8030 0000 0400 0001 time      Time value visible to Guest2862  0x8030 0000 0400 0002 compare   Time compare programmed by Guest2863  0x8030 0000 0400 0003 state     Time compare state (1 = ON or 0 = OFF)2864======================= ========= =============================================2865 2866RISC-V F-extension registers represent the single precision floating point2867state of a Guest VCPU and it has the following id bit patterns::2868 2869  0x8020 0000 05 <index into the __riscv_f_ext_state struct:24>2870 2871Following are the RISC-V F-extension registers:2872 2873======================= ========= =============================================2874    Encoding            Register  Description2875======================= ========= =============================================2876  0x8020 0000 0500 0000 f[0]      Floating point register 02877  ...2878  0x8020 0000 0500 001f f[31]     Floating point register 312879  0x8020 0000 0500 0020 fcsr      Floating point control and status register2880======================= ========= =============================================2881 2882RISC-V D-extension registers represent the double precision floating point2883state of a Guest VCPU and it has the following id bit patterns::2884 2885  0x8020 0000 06 <index into the __riscv_d_ext_state struct:24> (fcsr)2886  0x8030 0000 06 <index into the __riscv_d_ext_state struct:24> (non-fcsr)2887 2888Following are the RISC-V D-extension registers:2889 2890======================= ========= =============================================2891    Encoding            Register  Description2892======================= ========= =============================================2893  0x8030 0000 0600 0000 f[0]      Floating point register 02894  ...2895  0x8030 0000 0600 001f f[31]     Floating point register 312896  0x8020 0000 0600 0020 fcsr      Floating point control and status register2897======================= ========= =============================================2898 2899LoongArch registers are mapped using the lower 32 bits. The upper 16 bits of2900that is the register group type.2901 2902LoongArch csr registers are used to control guest cpu or get status of guest2903cpu, and they have the following id bit patterns::2904 2905  0x9030 0000 0001 00 <reg:5> <sel:3>   (64-bit)2906 2907LoongArch KVM control registers are used to implement some new defined functions2908such as set vcpu counter or reset vcpu, and they have the following id bit patterns::2909 2910  0x9030 0000 0002 <reg:16>2911 2912 29134.69 KVM_GET_ONE_REG2914--------------------2915 2916:Capability: KVM_CAP_ONE_REG2917:Architectures: all2918:Type: vcpu ioctl2919:Parameters: struct kvm_one_reg (in and out)2920:Returns: 0 on success, negative value on failure2921 2922Errors include:2923 2924  ======== ============================================================2925  ENOENT   no such register2926  EINVAL   invalid register ID, or no such register or used with VMs in2927           protected virtualization mode on s3902928  EPERM    (arm64) register access not allowed before vcpu finalization2929  ======== ============================================================2930 2931(These error codes are indicative only: do not rely on a specific error2932code being returned in a specific situation.)2933 2934This ioctl allows to receive the value of a single register implemented2935in a vcpu. The register to read is indicated by the "id" field of the2936kvm_one_reg struct passed in. On success, the register value can be found2937at the memory location pointed to by "addr".2938 2939The list of registers accessible using this interface is identical to the2940list in 4.68.2941 2942 29434.70 KVM_KVMCLOCK_CTRL2944----------------------2945 2946:Capability: KVM_CAP_KVMCLOCK_CTRL2947:Architectures: Any that implement pvclocks (currently x86 only)2948:Type: vcpu ioctl2949:Parameters: None2950:Returns: 0 on success, -1 on error2951 2952This ioctl sets a flag accessible to the guest indicating that the specified2953vCPU has been paused by the host userspace.2954 2955The host will set a flag in the pvclock structure that is checked from the2956soft lockup watchdog.  The flag is part of the pvclock structure that is2957shared between guest and host, specifically the second bit of the flags2958field of the pvclock_vcpu_time_info structure.  It will be set exclusively by2959the host and read/cleared exclusively by the guest.  The guest operation of2960checking and clearing the flag must be an atomic operation so2961load-link/store-conditional, or equivalent must be used.  There are two cases2962where the guest will clear the flag: when the soft lockup watchdog timer resets2963itself or when a soft lockup is detected.  This ioctl can be called any time2964after pausing the vcpu, but before it is resumed.2965 2966 29674.71 KVM_SIGNAL_MSI2968-------------------2969 2970:Capability: KVM_CAP_SIGNAL_MSI2971:Architectures: x86 arm642972:Type: vm ioctl2973:Parameters: struct kvm_msi (in)2974:Returns: >0 on delivery, 0 if guest blocked the MSI, and -1 on error2975 2976Directly inject a MSI message. Only valid with in-kernel irqchip that handles2977MSI messages.2978 2979::2980 2981  struct kvm_msi {2982	__u32 address_lo;2983	__u32 address_hi;2984	__u32 data;2985	__u32 flags;2986	__u32 devid;2987	__u8  pad[12];2988  };2989 2990flags:2991  KVM_MSI_VALID_DEVID: devid contains a valid value.  The per-VM2992  KVM_CAP_MSI_DEVID capability advertises the requirement to provide2993  the device ID.  If this capability is not available, userspace2994  should never set the KVM_MSI_VALID_DEVID flag as the ioctl might fail.2995 2996If KVM_MSI_VALID_DEVID is set, devid contains a unique device identifier2997for the device that wrote the MSI message.  For PCI, this is usually a2998BDF identifier in the lower 16 bits.2999 3000On x86, address_hi is ignored unless the KVM_X2APIC_API_USE_32BIT_IDS3001feature of KVM_CAP_X2APIC_API capability is enabled.  If it is enabled,3002address_hi bits 31-8 provide bits 31-8 of the destination id.  Bits 7-0 of3003address_hi must be zero.3004 3005 30064.71 KVM_CREATE_PIT23007--------------------3008 3009:Capability: KVM_CAP_PIT23010:Architectures: x863011:Type: vm ioctl3012:Parameters: struct kvm_pit_config (in)3013:Returns: 0 on success, -1 on error3014 3015Creates an in-kernel device model for the i8254 PIT. This call is only valid3016after enabling in-kernel irqchip support via KVM_CREATE_IRQCHIP. The following3017parameters have to be passed::3018 3019  struct kvm_pit_config {3020	__u32 flags;3021	__u32 pad[15];3022  };3023 3024Valid flags are::3025 3026  #define KVM_PIT_SPEAKER_DUMMY     1 /* emulate speaker port stub */3027 3028PIT timer interrupts may use a per-VM kernel thread for injection. If it3029exists, this thread will have a name of the following pattern::3030 3031  kvm-pit/<owner-process-pid>3032 3033When running a guest with elevated priorities, the scheduling parameters of3034this thread may have to be adjusted accordingly.3035 3036This IOCTL replaces the obsolete KVM_CREATE_PIT.3037 3038 30394.72 KVM_GET_PIT23040-----------------3041 3042:Capability: KVM_CAP_PIT_STATE23043:Architectures: x863044:Type: vm ioctl3045:Parameters: struct kvm_pit_state2 (out)3046:Returns: 0 on success, -1 on error3047 3048Retrieves the state of the in-kernel PIT model. Only valid after3049KVM_CREATE_PIT2. The state is returned in the following structure::3050 3051  struct kvm_pit_state2 {3052	struct kvm_pit_channel_state channels[3];3053	__u32 flags;3054	__u32 reserved[9];3055  };3056 3057Valid flags are::3058 3059  /* disable PIT in HPET legacy mode */3060  #define KVM_PIT_FLAGS_HPET_LEGACY     0x000000013061  /* speaker port data bit enabled */3062  #define KVM_PIT_FLAGS_SPEAKER_DATA_ON 0x000000023063 3064This IOCTL replaces the obsolete KVM_GET_PIT.3065 3066 30674.73 KVM_SET_PIT23068-----------------3069 3070:Capability: KVM_CAP_PIT_STATE23071:Architectures: x863072:Type: vm ioctl3073:Parameters: struct kvm_pit_state2 (in)3074:Returns: 0 on success, -1 on error3075 3076Sets the state of the in-kernel PIT model. Only valid after KVM_CREATE_PIT2.3077See KVM_GET_PIT2 for details on struct kvm_pit_state2.3078 3079This IOCTL replaces the obsolete KVM_SET_PIT.3080 3081 30824.74 KVM_PPC_GET_SMMU_INFO3083--------------------------3084 3085:Capability: KVM_CAP_PPC_GET_SMMU_INFO3086:Architectures: powerpc3087:Type: vm ioctl3088:Parameters: None3089:Returns: 0 on success, -1 on error3090 3091This populates and returns a structure describing the features of3092the "Server" class MMU emulation supported by KVM.3093This can in turn be used by userspace to generate the appropriate3094device-tree properties for the guest operating system.3095 3096The structure contains some global information, followed by an3097array of supported segment page sizes::3098 3099      struct kvm_ppc_smmu_info {3100	     __u64 flags;3101	     __u32 slb_size;3102	     __u32 pad;3103	     struct kvm_ppc_one_seg_page_size sps[KVM_PPC_PAGE_SIZES_MAX_SZ];3104      };3105 3106The supported flags are:3107 3108    - KVM_PPC_PAGE_SIZES_REAL:3109        When that flag is set, guest page sizes must "fit" the backing3110        store page sizes. When not set, any page size in the list can3111        be used regardless of how they are backed by userspace.3112 3113    - KVM_PPC_1T_SEGMENTS3114        The emulated MMU supports 1T segments in addition to the3115        standard 256M ones.3116 3117    - KVM_PPC_NO_HASH3118	This flag indicates that HPT guests are not supported by KVM,3119	thus all guests must use radix MMU mode.3120 3121The "slb_size" field indicates how many SLB entries are supported3122 3123The "sps" array contains 8 entries indicating the supported base3124page sizes for a segment in increasing order. Each entry is defined3125as follow::3126 3127   struct kvm_ppc_one_seg_page_size {3128	__u32 page_shift;	/* Base page shift of segment (or 0) */3129	__u32 slb_enc;		/* SLB encoding for BookS */3130	struct kvm_ppc_one_page_size enc[KVM_PPC_PAGE_SIZES_MAX_SZ];3131   };3132 3133An entry with a "page_shift" of 0 is unused. Because the array is3134organized in increasing order, a lookup can stop when encountering3135such an entry.3136 3137The "slb_enc" field provides the encoding to use in the SLB for the3138page size. The bits are in positions such as the value can directly3139be OR'ed into the "vsid" argument of the slbmte instruction.3140 3141The "enc" array is a list which for each of those segment base page3142size provides the list of supported actual page sizes (which can be3143only larger or equal to the base page size), along with the3144corresponding encoding in the hash PTE. Similarly, the array is31458 entries sorted by increasing sizes and an entry with a "0" shift3146is an empty entry and a terminator::3147 3148   struct kvm_ppc_one_page_size {3149	__u32 page_shift;	/* Page shift (or 0) */3150	__u32 pte_enc;		/* Encoding in the HPTE (>>12) */3151   };3152 3153The "pte_enc" field provides a value that can OR'ed into the hash3154PTE's RPN field (ie, it needs to be shifted left by 12 to OR it3155into the hash PTE second double word).3156 31574.75 KVM_IRQFD3158--------------3159 3160:Capability: KVM_CAP_IRQFD3161:Architectures: x86 s390 arm643162:Type: vm ioctl3163:Parameters: struct kvm_irqfd (in)3164:Returns: 0 on success, -1 on error3165 3166Allows setting an eventfd to directly trigger a guest interrupt.3167kvm_irqfd.fd specifies the file descriptor to use as the eventfd and3168kvm_irqfd.gsi specifies the irqchip pin toggled by this event.  When3169an event is triggered on the eventfd, an interrupt is injected into3170the guest using the specified gsi pin.  The irqfd is removed using3171the KVM_IRQFD_FLAG_DEASSIGN flag, specifying both kvm_irqfd.fd3172and kvm_irqfd.gsi.3173 3174With KVM_CAP_IRQFD_RESAMPLE, KVM_IRQFD supports a de-assert and notify3175mechanism allowing emulation of level-triggered, irqfd-based3176interrupts.  When KVM_IRQFD_FLAG_RESAMPLE is set the user must pass an3177additional eventfd in the kvm_irqfd.resamplefd field.  When operating3178in resample mode, posting of an interrupt through kvm_irq.fd asserts3179the specified gsi in the irqchip.  When the irqchip is resampled, such3180as from an EOI, the gsi is de-asserted and the user is notified via3181kvm_irqfd.resamplefd.  It is the user's responsibility to re-queue3182the interrupt if the device making use of it still requires service.3183Note that closing the resamplefd is not sufficient to disable the3184irqfd.  The KVM_IRQFD_FLAG_RESAMPLE is only necessary on assignment3185and need not be specified with KVM_IRQFD_FLAG_DEASSIGN.3186 3187On arm64, gsi routing being supported, the following can happen:3188 3189- in case no routing entry is associated to this gsi, injection fails3190- in case the gsi is associated to an irqchip routing entry,3191  irqchip.pin + 32 corresponds to the injected SPI ID.3192- in case the gsi is associated to an MSI routing entry, the MSI3193  message and device ID are translated into an LPI (support restricted3194  to GICv3 ITS in-kernel emulation).3195 31964.76 KVM_PPC_ALLOCATE_HTAB3197--------------------------3198 3199:Capability: KVM_CAP_PPC_ALLOC_HTAB3200:Architectures: powerpc3201:Type: vm ioctl3202:Parameters: Pointer to u32 containing hash table order (in/out)3203:Returns: 0 on success, -1 on error3204 3205This requests the host kernel to allocate an MMU hash table for a3206guest using the PAPR paravirtualization interface.  This only does3207anything if the kernel is configured to use the Book 3S HV style of3208virtualization.  Otherwise the capability doesn't exist and the ioctl3209returns an ENOTTY error.  The rest of this description assumes Book 3S3210HV.3211 3212There must be no vcpus running when this ioctl is called; if there3213are, it will do nothing and return an EBUSY error.3214 3215The parameter is a pointer to a 32-bit unsigned integer variable3216containing the order (log base 2) of the desired size of the hash3217table, which must be between 18 and 46.  On successful return from the3218ioctl, the value will not be changed by the kernel.3219 3220If no hash table has been allocated when any vcpu is asked to run3221(with the KVM_RUN ioctl), the host kernel will allocate a3222default-sized hash table (16 MB).3223 3224If this ioctl is called when a hash table has already been allocated,3225with a different order from the existing hash table, the existing hash3226table will be freed and a new one allocated.  If this is ioctl is3227called when a hash table has already been allocated of the same order3228as specified, the kernel will clear out the existing hash table (zero3229all HPTEs).  In either case, if the guest is using the virtualized3230real-mode area (VRMA) facility, the kernel will re-create the VMRA3231HPTEs on the next KVM_RUN of any vcpu.3232 32334.77 KVM_S390_INTERRUPT3234-----------------------3235 3236:Capability: basic3237:Architectures: s3903238:Type: vm ioctl, vcpu ioctl3239:Parameters: struct kvm_s390_interrupt (in)3240:Returns: 0 on success, -1 on error3241 3242Allows to inject an interrupt to the guest. Interrupts can be floating3243(vm ioctl) or per cpu (vcpu ioctl), depending on the interrupt type.3244 3245Interrupt parameters are passed via kvm_s390_interrupt::3246 3247  struct kvm_s390_interrupt {3248	__u32 type;3249	__u32 parm;3250	__u64 parm64;3251  };3252 3253type can be one of the following:3254 3255KVM_S390_SIGP_STOP (vcpu)3256    - sigp stop; optional flags in parm3257KVM_S390_PROGRAM_INT (vcpu)3258    - program check; code in parm3259KVM_S390_SIGP_SET_PREFIX (vcpu)3260    - sigp set prefix; prefix address in parm3261KVM_S390_RESTART (vcpu)3262    - restart3263KVM_S390_INT_CLOCK_COMP (vcpu)3264    - clock comparator interrupt3265KVM_S390_INT_CPU_TIMER (vcpu)3266    - CPU timer interrupt3267KVM_S390_INT_VIRTIO (vm)3268    - virtio external interrupt; external interrupt3269      parameters in parm and parm643270KVM_S390_INT_SERVICE (vm)3271    - sclp external interrupt; sclp parameter in parm3272KVM_S390_INT_EMERGENCY (vcpu)3273    - sigp emergency; source cpu in parm3274KVM_S390_INT_EXTERNAL_CALL (vcpu)3275    - sigp external call; source cpu in parm3276KVM_S390_INT_IO(ai,cssid,ssid,schid) (vm)3277    - compound value to indicate an3278      I/O interrupt (ai - adapter interrupt; cssid,ssid,schid - subchannel);3279      I/O interruption parameters in parm (subchannel) and parm64 (intparm,3280      interruption subclass)3281KVM_S390_MCHK (vm, vcpu)3282    - machine check interrupt; cr 14 bits in parm, machine check interrupt3283      code in parm64 (note that machine checks needing further payload are not3284      supported by this ioctl)3285 3286This is an asynchronous vcpu ioctl and can be invoked from any thread.3287 32884.78 KVM_PPC_GET_HTAB_FD3289------------------------3290 3291:Capability: KVM_CAP_PPC_HTAB_FD3292:Architectures: powerpc3293:Type: vm ioctl3294:Parameters: Pointer to struct kvm_get_htab_fd (in)3295:Returns: file descriptor number (>= 0) on success, -1 on error3296 3297This returns a file descriptor that can be used either to read out the3298entries in the guest's hashed page table (HPT), or to write entries to3299initialize the HPT.  The returned fd can only be written to if the3300KVM_GET_HTAB_WRITE bit is set in the flags field of the argument, and3301can only be read if that bit is clear.  The argument struct looks like3302this::3303 3304  /* For KVM_PPC_GET_HTAB_FD */3305  struct kvm_get_htab_fd {3306	__u64	flags;3307	__u64	start_index;3308	__u64	reserved[2];3309  };3310 3311  /* Values for kvm_get_htab_fd.flags */3312  #define KVM_GET_HTAB_BOLTED_ONLY	((__u64)0x1)3313  #define KVM_GET_HTAB_WRITE		((__u64)0x2)3314 3315The 'start_index' field gives the index in the HPT of the entry at3316which to start reading.  It is ignored when writing.3317 3318Reads on the fd will initially supply information about all3319"interesting" HPT entries.  Interesting entries are those with the3320bolted bit set, if the KVM_GET_HTAB_BOLTED_ONLY bit is set, otherwise3321all entries.  When the end of the HPT is reached, the read() will3322return.  If read() is called again on the fd, it will start again from3323the beginning of the HPT, but will only return HPT entries that have3324changed since they were last read.3325 3326Data read or written is structured as a header (8 bytes) followed by a3327series of valid HPT entries (16 bytes) each.  The header indicates how3328many valid HPT entries there are and how many invalid entries follow3329the valid entries.  The invalid entries are not represented explicitly3330in the stream.  The header format is::3331 3332  struct kvm_get_htab_header {3333	__u32	index;3334	__u16	n_valid;3335	__u16	n_invalid;3336  };3337 3338Writes to the fd create HPT entries starting at the index given in the3339header; first 'n_valid' valid entries with contents from the data3340written, then 'n_invalid' invalid entries, invalidating any previously3341valid entries found.3342 33434.79 KVM_CREATE_DEVICE3344----------------------3345 3346:Capability: KVM_CAP_DEVICE_CTRL3347:Architectures: all3348:Type: vm ioctl3349:Parameters: struct kvm_create_device (in/out)3350:Returns: 0 on success, -1 on error3351 3352Errors:3353 3354  ======  =======================================================3355  ENODEV  The device type is unknown or unsupported3356  EEXIST  Device already created, and this type of device may not3357          be instantiated multiple times3358  ======  =======================================================3359 3360  Other error conditions may be defined by individual device types or3361  have their standard meanings.3362 3363Creates an emulated device in the kernel.  The file descriptor returned3364in fd can be used with KVM_SET/GET/HAS_DEVICE_ATTR.3365 3366If the KVM_CREATE_DEVICE_TEST flag is set, only test whether the3367device type is supported (not necessarily whether it can be created3368in the current vm).3369 3370Individual devices should not define flags.  Attributes should be used3371for specifying any behavior that is not implied by the device type3372number.3373 3374::3375 3376  struct kvm_create_device {3377	__u32	type;	/* in: KVM_DEV_TYPE_xxx */3378	__u32	fd;	/* out: device handle */3379	__u32	flags;	/* in: KVM_CREATE_DEVICE_xxx */3380  };3381 33824.80 KVM_SET_DEVICE_ATTR/KVM_GET_DEVICE_ATTR3383--------------------------------------------3384 3385:Capability: KVM_CAP_DEVICE_CTRL, KVM_CAP_VM_ATTRIBUTES for vm device,3386             KVM_CAP_VCPU_ATTRIBUTES for vcpu device3387             KVM_CAP_SYS_ATTRIBUTES for system (/dev/kvm) device (no set)3388:Architectures: x86, arm64, s3903389:Type: device ioctl, vm ioctl, vcpu ioctl3390:Parameters: struct kvm_device_attr3391:Returns: 0 on success, -1 on error3392 3393Errors:3394 3395  =====   =============================================================3396  ENXIO   The group or attribute is unknown/unsupported for this device3397          or hardware support is missing.3398  EPERM   The attribute cannot (currently) be accessed this way3399          (e.g. read-only attribute, or attribute that only makes3400          sense when the device is in a different state)3401  =====   =============================================================3402 3403  Other error conditions may be defined by individual device types.3404 3405Gets/sets a specified piece of device configuration and/or state.  The3406semantics are device-specific.  See individual device documentation in3407the "devices" directory.  As with ONE_REG, the size of the data3408transferred is defined by the particular attribute.3409 3410::3411 3412  struct kvm_device_attr {3413	__u32	flags;		/* no flags currently defined */3414	__u32	group;		/* device-defined */3415	__u64	attr;		/* group-defined */3416	__u64	addr;		/* userspace address of attr data */3417  };3418 34194.81 KVM_HAS_DEVICE_ATTR3420------------------------3421 3422:Capability: KVM_CAP_DEVICE_CTRL, KVM_CAP_VM_ATTRIBUTES for vm device,3423             KVM_CAP_VCPU_ATTRIBUTES for vcpu device3424             KVM_CAP_SYS_ATTRIBUTES for system (/dev/kvm) device3425:Type: device ioctl, vm ioctl, vcpu ioctl3426:Parameters: struct kvm_device_attr3427:Returns: 0 on success, -1 on error3428 3429Errors:3430 3431  =====   =============================================================3432  ENXIO   The group or attribute is unknown/unsupported for this device3433          or hardware support is missing.3434  =====   =============================================================3435 3436Tests whether a device supports a particular attribute.  A successful3437return indicates the attribute is implemented.  It does not necessarily3438indicate that the attribute can be read or written in the device's3439current state.  "addr" is ignored.3440 3441.. _KVM_ARM_VCPU_INIT:3442 34434.82 KVM_ARM_VCPU_INIT3444----------------------3445 3446:Capability: basic3447:Architectures: arm643448:Type: vcpu ioctl3449:Parameters: struct kvm_vcpu_init (in)3450:Returns: 0 on success; -1 on error3451 3452Errors:3453 3454  ======     =================================================================3455  EINVAL     the target is unknown, or the combination of features is invalid.3456  ENOENT     a features bit specified is unknown.3457  ======     =================================================================3458 3459This tells KVM what type of CPU to present to the guest, and what3460optional features it should have.  This will cause a reset of the cpu3461registers to their initial values.  If this is not called, KVM_RUN will3462return ENOEXEC for that vcpu.3463 3464The initial values are defined as:3465	- Processor state:3466		* AArch64: EL1h, D, A, I and F bits set. All other bits3467		  are cleared.3468		* AArch32: SVC, A, I and F bits set. All other bits are3469		  cleared.3470	- General Purpose registers, including PC and SP: set to 03471	- FPSIMD/NEON registers: set to 03472	- SVE registers: set to 03473	- System registers: Reset to their architecturally defined3474	  values as for a warm reset to EL1 (resp. SVC)3475 3476Note that because some registers reflect machine topology, all vcpus3477should be created before this ioctl is invoked.3478 3479Userspace can call this function multiple times for a given vcpu, including3480after the vcpu has been run. This will reset the vcpu to its initial3481state. All calls to this function after the initial call must use the same3482target and same set of feature flags, otherwise EINVAL will be returned.3483 3484Possible features:3485 3486	- KVM_ARM_VCPU_POWER_OFF: Starts the CPU in a power-off state.3487	  Depends on KVM_CAP_ARM_PSCI.  If not set, the CPU will be powered on3488	  and execute guest code when KVM_RUN is called.3489	- KVM_ARM_VCPU_EL1_32BIT: Starts the CPU in a 32bit mode.3490	  Depends on KVM_CAP_ARM_EL1_32BIT (arm64 only).3491	- KVM_ARM_VCPU_PSCI_0_2: Emulate PSCI v0.2 (or a future revision3492          backward compatible with v0.2) for the CPU.3493	  Depends on KVM_CAP_ARM_PSCI_0_2.3494	- KVM_ARM_VCPU_PMU_V3: Emulate PMUv3 for the CPU.3495	  Depends on KVM_CAP_ARM_PMU_V3.3496 3497	- KVM_ARM_VCPU_PTRAUTH_ADDRESS: Enables Address Pointer authentication3498	  for arm64 only.3499	  Depends on KVM_CAP_ARM_PTRAUTH_ADDRESS.3500	  If KVM_CAP_ARM_PTRAUTH_ADDRESS and KVM_CAP_ARM_PTRAUTH_GENERIC are3501	  both present, then both KVM_ARM_VCPU_PTRAUTH_ADDRESS and3502	  KVM_ARM_VCPU_PTRAUTH_GENERIC must be requested or neither must be3503	  requested.3504 3505	- KVM_ARM_VCPU_PTRAUTH_GENERIC: Enables Generic Pointer authentication3506	  for arm64 only.3507	  Depends on KVM_CAP_ARM_PTRAUTH_GENERIC.3508	  If KVM_CAP_ARM_PTRAUTH_ADDRESS and KVM_CAP_ARM_PTRAUTH_GENERIC are3509	  both present, then both KVM_ARM_VCPU_PTRAUTH_ADDRESS and3510	  KVM_ARM_VCPU_PTRAUTH_GENERIC must be requested or neither must be3511	  requested.3512 3513	- KVM_ARM_VCPU_SVE: Enables SVE for the CPU (arm64 only).3514	  Depends on KVM_CAP_ARM_SVE.3515	  Requires KVM_ARM_VCPU_FINALIZE(KVM_ARM_VCPU_SVE):3516 3517	   * After KVM_ARM_VCPU_INIT:3518 3519	      - KVM_REG_ARM64_SVE_VLS may be read using KVM_GET_ONE_REG: the3520	        initial value of this pseudo-register indicates the best set of3521	        vector lengths possible for a vcpu on this host.3522 3523	   * Before KVM_ARM_VCPU_FINALIZE(KVM_ARM_VCPU_SVE):3524 3525	      - KVM_RUN and KVM_GET_REG_LIST are not available;3526 3527	      - KVM_GET_ONE_REG and KVM_SET_ONE_REG cannot be used to access3528	        the scalable architectural SVE registers3529	        KVM_REG_ARM64_SVE_ZREG(), KVM_REG_ARM64_SVE_PREG() or3530	        KVM_REG_ARM64_SVE_FFR;3531 3532	      - KVM_REG_ARM64_SVE_VLS may optionally be written using3533	        KVM_SET_ONE_REG, to modify the set of vector lengths available3534	        for the vcpu.3535 3536	   * After KVM_ARM_VCPU_FINALIZE(KVM_ARM_VCPU_SVE):3537 3538	      - the KVM_REG_ARM64_SVE_VLS pseudo-register is immutable, and can3539	        no longer be written using KVM_SET_ONE_REG.3540 35414.83 KVM_ARM_PREFERRED_TARGET3542-----------------------------3543 3544:Capability: basic3545:Architectures: arm643546:Type: vm ioctl3547:Parameters: struct kvm_vcpu_init (out)3548:Returns: 0 on success; -1 on error3549 3550Errors:3551 3552  ======     ==========================================3553  ENODEV     no preferred target available for the host3554  ======     ==========================================3555 3556This queries KVM for preferred CPU target type which can be emulated3557by KVM on underlying host.3558 3559The ioctl returns struct kvm_vcpu_init instance containing information3560about preferred CPU target type and recommended features for it.  The3561kvm_vcpu_init->features bitmap returned will have feature bits set if3562the preferred target recommends setting these features, but this is3563not mandatory.3564 3565The information returned by this ioctl can be used to prepare an instance3566of struct kvm_vcpu_init for KVM_ARM_VCPU_INIT ioctl which will result in3567VCPU matching underlying host.3568 3569 35704.84 KVM_GET_REG_LIST3571---------------------3572 3573:Capability: basic3574:Architectures: arm64, mips, riscv3575:Type: vcpu ioctl3576:Parameters: struct kvm_reg_list (in/out)3577:Returns: 0 on success; -1 on error3578 3579Errors:3580 3581  =====      ==============================================================3582  E2BIG      the reg index list is too big to fit in the array specified by3583             the user (the number required will be written into n).3584  =====      ==============================================================3585 3586::3587 3588  struct kvm_reg_list {3589	__u64 n; /* number of registers in reg[] */3590	__u64 reg[0];3591  };3592 3593This ioctl returns the guest registers that are supported for the3594KVM_GET_ONE_REG/KVM_SET_ONE_REG calls.3595 3596 35974.85 KVM_ARM_SET_DEVICE_ADDR (deprecated)3598-----------------------------------------3599 3600:Capability: KVM_CAP_ARM_SET_DEVICE_ADDR3601:Architectures: arm643602:Type: vm ioctl3603:Parameters: struct kvm_arm_device_address (in)3604:Returns: 0 on success, -1 on error3605 3606Errors:3607 3608  ======  ============================================3609  ENODEV  The device id is unknown3610  ENXIO   Device not supported on current system3611  EEXIST  Address already set3612  E2BIG   Address outside guest physical address space3613  EBUSY   Address overlaps with other device range3614  ======  ============================================3615 3616::3617 3618  struct kvm_arm_device_addr {3619	__u64 id;3620	__u64 addr;3621  };3622 3623Specify a device address in the guest's physical address space where guests3624can access emulated or directly exposed devices, which the host kernel needs3625to know about. The id field is an architecture specific identifier for a3626specific device.3627 3628arm64 divides the id field into two parts, a device id and an3629address type id specific to the individual device::3630 3631  bits:  | 63        ...       32 | 31    ...    16 | 15    ...    0 |3632  field: |        0x00000000      |     device id   |  addr type id  |3633 3634arm64 currently only require this when using the in-kernel GIC3635support for the hardware VGIC features, using KVM_ARM_DEVICE_VGIC_V23636as the device id.  When setting the base address for the guest's3637mapping of the VGIC virtual CPU and distributor interface, the ioctl3638must be called after calling KVM_CREATE_IRQCHIP, but before calling3639KVM_RUN on any of the VCPUs.  Calling this ioctl twice for any of the3640base addresses will return -EEXIST.3641 3642Note, this IOCTL is deprecated and the more flexible SET/GET_DEVICE_ATTR API3643should be used instead.3644 3645 36464.86 KVM_PPC_RTAS_DEFINE_TOKEN3647------------------------------3648 3649:Capability: KVM_CAP_PPC_RTAS3650:Architectures: ppc3651:Type: vm ioctl3652:Parameters: struct kvm_rtas_token_args3653:Returns: 0 on success, -1 on error3654 3655Defines a token value for a RTAS (Run Time Abstraction Services)3656service in order to allow it to be handled in the kernel.  The3657argument struct gives the name of the service, which must be the name3658of a service that has a kernel-side implementation.  If the token3659value is non-zero, it will be associated with that service, and3660subsequent RTAS calls by the guest specifying that token will be3661handled by the kernel.  If the token value is 0, then any token3662associated with the service will be forgotten, and subsequent RTAS3663calls by the guest for that service will be passed to userspace to be3664handled.3665 36664.87 KVM_SET_GUEST_DEBUG3667------------------------3668 3669:Capability: KVM_CAP_SET_GUEST_DEBUG3670:Architectures: x86, s390, ppc, arm643671:Type: vcpu ioctl3672:Parameters: struct kvm_guest_debug (in)3673:Returns: 0 on success; -1 on error3674 3675::3676 3677  struct kvm_guest_debug {3678       __u32 control;3679       __u32 pad;3680       struct kvm_guest_debug_arch arch;3681  };3682 3683Set up the processor specific debug registers and configure vcpu for3684handling guest debug events. There are two parts to the structure, the3685first a control bitfield indicates the type of debug events to handle3686when running. Common control bits are:3687 3688  - KVM_GUESTDBG_ENABLE:        guest debugging is enabled3689  - KVM_GUESTDBG_SINGLESTEP:    the next run should single-step3690 3691The top 16 bits of the control field are architecture specific control3692flags which can include the following:3693 3694  - KVM_GUESTDBG_USE_SW_BP:     using software breakpoints [x86, arm64]3695  - KVM_GUESTDBG_USE_HW_BP:     using hardware breakpoints [x86, s390]3696  - KVM_GUESTDBG_USE_HW:        using hardware debug events [arm64]3697  - KVM_GUESTDBG_INJECT_DB:     inject DB type exception [x86]3698  - KVM_GUESTDBG_INJECT_BP:     inject BP type exception [x86]3699  - KVM_GUESTDBG_EXIT_PENDING:  trigger an immediate guest exit [s390]3700  - KVM_GUESTDBG_BLOCKIRQ:      avoid injecting interrupts/NMI/SMI [x86]3701 3702For example KVM_GUESTDBG_USE_SW_BP indicates that software breakpoints3703are enabled in memory so we need to ensure breakpoint exceptions are3704correctly trapped and the KVM run loop exits at the breakpoint and not3705running off into the normal guest vector. For KVM_GUESTDBG_USE_HW_BP3706we need to ensure the guest vCPUs architecture specific registers are3707updated to the correct (supplied) values.3708 3709The second part of the structure is architecture specific and3710typically contains a set of debug registers.3711 3712For arm64 the number of debug registers is implementation defined and3713can be determined by querying the KVM_CAP_GUEST_DEBUG_HW_BPS and3714KVM_CAP_GUEST_DEBUG_HW_WPS capabilities which return a positive number3715indicating the number of supported registers.3716 3717For ppc, the KVM_CAP_PPC_GUEST_DEBUG_SSTEP capability indicates whether3718the single-step debug event (KVM_GUESTDBG_SINGLESTEP) is supported.3719 3720Also when supported, KVM_CAP_SET_GUEST_DEBUG2 capability indicates the3721supported KVM_GUESTDBG_* bits in the control field.3722 3723When debug events exit the main run loop with the reason3724KVM_EXIT_DEBUG with the kvm_debug_exit_arch part of the kvm_run3725structure containing architecture specific debug information.3726 37274.88 KVM_GET_EMULATED_CPUID3728---------------------------3729 3730:Capability: KVM_CAP_EXT_EMUL_CPUID3731:Architectures: x863732:Type: system ioctl3733:Parameters: struct kvm_cpuid2 (in/out)3734:Returns: 0 on success, -1 on error3735 3736::3737 3738  struct kvm_cpuid2 {3739	__u32 nent;3740	__u32 flags;3741	struct kvm_cpuid_entry2 entries[0];3742  };3743 3744The member 'flags' is used for passing flags from userspace.3745 3746::3747 3748  #define KVM_CPUID_FLAG_SIGNIFCANT_INDEX		BIT(0)3749  #define KVM_CPUID_FLAG_STATEFUL_FUNC		BIT(1) /* deprecated */3750  #define KVM_CPUID_FLAG_STATE_READ_NEXT		BIT(2) /* deprecated */3751 3752  struct kvm_cpuid_entry2 {3753	__u32 function;3754	__u32 index;3755	__u32 flags;3756	__u32 eax;3757	__u32 ebx;3758	__u32 ecx;3759	__u32 edx;3760	__u32 padding[3];3761  };3762 3763This ioctl returns x86 cpuid features which are emulated by3764kvm.Userspace can use the information returned by this ioctl to query3765which features are emulated by kvm instead of being present natively.3766 3767Userspace invokes KVM_GET_EMULATED_CPUID by passing a kvm_cpuid23768structure with the 'nent' field indicating the number of entries in3769the variable-size array 'entries'. If the number of entries is too low3770to describe the cpu capabilities, an error (E2BIG) is returned. If the3771number is too high, the 'nent' field is adjusted and an error (ENOMEM)3772is returned. If the number is just right, the 'nent' field is adjusted3773to the number of valid entries in the 'entries' array, which is then3774filled.3775 3776The entries returned are the set CPUID bits of the respective features3777which kvm emulates, as returned by the CPUID instruction, with unknown3778or unsupported feature bits cleared.3779 3780Features like x2apic, for example, may not be present in the host cpu3781but are exposed by kvm in KVM_GET_SUPPORTED_CPUID because they can be3782emulated efficiently and thus not included here.3783 3784The fields in each entry are defined as follows:3785 3786  function:3787	 the eax value used to obtain the entry3788  index:3789	 the ecx value used to obtain the entry (for entries that are3790         affected by ecx)3791  flags:3792    an OR of zero or more of the following:3793 3794        KVM_CPUID_FLAG_SIGNIFCANT_INDEX:3795           if the index field is valid3796 3797   eax, ebx, ecx, edx:3798 3799         the values returned by the cpuid instruction for3800         this function/index combination3801 38024.89 KVM_S390_MEM_OP3803--------------------3804 3805:Capability: KVM_CAP_S390_MEM_OP, KVM_CAP_S390_PROTECTED, KVM_CAP_S390_MEM_OP_EXTENSION3806:Architectures: s3903807:Type: vm ioctl, vcpu ioctl3808:Parameters: struct kvm_s390_mem_op (in)3809:Returns: = 0 on success,3810          < 0 on generic error (e.g. -EFAULT or -ENOMEM),3811          16 bit program exception code if the access causes such an exception3812 3813Read or write data from/to the VM's memory.3814The KVM_CAP_S390_MEM_OP_EXTENSION capability specifies what functionality is3815supported.3816 3817Parameters are specified via the following structure::3818 3819  struct kvm_s390_mem_op {3820	__u64 gaddr;		/* the guest address */3821	__u64 flags;		/* flags */3822	__u32 size;		/* amount of bytes */3823	__u32 op;		/* type of operation */3824	__u64 buf;		/* buffer in userspace */3825	union {3826		struct {3827			__u8 ar;	/* the access register number */3828			__u8 key;	/* access key, ignored if flag unset */3829			__u8 pad1[6];	/* ignored */3830			__u64 old_addr;	/* ignored if flag unset */3831		};3832		__u32 sida_offset; /* offset into the sida */3833		__u8 reserved[32]; /* ignored */3834	};3835  };3836 3837The start address of the memory region has to be specified in the "gaddr"3838field, and the length of the region in the "size" field (which must not3839be 0). The maximum value for "size" can be obtained by checking the3840KVM_CAP_S390_MEM_OP capability. "buf" is the buffer supplied by the3841userspace application where the read data should be written to for3842a read access, or where the data that should be written is stored for3843a write access.  The "reserved" field is meant for future extensions.3844Reserved and unused values are ignored. Future extension that add members must3845introduce new flags.3846 3847The type of operation is specified in the "op" field. Flags modifying3848their behavior can be set in the "flags" field. Undefined flag bits must3849be set to 0.3850 3851Possible operations are:3852  * ``KVM_S390_MEMOP_LOGICAL_READ``3853  * ``KVM_S390_MEMOP_LOGICAL_WRITE``3854  * ``KVM_S390_MEMOP_ABSOLUTE_READ``3855  * ``KVM_S390_MEMOP_ABSOLUTE_WRITE``3856  * ``KVM_S390_MEMOP_SIDA_READ``3857  * ``KVM_S390_MEMOP_SIDA_WRITE``3858  * ``KVM_S390_MEMOP_ABSOLUTE_CMPXCHG``3859 3860Logical read/write:3861^^^^^^^^^^^^^^^^^^^3862 3863Access logical memory, i.e. translate the given guest address to an absolute3864address given the state of the VCPU and use the absolute address as target of3865the access. "ar" designates the access register number to be used; the valid3866range is 0..15.3867Logical accesses are permitted for the VCPU ioctl only.3868Logical accesses are permitted for non-protected guests only.3869 3870Supported flags:3871  * ``KVM_S390_MEMOP_F_CHECK_ONLY``3872  * ``KVM_S390_MEMOP_F_INJECT_EXCEPTION``3873  * ``KVM_S390_MEMOP_F_SKEY_PROTECTION``3874 3875The KVM_S390_MEMOP_F_CHECK_ONLY flag can be set to check whether the3876corresponding memory access would cause an access exception; however,3877no actual access to the data in memory at the destination is performed.3878In this case, "buf" is unused and can be NULL.3879 3880In case an access exception occurred during the access (or would occur3881in case of KVM_S390_MEMOP_F_CHECK_ONLY), the ioctl returns a positive3882error number indicating the type of exception. This exception is also3883raised directly at the corresponding VCPU if the flag3884KVM_S390_MEMOP_F_INJECT_EXCEPTION is set.3885On protection exceptions, unless specified otherwise, the injected3886translation-exception identifier (TEID) indicates suppression.3887 3888If the KVM_S390_MEMOP_F_SKEY_PROTECTION flag is set, storage key3889protection is also in effect and may cause exceptions if accesses are3890prohibited given the access key designated by "key"; the valid range is 0..15.3891KVM_S390_MEMOP_F_SKEY_PROTECTION is available if KVM_CAP_S390_MEM_OP_EXTENSION3892is > 0.3893Since the accessed memory may span multiple pages and those pages might have3894different storage keys, it is possible that a protection exception occurs3895after memory has been modified. In this case, if the exception is injected,3896the TEID does not indicate suppression.3897 3898Absolute read/write:3899^^^^^^^^^^^^^^^^^^^^3900 3901Access absolute memory. This operation is intended to be used with the3902KVM_S390_MEMOP_F_SKEY_PROTECTION flag, to allow accessing memory and performing3903the checks required for storage key protection as one operation (as opposed to3904user space getting the storage keys, performing the checks, and accessing3905memory thereafter, which could lead to a delay between check and access).3906Absolute accesses are permitted for the VM ioctl if KVM_CAP_S390_MEM_OP_EXTENSION3907has the KVM_S390_MEMOP_EXTENSION_CAP_BASE bit set.3908Currently absolute accesses are not permitted for VCPU ioctls.3909Absolute accesses are permitted for non-protected guests only.3910 3911Supported flags:3912  * ``KVM_S390_MEMOP_F_CHECK_ONLY``3913  * ``KVM_S390_MEMOP_F_SKEY_PROTECTION``3914 3915The semantics of the flags common with logical accesses are as for logical3916accesses.3917 3918Absolute cmpxchg:3919^^^^^^^^^^^^^^^^^3920 3921Perform cmpxchg on absolute guest memory. Intended for use with the3922KVM_S390_MEMOP_F_SKEY_PROTECTION flag.3923Instead of doing an unconditional write, the access occurs only if the target3924location contains the value pointed to by "old_addr".3925This is performed as an atomic cmpxchg with the length specified by the "size"3926parameter. "size" must be a power of two up to and including 16.3927If the exchange did not take place because the target value doesn't match the3928old value, the value "old_addr" points to is replaced by the target value.3929User space can tell if an exchange took place by checking if this replacement3930occurred. The cmpxchg op is permitted for the VM ioctl if3931KVM_CAP_S390_MEM_OP_EXTENSION has flag KVM_S390_MEMOP_EXTENSION_CAP_CMPXCHG set.3932 3933Supported flags:3934  * ``KVM_S390_MEMOP_F_SKEY_PROTECTION``3935 3936SIDA read/write:3937^^^^^^^^^^^^^^^^3938 3939Access the secure instruction data area which contains memory operands necessary3940for instruction emulation for protected guests.3941SIDA accesses are available if the KVM_CAP_S390_PROTECTED capability is available.3942SIDA accesses are permitted for the VCPU ioctl only.3943SIDA accesses are permitted for protected guests only.3944 3945No flags are supported.3946 39474.90 KVM_S390_GET_SKEYS3948-----------------------3949 3950:Capability: KVM_CAP_S390_SKEYS3951:Architectures: s3903952:Type: vm ioctl3953:Parameters: struct kvm_s390_skeys3954:Returns: 0 on success, KVM_S390_GET_SKEYS_NONE if guest is not using storage3955          keys, negative value on error3956 3957This ioctl is used to get guest storage key values on the s3903958architecture. The ioctl takes parameters via the kvm_s390_skeys struct::3959 3960  struct kvm_s390_skeys {3961	__u64 start_gfn;3962	__u64 count;3963	__u64 skeydata_addr;3964	__u32 flags;3965	__u32 reserved[9];3966  };3967 3968The start_gfn field is the number of the first guest frame whose storage keys3969you want to get.3970 3971The count field is the number of consecutive frames (starting from start_gfn)3972whose storage keys to get. The count field must be at least 1 and the maximum3973allowed value is defined as KVM_S390_SKEYS_MAX. Values outside this range3974will cause the ioctl to return -EINVAL.3975 3976The skeydata_addr field is the address to a buffer large enough to hold count3977bytes. This buffer will be filled with storage key data by the ioctl.3978 39794.91 KVM_S390_SET_SKEYS3980-----------------------3981 3982:Capability: KVM_CAP_S390_SKEYS3983:Architectures: s3903984:Type: vm ioctl3985:Parameters: struct kvm_s390_skeys3986:Returns: 0 on success, negative value on error3987 3988This ioctl is used to set guest storage key values on the s3903989architecture. The ioctl takes parameters via the kvm_s390_skeys struct.3990See section on KVM_S390_GET_SKEYS for struct definition.3991 3992The start_gfn field is the number of the first guest frame whose storage keys3993you want to set.3994 3995The count field is the number of consecutive frames (starting from start_gfn)3996whose storage keys to get. The count field must be at least 1 and the maximum3997allowed value is defined as KVM_S390_SKEYS_MAX. Values outside this range3998will cause the ioctl to return -EINVAL.3999 4000The skeydata_addr field is the address to a buffer containing count bytes of4001storage keys. Each byte in the buffer will be set as the storage key for a4002single frame starting at start_gfn for count frames.4003 4004Note: If any architecturally invalid key value is found in the given data then4005the ioctl will return -EINVAL.4006 40074.92 KVM_S390_IRQ4008-----------------4009 4010:Capability: KVM_CAP_S390_INJECT_IRQ4011:Architectures: s3904012:Type: vcpu ioctl4013:Parameters: struct kvm_s390_irq (in)4014:Returns: 0 on success, -1 on error4015 4016Errors:4017 4018 4019  ======  =================================================================4020  EINVAL  interrupt type is invalid4021          type is KVM_S390_SIGP_STOP and flag parameter is invalid value,4022          type is KVM_S390_INT_EXTERNAL_CALL and code is bigger4023          than the maximum of VCPUs4024  EBUSY   type is KVM_S390_SIGP_SET_PREFIX and vcpu is not stopped,4025          type is KVM_S390_SIGP_STOP and a stop irq is already pending,4026          type is KVM_S390_INT_EXTERNAL_CALL and an external call interrupt4027          is already pending4028  ======  =================================================================4029 4030Allows to inject an interrupt to the guest.4031 4032Using struct kvm_s390_irq as a parameter allows4033to inject additional payload which is not4034possible via KVM_S390_INTERRUPT.4035 4036Interrupt parameters are passed via kvm_s390_irq::4037 4038  struct kvm_s390_irq {4039	__u64 type;4040	union {4041		struct kvm_s390_io_info io;4042		struct kvm_s390_ext_info ext;4043		struct kvm_s390_pgm_info pgm;4044		struct kvm_s390_emerg_info emerg;4045		struct kvm_s390_extcall_info extcall;4046		struct kvm_s390_prefix_info prefix;4047		struct kvm_s390_stop_info stop;4048		struct kvm_s390_mchk_info mchk;4049		char reserved[64];4050	} u;4051  };4052 4053type can be one of the following:4054 4055- KVM_S390_SIGP_STOP - sigp stop; parameter in .stop4056- KVM_S390_PROGRAM_INT - program check; parameters in .pgm4057- KVM_S390_SIGP_SET_PREFIX - sigp set prefix; parameters in .prefix4058- KVM_S390_RESTART - restart; no parameters4059- KVM_S390_INT_CLOCK_COMP - clock comparator interrupt; no parameters4060- KVM_S390_INT_CPU_TIMER - CPU timer interrupt; no parameters4061- KVM_S390_INT_EMERGENCY - sigp emergency; parameters in .emerg4062- KVM_S390_INT_EXTERNAL_CALL - sigp external call; parameters in .extcall4063- KVM_S390_MCHK - machine check interrupt; parameters in .mchk4064 4065This is an asynchronous vcpu ioctl and can be invoked from any thread.4066 40674.94 KVM_S390_GET_IRQ_STATE4068---------------------------4069 4070:Capability: KVM_CAP_S390_IRQ_STATE4071:Architectures: s3904072:Type: vcpu ioctl4073:Parameters: struct kvm_s390_irq_state (out)4074:Returns: >= number of bytes copied into buffer,4075          -EINVAL if buffer size is 0,4076          -ENOBUFS if buffer size is too small to fit all pending interrupts,4077          -EFAULT if the buffer address was invalid4078 4079This ioctl allows userspace to retrieve the complete state of all currently4080pending interrupts in a single buffer. Use cases include migration4081and introspection. The parameter structure contains the address of a4082userspace buffer and its length::4083 4084  struct kvm_s390_irq_state {4085	__u64 buf;4086	__u32 flags;        /* will stay unused for compatibility reasons */4087	__u32 len;4088	__u32 reserved[4];  /* will stay unused for compatibility reasons */4089  };4090 4091Userspace passes in the above struct and for each pending interrupt a4092struct kvm_s390_irq is copied to the provided buffer.4093 4094The structure contains a flags and a reserved field for future extensions. As4095the kernel never checked for flags == 0 and QEMU never pre-zeroed flags and4096reserved, these fields can not be used in the future without breaking4097compatibility.4098 4099If -ENOBUFS is returned the buffer provided was too small and userspace4100may retry with a bigger buffer.4101 41024.95 KVM_S390_SET_IRQ_STATE4103---------------------------4104 4105:Capability: KVM_CAP_S390_IRQ_STATE4106:Architectures: s3904107:Type: vcpu ioctl4108:Parameters: struct kvm_s390_irq_state (in)4109:Returns: 0 on success,4110          -EFAULT if the buffer address was invalid,4111          -EINVAL for an invalid buffer length (see below),4112          -EBUSY if there were already interrupts pending,4113          errors occurring when actually injecting the4114          interrupt. See KVM_S390_IRQ.4115 4116This ioctl allows userspace to set the complete state of all cpu-local4117interrupts currently pending for the vcpu. It is intended for restoring4118interrupt state after a migration. The input parameter is a userspace buffer4119containing a struct kvm_s390_irq_state::4120 4121  struct kvm_s390_irq_state {4122	__u64 buf;4123	__u32 flags;        /* will stay unused for compatibility reasons */4124	__u32 len;4125	__u32 reserved[4];  /* will stay unused for compatibility reasons */4126  };4127 4128The restrictions for flags and reserved apply as well.4129(see KVM_S390_GET_IRQ_STATE)4130 4131The userspace memory referenced by buf contains a struct kvm_s390_irq4132for each interrupt to be injected into the guest.4133If one of the interrupts could not be injected for some reason the4134ioctl aborts.4135 4136len must be a multiple of sizeof(struct kvm_s390_irq). It must be > 04137and it must not exceed (max_vcpus + 32) * sizeof(struct kvm_s390_irq),4138which is the maximum number of possibly pending cpu-local interrupts.4139 41404.96 KVM_SMI4141------------4142 4143:Capability: KVM_CAP_X86_SMM4144:Architectures: x864145:Type: vcpu ioctl4146:Parameters: none4147:Returns: 0 on success, -1 on error4148 4149Queues an SMI on the thread's vcpu.4150 41514.97 KVM_X86_SET_MSR_FILTER4152----------------------------4153 4154:Capability: KVM_CAP_X86_MSR_FILTER4155:Architectures: x864156:Type: vm ioctl4157:Parameters: struct kvm_msr_filter4158:Returns: 0 on success, < 0 on error4159 4160::4161 4162  struct kvm_msr_filter_range {4163  #define KVM_MSR_FILTER_READ  (1 << 0)4164  #define KVM_MSR_FILTER_WRITE (1 << 1)4165	__u32 flags;4166	__u32 nmsrs; /* number of msrs in bitmap */4167	__u32 base;  /* MSR index the bitmap starts at */4168	__u8 *bitmap; /* a 1 bit allows the operations in flags, 0 denies */4169  };4170 4171  #define KVM_MSR_FILTER_MAX_RANGES 164172  struct kvm_msr_filter {4173  #define KVM_MSR_FILTER_DEFAULT_ALLOW (0 << 0)4174  #define KVM_MSR_FILTER_DEFAULT_DENY  (1 << 0)4175	__u32 flags;4176	struct kvm_msr_filter_range ranges[KVM_MSR_FILTER_MAX_RANGES];4177  };4178 4179flags values for ``struct kvm_msr_filter_range``:4180 4181``KVM_MSR_FILTER_READ``4182 4183  Filter read accesses to MSRs using the given bitmap. A 0 in the bitmap4184  indicates that read accesses should be denied, while a 1 indicates that4185  a read for a particular MSR should be allowed regardless of the default4186  filter action.4187 4188``KVM_MSR_FILTER_WRITE``4189 4190  Filter write accesses to MSRs using the given bitmap. A 0 in the bitmap4191  indicates that write accesses should be denied, while a 1 indicates that4192  a write for a particular MSR should be allowed regardless of the default4193  filter action.4194 4195flags values for ``struct kvm_msr_filter``:4196 4197``KVM_MSR_FILTER_DEFAULT_ALLOW``4198 4199  If no filter range matches an MSR index that is getting accessed, KVM will4200  allow accesses to all MSRs by default.4201 4202``KVM_MSR_FILTER_DEFAULT_DENY``4203 4204  If no filter range matches an MSR index that is getting accessed, KVM will4205  deny accesses to all MSRs by default.4206 4207This ioctl allows userspace to define up to 16 bitmaps of MSR ranges to deny4208guest MSR accesses that would normally be allowed by KVM.  If an MSR is not4209covered by a specific range, the "default" filtering behavior applies.  Each4210bitmap range covers MSRs from [base .. base+nmsrs).4211 4212If an MSR access is denied by userspace, the resulting KVM behavior depends on4213whether or not KVM_CAP_X86_USER_SPACE_MSR's KVM_MSR_EXIT_REASON_FILTER is4214enabled.  If KVM_MSR_EXIT_REASON_FILTER is enabled, KVM will exit to userspace4215on denied accesses, i.e. userspace effectively intercepts the MSR access.  If4216KVM_MSR_EXIT_REASON_FILTER is not enabled, KVM will inject a #GP into the guest4217on denied accesses.  Note, if an MSR access is denied during emulation of MSR4218load/stores during VMX transitions, KVM ignores KVM_MSR_EXIT_REASON_FILTER.4219See the below warning for full details.4220 4221If an MSR access is allowed by userspace, KVM will emulate and/or virtualize4222the access in accordance with the vCPU model.  Note, KVM may still ultimately4223inject a #GP if an access is allowed by userspace, e.g. if KVM doesn't support4224the MSR, or to follow architectural behavior for the MSR.4225 4226By default, KVM operates in KVM_MSR_FILTER_DEFAULT_ALLOW mode with no MSR range4227filters.4228 4229Calling this ioctl with an empty set of ranges (all nmsrs == 0) disables MSR4230filtering. In that mode, ``KVM_MSR_FILTER_DEFAULT_DENY`` is invalid and causes4231an error.4232 4233.. warning::4234   MSR accesses that are side effects of instruction execution (emulated or4235   native) are not filtered as hardware does not honor MSR bitmaps outside of4236   RDMSR and WRMSR, and KVM mimics that behavior when emulating instructions4237   to avoid pointless divergence from hardware.  E.g. RDPID reads MSR_TSC_AUX,4238   SYSENTER reads the SYSENTER MSRs, etc.4239 4240   MSRs that are loaded/stored via dedicated VMCS fields are not filtered as4241   part of VM-Enter/VM-Exit emulation.4242 4243   MSRs that are loaded/store via VMX's load/store lists _are_ filtered as part4244   of VM-Enter/VM-Exit emulation.  If an MSR access is denied on VM-Enter, KVM4245   synthesizes a consistency check VM-Exit(EXIT_REASON_MSR_LOAD_FAIL).  If an4246   MSR access is denied on VM-Exit, KVM synthesizes a VM-Abort.  In short, KVM4247   extends Intel's architectural list of MSRs that cannot be loaded/saved via4248   the VM-Enter/VM-Exit MSR list.  It is platform owner's responsibility to4249   to communicate any such restrictions to their end users.4250 4251   x2APIC MSR accesses cannot be filtered (KVM silently ignores filters that4252   cover any x2APIC MSRs).4253 4254Note, invoking this ioctl while a vCPU is running is inherently racy.  However,4255KVM does guarantee that vCPUs will see either the previous filter or the new4256filter, e.g. MSRs with identical settings in both the old and new filter will4257have deterministic behavior.4258 4259Similarly, if userspace wishes to intercept on denied accesses,4260KVM_MSR_EXIT_REASON_FILTER must be enabled before activating any filters, and4261left enabled until after all filters are deactivated.  Failure to do so may4262result in KVM injecting a #GP instead of exiting to userspace.4263 42644.98 KVM_CREATE_SPAPR_TCE_644265----------------------------4266 4267:Capability: KVM_CAP_SPAPR_TCE_644268:Architectures: powerpc4269:Type: vm ioctl4270:Parameters: struct kvm_create_spapr_tce_64 (in)4271:Returns: file descriptor for manipulating the created TCE table4272 4273This is an extension for KVM_CAP_SPAPR_TCE which only supports 32bit4274windows, described in 4.62 KVM_CREATE_SPAPR_TCE4275 4276This capability uses extended struct in ioctl interface::4277 4278  /* for KVM_CAP_SPAPR_TCE_64 */4279  struct kvm_create_spapr_tce_64 {4280	__u64 liobn;4281	__u32 page_shift;4282	__u32 flags;4283	__u64 offset;	/* in pages */4284	__u64 size; 	/* in pages */4285  };4286 4287The aim of extension is to support an additional bigger DMA window with4288a variable page size.4289KVM_CREATE_SPAPR_TCE_64 receives a 64bit window size, an IOMMU page shift and4290a bus offset of the corresponding DMA window, @size and @offset are numbers4291of IOMMU pages.4292 4293@flags are not used at the moment.4294 4295The rest of functionality is identical to KVM_CREATE_SPAPR_TCE.4296 42974.99 KVM_REINJECT_CONTROL4298-------------------------4299 4300:Capability: KVM_CAP_REINJECT_CONTROL4301:Architectures: x864302:Type: vm ioctl4303:Parameters: struct kvm_reinject_control (in)4304:Returns: 0 on success,4305         -EFAULT if struct kvm_reinject_control cannot be read,4306         -ENXIO if KVM_CREATE_PIT or KVM_CREATE_PIT2 didn't succeed earlier.4307 4308i8254 (PIT) has two modes, reinject and !reinject.  The default is reinject,4309where KVM queues elapsed i8254 ticks and monitors completion of interrupt from4310vector(s) that i8254 injects.  Reinject mode dequeues a tick and injects its4311interrupt whenever there isn't a pending interrupt from i8254.4312!reinject mode injects an interrupt as soon as a tick arrives.4313 4314::4315 4316  struct kvm_reinject_control {4317	__u8 pit_reinject;4318	__u8 reserved[31];4319  };4320 4321pit_reinject = 0 (!reinject mode) is recommended, unless running an old4322operating system that uses the PIT for timing (e.g. Linux 2.4.x).4323 43244.100 KVM_PPC_CONFIGURE_V3_MMU4325------------------------------4326 4327:Capability: KVM_CAP_PPC_MMU_RADIX or KVM_CAP_PPC_MMU_HASH_V34328:Architectures: ppc4329:Type: vm ioctl4330:Parameters: struct kvm_ppc_mmuv3_cfg (in)4331:Returns: 0 on success,4332         -EFAULT if struct kvm_ppc_mmuv3_cfg cannot be read,4333         -EINVAL if the configuration is invalid4334 4335This ioctl controls whether the guest will use radix or HPT (hashed4336page table) translation, and sets the pointer to the process table for4337the guest.4338 4339::4340 4341  struct kvm_ppc_mmuv3_cfg {4342	__u64	flags;4343	__u64	process_table;4344  };4345 4346There are two bits that can be set in flags; KVM_PPC_MMUV3_RADIX and4347KVM_PPC_MMUV3_GTSE.  KVM_PPC_MMUV3_RADIX, if set, configures the guest4348to use radix tree translation, and if clear, to use HPT translation.4349KVM_PPC_MMUV3_GTSE, if set and if KVM permits it, configures the guest4350to be able to use the global TLB and SLB invalidation instructions;4351if clear, the guest may not use these instructions.4352 4353The process_table field specifies the address and size of the guest4354process table, which is in the guest's space.  This field is formatted4355as the second doubleword of the partition table entry, as defined in4356the Power ISA V3.00, Book III section 5.7.6.1.4357 43584.101 KVM_PPC_GET_RMMU_INFO4359---------------------------4360 4361:Capability: KVM_CAP_PPC_MMU_RADIX4362:Architectures: ppc4363:Type: vm ioctl4364:Parameters: struct kvm_ppc_rmmu_info (out)4365:Returns: 0 on success,4366	 -EFAULT if struct kvm_ppc_rmmu_info cannot be written,4367	 -EINVAL if no useful information can be returned4368 4369This ioctl returns a structure containing two things: (a) a list4370containing supported radix tree geometries, and (b) a list that maps4371page sizes to put in the "AP" (actual page size) field for the tlbie4372(TLB invalidate entry) instruction.4373 4374::4375 4376  struct kvm_ppc_rmmu_info {4377	struct kvm_ppc_radix_geom {4378		__u8	page_shift;4379		__u8	level_bits[4];4380		__u8	pad[3];4381	}	geometries[8];4382	__u32	ap_encodings[8];4383  };4384 4385The geometries[] field gives up to 8 supported geometries for the4386radix page table, in terms of the log base 2 of the smallest page4387size, and the number of bits indexed at each level of the tree, from4388the PTE level up to the PGD level in that order.  Any unused entries4389will have 0 in the page_shift field.4390 4391The ap_encodings gives the supported page sizes and their AP field4392encodings, encoded with the AP value in the top 3 bits and the log4393base 2 of the page size in the bottom 6 bits.4394 43954.102 KVM_PPC_RESIZE_HPT_PREPARE4396--------------------------------4397 4398:Capability: KVM_CAP_SPAPR_RESIZE_HPT4399:Architectures: powerpc4400:Type: vm ioctl4401:Parameters: struct kvm_ppc_resize_hpt (in)4402:Returns: 0 on successful completion,4403	 >0 if a new HPT is being prepared, the value is an estimated4404         number of milliseconds until preparation is complete,4405         -EFAULT if struct kvm_reinject_control cannot be read,4406	 -EINVAL if the supplied shift or flags are invalid,4407	 -ENOMEM if unable to allocate the new HPT,4408 4409Used to implement the PAPR extension for runtime resizing of a guest's4410Hashed Page Table (HPT).  Specifically this starts, stops or monitors4411the preparation of a new potential HPT for the guest, essentially4412implementing the H_RESIZE_HPT_PREPARE hypercall.4413 4414::4415 4416  struct kvm_ppc_resize_hpt {4417	__u64 flags;4418	__u32 shift;4419	__u32 pad;4420  };4421 4422If called with shift > 0 when there is no pending HPT for the guest,4423this begins preparation of a new pending HPT of size 2^(shift) bytes.4424It then returns a positive integer with the estimated number of4425milliseconds until preparation is complete.4426 4427If called when there is a pending HPT whose size does not match that4428requested in the parameters, discards the existing pending HPT and4429creates a new one as above.4430 4431If called when there is a pending HPT of the size requested, will:4432 4433  * If preparation of the pending HPT is already complete, return 04434  * If preparation of the pending HPT has failed, return an error4435    code, then discard the pending HPT.4436  * If preparation of the pending HPT is still in progress, return an4437    estimated number of milliseconds until preparation is complete.4438 4439If called with shift == 0, discards any currently pending HPT and4440returns 0 (i.e. cancels any in-progress preparation).4441 4442flags is reserved for future expansion, currently setting any bits in4443flags will result in an -EINVAL.4444 4445Normally this will be called repeatedly with the same parameters until4446it returns <= 0.  The first call will initiate preparation, subsequent4447ones will monitor preparation until it completes or fails.4448 44494.103 KVM_PPC_RESIZE_HPT_COMMIT4450-------------------------------4451 4452:Capability: KVM_CAP_SPAPR_RESIZE_HPT4453:Architectures: powerpc4454:Type: vm ioctl4455:Parameters: struct kvm_ppc_resize_hpt (in)4456:Returns: 0 on successful completion,4457         -EFAULT if struct kvm_reinject_control cannot be read,4458	 -EINVAL if the supplied shift or flags are invalid,4459	 -ENXIO is there is no pending HPT, or the pending HPT doesn't4460         have the requested size,4461	 -EBUSY if the pending HPT is not fully prepared,4462	 -ENOSPC if there was a hash collision when moving existing4463         HPT entries to the new HPT,4464	 -EIO on other error conditions4465 4466Used to implement the PAPR extension for runtime resizing of a guest's4467Hashed Page Table (HPT).  Specifically this requests that the guest be4468transferred to working with the new HPT, essentially implementing the4469H_RESIZE_HPT_COMMIT hypercall.4470 4471::4472 4473  struct kvm_ppc_resize_hpt {4474	__u64 flags;4475	__u32 shift;4476	__u32 pad;4477  };4478 4479This should only be called after KVM_PPC_RESIZE_HPT_PREPARE has4480returned 0 with the same parameters.  In other cases4481KVM_PPC_RESIZE_HPT_COMMIT will return an error (usually -ENXIO or4482-EBUSY, though others may be possible if the preparation was started,4483but failed).4484 4485This will have undefined effects on the guest if it has not already4486placed itself in a quiescent state where no vcpu will make MMU enabled4487memory accesses.4488 4489On successful completion, the pending HPT will become the guest's active4490HPT and the previous HPT will be discarded.4491 4492On failure, the guest will still be operating on its previous HPT.4493 44944.104 KVM_X86_GET_MCE_CAP_SUPPORTED4495-----------------------------------4496 4497:Capability: KVM_CAP_MCE4498:Architectures: x864499:Type: system ioctl4500:Parameters: u64 mce_cap (out)4501:Returns: 0 on success, -1 on error4502 4503Returns supported MCE capabilities. The u64 mce_cap parameter4504has the same format as the MSR_IA32_MCG_CAP register. Supported4505capabilities will have the corresponding bits set.4506 45074.105 KVM_X86_SETUP_MCE4508-----------------------4509 4510:Capability: KVM_CAP_MCE4511:Architectures: x864512:Type: vcpu ioctl4513:Parameters: u64 mcg_cap (in)4514:Returns: 0 on success,4515         -EFAULT if u64 mcg_cap cannot be read,4516         -EINVAL if the requested number of banks is invalid,4517         -EINVAL if requested MCE capability is not supported.4518 4519Initializes MCE support for use. The u64 mcg_cap parameter4520has the same format as the MSR_IA32_MCG_CAP register and4521specifies which capabilities should be enabled. The maximum4522supported number of error-reporting banks can be retrieved when4523checking for KVM_CAP_MCE. The supported capabilities can be4524retrieved with KVM_X86_GET_MCE_CAP_SUPPORTED.4525 45264.106 KVM_X86_SET_MCE4527---------------------4528 4529:Capability: KVM_CAP_MCE4530:Architectures: x864531:Type: vcpu ioctl4532:Parameters: struct kvm_x86_mce (in)4533:Returns: 0 on success,4534         -EFAULT if struct kvm_x86_mce cannot be read,4535         -EINVAL if the bank number is invalid,4536         -EINVAL if VAL bit is not set in status field.4537 4538Inject a machine check error (MCE) into the guest. The input4539parameter is::4540 4541  struct kvm_x86_mce {4542	__u64 status;4543	__u64 addr;4544	__u64 misc;4545	__u64 mcg_status;4546	__u8 bank;4547	__u8 pad1[7];4548	__u64 pad2[3];4549  };4550 4551If the MCE being reported is an uncorrected error, KVM will4552inject it as an MCE exception into the guest. If the guest4553MCG_STATUS register reports that an MCE is in progress, KVM4554causes an KVM_EXIT_SHUTDOWN vmexit.4555 4556Otherwise, if the MCE is a corrected error, KVM will just4557store it in the corresponding bank (provided this bank is4558not holding a previously reported uncorrected error).4559 45604.107 KVM_S390_GET_CMMA_BITS4561----------------------------4562 4563:Capability: KVM_CAP_S390_CMMA_MIGRATION4564:Architectures: s3904565:Type: vm ioctl4566:Parameters: struct kvm_s390_cmma_log (in, out)4567:Returns: 0 on success, a negative value on error4568 4569Errors:4570 4571  ======     =============================================================4572  ENOMEM     not enough memory can be allocated to complete the task4573  ENXIO      if CMMA is not enabled4574  EINVAL     if KVM_S390_CMMA_PEEK is not set but migration mode was not enabled4575  EINVAL     if KVM_S390_CMMA_PEEK is not set but dirty tracking has been4576             disabled (and thus migration mode was automatically disabled)4577  EFAULT     if the userspace address is invalid or if no page table is4578             present for the addresses (e.g. when using hugepages).4579  ======     =============================================================4580 4581This ioctl is used to get the values of the CMMA bits on the s3904582architecture. It is meant to be used in two scenarios:4583 4584- During live migration to save the CMMA values. Live migration needs4585  to be enabled via the KVM_REQ_START_MIGRATION VM property.4586- To non-destructively peek at the CMMA values, with the flag4587  KVM_S390_CMMA_PEEK set.4588 4589The ioctl takes parameters via the kvm_s390_cmma_log struct. The desired4590values are written to a buffer whose location is indicated via the "values"4591member in the kvm_s390_cmma_log struct.  The values in the input struct are4592also updated as needed.4593 4594Each CMMA value takes up one byte.4595 4596::4597 4598  struct kvm_s390_cmma_log {4599	__u64 start_gfn;4600	__u32 count;4601	__u32 flags;4602	union {4603		__u64 remaining;4604		__u64 mask;4605	};4606	__u64 values;4607  };4608 4609start_gfn is the number of the first guest frame whose CMMA values are4610to be retrieved,4611 4612count is the length of the buffer in bytes,4613 4614values points to the buffer where the result will be written to.4615 4616If count is greater than KVM_S390_SKEYS_MAX, then it is considered to be4617KVM_S390_SKEYS_MAX. KVM_S390_SKEYS_MAX is re-used for consistency with4618other ioctls.4619 4620The result is written in the buffer pointed to by the field values, and4621the values of the input parameter are updated as follows.4622 4623Depending on the flags, different actions are performed. The only4624supported flag so far is KVM_S390_CMMA_PEEK.4625 4626The default behaviour if KVM_S390_CMMA_PEEK is not set is:4627start_gfn will indicate the first page frame whose CMMA bits were dirty.4628It is not necessarily the same as the one passed as input, as clean pages4629are skipped.4630 4631count will indicate the number of bytes actually written in the buffer.4632It can (and very often will) be smaller than the input value, since the4633buffer is only filled until 16 bytes of clean values are found (which4634are then not copied in the buffer). Since a CMMA migration block needs4635the base address and the length, for a total of 16 bytes, we will send4636back some clean data if there is some dirty data afterwards, as long as4637the size of the clean data does not exceed the size of the header. This4638allows to minimize the amount of data to be saved or transferred over4639the network at the expense of more roundtrips to userspace. The next4640invocation of the ioctl will skip over all the clean values, saving4641potentially more than just the 16 bytes we found.4642 4643If KVM_S390_CMMA_PEEK is set:4644the existing storage attributes are read even when not in migration4645mode, and no other action is performed;4646 4647the output start_gfn will be equal to the input start_gfn,4648 4649the output count will be equal to the input count, except if the end of4650memory has been reached.4651 4652In both cases:4653the field "remaining" will indicate the total number of dirty CMMA values4654still remaining, or 0 if KVM_S390_CMMA_PEEK is set and migration mode is4655not enabled.4656 4657mask is unused.4658 4659values points to the userspace buffer where the result will be stored.4660 46614.108 KVM_S390_SET_CMMA_BITS4662----------------------------4663 4664:Capability: KVM_CAP_S390_CMMA_MIGRATION4665:Architectures: s3904666:Type: vm ioctl4667:Parameters: struct kvm_s390_cmma_log (in)4668:Returns: 0 on success, a negative value on error4669 4670This ioctl is used to set the values of the CMMA bits on the s3904671architecture. It is meant to be used during live migration to restore4672the CMMA values, but there are no restrictions on its use.4673The ioctl takes parameters via the kvm_s390_cmma_values struct.4674Each CMMA value takes up one byte.4675 4676::4677 4678  struct kvm_s390_cmma_log {4679	__u64 start_gfn;4680	__u32 count;4681	__u32 flags;4682	union {4683		__u64 remaining;4684		__u64 mask;4685 	};4686	__u64 values;4687  };4688 4689start_gfn indicates the starting guest frame number,4690 4691count indicates how many values are to be considered in the buffer,4692 4693flags is not used and must be 0.4694 4695mask indicates which PGSTE bits are to be considered.4696 4697remaining is not used.4698 4699values points to the buffer in userspace where to store the values.4700 4701This ioctl can fail with -ENOMEM if not enough memory can be allocated to4702complete the task, with -ENXIO if CMMA is not enabled, with -EINVAL if4703the count field is too large (e.g. more than KVM_S390_CMMA_SIZE_MAX) or4704if the flags field was not 0, with -EFAULT if the userspace address is4705invalid, if invalid pages are written to (e.g. after the end of memory)4706or if no page table is present for the addresses (e.g. when using4707hugepages).4708 47094.109 KVM_PPC_GET_CPU_CHAR4710--------------------------4711 4712:Capability: KVM_CAP_PPC_GET_CPU_CHAR4713:Architectures: powerpc4714:Type: vm ioctl4715:Parameters: struct kvm_ppc_cpu_char (out)4716:Returns: 0 on successful completion,4717	 -EFAULT if struct kvm_ppc_cpu_char cannot be written4718 4719This ioctl gives userspace information about certain characteristics4720of the CPU relating to speculative execution of instructions and4721possible information leakage resulting from speculative execution (see4722CVE-2017-5715, CVE-2017-5753 and CVE-2017-5754).  The information is4723returned in struct kvm_ppc_cpu_char, which looks like this::4724 4725  struct kvm_ppc_cpu_char {4726	__u64	character;		/* characteristics of the CPU */4727	__u64	behaviour;		/* recommended software behaviour */4728	__u64	character_mask;		/* valid bits in character */4729	__u64	behaviour_mask;		/* valid bits in behaviour */4730  };4731 4732For extensibility, the character_mask and behaviour_mask fields4733indicate which bits of character and behaviour have been filled in by4734the kernel.  If the set of defined bits is extended in future then4735userspace will be able to tell whether it is running on a kernel that4736knows about the new bits.4737 4738The character field describes attributes of the CPU which can help4739with preventing inadvertent information disclosure - specifically,4740whether there is an instruction to flash-invalidate the L1 data cache4741(ori 30,30,0 or mtspr SPRN_TRIG2,rN), whether the L1 data cache is set4742to a mode where entries can only be used by the thread that created4743them, whether the bcctr[l] instruction prevents speculation, and4744whether a speculation barrier instruction (ori 31,31,0) is provided.4745 4746The behaviour field describes actions that software should take to4747prevent inadvertent information disclosure, and thus describes which4748vulnerabilities the hardware is subject to; specifically whether the4749L1 data cache should be flushed when returning to user mode from the4750kernel, and whether a speculation barrier should be placed between an4751array bounds check and the array access.4752 4753These fields use the same bit definitions as the new4754H_GET_CPU_CHARACTERISTICS hypercall.4755 47564.110 KVM_MEMORY_ENCRYPT_OP4757---------------------------4758 4759:Capability: basic4760:Architectures: x864761:Type: vm4762:Parameters: an opaque platform specific structure (in/out)4763:Returns: 0 on success; -1 on error4764 4765If the platform supports creating encrypted VMs then this ioctl can be used4766for issuing platform-specific memory encryption commands to manage those4767encrypted VMs.4768 4769Currently, this ioctl is used for issuing Secure Encrypted Virtualization4770(SEV) commands on AMD Processors. The SEV commands are defined in4771Documentation/virt/kvm/x86/amd-memory-encryption.rst.4772 47734.111 KVM_MEMORY_ENCRYPT_REG_REGION4774-----------------------------------4775 4776:Capability: basic4777:Architectures: x864778:Type: system4779:Parameters: struct kvm_enc_region (in)4780:Returns: 0 on success; -1 on error4781 4782This ioctl can be used to register a guest memory region which may4783contain encrypted data (e.g. guest RAM, SMRAM etc).4784 4785It is used in the SEV-enabled guest. When encryption is enabled, a guest4786memory region may contain encrypted data. The SEV memory encryption4787engine uses a tweak such that two identical plaintext pages, each at4788different locations will have differing ciphertexts. So swapping or4789moving ciphertext of those pages will not result in plaintext being4790swapped. So relocating (or migrating) physical backing pages for the SEV4791guest will require some additional steps.4792 4793Note: The current SEV key management spec does not provide commands to4794swap or migrate (move) ciphertext pages. Hence, for now we pin the guest4795memory region registered with the ioctl.4796 47974.112 KVM_MEMORY_ENCRYPT_UNREG_REGION4798-------------------------------------4799 4800:Capability: basic4801:Architectures: x864802:Type: system4803:Parameters: struct kvm_enc_region (in)4804:Returns: 0 on success; -1 on error4805 4806This ioctl can be used to unregister the guest memory region registered4807with KVM_MEMORY_ENCRYPT_REG_REGION ioctl above.4808 48094.113 KVM_HYPERV_EVENTFD4810------------------------4811 4812:Capability: KVM_CAP_HYPERV_EVENTFD4813:Architectures: x864814:Type: vm ioctl4815:Parameters: struct kvm_hyperv_eventfd (in)4816 4817This ioctl (un)registers an eventfd to receive notifications from the guest on4818the specified Hyper-V connection id through the SIGNAL_EVENT hypercall, without4819causing a user exit.  SIGNAL_EVENT hypercall with non-zero event flag number4820(bits 24-31) still triggers a KVM_EXIT_HYPERV_HCALL user exit.4821 4822::4823 4824  struct kvm_hyperv_eventfd {4825	__u32 conn_id;4826	__s32 fd;4827	__u32 flags;4828	__u32 padding[3];4829  };4830 4831The conn_id field should fit within 24 bits::4832 4833  #define KVM_HYPERV_CONN_ID_MASK		0x00ffffff4834 4835The acceptable values for the flags field are::4836 4837  #define KVM_HYPERV_EVENTFD_DEASSIGN	(1 << 0)4838 4839:Returns: 0 on success,4840 	  -EINVAL if conn_id or flags is outside the allowed range,4841	  -ENOENT on deassign if the conn_id isn't registered,4842	  -EEXIST on assign if the conn_id is already registered4843 48444.114 KVM_GET_NESTED_STATE4845--------------------------4846 4847:Capability: KVM_CAP_NESTED_STATE4848:Architectures: x864849:Type: vcpu ioctl4850:Parameters: struct kvm_nested_state (in/out)4851:Returns: 0 on success, -1 on error4852 4853Errors:4854 4855  =====      =============================================================4856  E2BIG      the total state size exceeds the value of 'size' specified by4857             the user; the size required will be written into size.4858  =====      =============================================================4859 4860::4861 4862  struct kvm_nested_state {4863	__u16 flags;4864	__u16 format;4865	__u32 size;4866 4867	union {4868		struct kvm_vmx_nested_state_hdr vmx;4869		struct kvm_svm_nested_state_hdr svm;4870 4871		/* Pad the header to 128 bytes.  */4872		__u8 pad[120];4873	} hdr;4874 4875	union {4876		struct kvm_vmx_nested_state_data vmx[0];4877		struct kvm_svm_nested_state_data svm[0];4878	} data;4879  };4880 4881  #define KVM_STATE_NESTED_GUEST_MODE		0x000000014882  #define KVM_STATE_NESTED_RUN_PENDING		0x000000024883  #define KVM_STATE_NESTED_EVMCS		0x000000044884 4885  #define KVM_STATE_NESTED_FORMAT_VMX		04886  #define KVM_STATE_NESTED_FORMAT_SVM		14887 4888  #define KVM_STATE_NESTED_VMX_VMCS_SIZE	0x10004889 4890  #define KVM_STATE_NESTED_VMX_SMM_GUEST_MODE	0x000000014891  #define KVM_STATE_NESTED_VMX_SMM_VMXON	0x000000024892 4893  #define KVM_STATE_VMX_PREEMPTION_TIMER_DEADLINE 0x000000014894 4895  struct kvm_vmx_nested_state_hdr {4896	__u64 vmxon_pa;4897	__u64 vmcs12_pa;4898 4899	struct {4900		__u16 flags;4901	} smm;4902 4903	__u32 flags;4904	__u64 preemption_timer_deadline;4905  };4906 4907  struct kvm_vmx_nested_state_data {4908	__u8 vmcs12[KVM_STATE_NESTED_VMX_VMCS_SIZE];4909	__u8 shadow_vmcs12[KVM_STATE_NESTED_VMX_VMCS_SIZE];4910  };4911 4912This ioctl copies the vcpu's nested virtualization state from the kernel to4913userspace.4914 4915The maximum size of the state can be retrieved by passing KVM_CAP_NESTED_STATE4916to the KVM_CHECK_EXTENSION ioctl().4917 49184.115 KVM_SET_NESTED_STATE4919--------------------------4920 4921:Capability: KVM_CAP_NESTED_STATE4922:Architectures: x864923:Type: vcpu ioctl4924:Parameters: struct kvm_nested_state (in)4925:Returns: 0 on success, -1 on error4926 4927This copies the vcpu's kvm_nested_state struct from userspace to the kernel.4928For the definition of struct kvm_nested_state, see KVM_GET_NESTED_STATE.4929 49304.116 KVM_(UN)REGISTER_COALESCED_MMIO4931-------------------------------------4932 4933:Capability: KVM_CAP_COALESCED_MMIO (for coalesced mmio)4934	     KVM_CAP_COALESCED_PIO (for coalesced pio)4935:Architectures: all4936:Type: vm ioctl4937:Parameters: struct kvm_coalesced_mmio_zone4938:Returns: 0 on success, < 0 on error4939 4940Coalesced I/O is a performance optimization that defers hardware4941register write emulation so that userspace exits are avoided.  It is4942typically used to reduce the overhead of emulating frequently accessed4943hardware registers.4944 4945When a hardware register is configured for coalesced I/O, write accesses4946do not exit to userspace and their value is recorded in a ring buffer4947that is shared between kernel and userspace.4948 4949Coalesced I/O is used if one or more write accesses to a hardware4950register can be deferred until a read or a write to another hardware4951register on the same device.  This last access will cause a vmexit and4952userspace will process accesses from the ring buffer before emulating4953it. That will avoid exiting to userspace on repeated writes.4954 4955Coalesced pio is based on coalesced mmio. There is little difference4956between coalesced mmio and pio except that coalesced pio records accesses4957to I/O ports.4958 49594.117 KVM_CLEAR_DIRTY_LOG (vm ioctl)4960------------------------------------4961 4962:Capability: KVM_CAP_MANUAL_DIRTY_LOG_PROTECT24963:Architectures: x86, arm64, mips4964:Type: vm ioctl4965:Parameters: struct kvm_clear_dirty_log (in)4966:Returns: 0 on success, -1 on error4967 4968::4969 4970  /* for KVM_CLEAR_DIRTY_LOG */4971  struct kvm_clear_dirty_log {4972	__u32 slot;4973	__u32 num_pages;4974	__u64 first_page;4975	union {4976		void __user *dirty_bitmap; /* one bit per page */4977		__u64 padding;4978	};4979  };4980 4981The ioctl clears the dirty status of pages in a memory slot, according to4982the bitmap that is passed in struct kvm_clear_dirty_log's dirty_bitmap4983field.  Bit 0 of the bitmap corresponds to page "first_page" in the4984memory slot, and num_pages is the size in bits of the input bitmap.4985first_page must be a multiple of 64; num_pages must also be a multiple of498664 unless first_page + num_pages is the size of the memory slot.  For each4987bit that is set in the input bitmap, the corresponding page is marked "clean"4988in KVM's dirty bitmap, and dirty tracking is re-enabled for that page4989(for example via write-protection, or by clearing the dirty bit in4990a page table entry).4991 4992If KVM_CAP_MULTI_ADDRESS_SPACE is available, bits 16-31 of slot field specifies4993the address space for which you want to clear the dirty status.  See4994KVM_SET_USER_MEMORY_REGION for details on the usage of slot field.4995 4996This ioctl is mostly useful when KVM_CAP_MANUAL_DIRTY_LOG_PROTECT24997is enabled; for more information, see the description of the capability.4998However, it can always be used as long as KVM_CHECK_EXTENSION confirms4999that KVM_CAP_MANUAL_DIRTY_LOG_PROTECT2 is present.5000 50014.118 KVM_GET_SUPPORTED_HV_CPUID5002--------------------------------5003 5004:Capability: KVM_CAP_HYPERV_CPUID (vcpu), KVM_CAP_SYS_HYPERV_CPUID (system)5005:Architectures: x865006:Type: system ioctl, vcpu ioctl5007:Parameters: struct kvm_cpuid2 (in/out)5008:Returns: 0 on success, -1 on error5009 5010::5011 5012  struct kvm_cpuid2 {5013	__u32 nent;5014	__u32 padding;5015	struct kvm_cpuid_entry2 entries[0];5016  };5017 5018  struct kvm_cpuid_entry2 {5019	__u32 function;5020	__u32 index;5021	__u32 flags;5022	__u32 eax;5023	__u32 ebx;5024	__u32 ecx;5025	__u32 edx;5026	__u32 padding[3];5027  };5028 5029This ioctl returns x86 cpuid features leaves related to Hyper-V emulation in5030KVM.  Userspace can use the information returned by this ioctl to construct5031cpuid information presented to guests consuming Hyper-V enlightenments (e.g.5032Windows or Hyper-V guests).5033 5034CPUID feature leaves returned by this ioctl are defined by Hyper-V Top Level5035Functional Specification (TLFS). These leaves can't be obtained with5036KVM_GET_SUPPORTED_CPUID ioctl because some of them intersect with KVM feature5037leaves (0x40000000, 0x40000001).5038 5039Currently, the following list of CPUID leaves are returned:5040 5041 - HYPERV_CPUID_VENDOR_AND_MAX_FUNCTIONS5042 - HYPERV_CPUID_INTERFACE5043 - HYPERV_CPUID_VERSION5044 - HYPERV_CPUID_FEATURES5045 - HYPERV_CPUID_ENLIGHTMENT_INFO5046 - HYPERV_CPUID_IMPLEMENT_LIMITS5047 - HYPERV_CPUID_NESTED_FEATURES5048 - HYPERV_CPUID_SYNDBG_VENDOR_AND_MAX_FUNCTIONS5049 - HYPERV_CPUID_SYNDBG_INTERFACE5050 - HYPERV_CPUID_SYNDBG_PLATFORM_CAPABILITIES5051 5052Userspace invokes KVM_GET_SUPPORTED_HV_CPUID by passing a kvm_cpuid2 structure5053with the 'nent' field indicating the number of entries in the variable-size5054array 'entries'.  If the number of entries is too low to describe all Hyper-V5055feature leaves, an error (E2BIG) is returned. If the number is more or equal5056to the number of Hyper-V feature leaves, the 'nent' field is adjusted to the5057number of valid entries in the 'entries' array, which is then filled.5058 5059'index' and 'flags' fields in 'struct kvm_cpuid_entry2' are currently reserved,5060userspace should not expect to get any particular value there.5061 5062Note, vcpu version of KVM_GET_SUPPORTED_HV_CPUID is currently deprecated. Unlike5063system ioctl which exposes all supported feature bits unconditionally, vcpu5064version has the following quirks:5065 5066- HYPERV_CPUID_NESTED_FEATURES leaf and HV_X64_ENLIGHTENED_VMCS_RECOMMENDED5067  feature bit are only exposed when Enlightened VMCS was previously enabled5068  on the corresponding vCPU (KVM_CAP_HYPERV_ENLIGHTENED_VMCS).5069- HV_STIMER_DIRECT_MODE_AVAILABLE bit is only exposed with in-kernel LAPIC.5070  (presumes KVM_CREATE_IRQCHIP has already been called).5071 50724.119 KVM_ARM_VCPU_FINALIZE5073---------------------------5074 5075:Architectures: arm645076:Type: vcpu ioctl5077:Parameters: int feature (in)5078:Returns: 0 on success, -1 on error5079 5080Errors:5081 5082  ======     ==============================================================5083  EPERM      feature not enabled, needs configuration, or already finalized5084  EINVAL     feature unknown or not present5085  ======     ==============================================================5086 5087Recognised values for feature:5088 5089  =====      ===========================================5090  arm64      KVM_ARM_VCPU_SVE (requires KVM_CAP_ARM_SVE)5091  =====      ===========================================5092 5093Finalizes the configuration of the specified vcpu feature.5094 5095The vcpu must already have been initialised, enabling the affected feature, by5096means of a successful KVM_ARM_VCPU_INIT call with the appropriate flag set in5097features[].5098 5099For affected vcpu features, this is a mandatory step that must be performed5100before the vcpu is fully usable.5101 5102Between KVM_ARM_VCPU_INIT and KVM_ARM_VCPU_FINALIZE, the feature may be5103configured by use of ioctls such as KVM_SET_ONE_REG.  The exact configuration5104that should be performed and how to do it are feature-dependent.5105 5106Other calls that depend on a particular feature being finalized, such as5107KVM_RUN, KVM_GET_REG_LIST, KVM_GET_ONE_REG and KVM_SET_ONE_REG, will fail with5108-EPERM unless the feature has already been finalized by means of a5109KVM_ARM_VCPU_FINALIZE call.5110 5111See KVM_ARM_VCPU_INIT for details of vcpu features that require finalization5112using this ioctl.5113 51144.120 KVM_SET_PMU_EVENT_FILTER5115------------------------------5116 5117:Capability: KVM_CAP_PMU_EVENT_FILTER5118:Architectures: x865119:Type: vm ioctl5120:Parameters: struct kvm_pmu_event_filter (in)5121:Returns: 0 on success, -1 on error5122 5123Errors:5124 5125  ======     ============================================================5126  EFAULT     args[0] cannot be accessed5127  EINVAL     args[0] contains invalid data in the filter or filter events5128  E2BIG      nevents is too large5129  EBUSY      not enough memory to allocate the filter5130  ======     ============================================================5131 5132::5133 5134  struct kvm_pmu_event_filter {5135	__u32 action;5136	__u32 nevents;5137	__u32 fixed_counter_bitmap;5138	__u32 flags;5139	__u32 pad[4];5140	__u64 events[0];5141  };5142 5143This ioctl restricts the set of PMU events the guest can program by limiting5144which event select and unit mask combinations are permitted.5145 5146The argument holds a list of filter events which will be allowed or denied.5147 5148Filter events only control general purpose counters; fixed purpose counters5149are controlled by the fixed_counter_bitmap.5150 5151Valid values for 'flags'::5152 5153``0``5154 5155To use this mode, clear the 'flags' field.5156 5157In this mode each event will contain an event select + unit mask.5158 5159When the guest attempts to program the PMU the guest's event select +5160unit mask is compared against the filter events to determine whether the5161guest should have access.5162 5163``KVM_PMU_EVENT_FLAG_MASKED_EVENTS``5164:Capability: KVM_CAP_PMU_EVENT_MASKED_EVENTS5165 5166In this mode each filter event will contain an event select, mask, match, and5167exclude value.  To encode a masked event use::5168 5169  KVM_PMU_ENCODE_MASKED_ENTRY()5170 5171An encoded event will follow this layout::5172 5173  Bits   Description5174  ----   -----------5175  7:0    event select (low bits)5176  15:8   umask match5177  31:16  unused5178  35:32  event select (high bits)5179  36:54  unused5180  55     exclude bit5181  63:56  umask mask5182 5183When the guest attempts to program the PMU, these steps are followed in5184determining if the guest should have access:5185 5186 1. Match the event select from the guest against the filter events.5187 2. If a match is found, match the guest's unit mask to the mask and match5188    values of the included filter events.5189    I.e. (unit mask & mask) == match && !exclude.5190 3. If a match is found, match the guest's unit mask to the mask and match5191    values of the excluded filter events.5192    I.e. (unit mask & mask) == match && exclude.5193 4.5194   a. If an included match is found and an excluded match is not found, filter5195      the event.5196   b. For everything else, do not filter the event.5197 5.5198   a. If the event is filtered and it's an allow list, allow the guest to5199      program the event.5200   b. If the event is filtered and it's a deny list, do not allow the guest to5201      program the event.5202 5203When setting a new pmu event filter, -EINVAL will be returned if any of the5204unused fields are set or if any of the high bits (35:32) in the event5205select are set when called on Intel.5206 5207Valid values for 'action'::5208 5209  #define KVM_PMU_EVENT_ALLOW 05210  #define KVM_PMU_EVENT_DENY 15211 5212Via this API, KVM userspace can also control the behavior of the VM's fixed5213counters (if any) by configuring the "action" and "fixed_counter_bitmap" fields.5214 5215Specifically, KVM follows the following pseudo-code when determining whether to5216allow the guest FixCtr[i] to count its pre-defined fixed event::5217 5218  FixCtr[i]_is_allowed = (action == ALLOW) && (bitmap & BIT(i)) ||5219    (action == DENY) && !(bitmap & BIT(i));5220  FixCtr[i]_is_denied = !FixCtr[i]_is_allowed;5221 5222KVM always consumes fixed_counter_bitmap, it's userspace's responsibility to5223ensure fixed_counter_bitmap is set correctly, e.g. if userspace wants to define5224a filter that only affects general purpose counters.5225 5226Note, the "events" field also applies to fixed counters' hardcoded event_select5227and unit_mask values.  "fixed_counter_bitmap" has higher priority than "events"5228if there is a contradiction between the two.5229 52304.121 KVM_PPC_SVM_OFF5231---------------------5232 5233:Capability: basic5234:Architectures: powerpc5235:Type: vm ioctl5236:Parameters: none5237:Returns: 0 on successful completion,5238 5239Errors:5240 5241  ======     ================================================================5242  EINVAL     if ultravisor failed to terminate the secure guest5243  ENOMEM     if hypervisor failed to allocate new radix page tables for guest5244  ======     ================================================================5245 5246This ioctl is used to turn off the secure mode of the guest or transition5247the guest from secure mode to normal mode. This is invoked when the guest5248is reset. This has no effect if called for a normal guest.5249 5250This ioctl issues an ultravisor call to terminate the secure guest,5251unpins the VPA pages and releases all the device pages that are used to5252track the secure pages by hypervisor.5253 52544.122 KVM_S390_NORMAL_RESET5255---------------------------5256 5257:Capability: KVM_CAP_S390_VCPU_RESETS5258:Architectures: s3905259:Type: vcpu ioctl5260:Parameters: none5261:Returns: 05262 5263This ioctl resets VCPU registers and control structures according to5264the cpu reset definition in the POP (Principles Of Operation).5265 52664.123 KVM_S390_INITIAL_RESET5267----------------------------5268 5269:Capability: none5270:Architectures: s3905271:Type: vcpu ioctl5272:Parameters: none5273:Returns: 05274 5275This ioctl resets VCPU registers and control structures according to5276the initial cpu reset definition in the POP. However, the cpu is not5277put into ESA mode. This reset is a superset of the normal reset.5278 52794.124 KVM_S390_CLEAR_RESET5280--------------------------5281 5282:Capability: KVM_CAP_S390_VCPU_RESETS5283:Architectures: s3905284:Type: vcpu ioctl5285:Parameters: none5286:Returns: 05287 5288This ioctl resets VCPU registers and control structures according to5289the clear cpu reset definition in the POP. However, the cpu is not put5290into ESA mode. This reset is a superset of the initial reset.5291 5292 52934.125 KVM_S390_PV_COMMAND5294-------------------------5295 5296:Capability: KVM_CAP_S390_PROTECTED5297:Architectures: s3905298:Type: vm ioctl5299:Parameters: struct kvm_pv_cmd5300:Returns: 0 on success, < 0 on error5301 5302::5303 5304  struct kvm_pv_cmd {5305	__u32 cmd;	/* Command to be executed */5306	__u16 rc;	/* Ultravisor return code */5307	__u16 rrc;	/* Ultravisor return reason code */5308	__u64 data;	/* Data or address */5309	__u32 flags;    /* flags for future extensions. Must be 0 for now */5310	__u32 reserved[3];5311  };5312 5313**Ultravisor return codes**5314The Ultravisor return (reason) codes are provided by the kernel if a5315Ultravisor call has been executed to achieve the results expected by5316the command. Therefore they are independent of the IOCTL return5317code. If KVM changes `rc`, its value will always be greater than 05318hence setting it to 0 before issuing a PV command is advised to be5319able to detect a change of `rc`.5320 5321**cmd values:**5322 5323KVM_PV_ENABLE5324  Allocate memory and register the VM with the Ultravisor, thereby5325  donating memory to the Ultravisor that will become inaccessible to5326  KVM. All existing CPUs are converted to protected ones. After this5327  command has succeeded, any CPU added via hotplug will become5328  protected during its creation as well.5329 5330  Errors:5331 5332  =====      =============================5333  EINTR      an unmasked signal is pending5334  =====      =============================5335 5336KVM_PV_DISABLE5337  Deregister the VM from the Ultravisor and reclaim the memory that had5338  been donated to the Ultravisor, making it usable by the kernel again.5339  All registered VCPUs are converted back to non-protected ones. If a5340  previous protected VM had been prepared for asynchronous teardown with5341  KVM_PV_ASYNC_CLEANUP_PREPARE and not subsequently torn down with5342  KVM_PV_ASYNC_CLEANUP_PERFORM, it will be torn down in this call5343  together with the current protected VM.5344 5345KVM_PV_VM_SET_SEC_PARMS5346  Pass the image header from VM memory to the Ultravisor in5347  preparation of image unpacking and verification.5348 5349KVM_PV_VM_UNPACK5350  Unpack (protect and decrypt) a page of the encrypted boot image.5351 5352KVM_PV_VM_VERIFY5353  Verify the integrity of the unpacked image. Only if this succeeds,5354  KVM is allowed to start protected VCPUs.5355 5356KVM_PV_INFO5357  :Capability: KVM_CAP_S390_PROTECTED_DUMP5358 5359  Presents an API that provides Ultravisor related data to userspace5360  via subcommands. len_max is the size of the user space buffer,5361  len_written is KVM's indication of how much bytes of that buffer5362  were actually written to. len_written can be used to determine the5363  valid fields if more response fields are added in the future.5364 5365  ::5366 5367     enum pv_cmd_info_id {5368	KVM_PV_INFO_VM,5369	KVM_PV_INFO_DUMP,5370     };5371 5372     struct kvm_s390_pv_info_header {5373	__u32 id;5374	__u32 len_max;5375	__u32 len_written;5376	__u32 reserved;5377     };5378 5379     struct kvm_s390_pv_info {5380	struct kvm_s390_pv_info_header header;5381	struct kvm_s390_pv_info_dump dump;5382	struct kvm_s390_pv_info_vm vm;5383     };5384 5385**subcommands:**5386 5387  KVM_PV_INFO_VM5388    This subcommand provides basic Ultravisor information for PV5389    hosts. These values are likely also exported as files in the sysfs5390    firmware UV query interface but they are more easily available to5391    programs in this API.5392 5393    The installed calls and feature_indication members provide the5394    installed UV calls and the UV's other feature indications.5395 5396    The max_* members provide information about the maximum number of PV5397    vcpus, PV guests and PV guest memory size.5398 5399    ::5400 5401      struct kvm_s390_pv_info_vm {5402	__u64 inst_calls_list[4];5403	__u64 max_cpus;5404	__u64 max_guests;5405	__u64 max_guest_addr;5406	__u64 feature_indication;5407      };5408 5409 5410  KVM_PV_INFO_DUMP5411    This subcommand provides information related to dumping PV guests.5412 5413    ::5414 5415      struct kvm_s390_pv_info_dump {5416	__u64 dump_cpu_buffer_len;5417	__u64 dump_config_mem_buffer_per_1m;5418	__u64 dump_config_finalize_len;5419      };5420 5421KVM_PV_DUMP5422  :Capability: KVM_CAP_S390_PROTECTED_DUMP5423 5424  Presents an API that provides calls which facilitate dumping a5425  protected VM.5426 5427  ::5428 5429    struct kvm_s390_pv_dmp {5430      __u64 subcmd;5431      __u64 buff_addr;5432      __u64 buff_len;5433      __u64 gaddr;		/* For dump storage state */5434    };5435 5436  **subcommands:**5437 5438  KVM_PV_DUMP_INIT5439    Initializes the dump process of a protected VM. If this call does5440    not succeed all other subcommands will fail with -EINVAL. This5441    subcommand will return -EINVAL if a dump process has not yet been5442    completed.5443 5444    Not all PV vms can be dumped, the owner needs to set `dump5445    allowed` PCF bit 34 in the SE header to allow dumping.5446 5447  KVM_PV_DUMP_CONFIG_STOR_STATE5448     Stores `buff_len` bytes of tweak component values starting with5449     the 1MB block specified by the absolute guest address5450     (`gaddr`). `buff_len` needs to be `conf_dump_storage_state_len`5451     aligned and at least >= the `conf_dump_storage_state_len` value5452     provided by the dump uv_info data. buff_user might be written to5453     even if an error rc is returned. For instance if we encounter a5454     fault after writing the first page of data.5455 5456  KVM_PV_DUMP_COMPLETE5457    If the subcommand succeeds it completes the dump process and lets5458    KVM_PV_DUMP_INIT be called again.5459 5460    On success `conf_dump_finalize_len` bytes of completion data will be5461    stored to the `buff_addr`. The completion data contains a key5462    derivation seed, IV, tweak nonce and encryption keys as well as an5463    authentication tag all of which are needed to decrypt the dump at a5464    later time.5465 5466KVM_PV_ASYNC_CLEANUP_PREPARE5467  :Capability: KVM_CAP_S390_PROTECTED_ASYNC_DISABLE5468 5469  Prepare the current protected VM for asynchronous teardown. Most5470  resources used by the current protected VM will be set aside for a5471  subsequent asynchronous teardown. The current protected VM will then5472  resume execution immediately as non-protected. There can be at most5473  one protected VM prepared for asynchronous teardown at any time. If5474  a protected VM had already been prepared for teardown without5475  subsequently calling KVM_PV_ASYNC_CLEANUP_PERFORM, this call will5476  fail. In that case, the userspace process should issue a normal5477  KVM_PV_DISABLE. The resources set aside with this call will need to5478  be cleaned up with a subsequent call to KVM_PV_ASYNC_CLEANUP_PERFORM5479  or KVM_PV_DISABLE, otherwise they will be cleaned up when KVM5480  terminates. KVM_PV_ASYNC_CLEANUP_PREPARE can be called again as soon5481  as cleanup starts, i.e. before KVM_PV_ASYNC_CLEANUP_PERFORM finishes.5482 5483KVM_PV_ASYNC_CLEANUP_PERFORM5484  :Capability: KVM_CAP_S390_PROTECTED_ASYNC_DISABLE5485 5486  Tear down the protected VM previously prepared for teardown with5487  KVM_PV_ASYNC_CLEANUP_PREPARE. The resources that had been set aside5488  will be freed during the execution of this command. This PV command5489  should ideally be issued by userspace from a separate thread. If a5490  fatal signal is received (or the process terminates naturally), the5491  command will terminate immediately without completing, and the normal5492  KVM shutdown procedure will take care of cleaning up all remaining5493  protected VMs, including the ones whose teardown was interrupted by5494  process termination.5495 54964.126 KVM_XEN_HVM_SET_ATTR5497--------------------------5498 5499:Capability: KVM_CAP_XEN_HVM / KVM_XEN_HVM_CONFIG_SHARED_INFO5500:Architectures: x865501:Type: vm ioctl5502:Parameters: struct kvm_xen_hvm_attr5503:Returns: 0 on success, < 0 on error5504 5505::5506 5507  struct kvm_xen_hvm_attr {5508	__u16 type;5509	__u16 pad[3];5510	union {5511		__u8 long_mode;5512		__u8 vector;5513		__u8 runstate_update_flag;5514		union {5515			__u64 gfn;5516			__u64 hva;5517		} shared_info;5518		struct {5519			__u32 send_port;5520			__u32 type; /* EVTCHNSTAT_ipi / EVTCHNSTAT_interdomain */5521			__u32 flags;5522			union {5523				struct {5524					__u32 port;5525					__u32 vcpu;5526					__u32 priority;5527				} port;5528				struct {5529					__u32 port; /* Zero for eventfd */5530					__s32 fd;5531				} eventfd;5532				__u32 padding[4];5533			} deliver;5534		} evtchn;5535		__u32 xen_version;5536		__u64 pad[8];5537	} u;5538  };5539 5540type values:5541 5542KVM_XEN_ATTR_TYPE_LONG_MODE5543  Sets the ABI mode of the VM to 32-bit or 64-bit (long mode). This5544  determines the layout of the shared_info page exposed to the VM.5545 5546KVM_XEN_ATTR_TYPE_SHARED_INFO5547  Sets the guest physical frame number at which the Xen shared_info5548  page resides. Note that although Xen places vcpu_info for the first5549  32 vCPUs in the shared_info page, KVM does not automatically do so5550  and instead requires that KVM_XEN_VCPU_ATTR_TYPE_VCPU_INFO or5551  KVM_XEN_VCPU_ATTR_TYPE_VCPU_INFO_HVA be used explicitly even when5552  the vcpu_info for a given vCPU resides at the "default" location5553  in the shared_info page. This is because KVM may not be aware of5554  the Xen CPU id which is used as the index into the vcpu_info[]5555  array, so may know the correct default location.5556 5557  Note that the shared_info page may be constantly written to by KVM;5558  it contains the event channel bitmap used to deliver interrupts to5559  a Xen guest, amongst other things. It is exempt from dirty tracking5560  mechanisms — KVM will not explicitly mark the page as dirty each5561  time an event channel interrupt is delivered to the guest! Thus,5562  userspace should always assume that the designated GFN is dirty if5563  any vCPU has been running or any event channel interrupts can be5564  routed to the guest.5565 5566  Setting the gfn to KVM_XEN_INVALID_GFN will disable the shared_info5567  page.5568 5569KVM_XEN_ATTR_TYPE_SHARED_INFO_HVA5570  If the KVM_XEN_HVM_CONFIG_SHARED_INFO_HVA flag is also set in the5571  Xen capabilities, then this attribute may be used to set the5572  userspace address at which the shared_info page resides, which5573  will always be fixed in the VMM regardless of where it is mapped5574  in guest physical address space. This attribute should be used in5575  preference to KVM_XEN_ATTR_TYPE_SHARED_INFO as it avoids5576  unnecessary invalidation of an internal cache when the page is5577  re-mapped in guest physcial address space.5578 5579  Setting the hva to zero will disable the shared_info page.5580 5581KVM_XEN_ATTR_TYPE_UPCALL_VECTOR5582  Sets the exception vector used to deliver Xen event channel upcalls.5583  This is the HVM-wide vector injected directly by the hypervisor5584  (not through the local APIC), typically configured by a guest via5585  HVM_PARAM_CALLBACK_IRQ. This can be disabled again (e.g. for guest5586  SHUTDOWN_soft_reset) by setting it to zero.5587 5588KVM_XEN_ATTR_TYPE_EVTCHN5589  This attribute is available when the KVM_CAP_XEN_HVM ioctl indicates5590  support for KVM_XEN_HVM_CONFIG_EVTCHN_SEND features. It configures5591  an outbound port number for interception of EVTCHNOP_send requests5592  from the guest. A given sending port number may be directed back to5593  a specified vCPU (by APIC ID) / port / priority on the guest, or to5594  trigger events on an eventfd. The vCPU and priority can be changed5595  by setting KVM_XEN_EVTCHN_UPDATE in a subsequent call, but other5596  fields cannot change for a given sending port. A port mapping is5597  removed by using KVM_XEN_EVTCHN_DEASSIGN in the flags field. Passing5598  KVM_XEN_EVTCHN_RESET in the flags field removes all interception of5599  outbound event channels. The values of the flags field are mutually5600  exclusive and cannot be combined as a bitmask.5601 5602KVM_XEN_ATTR_TYPE_XEN_VERSION5603  This attribute is available when the KVM_CAP_XEN_HVM ioctl indicates5604  support for KVM_XEN_HVM_CONFIG_EVTCHN_SEND features. It configures5605  the 32-bit version code returned to the guest when it invokes the5606  XENVER_version call; typically (XEN_MAJOR << 16 | XEN_MINOR). PV5607  Xen guests will often use this to as a dummy hypercall to trigger5608  event channel delivery, so responding within the kernel without5609  exiting to userspace is beneficial.5610 5611KVM_XEN_ATTR_TYPE_RUNSTATE_UPDATE_FLAG5612  This attribute is available when the KVM_CAP_XEN_HVM ioctl indicates5613  support for KVM_XEN_HVM_CONFIG_RUNSTATE_UPDATE_FLAG. It enables the5614  XEN_RUNSTATE_UPDATE flag which allows guest vCPUs to safely read5615  other vCPUs' vcpu_runstate_info. Xen guests enable this feature via5616  the VMASST_TYPE_runstate_update_flag of the HYPERVISOR_vm_assist5617  hypercall.5618 56194.127 KVM_XEN_HVM_GET_ATTR5620--------------------------5621 5622:Capability: KVM_CAP_XEN_HVM / KVM_XEN_HVM_CONFIG_SHARED_INFO5623:Architectures: x865624:Type: vm ioctl5625:Parameters: struct kvm_xen_hvm_attr5626:Returns: 0 on success, < 0 on error5627 5628Allows Xen VM attributes to be read. For the structure and types,5629see KVM_XEN_HVM_SET_ATTR above. The KVM_XEN_ATTR_TYPE_EVTCHN5630attribute cannot be read.5631 56324.128 KVM_XEN_VCPU_SET_ATTR5633---------------------------5634 5635:Capability: KVM_CAP_XEN_HVM / KVM_XEN_HVM_CONFIG_SHARED_INFO5636:Architectures: x865637:Type: vcpu ioctl5638:Parameters: struct kvm_xen_vcpu_attr5639:Returns: 0 on success, < 0 on error5640 5641::5642 5643  struct kvm_xen_vcpu_attr {5644	__u16 type;5645	__u16 pad[3];5646	union {5647		__u64 gpa;5648		__u64 pad[4];5649		struct {5650			__u64 state;5651			__u64 state_entry_time;5652			__u64 time_running;5653			__u64 time_runnable;5654			__u64 time_blocked;5655			__u64 time_offline;5656		} runstate;5657		__u32 vcpu_id;5658		struct {5659			__u32 port;5660			__u32 priority;5661			__u64 expires_ns;5662		} timer;5663		__u8 vector;5664	} u;5665  };5666 5667type values:5668 5669KVM_XEN_VCPU_ATTR_TYPE_VCPU_INFO5670  Sets the guest physical address of the vcpu_info for a given vCPU.5671  As with the shared_info page for the VM, the corresponding page may be5672  dirtied at any time if event channel interrupt delivery is enabled, so5673  userspace should always assume that the page is dirty without relying5674  on dirty logging. Setting the gpa to KVM_XEN_INVALID_GPA will disable5675  the vcpu_info.5676 5677KVM_XEN_VCPU_ATTR_TYPE_VCPU_INFO_HVA5678  If the KVM_XEN_HVM_CONFIG_SHARED_INFO_HVA flag is also set in the5679  Xen capabilities, then this attribute may be used to set the5680  userspace address of the vcpu_info for a given vCPU. It should5681  only be used when the vcpu_info resides at the "default" location5682  in the shared_info page. In this case it is safe to assume the5683  userspace address will not change, because the shared_info page is5684  an overlay on guest memory and remains at a fixed host address5685  regardless of where it is mapped in guest physical address space5686  and hence unnecessary invalidation of an internal cache may be5687  avoided if the guest memory layout is modified.5688  If the vcpu_info does not reside at the "default" location then5689  it is not guaranteed to remain at the same host address and5690  hence the aforementioned cache invalidation is required.5691 5692KVM_XEN_VCPU_ATTR_TYPE_VCPU_TIME_INFO5693  Sets the guest physical address of an additional pvclock structure5694  for a given vCPU. This is typically used for guest vsyscall support.5695  Setting the gpa to KVM_XEN_INVALID_GPA will disable the structure.5696 5697KVM_XEN_VCPU_ATTR_TYPE_RUNSTATE_ADDR5698  Sets the guest physical address of the vcpu_runstate_info for a given5699  vCPU. This is how a Xen guest tracks CPU state such as steal time.5700  Setting the gpa to KVM_XEN_INVALID_GPA will disable the runstate area.5701 5702KVM_XEN_VCPU_ATTR_TYPE_RUNSTATE_CURRENT5703  Sets the runstate (RUNSTATE_running/_runnable/_blocked/_offline) of5704  the given vCPU from the .u.runstate.state member of the structure.5705  KVM automatically accounts running and runnable time but blocked5706  and offline states are only entered explicitly.5707 5708KVM_XEN_VCPU_ATTR_TYPE_RUNSTATE_DATA5709  Sets all fields of the vCPU runstate data from the .u.runstate member5710  of the structure, including the current runstate. The state_entry_time5711  must equal the sum of the other four times.5712 5713KVM_XEN_VCPU_ATTR_TYPE_RUNSTATE_ADJUST5714  This *adds* the contents of the .u.runstate members of the structure5715  to the corresponding members of the given vCPU's runstate data, thus5716  permitting atomic adjustments to the runstate times. The adjustment5717  to the state_entry_time must equal the sum of the adjustments to the5718  other four times. The state field must be set to -1, or to a valid5719  runstate value (RUNSTATE_running, RUNSTATE_runnable, RUNSTATE_blocked5720  or RUNSTATE_offline) to set the current accounted state as of the5721  adjusted state_entry_time.5722 5723KVM_XEN_VCPU_ATTR_TYPE_VCPU_ID5724  This attribute is available when the KVM_CAP_XEN_HVM ioctl indicates5725  support for KVM_XEN_HVM_CONFIG_EVTCHN_SEND features. It sets the Xen5726  vCPU ID of the given vCPU, to allow timer-related VCPU operations to5727  be intercepted by KVM.5728 5729KVM_XEN_VCPU_ATTR_TYPE_TIMER5730  This attribute is available when the KVM_CAP_XEN_HVM ioctl indicates5731  support for KVM_XEN_HVM_CONFIG_EVTCHN_SEND features. It sets the5732  event channel port/priority for the VIRQ_TIMER of the vCPU, as well5733  as allowing a pending timer to be saved/restored. Setting the timer5734  port to zero disables kernel handling of the singleshot timer.5735 5736KVM_XEN_VCPU_ATTR_TYPE_UPCALL_VECTOR5737  This attribute is available when the KVM_CAP_XEN_HVM ioctl indicates5738  support for KVM_XEN_HVM_CONFIG_EVTCHN_SEND features. It sets the5739  per-vCPU local APIC upcall vector, configured by a Xen guest with5740  the HVMOP_set_evtchn_upcall_vector hypercall. This is typically5741  used by Windows guests, and is distinct from the HVM-wide upcall5742  vector configured with HVM_PARAM_CALLBACK_IRQ. It is disabled by5743  setting the vector to zero.5744 5745 57464.129 KVM_XEN_VCPU_GET_ATTR5747---------------------------5748 5749:Capability: KVM_CAP_XEN_HVM / KVM_XEN_HVM_CONFIG_SHARED_INFO5750:Architectures: x865751:Type: vcpu ioctl5752:Parameters: struct kvm_xen_vcpu_attr5753:Returns: 0 on success, < 0 on error5754 5755Allows Xen vCPU attributes to be read. For the structure and types,5756see KVM_XEN_VCPU_SET_ATTR above.5757 5758The KVM_XEN_VCPU_ATTR_TYPE_RUNSTATE_ADJUST type may not be used5759with the KVM_XEN_VCPU_GET_ATTR ioctl.5760 57614.130 KVM_ARM_MTE_COPY_TAGS5762---------------------------5763 5764:Capability: KVM_CAP_ARM_MTE5765:Architectures: arm645766:Type: vm ioctl5767:Parameters: struct kvm_arm_copy_mte_tags5768:Returns: number of bytes copied, < 0 on error (-EINVAL for incorrect5769          arguments, -EFAULT if memory cannot be accessed).5770 5771::5772 5773  struct kvm_arm_copy_mte_tags {5774	__u64 guest_ipa;5775	__u64 length;5776	void __user *addr;5777	__u64 flags;5778	__u64 reserved[2];5779  };5780 5781Copies Memory Tagging Extension (MTE) tags to/from guest tag memory. The5782``guest_ipa`` and ``length`` fields must be ``PAGE_SIZE`` aligned.5783``length`` must not be bigger than 2^31 - PAGE_SIZE bytes. The ``addr``5784field must point to a buffer which the tags will be copied to or from.5785 5786``flags`` specifies the direction of copy, either ``KVM_ARM_TAGS_TO_GUEST`` or5787``KVM_ARM_TAGS_FROM_GUEST``.5788 5789The size of the buffer to store the tags is ``(length / 16)`` bytes5790(granules in MTE are 16 bytes long). Each byte contains a single tag5791value. This matches the format of ``PTRACE_PEEKMTETAGS`` and5792``PTRACE_POKEMTETAGS``.5793 5794If an error occurs before any data is copied then a negative error code is5795returned. If some tags have been copied before an error occurs then the number5796of bytes successfully copied is returned. If the call completes successfully5797then ``length`` is returned.5798 57994.131 KVM_GET_SREGS25800--------------------5801 5802:Capability: KVM_CAP_SREGS25803:Architectures: x865804:Type: vcpu ioctl5805:Parameters: struct kvm_sregs2 (out)5806:Returns: 0 on success, -1 on error5807 5808Reads special registers from the vcpu.5809This ioctl (when supported) replaces the KVM_GET_SREGS.5810 5811::5812 5813        struct kvm_sregs2 {5814                /* out (KVM_GET_SREGS2) / in (KVM_SET_SREGS2) */5815                struct kvm_segment cs, ds, es, fs, gs, ss;5816                struct kvm_segment tr, ldt;5817                struct kvm_dtable gdt, idt;5818                __u64 cr0, cr2, cr3, cr4, cr8;5819                __u64 efer;5820                __u64 apic_base;5821                __u64 flags;5822                __u64 pdptrs[4];5823        };5824 5825flags values for ``kvm_sregs2``:5826 5827``KVM_SREGS2_FLAGS_PDPTRS_VALID``5828 5829  Indicates that the struct contains valid PDPTR values.5830 5831 58324.132 KVM_SET_SREGS25833--------------------5834 5835:Capability: KVM_CAP_SREGS25836:Architectures: x865837:Type: vcpu ioctl5838:Parameters: struct kvm_sregs2 (in)5839:Returns: 0 on success, -1 on error5840 5841Writes special registers into the vcpu.5842See KVM_GET_SREGS2 for the data structures.5843This ioctl (when supported) replaces the KVM_SET_SREGS.5844 58454.133 KVM_GET_STATS_FD5846----------------------5847 5848:Capability: KVM_CAP_STATS_BINARY_FD5849:Architectures: all5850:Type: vm ioctl, vcpu ioctl5851:Parameters: none5852:Returns: statistics file descriptor on success, < 0 on error5853 5854Errors:5855 5856  ======     ======================================================5857  ENOMEM     if the fd could not be created due to lack of memory5858  EMFILE     if the number of opened files exceeds the limit5859  ======     ======================================================5860 5861The returned file descriptor can be used to read VM/vCPU statistics data in5862binary format. The data in the file descriptor consists of four blocks5863organized as follows:5864 5865+-------------+5866|   Header    |5867+-------------+5868|  id string  |5869+-------------+5870| Descriptors |5871+-------------+5872| Stats Data  |5873+-------------+5874 5875Apart from the header starting at offset 0, please be aware that it is5876not guaranteed that the four blocks are adjacent or in the above order;5877the offsets of the id, descriptors and data blocks are found in the5878header.  However, all four blocks are aligned to 64 bit offsets in the5879file and they do not overlap.5880 5881All blocks except the data block are immutable.  Userspace can read them5882only one time after retrieving the file descriptor, and then use ``pread`` or5883``lseek`` to read the statistics repeatedly.5884 5885All data is in system endianness.5886 5887The format of the header is as follows::5888 5889	struct kvm_stats_header {5890		__u32 flags;5891		__u32 name_size;5892		__u32 num_desc;5893		__u32 id_offset;5894		__u32 desc_offset;5895		__u32 data_offset;5896	};5897 5898The ``flags`` field is not used at the moment. It is always read as 0.5899 5900The ``name_size`` field is the size (in byte) of the statistics name string5901(including trailing '\0') which is contained in the "id string" block and5902appended at the end of every descriptor.5903 5904The ``num_desc`` field is the number of descriptors that are included in the5905descriptor block.  (The actual number of values in the data block may be5906larger, since each descriptor may comprise more than one value).5907 5908The ``id_offset`` field is the offset of the id string from the start of the5909file indicated by the file descriptor. It is a multiple of 8.5910 5911The ``desc_offset`` field is the offset of the Descriptors block from the start5912of the file indicated by the file descriptor. It is a multiple of 8.5913 5914The ``data_offset`` field is the offset of the Stats Data block from the start5915of the file indicated by the file descriptor. It is a multiple of 8.5916 5917The id string block contains a string which identifies the file descriptor on5918which KVM_GET_STATS_FD was invoked.  The size of the block, including the5919trailing ``'\0'``, is indicated by the ``name_size`` field in the header.5920 5921The descriptors block is only needed to be read once for the lifetime of the5922file descriptor contains a sequence of ``struct kvm_stats_desc``, each followed5923by a string of size ``name_size``.5924::5925 5926	#define KVM_STATS_TYPE_SHIFT		05927	#define KVM_STATS_TYPE_MASK		(0xF << KVM_STATS_TYPE_SHIFT)5928	#define KVM_STATS_TYPE_CUMULATIVE	(0x0 << KVM_STATS_TYPE_SHIFT)5929	#define KVM_STATS_TYPE_INSTANT		(0x1 << KVM_STATS_TYPE_SHIFT)5930	#define KVM_STATS_TYPE_PEAK		(0x2 << KVM_STATS_TYPE_SHIFT)5931	#define KVM_STATS_TYPE_LINEAR_HIST	(0x3 << KVM_STATS_TYPE_SHIFT)5932	#define KVM_STATS_TYPE_LOG_HIST		(0x4 << KVM_STATS_TYPE_SHIFT)5933	#define KVM_STATS_TYPE_MAX		KVM_STATS_TYPE_LOG_HIST5934 5935	#define KVM_STATS_UNIT_SHIFT		45936	#define KVM_STATS_UNIT_MASK		(0xF << KVM_STATS_UNIT_SHIFT)5937	#define KVM_STATS_UNIT_NONE		(0x0 << KVM_STATS_UNIT_SHIFT)5938	#define KVM_STATS_UNIT_BYTES		(0x1 << KVM_STATS_UNIT_SHIFT)5939	#define KVM_STATS_UNIT_SECONDS		(0x2 << KVM_STATS_UNIT_SHIFT)5940	#define KVM_STATS_UNIT_CYCLES		(0x3 << KVM_STATS_UNIT_SHIFT)5941	#define KVM_STATS_UNIT_BOOLEAN		(0x4 << KVM_STATS_UNIT_SHIFT)5942	#define KVM_STATS_UNIT_MAX		KVM_STATS_UNIT_BOOLEAN5943 5944	#define KVM_STATS_BASE_SHIFT		85945	#define KVM_STATS_BASE_MASK		(0xF << KVM_STATS_BASE_SHIFT)5946	#define KVM_STATS_BASE_POW10		(0x0 << KVM_STATS_BASE_SHIFT)5947	#define KVM_STATS_BASE_POW2		(0x1 << KVM_STATS_BASE_SHIFT)5948	#define KVM_STATS_BASE_MAX		KVM_STATS_BASE_POW25949 5950	struct kvm_stats_desc {5951		__u32 flags;5952		__s16 exponent;5953		__u16 size;5954		__u32 offset;5955		__u32 bucket_size;5956		char name[];5957	};5958 5959The ``flags`` field contains the type and unit of the statistics data described5960by this descriptor. Its endianness is CPU native.5961The following flags are supported:5962 5963Bits 0-3 of ``flags`` encode the type:5964 5965  * ``KVM_STATS_TYPE_CUMULATIVE``5966    The statistics reports a cumulative count. The value of data can only be increased.5967    Most of the counters used in KVM are of this type.5968    The corresponding ``size`` field for this type is always 1.5969    All cumulative statistics data are read/write.5970  * ``KVM_STATS_TYPE_INSTANT``5971    The statistics reports an instantaneous value. Its value can be increased or5972    decreased. This type is usually used as a measurement of some resources,5973    like the number of dirty pages, the number of large pages, etc.5974    All instant statistics are read only.5975    The corresponding ``size`` field for this type is always 1.5976  * ``KVM_STATS_TYPE_PEAK``5977    The statistics data reports a peak value, for example the maximum number5978    of items in a hash table bucket, the longest time waited and so on.5979    The value of data can only be increased.5980    The corresponding ``size`` field for this type is always 1.5981  * ``KVM_STATS_TYPE_LINEAR_HIST``5982    The statistic is reported as a linear histogram. The number of5983    buckets is specified by the ``size`` field. The size of buckets is specified5984    by the ``hist_param`` field. The range of the Nth bucket (1 <= N < ``size``)5985    is [``hist_param``*(N-1), ``hist_param``*N), while the range of the last5986    bucket is [``hist_param``*(``size``-1), +INF). (+INF means positive infinity5987    value.)5988  * ``KVM_STATS_TYPE_LOG_HIST``5989    The statistic is reported as a logarithmic histogram. The number of5990    buckets is specified by the ``size`` field. The range of the first bucket is5991    [0, 1), while the range of the last bucket is [pow(2, ``size``-2), +INF).5992    Otherwise, The Nth bucket (1 < N < ``size``) covers5993    [pow(2, N-2), pow(2, N-1)).5994 5995Bits 4-7 of ``flags`` encode the unit:5996 5997  * ``KVM_STATS_UNIT_NONE``5998    There is no unit for the value of statistics data. This usually means that5999    the value is a simple counter of an event.6000  * ``KVM_STATS_UNIT_BYTES``6001    It indicates that the statistics data is used to measure memory size, in the6002    unit of Byte, KiByte, MiByte, GiByte, etc. The unit of the data is6003    determined by the ``exponent`` field in the descriptor.6004  * ``KVM_STATS_UNIT_SECONDS``6005    It indicates that the statistics data is used to measure time or latency.6006  * ``KVM_STATS_UNIT_CYCLES``6007    It indicates that the statistics data is used to measure CPU clock cycles.6008  * ``KVM_STATS_UNIT_BOOLEAN``6009    It indicates that the statistic will always be either 0 or 1.  Boolean6010    statistics of "peak" type will never go back from 1 to 0.  Boolean6011    statistics can be linear histograms (with two buckets) but not logarithmic6012    histograms.6013 6014Note that, in the case of histograms, the unit applies to the bucket6015ranges, while the bucket value indicates how many samples fell in the6016bucket's range.6017 6018Bits 8-11 of ``flags``, together with ``exponent``, encode the scale of the6019unit:6020 6021  * ``KVM_STATS_BASE_POW10``6022    The scale is based on power of 10. It is used for measurement of time and6023    CPU clock cycles.  For example, an exponent of -9 can be used with6024    ``KVM_STATS_UNIT_SECONDS`` to express that the unit is nanoseconds.6025  * ``KVM_STATS_BASE_POW2``6026    The scale is based on power of 2. It is used for measurement of memory size.6027    For example, an exponent of 20 can be used with ``KVM_STATS_UNIT_BYTES`` to6028    express that the unit is MiB.6029 6030The ``size`` field is the number of values of this statistics data. Its6031value is usually 1 for most of simple statistics. 1 means it contains an6032unsigned 64bit data.6033 6034The ``offset`` field is the offset from the start of Data Block to the start of6035the corresponding statistics data.6036 6037The ``bucket_size`` field is used as a parameter for histogram statistics data.6038It is only used by linear histogram statistics data, specifying the size of a6039bucket in the unit expressed by bits 4-11 of ``flags`` together with ``exponent``.6040 6041The ``name`` field is the name string of the statistics data. The name string6042starts at the end of ``struct kvm_stats_desc``.  The maximum length including6043the trailing ``'\0'``, is indicated by ``name_size`` in the header.6044 6045The Stats Data block contains an array of 64-bit values in the same order6046as the descriptors in Descriptors block.6047 60484.134 KVM_GET_XSAVE26049--------------------6050 6051:Capability: KVM_CAP_XSAVE26052:Architectures: x866053:Type: vcpu ioctl6054:Parameters: struct kvm_xsave (out)6055:Returns: 0 on success, -1 on error6056 6057 6058::6059 6060  struct kvm_xsave {6061	__u32 region[1024];6062	__u32 extra[0];6063  };6064 6065This ioctl would copy current vcpu's xsave struct to the userspace. It6066copies as many bytes as are returned by KVM_CHECK_EXTENSION(KVM_CAP_XSAVE2)6067when invoked on the vm file descriptor. The size value returned by6068KVM_CHECK_EXTENSION(KVM_CAP_XSAVE2) will always be at least 4096.6069Currently, it is only greater than 4096 if a dynamic feature has been6070enabled with ``arch_prctl()``, but this may change in the future.6071 6072The offsets of the state save areas in struct kvm_xsave follow the contents6073of CPUID leaf 0xD on the host.6074 60754.135 KVM_XEN_HVM_EVTCHN_SEND6076-----------------------------6077 6078:Capability: KVM_CAP_XEN_HVM / KVM_XEN_HVM_CONFIG_EVTCHN_SEND6079:Architectures: x866080:Type: vm ioctl6081:Parameters: struct kvm_irq_routing_xen_evtchn6082:Returns: 0 on success, < 0 on error6083 6084 6085::6086 6087   struct kvm_irq_routing_xen_evtchn {6088	__u32 port;6089	__u32 vcpu;6090	__u32 priority;6091   };6092 6093This ioctl injects an event channel interrupt directly to the guest vCPU.6094 60954.136 KVM_S390_PV_CPU_COMMAND6096-----------------------------6097 6098:Capability: KVM_CAP_S390_PROTECTED_DUMP6099:Architectures: s3906100:Type: vcpu ioctl6101:Parameters: none6102:Returns: 0 on success, < 0 on error6103 6104This ioctl closely mirrors `KVM_S390_PV_COMMAND` but handles requests6105for vcpus. It re-uses the kvm_s390_pv_dmp struct and hence also shares6106the command ids.6107 6108**command:**6109 6110KVM_PV_DUMP6111  Presents an API that provides calls which facilitate dumping a vcpu6112  of a protected VM.6113 6114**subcommand:**6115 6116KVM_PV_DUMP_CPU6117  Provides encrypted dump data like register values.6118  The length of the returned data is provided by uv_info.guest_cpu_stor_len.6119 61204.137 KVM_S390_ZPCI_OP6121----------------------6122 6123:Capability: KVM_CAP_S390_ZPCI_OP6124:Architectures: s3906125:Type: vm ioctl6126:Parameters: struct kvm_s390_zpci_op (in)6127:Returns: 0 on success, <0 on error6128 6129Used to manage hardware-assisted virtualization features for zPCI devices.6130 6131Parameters are specified via the following structure::6132 6133  struct kvm_s390_zpci_op {6134	/* in */6135	__u32 fh;		/* target device */6136	__u8  op;		/* operation to perform */6137	__u8  pad[3];6138	union {6139		/* for KVM_S390_ZPCIOP_REG_AEN */6140		struct {6141			__u64 ibv;	/* Guest addr of interrupt bit vector */6142			__u64 sb;	/* Guest addr of summary bit */6143			__u32 flags;6144			__u32 noi;	/* Number of interrupts */6145			__u8 isc;	/* Guest interrupt subclass */6146			__u8 sbo;	/* Offset of guest summary bit vector */6147			__u16 pad;6148		} reg_aen;6149		__u64 reserved[8];6150	} u;6151  };6152 6153The type of operation is specified in the "op" field.6154KVM_S390_ZPCIOP_REG_AEN is used to register the VM for adapter event6155notification interpretation, which will allow firmware delivery of adapter6156events directly to the vm, with KVM providing a backup delivery mechanism;6157KVM_S390_ZPCIOP_DEREG_AEN is used to subsequently disable interpretation of6158adapter event notifications.6159 6160The target zPCI function must also be specified via the "fh" field.  For the6161KVM_S390_ZPCIOP_REG_AEN operation, additional information to establish firmware6162delivery must be provided via the "reg_aen" struct.6163 6164The "pad" and "reserved" fields may be used for future extensions and should be6165set to 0s by userspace.6166 61674.138 KVM_ARM_SET_COUNTER_OFFSET6168--------------------------------6169 6170:Capability: KVM_CAP_COUNTER_OFFSET6171:Architectures: arm646172:Type: vm ioctl6173:Parameters: struct kvm_arm_counter_offset (in)6174:Returns: 0 on success, < 0 on error6175 6176This capability indicates that userspace is able to apply a single VM-wide6177offset to both the virtual and physical counters as viewed by the guest6178using the KVM_ARM_SET_CNT_OFFSET ioctl and the following data structure:6179 6180::6181 6182	struct kvm_arm_counter_offset {6183		__u64 counter_offset;6184		__u64 reserved;6185	};6186 6187The offset describes a number of counter cycles that are subtracted from6188both virtual and physical counter views (similar to the effects of the6189CNTVOFF_EL2 and CNTPOFF_EL2 system registers, but only global). The offset6190always applies to all vcpus (already created or created after this ioctl)6191for this VM.6192 6193It is userspace's responsibility to compute the offset based, for example,6194on previous values of the guest counters.6195 6196Any value other than 0 for the "reserved" field may result in an error6197(-EINVAL) being returned. This ioctl can also return -EBUSY if any vcpu6198ioctl is issued concurrently.6199 6200Note that using this ioctl results in KVM ignoring subsequent userspace6201writes to the CNTVCT_EL0 and CNTPCT_EL0 registers using the SET_ONE_REG6202interface. No error will be returned, but the resulting offset will not be6203applied.6204 6205.. _KVM_ARM_GET_REG_WRITABLE_MASKS:6206 62074.139 KVM_ARM_GET_REG_WRITABLE_MASKS6208-------------------------------------------6209 6210:Capability: KVM_CAP_ARM_SUPPORTED_REG_MASK_RANGES6211:Architectures: arm646212:Type: vm ioctl6213:Parameters: struct reg_mask_range (in/out)6214:Returns: 0 on success, < 0 on error6215 6216 6217::6218 6219        #define KVM_ARM_FEATURE_ID_RANGE	06220        #define KVM_ARM_FEATURE_ID_RANGE_SIZE	(3 * 8 * 8)6221 6222        struct reg_mask_range {6223                __u64 addr;             /* Pointer to mask array */6224                __u32 range;            /* Requested range */6225                __u32 reserved[13];6226        };6227 6228This ioctl copies the writable masks for a selected range of registers to6229userspace.6230 6231The ``addr`` field is a pointer to the destination array where KVM copies6232the writable masks.6233 6234The ``range`` field indicates the requested range of registers.6235``KVM_CHECK_EXTENSION`` for the ``KVM_CAP_ARM_SUPPORTED_REG_MASK_RANGES``6236capability returns the supported ranges, expressed as a set of flags. Each6237flag's bit index represents a possible value for the ``range`` field.6238All other values are reserved for future use and KVM may return an error.6239 6240The ``reserved[13]`` array is reserved for future use and should be 0, or6241KVM may return an error.6242 6243KVM_ARM_FEATURE_ID_RANGE (0)6244^^^^^^^^^^^^^^^^^^^^^^^^^^^^6245 6246The Feature ID range is defined as the AArch64 System register space with6247op0==3, op1=={0, 1, 3}, CRn==0, CRm=={0-7}, op2=={0-7}.6248 6249The mask returned array pointed to by ``addr`` is indexed by the macro6250``ARM64_FEATURE_ID_RANGE_IDX(op0, op1, crn, crm, op2)``, allowing userspace6251to know what fields can be changed for the system register described by6252``op0, op1, crn, crm, op2``. KVM rejects ID register values that describe a6253superset of the features supported by the system.6254 62554.140 KVM_SET_USER_MEMORY_REGION26256---------------------------------6257 6258:Capability: KVM_CAP_USER_MEMORY26259:Architectures: all6260:Type: vm ioctl6261:Parameters: struct kvm_userspace_memory_region2 (in)6262:Returns: 0 on success, -1 on error6263 6264KVM_SET_USER_MEMORY_REGION2 is an extension to KVM_SET_USER_MEMORY_REGION that6265allows mapping guest_memfd memory into a guest.  All fields shared with6266KVM_SET_USER_MEMORY_REGION identically.  Userspace can set KVM_MEM_GUEST_MEMFD6267in flags to have KVM bind the memory region to a given guest_memfd range of6268[guest_memfd_offset, guest_memfd_offset + memory_size].  The target guest_memfd6269must point at a file created via KVM_CREATE_GUEST_MEMFD on the current VM, and6270the target range must not be bound to any other memory region.  All standard6271bounds checks apply (use common sense).6272 6273::6274 6275  struct kvm_userspace_memory_region2 {6276	__u32 slot;6277	__u32 flags;6278	__u64 guest_phys_addr;6279	__u64 memory_size; /* bytes */6280	__u64 userspace_addr; /* start of the userspace allocated memory */6281	__u64 guest_memfd_offset;6282	__u32 guest_memfd;6283	__u32 pad1;6284	__u64 pad2[14];6285  };6286 6287A KVM_MEM_GUEST_MEMFD region _must_ have a valid guest_memfd (private memory) and6288userspace_addr (shared memory).  However, "valid" for userspace_addr simply6289means that the address itself must be a legal userspace address.  The backing6290mapping for userspace_addr is not required to be valid/populated at the time of6291KVM_SET_USER_MEMORY_REGION2, e.g. shared memory can be lazily mapped/allocated6292on-demand.6293 6294When mapping a gfn into the guest, KVM selects shared vs. private, i.e consumes6295userspace_addr vs. guest_memfd, based on the gfn's KVM_MEMORY_ATTRIBUTE_PRIVATE6296state.  At VM creation time, all memory is shared, i.e. the PRIVATE attribute6297is '0' for all gfns.  Userspace can control whether memory is shared/private by6298toggling KVM_MEMORY_ATTRIBUTE_PRIVATE via KVM_SET_MEMORY_ATTRIBUTES as needed.6299 6300S390:6301^^^^^6302 6303Returns -EINVAL if the VM has the KVM_VM_S390_UCONTROL flag set.6304Returns -EINVAL if called on a protected VM.6305 63064.141 KVM_SET_MEMORY_ATTRIBUTES6307-------------------------------6308 6309:Capability: KVM_CAP_MEMORY_ATTRIBUTES6310:Architectures: x866311:Type: vm ioctl6312:Parameters: struct kvm_memory_attributes (in)6313:Returns: 0 on success, <0 on error6314 6315KVM_SET_MEMORY_ATTRIBUTES allows userspace to set memory attributes for a range6316of guest physical memory.6317 6318::6319 6320  struct kvm_memory_attributes {6321	__u64 address;6322	__u64 size;6323	__u64 attributes;6324	__u64 flags;6325  };6326 6327  #define KVM_MEMORY_ATTRIBUTE_PRIVATE           (1ULL << 3)6328 6329The address and size must be page aligned.  The supported attributes can be6330retrieved via ioctl(KVM_CHECK_EXTENSION) on KVM_CAP_MEMORY_ATTRIBUTES.  If6331executed on a VM, KVM_CAP_MEMORY_ATTRIBUTES precisely returns the attributes6332supported by that VM.  If executed at system scope, KVM_CAP_MEMORY_ATTRIBUTES6333returns all attributes supported by KVM.  The only attribute defined at this6334time is KVM_MEMORY_ATTRIBUTE_PRIVATE, which marks the associated gfn as being6335guest private memory.6336 6337Note, there is no "get" API.  Userspace is responsible for explicitly tracking6338the state of a gfn/page as needed.6339 6340The "flags" field is reserved for future extensions and must be '0'.6341 63424.142 KVM_CREATE_GUEST_MEMFD6343----------------------------6344 6345:Capability: KVM_CAP_GUEST_MEMFD6346:Architectures: none6347:Type: vm ioctl6348:Parameters: struct kvm_create_guest_memfd(in)6349:Returns: A file descriptor on success, <0 on error6350 6351KVM_CREATE_GUEST_MEMFD creates an anonymous file and returns a file descriptor6352that refers to it.  guest_memfd files are roughly analogous to files created6353via memfd_create(), e.g. guest_memfd files live in RAM, have volatile storage,6354and are automatically released when the last reference is dropped.  Unlike6355"regular" memfd_create() files, guest_memfd files are bound to their owning6356virtual machine (see below), cannot be mapped, read, or written by userspace,6357and cannot be resized  (guest_memfd files do however support PUNCH_HOLE).6358 6359::6360 6361  struct kvm_create_guest_memfd {6362	__u64 size;6363	__u64 flags;6364	__u64 reserved[6];6365  };6366 6367Conceptually, the inode backing a guest_memfd file represents physical memory,6368i.e. is coupled to the virtual machine as a thing, not to a "struct kvm".  The6369file itself, which is bound to a "struct kvm", is that instance's view of the6370underlying memory, e.g. effectively provides the translation of guest addresses6371to host memory.  This allows for use cases where multiple KVM structures are6372used to manage a single virtual machine, e.g. when performing intrahost6373migration of a virtual machine.6374 6375KVM currently only supports mapping guest_memfd via KVM_SET_USER_MEMORY_REGION2,6376and more specifically via the guest_memfd and guest_memfd_offset fields in6377"struct kvm_userspace_memory_region2", where guest_memfd_offset is the offset6378into the guest_memfd instance.  For a given guest_memfd file, there can be at6379most one mapping per page, i.e. binding multiple memory regions to a single6380guest_memfd range is not allowed (any number of memory regions can be bound to6381a single guest_memfd file, but the bound ranges must not overlap).6382 6383See KVM_SET_USER_MEMORY_REGION2 for additional details.6384 63854.143 KVM_PRE_FAULT_MEMORY6386---------------------------6387 6388:Capability: KVM_CAP_PRE_FAULT_MEMORY6389:Architectures: none6390:Type: vcpu ioctl6391:Parameters: struct kvm_pre_fault_memory (in/out)6392:Returns: 0 if at least one page is processed, < 0 on error6393 6394Errors:6395 6396  ========== ===============================================================6397  EINVAL     The specified `gpa` and `size` were invalid (e.g. not6398             page aligned, causes an overflow, or size is zero).6399  ENOENT     The specified `gpa` is outside defined memslots.6400  EINTR      An unmasked signal is pending and no page was processed.6401  EFAULT     The parameter address was invalid.6402  EOPNOTSUPP Mapping memory for a GPA is unsupported by the6403             hypervisor, and/or for the current vCPU state/mode.6404  EIO        unexpected error conditions (also causes a WARN)6405  ========== ===============================================================6406 6407::6408 6409  struct kvm_pre_fault_memory {6410	/* in/out */6411	__u64 gpa;6412	__u64 size;6413	/* in */6414	__u64 flags;6415	__u64 padding[5];6416  };6417 6418KVM_PRE_FAULT_MEMORY populates KVM's stage-2 page tables used to map memory6419for the current vCPU state.  KVM maps memory as if the vCPU generated a6420stage-2 read page fault, e.g. faults in memory as needed, but doesn't break6421CoW.  However, KVM does not mark any newly created stage-2 PTE as Accessed.6422 6423In the case of confidential VM types where there is an initial set up of6424private guest memory before the guest is 'finalized'/measured, this ioctl6425should only be issued after completing all the necessary setup to put the6426guest into a 'finalized' state so that the above semantics can be reliably6427ensured.6428 6429In some cases, multiple vCPUs might share the page tables.  In this6430case, the ioctl can be called in parallel.6431 6432When the ioctl returns, the input values are updated to point to the6433remaining range.  If `size` > 0 on return, the caller can just issue6434the ioctl again with the same `struct kvm_map_memory` argument.6435 6436Shadow page tables cannot support this ioctl because they6437are indexed by virtual address or nested guest physical address.6438Calling this ioctl when the guest is using shadow page tables (for6439example because it is running a nested guest with nested page tables)6440will fail with `EOPNOTSUPP` even if `KVM_CHECK_EXTENSION` reports6441the capability to be present.6442 6443`flags` must currently be zero.6444 6445 64465. The kvm_run structure6447========================6448 6449Application code obtains a pointer to the kvm_run structure by6450mmap()ing a vcpu fd.  From that point, application code can control6451execution by changing fields in kvm_run prior to calling the KVM_RUN6452ioctl, and obtain information about the reason KVM_RUN returned by6453looking up structure members.6454 6455::6456 6457  struct kvm_run {6458	/* in */6459	__u8 request_interrupt_window;6460 6461Request that KVM_RUN return when it becomes possible to inject external6462interrupts into the guest.  Useful in conjunction with KVM_INTERRUPT.6463 6464::6465 6466	__u8 immediate_exit;6467 6468This field is polled once when KVM_RUN starts; if non-zero, KVM_RUN6469exits immediately, returning -EINTR.  In the common scenario where a6470signal is used to "kick" a VCPU out of KVM_RUN, this field can be used6471to avoid usage of KVM_SET_SIGNAL_MASK, which has worse scalability.6472Rather than blocking the signal outside KVM_RUN, userspace can set up6473a signal handler that sets run->immediate_exit to a non-zero value.6474 6475This field is ignored if KVM_CAP_IMMEDIATE_EXIT is not available.6476 6477::6478 6479	__u8 padding1[6];6480 6481	/* out */6482	__u32 exit_reason;6483 6484When KVM_RUN has returned successfully (return value 0), this informs6485application code why KVM_RUN has returned.  Allowable values for this6486field are detailed below.6487 6488::6489 6490	__u8 ready_for_interrupt_injection;6491 6492If request_interrupt_window has been specified, this field indicates6493an interrupt can be injected now with KVM_INTERRUPT.6494 6495::6496 6497	__u8 if_flag;6498 6499The value of the current interrupt flag.  Only valid if in-kernel6500local APIC is not used.6501 6502::6503 6504	__u16 flags;6505 6506More architecture-specific flags detailing state of the VCPU that may6507affect the device's behavior. Current defined flags::6508 6509  /* x86, set if the VCPU is in system management mode */6510  #define KVM_RUN_X86_SMM          (1 << 0)6511  /* x86, set if bus lock detected in VM */6512  #define KVM_RUN_X86_BUS_LOCK     (1 << 1)6513  /* x86, set if the VCPU is executing a nested (L2) guest */6514  #define KVM_RUN_X86_GUEST_MODE   (1 << 2)6515 6516  /* arm64, set for KVM_EXIT_DEBUG */6517  #define KVM_DEBUG_ARCH_HSR_HIGH_VALID  (1 << 0)6518 6519::6520 6521	/* in (pre_kvm_run), out (post_kvm_run) */6522	__u64 cr8;6523 6524The value of the cr8 register.  Only valid if in-kernel local APIC is6525not used.  Both input and output.6526 6527::6528 6529	__u64 apic_base;6530 6531The value of the APIC BASE msr.  Only valid if in-kernel local6532APIC is not used.  Both input and output.6533 6534::6535 6536	union {6537		/* KVM_EXIT_UNKNOWN */6538		struct {6539			__u64 hardware_exit_reason;6540		} hw;6541 6542If exit_reason is KVM_EXIT_UNKNOWN, the vcpu has exited due to unknown6543reasons.  Further architecture-specific information is available in6544hardware_exit_reason.6545 6546::6547 6548		/* KVM_EXIT_FAIL_ENTRY */6549		struct {6550			__u64 hardware_entry_failure_reason;6551			__u32 cpu; /* if KVM_LAST_CPU */6552		} fail_entry;6553 6554If exit_reason is KVM_EXIT_FAIL_ENTRY, the vcpu could not be run due6555to unknown reasons.  Further architecture-specific information is6556available in hardware_entry_failure_reason.6557 6558::6559 6560		/* KVM_EXIT_EXCEPTION */6561		struct {6562			__u32 exception;6563			__u32 error_code;6564		} ex;6565 6566Unused.6567 6568::6569 6570		/* KVM_EXIT_IO */6571		struct {6572  #define KVM_EXIT_IO_IN  06573  #define KVM_EXIT_IO_OUT 16574			__u8 direction;6575			__u8 size; /* bytes */6576			__u16 port;6577			__u32 count;6578			__u64 data_offset; /* relative to kvm_run start */6579		} io;6580 6581If exit_reason is KVM_EXIT_IO, then the vcpu has6582executed a port I/O instruction which could not be satisfied by kvm.6583data_offset describes where the data is located (KVM_EXIT_IO_OUT) or6584where kvm expects application code to place the data for the next6585KVM_RUN invocation (KVM_EXIT_IO_IN).  Data format is a packed array.6586 6587::6588 6589		/* KVM_EXIT_DEBUG */6590		struct {6591			struct kvm_debug_exit_arch arch;6592		} debug;6593 6594If the exit_reason is KVM_EXIT_DEBUG, then a vcpu is processing a debug event6595for which architecture specific information is returned.6596 6597::6598 6599		/* KVM_EXIT_MMIO */6600		struct {6601			__u64 phys_addr;6602			__u8  data[8];6603			__u32 len;6604			__u8  is_write;6605		} mmio;6606 6607If exit_reason is KVM_EXIT_MMIO, then the vcpu has6608executed a memory-mapped I/O instruction which could not be satisfied6609by kvm.  The 'data' member contains the written data if 'is_write' is6610true, and should be filled by application code otherwise.6611 6612The 'data' member contains, in its first 'len' bytes, the value as it would6613appear if the VCPU performed a load or store of the appropriate width directly6614to the byte array.6615 6616.. note::6617 6618      For KVM_EXIT_IO, KVM_EXIT_MMIO, KVM_EXIT_OSI, KVM_EXIT_PAPR, KVM_EXIT_XEN,6619      KVM_EXIT_EPR, KVM_EXIT_X86_RDMSR and KVM_EXIT_X86_WRMSR the corresponding6620      operations are complete (and guest state is consistent) only after userspace6621      has re-entered the kernel with KVM_RUN.  The kernel side will first finish6622      incomplete operations and then check for pending signals.6623 6624      The pending state of the operation is not preserved in state which is6625      visible to userspace, thus userspace should ensure that the operation is6626      completed before performing a live migration.  Userspace can re-enter the6627      guest with an unmasked signal pending or with the immediate_exit field set6628      to complete pending operations without allowing any further instructions6629      to be executed.6630 6631::6632 6633		/* KVM_EXIT_HYPERCALL */6634		struct {6635			__u64 nr;6636			__u64 args[6];6637			__u64 ret;6638			__u64 flags;6639		} hypercall;6640 6641 6642It is strongly recommended that userspace use ``KVM_EXIT_IO`` (x86) or6643``KVM_EXIT_MMIO`` (all except s390) to implement functionality that6644requires a guest to interact with host userspace.6645 6646.. note:: KVM_EXIT_IO is significantly faster than KVM_EXIT_MMIO.6647 6648For arm64:6649----------6650 6651SMCCC exits can be enabled depending on the configuration of the SMCCC6652filter. See the Documentation/virt/kvm/devices/vm.rst6653``KVM_ARM_SMCCC_FILTER`` for more details.6654 6655``nr`` contains the function ID of the guest's SMCCC call. Userspace is6656expected to use the ``KVM_GET_ONE_REG`` ioctl to retrieve the call6657parameters from the vCPU's GPRs.6658 6659Definition of ``flags``:6660 - ``KVM_HYPERCALL_EXIT_SMC``: Indicates that the guest used the SMC6661   conduit to initiate the SMCCC call. If this bit is 0 then the guest6662   used the HVC conduit for the SMCCC call.6663 6664 - ``KVM_HYPERCALL_EXIT_16BIT``: Indicates that the guest used a 16bit6665   instruction to initiate the SMCCC call. If this bit is 0 then the6666   guest used a 32bit instruction. An AArch64 guest always has this6667   bit set to 0.6668 6669At the point of exit, PC points to the instruction immediately following6670the trapping instruction.6671 6672::6673 6674		/* KVM_EXIT_TPR_ACCESS */6675		struct {6676			__u64 rip;6677			__u32 is_write;6678			__u32 pad;6679		} tpr_access;6680 6681To be documented (KVM_TPR_ACCESS_REPORTING).6682 6683::6684 6685		/* KVM_EXIT_S390_SIEIC */6686		struct {6687			__u8 icptcode;6688			__u64 mask; /* psw upper half */6689			__u64 addr; /* psw lower half */6690			__u16 ipa;6691			__u32 ipb;6692		} s390_sieic;6693 6694s390 specific.6695 6696::6697 6698		/* KVM_EXIT_S390_RESET */6699  #define KVM_S390_RESET_POR       16700  #define KVM_S390_RESET_CLEAR     26701  #define KVM_S390_RESET_SUBSYSTEM 46702  #define KVM_S390_RESET_CPU_INIT  86703  #define KVM_S390_RESET_IPL       166704		__u64 s390_reset_flags;6705 6706s390 specific.6707 6708::6709 6710		/* KVM_EXIT_S390_UCONTROL */6711		struct {6712			__u64 trans_exc_code;6713			__u32 pgm_code;6714		} s390_ucontrol;6715 6716s390 specific. A page fault has occurred for a user controlled virtual6717machine (KVM_VM_S390_UNCONTROL) on its host page table that cannot be6718resolved by the kernel.6719The program code and the translation exception code that were placed6720in the cpu's lowcore are presented here as defined by the z Architecture6721Principles of Operation Book in the Chapter for Dynamic Address Translation6722(DAT)6723 6724::6725 6726		/* KVM_EXIT_DCR */6727		struct {6728			__u32 dcrn;6729			__u32 data;6730			__u8  is_write;6731		} dcr;6732 6733Deprecated - was used for 440 KVM.6734 6735::6736 6737		/* KVM_EXIT_OSI */6738		struct {6739			__u64 gprs[32];6740		} osi;6741 6742MOL uses a special hypercall interface it calls 'OSI'. To enable it, we catch6743hypercalls and exit with this exit struct that contains all the guest gprs.6744 6745If exit_reason is KVM_EXIT_OSI, then the vcpu has triggered such a hypercall.6746Userspace can now handle the hypercall and when it's done modify the gprs as6747necessary. Upon guest entry all guest GPRs will then be replaced by the values6748in this struct.6749 6750::6751 6752		/* KVM_EXIT_PAPR_HCALL */6753		struct {6754			__u64 nr;6755			__u64 ret;6756			__u64 args[9];6757		} papr_hcall;6758 6759This is used on 64-bit PowerPC when emulating a pSeries partition,6760e.g. with the 'pseries' machine type in qemu.  It occurs when the6761guest does a hypercall using the 'sc 1' instruction.  The 'nr' field6762contains the hypercall number (from the guest R3), and 'args' contains6763the arguments (from the guest R4 - R12).  Userspace should put the6764return code in 'ret' and any extra returned values in args[].6765The possible hypercalls are defined in the Power Architecture Platform6766Requirements (PAPR) document available from www.power.org (free6767developer registration required to access it).6768 6769::6770 6771		/* KVM_EXIT_S390_TSCH */6772		struct {6773			__u16 subchannel_id;6774			__u16 subchannel_nr;6775			__u32 io_int_parm;6776			__u32 io_int_word;6777			__u32 ipb;6778			__u8 dequeued;6779		} s390_tsch;6780 6781s390 specific. This exit occurs when KVM_CAP_S390_CSS_SUPPORT has been enabled6782and TEST SUBCHANNEL was intercepted. If dequeued is set, a pending I/O6783interrupt for the target subchannel has been dequeued and subchannel_id,6784subchannel_nr, io_int_parm and io_int_word contain the parameters for that6785interrupt. ipb is needed for instruction parameter decoding.6786 6787::6788 6789		/* KVM_EXIT_EPR */6790		struct {6791			__u32 epr;6792		} epr;6793 6794On FSL BookE PowerPC chips, the interrupt controller has a fast patch6795interrupt acknowledge path to the core. When the core successfully6796delivers an interrupt, it automatically populates the EPR register with6797the interrupt vector number and acknowledges the interrupt inside6798the interrupt controller.6799 6800In case the interrupt controller lives in user space, we need to do6801the interrupt acknowledge cycle through it to fetch the next to be6802delivered interrupt vector using this exit.6803 6804It gets triggered whenever both KVM_CAP_PPC_EPR are enabled and an6805external interrupt has just been delivered into the guest. User space6806should put the acknowledged interrupt vector into the 'epr' field.6807 6808::6809 6810		/* KVM_EXIT_SYSTEM_EVENT */6811		struct {6812  #define KVM_SYSTEM_EVENT_SHUTDOWN       16813  #define KVM_SYSTEM_EVENT_RESET          26814  #define KVM_SYSTEM_EVENT_CRASH          36815  #define KVM_SYSTEM_EVENT_WAKEUP         46816  #define KVM_SYSTEM_EVENT_SUSPEND        56817  #define KVM_SYSTEM_EVENT_SEV_TERM       66818			__u32 type;6819                        __u32 ndata;6820                        __u64 data[16];6821		} system_event;6822 6823If exit_reason is KVM_EXIT_SYSTEM_EVENT then the vcpu has triggered6824a system-level event using some architecture specific mechanism (hypercall6825or some special instruction). In case of ARM64, this is triggered using6826HVC instruction based PSCI call from the vcpu.6827 6828The 'type' field describes the system-level event type.6829Valid values for 'type' are:6830 6831 - KVM_SYSTEM_EVENT_SHUTDOWN -- the guest has requested a shutdown of the6832   VM. Userspace is not obliged to honour this, and if it does honour6833   this does not need to destroy the VM synchronously (ie it may call6834   KVM_RUN again before shutdown finally occurs).6835 - KVM_SYSTEM_EVENT_RESET -- the guest has requested a reset of the VM.6836   As with SHUTDOWN, userspace can choose to ignore the request, or6837   to schedule the reset to occur in the future and may call KVM_RUN again.6838 - KVM_SYSTEM_EVENT_CRASH -- the guest crash occurred and the guest6839   has requested a crash condition maintenance. Userspace can choose6840   to ignore the request, or to gather VM memory core dump and/or6841   reset/shutdown of the VM.6842 - KVM_SYSTEM_EVENT_SEV_TERM -- an AMD SEV guest requested termination.6843   The guest physical address of the guest's GHCB is stored in `data[0]`.6844 - KVM_SYSTEM_EVENT_WAKEUP -- the exiting vCPU is in a suspended state and6845   KVM has recognized a wakeup event. Userspace may honor this event by6846   marking the exiting vCPU as runnable, or deny it and call KVM_RUN again.6847 - KVM_SYSTEM_EVENT_SUSPEND -- the guest has requested a suspension of6848   the VM.6849 6850If KVM_CAP_SYSTEM_EVENT_DATA is present, the 'data' field can contain6851architecture specific information for the system-level event.  Only6852the first `ndata` items (possibly zero) of the data array are valid.6853 6854 - for arm64, data[0] is set to KVM_SYSTEM_EVENT_RESET_FLAG_PSCI_RESET2 if6855   the guest issued a SYSTEM_RESET2 call according to v1.1 of the PSCI6856   specification.6857 6858 - for RISC-V, data[0] is set to the value of the second argument of the6859   ``sbi_system_reset`` call.6860 6861Previous versions of Linux defined a `flags` member in this struct.  The6862field is now aliased to `data[0]`.  Userspace can assume that it is only6863written if ndata is greater than 0.6864 6865For arm/arm64:6866--------------6867 6868KVM_SYSTEM_EVENT_SUSPEND exits are enabled with the6869KVM_CAP_ARM_SYSTEM_SUSPEND VM capability. If a guest invokes the PSCI6870SYSTEM_SUSPEND function, KVM will exit to userspace with this event6871type.6872 6873It is the sole responsibility of userspace to implement the PSCI6874SYSTEM_SUSPEND call according to ARM DEN0022D.b 5.19 "SYSTEM_SUSPEND".6875KVM does not change the vCPU's state before exiting to userspace, so6876the call parameters are left in-place in the vCPU registers.6877 6878Userspace is _required_ to take action for such an exit. It must6879either:6880 6881 - Honor the guest request to suspend the VM. Userspace can request6882   in-kernel emulation of suspension by setting the calling vCPU's6883   state to KVM_MP_STATE_SUSPENDED. Userspace must configure the vCPU's6884   state according to the parameters passed to the PSCI function when6885   the calling vCPU is resumed. See ARM DEN0022D.b 5.19.1 "Intended use"6886   for details on the function parameters.6887 6888 - Deny the guest request to suspend the VM. See ARM DEN0022D.b 5.19.26889   "Caller responsibilities" for possible return values.6890 6891::6892 6893		/* KVM_EXIT_IOAPIC_EOI */6894		struct {6895			__u8 vector;6896		} eoi;6897 6898Indicates that the VCPU's in-kernel local APIC received an EOI for a6899level-triggered IOAPIC interrupt.  This exit only triggers when the6900IOAPIC is implemented in userspace (i.e. KVM_CAP_SPLIT_IRQCHIP is enabled);6901the userspace IOAPIC should process the EOI and retrigger the interrupt if6902it is still asserted.  Vector is the LAPIC interrupt vector for which the6903EOI was received.6904 6905::6906 6907		struct kvm_hyperv_exit {6908  #define KVM_EXIT_HYPERV_SYNIC          16909  #define KVM_EXIT_HYPERV_HCALL          26910  #define KVM_EXIT_HYPERV_SYNDBG         36911			__u32 type;6912			__u32 pad1;6913			union {6914				struct {6915					__u32 msr;6916					__u32 pad2;6917					__u64 control;6918					__u64 evt_page;6919					__u64 msg_page;6920				} synic;6921				struct {6922					__u64 input;6923					__u64 result;6924					__u64 params[2];6925				} hcall;6926				struct {6927					__u32 msr;6928					__u32 pad2;6929					__u64 control;6930					__u64 status;6931					__u64 send_page;6932					__u64 recv_page;6933					__u64 pending_page;6934				} syndbg;6935			} u;6936		};6937		/* KVM_EXIT_HYPERV */6938                struct kvm_hyperv_exit hyperv;6939 6940Indicates that the VCPU exits into userspace to process some tasks6941related to Hyper-V emulation.6942 6943Valid values for 'type' are:6944 6945	- KVM_EXIT_HYPERV_SYNIC -- synchronously notify user-space about6946 6947Hyper-V SynIC state change. Notification is used to remap SynIC6948event/message pages and to enable/disable SynIC messages/events processing6949in userspace.6950 6951	- KVM_EXIT_HYPERV_SYNDBG -- synchronously notify user-space about6952 6953Hyper-V Synthetic debugger state change. Notification is used to either update6954the pending_page location or to send a control command (send the buffer located6955in send_page or recv a buffer to recv_page).6956 6957::6958 6959		/* KVM_EXIT_ARM_NISV */6960		struct {6961			__u64 esr_iss;6962			__u64 fault_ipa;6963		} arm_nisv;6964 6965Used on arm64 systems. If a guest accesses memory not in a memslot,6966KVM will typically return to userspace and ask it to do MMIO emulation on its6967behalf. However, for certain classes of instructions, no instruction decode6968(direction, length of memory access) is provided, and fetching and decoding6969the instruction from the VM is overly complicated to live in the kernel.6970 6971Historically, when this situation occurred, KVM would print a warning and kill6972the VM. KVM assumed that if the guest accessed non-memslot memory, it was6973trying to do I/O, which just couldn't be emulated, and the warning message was6974phrased accordingly. However, what happened more often was that a guest bug6975caused access outside the guest memory areas which should lead to a more6976meaningful warning message and an external abort in the guest, if the access6977did not fall within an I/O window.6978 6979Userspace implementations can query for KVM_CAP_ARM_NISV_TO_USER, and enable6980this capability at VM creation. Once this is done, these types of errors will6981instead return to userspace with KVM_EXIT_ARM_NISV, with the valid bits from6982the ESR_EL2 in the esr_iss field, and the faulting IPA in the fault_ipa field.6983Userspace can either fix up the access if it's actually an I/O access by6984decoding the instruction from guest memory (if it's very brave) and continue6985executing the guest, or it can decide to suspend, dump, or restart the guest.6986 6987Note that KVM does not skip the faulting instruction as it does for6988KVM_EXIT_MMIO, but userspace has to emulate any change to the processing state6989if it decides to decode and emulate the instruction.6990 6991This feature isn't available to protected VMs, as userspace does not6992have access to the state that is required to perform the emulation.6993Instead, a data abort exception is directly injected in the guest.6994Note that although KVM_CAP_ARM_NISV_TO_USER will be reported if6995queried outside of a protected VM context, the feature will not be6996exposed if queried on a protected VM file descriptor.6997 6998::6999 7000		/* KVM_EXIT_X86_RDMSR / KVM_EXIT_X86_WRMSR */7001		struct {7002			__u8 error; /* user -> kernel */7003			__u8 pad[7];7004			__u32 reason; /* kernel -> user */7005			__u32 index; /* kernel -> user */7006			__u64 data; /* kernel <-> user */7007		} msr;7008 7009Used on x86 systems. When the VM capability KVM_CAP_X86_USER_SPACE_MSR is7010enabled, MSR accesses to registers that would invoke a #GP by KVM kernel code7011may instead trigger a KVM_EXIT_X86_RDMSR exit for reads and KVM_EXIT_X86_WRMSR7012exit for writes.7013 7014The "reason" field specifies why the MSR interception occurred. Userspace will7015only receive MSR exits when a particular reason was requested during through7016ENABLE_CAP. Currently valid exit reasons are:7017 7018============================ ========================================7019 KVM_MSR_EXIT_REASON_UNKNOWN access to MSR that is unknown to KVM7020 KVM_MSR_EXIT_REASON_INVAL   access to invalid MSRs or reserved bits7021 KVM_MSR_EXIT_REASON_FILTER  access blocked by KVM_X86_SET_MSR_FILTER7022============================ ========================================7023 7024For KVM_EXIT_X86_RDMSR, the "index" field tells userspace which MSR the guest7025wants to read. To respond to this request with a successful read, userspace7026writes the respective data into the "data" field and must continue guest7027execution to ensure the read data is transferred into guest register state.7028 7029If the RDMSR request was unsuccessful, userspace indicates that with a "1" in7030the "error" field. This will inject a #GP into the guest when the VCPU is7031executed again.7032 7033For KVM_EXIT_X86_WRMSR, the "index" field tells userspace which MSR the guest7034wants to write. Once finished processing the event, userspace must continue7035vCPU execution. If the MSR write was unsuccessful, userspace also sets the7036"error" field to "1".7037 7038See KVM_X86_SET_MSR_FILTER for details on the interaction with MSR filtering.7039 7040::7041 7042 7043		struct kvm_xen_exit {7044  #define KVM_EXIT_XEN_HCALL          17045			__u32 type;7046			union {7047				struct {7048					__u32 longmode;7049					__u32 cpl;7050					__u64 input;7051					__u64 result;7052					__u64 params[6];7053				} hcall;7054			} u;7055		};7056		/* KVM_EXIT_XEN */7057                struct kvm_hyperv_exit xen;7058 7059Indicates that the VCPU exits into userspace to process some tasks7060related to Xen emulation.7061 7062Valid values for 'type' are:7063 7064  - KVM_EXIT_XEN_HCALL -- synchronously notify user-space about Xen hypercall.7065    Userspace is expected to place the hypercall result into the appropriate7066    field before invoking KVM_RUN again.7067 7068::7069 7070		/* KVM_EXIT_RISCV_SBI */7071		struct {7072			unsigned long extension_id;7073			unsigned long function_id;7074			unsigned long args[6];7075			unsigned long ret[2];7076		} riscv_sbi;7077 7078If exit reason is KVM_EXIT_RISCV_SBI then it indicates that the VCPU has7079done a SBI call which is not handled by KVM RISC-V kernel module. The details7080of the SBI call are available in 'riscv_sbi' member of kvm_run structure. The7081'extension_id' field of 'riscv_sbi' represents SBI extension ID whereas the7082'function_id' field represents function ID of given SBI extension. The 'args'7083array field of 'riscv_sbi' represents parameters for the SBI call and 'ret'7084array field represents return values. The userspace should update the return7085values of SBI call before resuming the VCPU. For more details on RISC-V SBI7086spec refer, https://github.com/riscv/riscv-sbi-doc.7087 7088::7089 7090		/* KVM_EXIT_MEMORY_FAULT */7091		struct {7092  #define KVM_MEMORY_EXIT_FLAG_PRIVATE	(1ULL << 3)7093			__u64 flags;7094			__u64 gpa;7095			__u64 size;7096		} memory_fault;7097 7098KVM_EXIT_MEMORY_FAULT indicates the vCPU has encountered a memory fault that7099could not be resolved by KVM.  The 'gpa' and 'size' (in bytes) describe the7100guest physical address range [gpa, gpa + size) of the fault.  The 'flags' field7101describes properties of the faulting access that are likely pertinent:7102 7103 - KVM_MEMORY_EXIT_FLAG_PRIVATE - When set, indicates the memory fault occurred7104   on a private memory access.  When clear, indicates the fault occurred on a7105   shared access.7106 7107Note!  KVM_EXIT_MEMORY_FAULT is unique among all KVM exit reasons in that it7108accompanies a return code of '-1', not '0'!  errno will always be set to EFAULT7109or EHWPOISON when KVM exits with KVM_EXIT_MEMORY_FAULT, userspace should assume7110kvm_run.exit_reason is stale/undefined for all other error numbers.7111 7112::7113 7114    /* KVM_EXIT_NOTIFY */7115    struct {7116  #define KVM_NOTIFY_CONTEXT_INVALID	(1 << 0)7117      __u32 flags;7118    } notify;7119 7120Used on x86 systems. When the VM capability KVM_CAP_X86_NOTIFY_VMEXIT is7121enabled, a VM exit generated if no event window occurs in VM non-root mode7122for a specified amount of time. Once KVM_X86_NOTIFY_VMEXIT_USER is set when7123enabling the cap, it would exit to userspace with the exit reason7124KVM_EXIT_NOTIFY for further handling. The "flags" field contains more7125detailed info.7126 7127The valid value for 'flags' is:7128 7129  - KVM_NOTIFY_CONTEXT_INVALID -- the VM context is corrupted and not valid7130    in VMCS. It would run into unknown result if resume the target VM.7131 7132::7133 7134		/* Fix the size of the union. */7135		char padding[256];7136	};7137 7138	/*7139	 * shared registers between kvm and userspace.7140	 * kvm_valid_regs specifies the register classes set by the host7141	 * kvm_dirty_regs specified the register classes dirtied by userspace7142	 * struct kvm_sync_regs is architecture specific, as well as the7143	 * bits for kvm_valid_regs and kvm_dirty_regs7144	 */7145	__u64 kvm_valid_regs;7146	__u64 kvm_dirty_regs;7147	union {7148		struct kvm_sync_regs regs;7149		char padding[SYNC_REGS_SIZE_BYTES];7150	} s;7151 7152If KVM_CAP_SYNC_REGS is defined, these fields allow userspace to access7153certain guest registers without having to call SET/GET_*REGS. Thus we can7154avoid some system call overhead if userspace has to handle the exit.7155Userspace can query the validity of the structure by checking7156kvm_valid_regs for specific bits. These bits are architecture specific7157and usually define the validity of a groups of registers. (e.g. one bit7158for general purpose registers)7159 7160Please note that the kernel is allowed to use the kvm_run structure as the7161primary storage for certain register types. Therefore, the kernel may use the7162values in kvm_run even if the corresponding bit in kvm_dirty_regs is not set.7163 7164 71656. Capabilities that can be enabled on vCPUs7166============================================7167 7168There are certain capabilities that change the behavior of the virtual CPU or7169the virtual machine when enabled. To enable them, please see section 4.37.7170Below you can find a list of capabilities and what their effect on the vCPU or7171the virtual machine is when enabling them.7172 7173The following information is provided along with the description:7174 7175  Architectures:7176      which instruction set architectures provide this ioctl.7177      x86 includes both i386 and x86_64.7178 7179  Target:7180      whether this is a per-vcpu or per-vm capability.7181 7182  Parameters:7183      what parameters are accepted by the capability.7184 7185  Returns:7186      the return value.  General error numbers (EBADF, ENOMEM, EINVAL)7187      are not detailed, but errors with specific meanings are.7188 7189 71906.1 KVM_CAP_PPC_OSI7191-------------------7192 7193:Architectures: ppc7194:Target: vcpu7195:Parameters: none7196:Returns: 0 on success; -1 on error7197 7198This capability enables interception of OSI hypercalls that otherwise would7199be treated as normal system calls to be injected into the guest. OSI hypercalls7200were invented by Mac-on-Linux to have a standardized communication mechanism7201between the guest and the host.7202 7203When this capability is enabled, KVM_EXIT_OSI can occur.7204 7205 72066.2 KVM_CAP_PPC_PAPR7207--------------------7208 7209:Architectures: ppc7210:Target: vcpu7211:Parameters: none7212:Returns: 0 on success; -1 on error7213 7214This capability enables interception of PAPR hypercalls. PAPR hypercalls are7215done using the hypercall instruction "sc 1".7216 7217It also sets the guest privilege level to "supervisor" mode. Usually the guest7218runs in "hypervisor" privilege mode with a few missing features.7219 7220In addition to the above, it changes the semantics of SDR1. In this mode, the7221HTAB address part of SDR1 contains an HVA instead of a GPA, as PAPR keeps the7222HTAB invisible to the guest.7223 7224When this capability is enabled, KVM_EXIT_PAPR_HCALL can occur.7225 7226 72276.3 KVM_CAP_SW_TLB7228------------------7229 7230:Architectures: ppc7231:Target: vcpu7232:Parameters: args[0] is the address of a struct kvm_config_tlb7233:Returns: 0 on success; -1 on error7234 7235::7236 7237  struct kvm_config_tlb {7238	__u64 params;7239	__u64 array;7240	__u32 mmu_type;7241	__u32 array_len;7242  };7243 7244Configures the virtual CPU's TLB array, establishing a shared memory area7245between userspace and KVM.  The "params" and "array" fields are userspace7246addresses of mmu-type-specific data structures.  The "array_len" field is an7247safety mechanism, and should be set to the size in bytes of the memory that7248userspace has reserved for the array.  It must be at least the size dictated7249by "mmu_type" and "params".7250 7251While KVM_RUN is active, the shared region is under control of KVM.  Its7252contents are undefined, and any modification by userspace results in7253boundedly undefined behavior.7254 7255On return from KVM_RUN, the shared region will reflect the current state of7256the guest's TLB.  If userspace makes any changes, it must call KVM_DIRTY_TLB7257to tell KVM which entries have been changed, prior to calling KVM_RUN again7258on this vcpu.7259 7260For mmu types KVM_MMU_FSL_BOOKE_NOHV and KVM_MMU_FSL_BOOKE_HV:7261 7262 - The "params" field is of type "struct kvm_book3e_206_tlb_params".7263 - The "array" field points to an array of type "struct7264   kvm_book3e_206_tlb_entry".7265 - The array consists of all entries in the first TLB, followed by all7266   entries in the second TLB.7267 - Within a TLB, entries are ordered first by increasing set number.  Within a7268   set, entries are ordered by way (increasing ESEL).7269 - The hash for determining set number in TLB0 is: (MAS2 >> 12) & (num_sets - 1)7270   where "num_sets" is the tlb_sizes[] value divided by the tlb_ways[] value.7271 - The tsize field of mas1 shall be set to 4K on TLB0, even though the7272   hardware ignores this value for TLB0.7273 72746.4 KVM_CAP_S390_CSS_SUPPORT7275----------------------------7276 7277:Architectures: s3907278:Target: vcpu7279:Parameters: none7280:Returns: 0 on success; -1 on error7281 7282This capability enables support for handling of channel I/O instructions.7283 7284TEST PENDING INTERRUPTION and the interrupt portion of TEST SUBCHANNEL are7285handled in-kernel, while the other I/O instructions are passed to userspace.7286 7287When this capability is enabled, KVM_EXIT_S390_TSCH will occur on TEST7288SUBCHANNEL intercepts.7289 7290Note that even though this capability is enabled per-vcpu, the complete7291virtual machine is affected.7292 72936.5 KVM_CAP_PPC_EPR7294-------------------7295 7296:Architectures: ppc7297:Target: vcpu7298:Parameters: args[0] defines whether the proxy facility is active7299:Returns: 0 on success; -1 on error7300 7301This capability enables or disables the delivery of interrupts through the7302external proxy facility.7303 7304When enabled (args[0] != 0), every time the guest gets an external interrupt7305delivered, it automatically exits into user space with a KVM_EXIT_EPR exit7306to receive the topmost interrupt vector.7307 7308When disabled (args[0] == 0), behavior is as if this facility is unsupported.7309 7310When this capability is enabled, KVM_EXIT_EPR can occur.7311 73126.6 KVM_CAP_IRQ_MPIC7313--------------------7314 7315:Architectures: ppc7316:Parameters: args[0] is the MPIC device fd;7317             args[1] is the MPIC CPU number for this vcpu7318 7319This capability connects the vcpu to an in-kernel MPIC device.7320 73216.7 KVM_CAP_IRQ_XICS7322--------------------7323 7324:Architectures: ppc7325:Target: vcpu7326:Parameters: args[0] is the XICS device fd;7327             args[1] is the XICS CPU number (server ID) for this vcpu7328 7329This capability connects the vcpu to an in-kernel XICS device.7330 73316.8 KVM_CAP_S390_IRQCHIP7332------------------------7333 7334:Architectures: s3907335:Target: vm7336:Parameters: none7337 7338This capability enables the in-kernel irqchip for s390. Please refer to7339"4.24 KVM_CREATE_IRQCHIP" for details.7340 73416.9 KVM_CAP_MIPS_FPU7342--------------------7343 7344:Architectures: mips7345:Target: vcpu7346:Parameters: args[0] is reserved for future use (should be 0).7347 7348This capability allows the use of the host Floating Point Unit by the guest. It7349allows the Config1.FP bit to be set to enable the FPU in the guest. Once this is7350done the ``KVM_REG_MIPS_FPR_*`` and ``KVM_REG_MIPS_FCR_*`` registers can be7351accessed (depending on the current guest FPU register mode), and the Status.FR,7352Config5.FRE bits are accessible via the KVM API and also from the guest,7353depending on them being supported by the FPU.7354 73556.10 KVM_CAP_MIPS_MSA7356---------------------7357 7358:Architectures: mips7359:Target: vcpu7360:Parameters: args[0] is reserved for future use (should be 0).7361 7362This capability allows the use of the MIPS SIMD Architecture (MSA) by the guest.7363It allows the Config3.MSAP bit to be set to enable the use of MSA by the guest.7364Once this is done the ``KVM_REG_MIPS_VEC_*`` and ``KVM_REG_MIPS_MSA_*``7365registers can be accessed, and the Config5.MSAEn bit is accessible via the7366KVM API and also from the guest.7367 73686.74 KVM_CAP_SYNC_REGS7369----------------------7370 7371:Architectures: s390, x867372:Target: s390: always enabled, x86: vcpu7373:Parameters: none7374:Returns: x86: KVM_CHECK_EXTENSION returns a bit-array indicating which register7375          sets are supported7376          (bitfields defined in arch/x86/include/uapi/asm/kvm.h).7377 7378As described above in the kvm_sync_regs struct info in section 5 (kvm_run):7379KVM_CAP_SYNC_REGS "allow[s] userspace to access certain guest registers7380without having to call SET/GET_*REGS". This reduces overhead by eliminating7381repeated ioctl calls for setting and/or getting register values. This is7382particularly important when userspace is making synchronous guest state7383modifications, e.g. when emulating and/or intercepting instructions in7384userspace.7385 7386For s390 specifics, please refer to the source code.7387 7388For x86:7389 7390- the register sets to be copied out to kvm_run are selectable7391  by userspace (rather that all sets being copied out for every exit).7392- vcpu_events are available in addition to regs and sregs.7393 7394For x86, the 'kvm_valid_regs' field of struct kvm_run is overloaded to7395function as an input bit-array field set by userspace to indicate the7396specific register sets to be copied out on the next exit.7397 7398To indicate when userspace has modified values that should be copied into7399the vCPU, the all architecture bitarray field, 'kvm_dirty_regs' must be set.7400This is done using the same bitflags as for the 'kvm_valid_regs' field.7401If the dirty bit is not set, then the register set values will not be copied7402into the vCPU even if they've been modified.7403 7404Unused bitfields in the bitarrays must be set to zero.7405 7406::7407 7408  struct kvm_sync_regs {7409        struct kvm_regs regs;7410        struct kvm_sregs sregs;7411        struct kvm_vcpu_events events;7412  };7413 74146.75 KVM_CAP_PPC_IRQ_XIVE7415-------------------------7416 7417:Architectures: ppc7418:Target: vcpu7419:Parameters: args[0] is the XIVE device fd;7420             args[1] is the XIVE CPU number (server ID) for this vcpu7421 7422This capability connects the vcpu to an in-kernel XIVE device.7423 74247. Capabilities that can be enabled on VMs7425==========================================7426 7427There are certain capabilities that change the behavior of the virtual7428machine when enabled. To enable them, please see section 4.37. Below7429you can find a list of capabilities and what their effect on the VM7430is when enabling them.7431 7432The following information is provided along with the description:7433 7434  Architectures:7435      which instruction set architectures provide this ioctl.7436      x86 includes both i386 and x86_64.7437 7438  Parameters:7439      what parameters are accepted by the capability.7440 7441  Returns:7442      the return value.  General error numbers (EBADF, ENOMEM, EINVAL)7443      are not detailed, but errors with specific meanings are.7444 7445 74467.1 KVM_CAP_PPC_ENABLE_HCALL7447----------------------------7448 7449:Architectures: ppc7450:Parameters: args[0] is the sPAPR hcall number;7451	     args[1] is 0 to disable, 1 to enable in-kernel handling7452 7453This capability controls whether individual sPAPR hypercalls (hcalls)7454get handled by the kernel or not.  Enabling or disabling in-kernel7455handling of an hcall is effective across the VM.  On creation, an7456initial set of hcalls are enabled for in-kernel handling, which7457consists of those hcalls for which in-kernel handlers were implemented7458before this capability was implemented.  If disabled, the kernel will7459not to attempt to handle the hcall, but will always exit to userspace7460to handle it.  Note that it may not make sense to enable some and7461disable others of a group of related hcalls, but KVM does not prevent7462userspace from doing that.7463 7464If the hcall number specified is not one that has an in-kernel7465implementation, the KVM_ENABLE_CAP ioctl will fail with an EINVAL7466error.7467 74687.2 KVM_CAP_S390_USER_SIGP7469--------------------------7470 7471:Architectures: s3907472:Parameters: none7473 7474This capability controls which SIGP orders will be handled completely in user7475space. With this capability enabled, all fast orders will be handled completely7476in the kernel:7477 7478- SENSE7479- SENSE RUNNING7480- EXTERNAL CALL7481- EMERGENCY SIGNAL7482- CONDITIONAL EMERGENCY SIGNAL7483 7484All other orders will be handled completely in user space.7485 7486Only privileged operation exceptions will be checked for in the kernel (or even7487in the hardware prior to interception). If this capability is not enabled, the7488old way of handling SIGP orders is used (partially in kernel and user space).7489 74907.3 KVM_CAP_S390_VECTOR_REGISTERS7491---------------------------------7492 7493:Architectures: s3907494:Parameters: none7495:Returns: 0 on success, negative value on error7496 7497Allows use of the vector registers introduced with z13 processor, and7498provides for the synchronization between host and user space.  Will7499return -EINVAL if the machine does not support vectors.7500 75017.4 KVM_CAP_S390_USER_STSI7502--------------------------7503 7504:Architectures: s3907505:Parameters: none7506 7507This capability allows post-handlers for the STSI instruction. After7508initial handling in the kernel, KVM exits to user space with7509KVM_EXIT_S390_STSI to allow user space to insert further data.7510 7511Before exiting to userspace, kvm handlers should fill in s390_stsi field of7512vcpu->run::7513 7514  struct {7515	__u64 addr;7516	__u8 ar;7517	__u8 reserved;7518	__u8 fc;7519	__u8 sel1;7520	__u16 sel2;7521  } s390_stsi;7522 7523  @addr - guest address of STSI SYSIB7524  @fc   - function code7525  @sel1 - selector 17526  @sel2 - selector 27527  @ar   - access register number7528 7529KVM handlers should exit to userspace with rc = -EREMOTE.7530 75317.5 KVM_CAP_SPLIT_IRQCHIP7532-------------------------7533 7534:Architectures: x867535:Parameters: args[0] - number of routes reserved for userspace IOAPICs7536:Returns: 0 on success, -1 on error7537 7538Create a local apic for each processor in the kernel. This can be used7539instead of KVM_CREATE_IRQCHIP if the userspace VMM wishes to emulate the7540IOAPIC and PIC (and also the PIT, even though this has to be enabled7541separately).7542 7543This capability also enables in kernel routing of interrupt requests;7544when KVM_CAP_SPLIT_IRQCHIP only routes of KVM_IRQ_ROUTING_MSI type are7545used in the IRQ routing table.  The first args[0] MSI routes are reserved7546for the IOAPIC pins.  Whenever the LAPIC receives an EOI for these routes,7547a KVM_EXIT_IOAPIC_EOI vmexit will be reported to userspace.7548 7549Fails if VCPU has already been created, or if the irqchip is already in the7550kernel (i.e. KVM_CREATE_IRQCHIP has already been called).7551 75527.6 KVM_CAP_S390_RI7553-------------------7554 7555:Architectures: s3907556:Parameters: none7557 7558Allows use of runtime-instrumentation introduced with zEC12 processor.7559Will return -EINVAL if the machine does not support runtime-instrumentation.7560Will return -EBUSY if a VCPU has already been created.7561 75627.7 KVM_CAP_X2APIC_API7563----------------------7564 7565:Architectures: x867566:Parameters: args[0] - features that should be enabled7567:Returns: 0 on success, -EINVAL when args[0] contains invalid features7568 7569Valid feature flags in args[0] are::7570 7571  #define KVM_X2APIC_API_USE_32BIT_IDS            (1ULL << 0)7572  #define KVM_X2APIC_API_DISABLE_BROADCAST_QUIRK  (1ULL << 1)7573 7574Enabling KVM_X2APIC_API_USE_32BIT_IDS changes the behavior of7575KVM_SET_GSI_ROUTING, KVM_SIGNAL_MSI, KVM_SET_LAPIC, and KVM_GET_LAPIC,7576allowing the use of 32-bit APIC IDs.  See KVM_CAP_X2APIC_API in their7577respective sections.7578 7579KVM_X2APIC_API_DISABLE_BROADCAST_QUIRK must be enabled for x2APIC to work7580in logical mode or with more than 255 VCPUs.  Otherwise, KVM treats 0xff7581as a broadcast even in x2APIC mode in order to support physical x2APIC7582without interrupt remapping.  This is undesirable in logical mode,7583where 0xff represents CPUs 0-7 in cluster 0.7584 75857.8 KVM_CAP_S390_USER_INSTR07586----------------------------7587 7588:Architectures: s3907589:Parameters: none7590 7591With this capability enabled, all illegal instructions 0x0000 (2 bytes) will7592be intercepted and forwarded to user space. User space can use this7593mechanism e.g. to realize 2-byte software breakpoints. The kernel will7594not inject an operating exception for these instructions, user space has7595to take care of that.7596 7597This capability can be enabled dynamically even if VCPUs were already7598created and are running.7599 76007.9 KVM_CAP_S390_GS7601-------------------7602 7603:Architectures: s3907604:Parameters: none7605:Returns: 0 on success; -EINVAL if the machine does not support7606          guarded storage; -EBUSY if a VCPU has already been created.7607 7608Allows use of guarded storage for the KVM guest.7609 76107.10 KVM_CAP_S390_AIS7611---------------------7612 7613:Architectures: s3907614:Parameters: none7615 7616Allow use of adapter-interruption suppression.7617:Returns: 0 on success; -EBUSY if a VCPU has already been created.7618 76197.11 KVM_CAP_PPC_SMT7620--------------------7621 7622:Architectures: ppc7623:Parameters: vsmt_mode, flags7624 7625Enabling this capability on a VM provides userspace with a way to set7626the desired virtual SMT mode (i.e. the number of virtual CPUs per7627virtual core).  The virtual SMT mode, vsmt_mode, must be a power of 27628between 1 and 8.  On POWER8, vsmt_mode must also be no greater than7629the number of threads per subcore for the host.  Currently flags must7630be 0.  A successful call to enable this capability will result in7631vsmt_mode being returned when the KVM_CAP_PPC_SMT capability is7632subsequently queried for the VM.  This capability is only supported by7633HV KVM, and can only be set before any VCPUs have been created.7634The KVM_CAP_PPC_SMT_POSSIBLE capability indicates which virtual SMT7635modes are available.7636 76377.12 KVM_CAP_PPC_FWNMI7638----------------------7639 7640:Architectures: ppc7641:Parameters: none7642 7643With this capability a machine check exception in the guest address7644space will cause KVM to exit the guest with NMI exit reason. This7645enables QEMU to build error log and branch to guest kernel registered7646machine check handling routine. Without this capability KVM will7647branch to guests' 0x200 interrupt vector.7648 76497.13 KVM_CAP_X86_DISABLE_EXITS7650------------------------------7651 7652:Architectures: x867653:Parameters: args[0] defines which exits are disabled7654:Returns: 0 on success, -EINVAL when args[0] contains invalid exits7655 7656Valid bits in args[0] are::7657 7658  #define KVM_X86_DISABLE_EXITS_MWAIT            (1 << 0)7659  #define KVM_X86_DISABLE_EXITS_HLT              (1 << 1)7660  #define KVM_X86_DISABLE_EXITS_PAUSE            (1 << 2)7661  #define KVM_X86_DISABLE_EXITS_CSTATE           (1 << 3)7662 7663Enabling this capability on a VM provides userspace with a way to no7664longer intercept some instructions for improved latency in some7665workloads, and is suggested when vCPUs are associated to dedicated7666physical CPUs.  More bits can be added in the future; userspace can7667just pass the KVM_CHECK_EXTENSION result to KVM_ENABLE_CAP to disable7668all such vmexits.7669 7670Do not enable KVM_FEATURE_PV_UNHALT if you disable HLT exits.7671 76727.14 KVM_CAP_S390_HPAGE_1M7673--------------------------7674 7675:Architectures: s3907676:Parameters: none7677:Returns: 0 on success, -EINVAL if hpage module parameter was not set7678	  or cmma is enabled, or the VM has the KVM_VM_S390_UCONTROL7679	  flag set7680 7681With this capability the KVM support for memory backing with 1m pages7682through hugetlbfs can be enabled for a VM. After the capability is7683enabled, cmma can't be enabled anymore and pfmfi and the storage key7684interpretation are disabled. If cmma has already been enabled or the7685hpage module parameter is not set to 1, -EINVAL is returned.7686 7687While it is generally possible to create a huge page backed VM without7688this capability, the VM will not be able to run.7689 76907.15 KVM_CAP_MSR_PLATFORM_INFO7691------------------------------7692 7693:Architectures: x867694:Parameters: args[0] whether feature should be enabled or not7695 7696With this capability, a guest may read the MSR_PLATFORM_INFO MSR. Otherwise,7697a #GP would be raised when the guest tries to access. Currently, this7698capability does not enable write permissions of this MSR for the guest.7699 77007.16 KVM_CAP_PPC_NESTED_HV7701--------------------------7702 7703:Architectures: ppc7704:Parameters: none7705:Returns: 0 on success, -EINVAL when the implementation doesn't support7706	  nested-HV virtualization.7707 7708HV-KVM on POWER9 and later systems allows for "nested-HV"7709virtualization, which provides a way for a guest VM to run guests that7710can run using the CPU's supervisor mode (privileged non-hypervisor7711state).  Enabling this capability on a VM depends on the CPU having7712the necessary functionality and on the facility being enabled with a7713kvm-hv module parameter.7714 77157.17 KVM_CAP_EXCEPTION_PAYLOAD7716------------------------------7717 7718:Architectures: x867719:Parameters: args[0] whether feature should be enabled or not7720 7721With this capability enabled, CR2 will not be modified prior to the7722emulated VM-exit when L1 intercepts a #PF exception that occurs in7723L2. Similarly, for kvm-intel only, DR6 will not be modified prior to7724the emulated VM-exit when L1 intercepts a #DB exception that occurs in7725L2. As a result, when KVM_GET_VCPU_EVENTS reports a pending #PF (or7726#DB) exception for L2, exception.has_payload will be set and the7727faulting address (or the new DR6 bits*) will be reported in the7728exception_payload field. Similarly, when userspace injects a #PF (or7729#DB) into L2 using KVM_SET_VCPU_EVENTS, it is expected to set7730exception.has_payload and to put the faulting address - or the new DR67731bits\ [#]_ - in the exception_payload field.7732 7733This capability also enables exception.pending in struct7734kvm_vcpu_events, which allows userspace to distinguish between pending7735and injected exceptions.7736 7737 7738.. [#] For the new DR6 bits, note that bit 16 is set iff the #DB exception7739       will clear DR6.RTM.7740 77417.18 KVM_CAP_MANUAL_DIRTY_LOG_PROTECT27742--------------------------------------7743 7744:Architectures: x86, arm64, mips7745:Parameters: args[0] whether feature should be enabled or not7746 7747Valid flags are::7748 7749  #define KVM_DIRTY_LOG_MANUAL_PROTECT_ENABLE   (1 << 0)7750  #define KVM_DIRTY_LOG_INITIALLY_SET           (1 << 1)7751 7752With KVM_DIRTY_LOG_MANUAL_PROTECT_ENABLE is set, KVM_GET_DIRTY_LOG will not7753automatically clear and write-protect all pages that are returned as dirty.7754Rather, userspace will have to do this operation separately using7755KVM_CLEAR_DIRTY_LOG.7756 7757At the cost of a slightly more complicated operation, this provides better7758scalability and responsiveness for two reasons.  First,7759KVM_CLEAR_DIRTY_LOG ioctl can operate on a 64-page granularity rather7760than requiring to sync a full memslot; this ensures that KVM does not7761take spinlocks for an extended period of time.  Second, in some cases a7762large amount of time can pass between a call to KVM_GET_DIRTY_LOG and7763userspace actually using the data in the page.  Pages can be modified7764during this time, which is inefficient for both the guest and userspace:7765the guest will incur a higher penalty due to write protection faults,7766while userspace can see false reports of dirty pages.  Manual reprotection7767helps reducing this time, improving guest performance and reducing the7768number of dirty log false positives.7769 7770With KVM_DIRTY_LOG_INITIALLY_SET set, all the bits of the dirty bitmap7771will be initialized to 1 when created.  This also improves performance because7772dirty logging can be enabled gradually in small chunks on the first call7773to KVM_CLEAR_DIRTY_LOG.  KVM_DIRTY_LOG_INITIALLY_SET depends on7774KVM_DIRTY_LOG_MANUAL_PROTECT_ENABLE (it is also only available on7775x86 and arm64 for now).7776 7777KVM_CAP_MANUAL_DIRTY_LOG_PROTECT2 was previously available under the name7778KVM_CAP_MANUAL_DIRTY_LOG_PROTECT, but the implementation had bugs that make7779it hard or impossible to use it correctly.  The availability of7780KVM_CAP_MANUAL_DIRTY_LOG_PROTECT2 signals that those bugs are fixed.7781Userspace should not try to use KVM_CAP_MANUAL_DIRTY_LOG_PROTECT.7782 77837.19 KVM_CAP_PPC_SECURE_GUEST7784------------------------------7785 7786:Architectures: ppc7787 7788This capability indicates that KVM is running on a host that has7789ultravisor firmware and thus can support a secure guest.  On such a7790system, a guest can ask the ultravisor to make it a secure guest,7791one whose memory is inaccessible to the host except for pages which7792are explicitly requested to be shared with the host.  The ultravisor7793notifies KVM when a guest requests to become a secure guest, and KVM7794has the opportunity to veto the transition.7795 7796If present, this capability can be enabled for a VM, meaning that KVM7797will allow the transition to secure guest mode.  Otherwise KVM will7798veto the transition.7799 78007.20 KVM_CAP_HALT_POLL7801----------------------7802 7803:Architectures: all7804:Target: VM7805:Parameters: args[0] is the maximum poll time in nanoseconds7806:Returns: 0 on success; -1 on error7807 7808KVM_CAP_HALT_POLL overrides the kvm.halt_poll_ns module parameter to set the7809maximum halt-polling time for all vCPUs in the target VM. This capability can7810be invoked at any time and any number of times to dynamically change the7811maximum halt-polling time.7812 7813See Documentation/virt/kvm/halt-polling.rst for more information on halt7814polling.7815 78167.21 KVM_CAP_X86_USER_SPACE_MSR7817-------------------------------7818 7819:Architectures: x867820:Target: VM7821:Parameters: args[0] contains the mask of KVM_MSR_EXIT_REASON_* events to report7822:Returns: 0 on success; -1 on error7823 7824This capability allows userspace to intercept RDMSR and WRMSR instructions if7825access to an MSR is denied.  By default, KVM injects #GP on denied accesses.7826 7827When a guest requests to read or write an MSR, KVM may not implement all MSRs7828that are relevant to a respective system. It also does not differentiate by7829CPU type.7830 7831To allow more fine grained control over MSR handling, userspace may enable7832this capability. With it enabled, MSR accesses that match the mask specified in7833args[0] and would trigger a #GP inside the guest will instead trigger7834KVM_EXIT_X86_RDMSR and KVM_EXIT_X86_WRMSR exit notifications.  Userspace7835can then implement model specific MSR handling and/or user notifications7836to inform a user that an MSR was not emulated/virtualized by KVM.7837 7838The valid mask flags are:7839 7840============================ ===============================================7841 KVM_MSR_EXIT_REASON_UNKNOWN intercept accesses to unknown (to KVM) MSRs7842 KVM_MSR_EXIT_REASON_INVAL   intercept accesses that are architecturally7843                             invalid according to the vCPU model and/or mode7844 KVM_MSR_EXIT_REASON_FILTER  intercept accesses that are denied by userspace7845                             via KVM_X86_SET_MSR_FILTER7846============================ ===============================================7847 78487.22 KVM_CAP_X86_BUS_LOCK_EXIT7849-------------------------------7850 7851:Architectures: x867852:Target: VM7853:Parameters: args[0] defines the policy used when bus locks detected in guest7854:Returns: 0 on success, -EINVAL when args[0] contains invalid bits7855 7856Valid bits in args[0] are::7857 7858  #define KVM_BUS_LOCK_DETECTION_OFF      (1 << 0)7859  #define KVM_BUS_LOCK_DETECTION_EXIT     (1 << 1)7860 7861Enabling this capability on a VM provides userspace with a way to select a7862policy to handle the bus locks detected in guest. Userspace can obtain the7863supported modes from the result of KVM_CHECK_EXTENSION and define it through7864the KVM_ENABLE_CAP. The supported modes are mutually-exclusive.7865 7866This capability allows userspace to force VM exits on bus locks detected in the7867guest, irrespective whether or not the host has enabled split-lock detection7868(which triggers an #AC exception that KVM intercepts). This capability is7869intended to mitigate attacks where a malicious/buggy guest can exploit bus7870locks to degrade the performance of the whole system.7871 7872If KVM_BUS_LOCK_DETECTION_OFF is set, KVM doesn't force guest bus locks to VM7873exit, although the host kernel's split-lock #AC detection still applies, if7874enabled.7875 7876If KVM_BUS_LOCK_DETECTION_EXIT is set, KVM enables a CPU feature that ensures7877bus locks in the guest trigger a VM exit, and KVM exits to userspace for all7878such VM exits, e.g. to allow userspace to throttle the offending guest and/or7879apply some other policy-based mitigation. When exiting to userspace, KVM sets7880KVM_RUN_X86_BUS_LOCK in vcpu-run->flags, and conditionally sets the exit_reason7881to KVM_EXIT_X86_BUS_LOCK.7882 7883Note! Detected bus locks may be coincident with other exits to userspace, i.e.7884KVM_RUN_X86_BUS_LOCK should be checked regardless of the primary exit reason if7885userspace wants to take action on all detected bus locks.7886 78877.23 KVM_CAP_PPC_DAWR17888----------------------7889 7890:Architectures: ppc7891:Parameters: none7892:Returns: 0 on success, -EINVAL when CPU doesn't support 2nd DAWR7893 7894This capability can be used to check / enable 2nd DAWR feature provided7895by POWER10 processor.7896 7897 78987.24 KVM_CAP_VM_COPY_ENC_CONTEXT_FROM7899-------------------------------------7900 7901Architectures: x86 SEV enabled7902Type: vm7903Parameters: args[0] is the fd of the source vm7904Returns: 0 on success; ENOTTY on error7905 7906This capability enables userspace to copy encryption context from the vm7907indicated by the fd to the vm this is called on.7908 7909This is intended to support in-guest workloads scheduled by the host. This7910allows the in-guest workload to maintain its own NPTs and keeps the two vms7911from accidentally clobbering each other with interrupts and the like (separate7912APIC/MSRs/etc).7913 79147.25 KVM_CAP_SGX_ATTRIBUTE7915--------------------------7916 7917:Architectures: x867918:Target: VM7919:Parameters: args[0] is a file handle of a SGX attribute file in securityfs7920:Returns: 0 on success, -EINVAL if the file handle is invalid or if a requested7921          attribute is not supported by KVM.7922 7923KVM_CAP_SGX_ATTRIBUTE enables a userspace VMM to grant a VM access to one or7924more privileged enclave attributes.  args[0] must hold a file handle to a valid7925SGX attribute file corresponding to an attribute that is supported/restricted7926by KVM (currently only PROVISIONKEY).7927 7928The SGX subsystem restricts access to a subset of enclave attributes to provide7929additional security for an uncompromised kernel, e.g. use of the PROVISIONKEY7930is restricted to deter malware from using the PROVISIONKEY to obtain a stable7931system fingerprint.  To prevent userspace from circumventing such restrictions7932by running an enclave in a VM, KVM prevents access to privileged attributes by7933default.7934 7935See Documentation/arch/x86/sgx.rst for more details.7936 79377.26 KVM_CAP_PPC_RPT_INVALIDATE7938-------------------------------7939 7940:Capability: KVM_CAP_PPC_RPT_INVALIDATE7941:Architectures: ppc7942:Type: vm7943 7944This capability indicates that the kernel is capable of handling7945H_RPT_INVALIDATE hcall.7946 7947In order to enable the use of H_RPT_INVALIDATE in the guest,7948user space might have to advertise it for the guest. For example,7949IBM pSeries (sPAPR) guest starts using it if "hcall-rpt-invalidate" is7950present in the "ibm,hypertas-functions" device-tree property.7951 7952This capability is enabled for hypervisors on platforms like POWER97953that support radix MMU.7954 79557.27 KVM_CAP_EXIT_ON_EMULATION_FAILURE7956--------------------------------------7957 7958:Architectures: x867959:Parameters: args[0] whether the feature should be enabled or not7960 7961When this capability is enabled, an emulation failure will result in an exit7962to userspace with KVM_INTERNAL_ERROR (except when the emulator was invoked7963to handle a VMware backdoor instruction). Furthermore, KVM will now provide up7964to 15 instruction bytes for any exit to userspace resulting from an emulation7965failure.  When these exits to userspace occur use the emulation_failure struct7966instead of the internal struct.  They both have the same layout, but the7967emulation_failure struct matches the content better.  It also explicitly7968defines the 'flags' field which is used to describe the fields in the struct7969that are valid (ie: if KVM_INTERNAL_ERROR_EMULATION_FLAG_INSTRUCTION_BYTES is7970set in the 'flags' field then both 'insn_size' and 'insn_bytes' have valid data7971in them.)7972 79737.28 KVM_CAP_ARM_MTE7974--------------------7975 7976:Architectures: arm647977:Parameters: none7978 7979This capability indicates that KVM (and the hardware) supports exposing the7980Memory Tagging Extensions (MTE) to the guest. It must also be enabled by the7981VMM before creating any VCPUs to allow the guest access. Note that MTE is only7982available to a guest running in AArch64 mode and enabling this capability will7983cause attempts to create AArch32 VCPUs to fail.7984 7985When enabled the guest is able to access tags associated with any memory given7986to the guest. KVM will ensure that the tags are maintained during swap or7987hibernation of the host; however the VMM needs to manually save/restore the7988tags as appropriate if the VM is migrated.7989 7990When this capability is enabled all memory in memslots must be mapped as7991``MAP_ANONYMOUS`` or with a RAM-based file mapping (``tmpfs``, ``memfd``),7992attempts to create a memslot with an invalid mmap will result in an7993-EINVAL return.7994 7995When enabled the VMM may make use of the ``KVM_ARM_MTE_COPY_TAGS`` ioctl to7996perform a bulk copy of tags to/from the guest.7997 79987.29 KVM_CAP_VM_MOVE_ENC_CONTEXT_FROM7999-------------------------------------8000 8001:Architectures: x86 SEV enabled8002:Type: vm8003:Parameters: args[0] is the fd of the source vm8004:Returns: 0 on success8005 8006This capability enables userspace to migrate the encryption context from the VM8007indicated by the fd to the VM this is called on.8008 8009This is intended to support intra-host migration of VMs between userspace VMMs,8010upgrading the VMM process without interrupting the guest.8011 80127.30 KVM_CAP_PPC_AIL_MODE_38013-------------------------------8014 8015:Capability: KVM_CAP_PPC_AIL_MODE_38016:Architectures: ppc8017:Type: vm8018 8019This capability indicates that the kernel supports the mode 3 setting for the8020"Address Translation Mode on Interrupt" aka "Alternate Interrupt Location"8021resource that is controlled with the H_SET_MODE hypercall.8022 8023This capability allows a guest kernel to use a better-performance mode for8024handling interrupts and system calls.8025 80267.31 KVM_CAP_DISABLE_QUIRKS28027----------------------------8028 8029:Capability: KVM_CAP_DISABLE_QUIRKS28030:Parameters: args[0] - set of KVM quirks to disable8031:Architectures: x868032:Type: vm8033 8034This capability, if enabled, will cause KVM to disable some behavior8035quirks.8036 8037Calling KVM_CHECK_EXTENSION for this capability returns a bitmask of8038quirks that can be disabled in KVM.8039 8040The argument to KVM_ENABLE_CAP for this capability is a bitmask of8041quirks to disable, and must be a subset of the bitmask returned by8042KVM_CHECK_EXTENSION.8043 8044The valid bits in cap.args[0] are:8045 8046=================================== ============================================8047 KVM_X86_QUIRK_LINT0_REENABLED      By default, the reset value for the LVT8048                                    LINT0 register is 0x700 (APIC_MODE_EXTINT).8049                                    When this quirk is disabled, the reset value8050                                    is 0x10000 (APIC_LVT_MASKED).8051 8052 KVM_X86_QUIRK_CD_NW_CLEARED        By default, KVM clears CR0.CD and CR0.NW on8053                                    AMD CPUs to workaround buggy guest firmware8054                                    that runs in perpetuity with CR0.CD, i.e.8055                                    with caches in "no fill" mode.8056 8057                                    When this quirk is disabled, KVM does not8058                                    change the value of CR0.CD and CR0.NW.8059 8060 KVM_X86_QUIRK_LAPIC_MMIO_HOLE      By default, the MMIO LAPIC interface is8061                                    available even when configured for x2APIC8062                                    mode. When this quirk is disabled, KVM8063                                    disables the MMIO LAPIC interface if the8064                                    LAPIC is in x2APIC mode.8065 8066 KVM_X86_QUIRK_OUT_7E_INC_RIP       By default, KVM pre-increments %rip before8067                                    exiting to userspace for an OUT instruction8068                                    to port 0x7e. When this quirk is disabled,8069                                    KVM does not pre-increment %rip before8070                                    exiting to userspace.8071 8072 KVM_X86_QUIRK_MISC_ENABLE_NO_MWAIT When this quirk is disabled, KVM sets8073                                    CPUID.01H:ECX[bit 3] (MONITOR/MWAIT) if8074                                    IA32_MISC_ENABLE[bit 18] (MWAIT) is set.8075                                    Additionally, when this quirk is disabled,8076                                    KVM clears CPUID.01H:ECX[bit 3] if8077                                    IA32_MISC_ENABLE[bit 18] is cleared.8078 8079 KVM_X86_QUIRK_FIX_HYPERCALL_INSN   By default, KVM rewrites guest8080                                    VMMCALL/VMCALL instructions to match the8081                                    vendor's hypercall instruction for the8082                                    system. When this quirk is disabled, KVM8083                                    will no longer rewrite invalid guest8084                                    hypercall instructions. Executing the8085                                    incorrect hypercall instruction will8086                                    generate a #UD within the guest.8087 8088KVM_X86_QUIRK_MWAIT_NEVER_UD_FAULTS By default, KVM emulates MONITOR/MWAIT (if8089                                    they are intercepted) as NOPs regardless of8090                                    whether or not MONITOR/MWAIT are supported8091                                    according to guest CPUID.  When this quirk8092                                    is disabled and KVM_X86_DISABLE_EXITS_MWAIT8093                                    is not set (MONITOR/MWAIT are intercepted),8094                                    KVM will inject a #UD on MONITOR/MWAIT if8095                                    they're unsupported per guest CPUID.  Note,8096                                    KVM will modify MONITOR/MWAIT support in8097                                    guest CPUID on writes to MISC_ENABLE if8098                                    KVM_X86_QUIRK_MISC_ENABLE_NO_MWAIT is8099                                    disabled.8100 8101KVM_X86_QUIRK_SLOT_ZAP_ALL          By default, for KVM_X86_DEFAULT_VM VMs, KVM8102                                    invalidates all SPTEs in all memslots and8103                                    address spaces when a memslot is deleted or8104                                    moved.  When this quirk is disabled (or the8105                                    VM type isn't KVM_X86_DEFAULT_VM), KVM only8106                                    ensures the backing memory of the deleted8107                                    or moved memslot isn't reachable, i.e KVM8108                                    _may_ invalidate only SPTEs related to the8109                                    memslot.8110=================================== ============================================8111 81127.32 KVM_CAP_MAX_VCPU_ID8113------------------------8114 8115:Architectures: x868116:Target: VM8117:Parameters: args[0] - maximum APIC ID value set for current VM8118:Returns: 0 on success, -EINVAL if args[0] is beyond KVM_MAX_VCPU_IDS8119          supported in KVM or if it has been set.8120 8121This capability allows userspace to specify maximum possible APIC ID8122assigned for current VM session prior to the creation of vCPUs, saving8123memory for data structures indexed by the APIC ID.  Userspace is able8124to calculate the limit to APIC ID values from designated8125CPU topology.8126 8127The value can be changed only until KVM_ENABLE_CAP is set to a nonzero8128value or until a vCPU is created.  Upon creation of the first vCPU,8129if the value was set to zero or KVM_ENABLE_CAP was not invoked, KVM8130uses the return value of KVM_CHECK_EXTENSION(KVM_CAP_MAX_VCPU_ID) as8131the maximum APIC ID.8132 81337.33 KVM_CAP_X86_NOTIFY_VMEXIT8134------------------------------8135 8136:Architectures: x868137:Target: VM8138:Parameters: args[0] is the value of notify window as well as some flags8139:Returns: 0 on success, -EINVAL if args[0] contains invalid flags or notify8140          VM exit is unsupported.8141 8142Bits 63:32 of args[0] are used for notify window.8143Bits 31:0 of args[0] are for some flags. Valid bits are::8144 8145  #define KVM_X86_NOTIFY_VMEXIT_ENABLED    (1 << 0)8146  #define KVM_X86_NOTIFY_VMEXIT_USER       (1 << 1)8147 8148This capability allows userspace to configure the notify VM exit on/off8149in per-VM scope during VM creation. Notify VM exit is disabled by default.8150When userspace sets KVM_X86_NOTIFY_VMEXIT_ENABLED bit in args[0], VMM will8151enable this feature with the notify window provided, which will generate8152a VM exit if no event window occurs in VM non-root mode for a specified of8153time (notify window).8154 8155If KVM_X86_NOTIFY_VMEXIT_USER is set in args[0], upon notify VM exits happen,8156KVM would exit to userspace for handling.8157 8158This capability is aimed to mitigate the threat that malicious VMs can8159cause CPU stuck (due to event windows don't open up) and make the CPU8160unavailable to host or other VMs.8161 81627.34 KVM_CAP_MEMORY_FAULT_INFO8163------------------------------8164 8165:Architectures: x868166:Returns: Informational only, -EINVAL on direct KVM_ENABLE_CAP.8167 8168The presence of this capability indicates that KVM_RUN will fill8169kvm_run.memory_fault if KVM cannot resolve a guest page fault VM-Exit, e.g. if8170there is a valid memslot but no backing VMA for the corresponding host virtual8171address.8172 8173The information in kvm_run.memory_fault is valid if and only if KVM_RUN returns8174an error with errno=EFAULT or errno=EHWPOISON *and* kvm_run.exit_reason is set8175to KVM_EXIT_MEMORY_FAULT.8176 8177Note: Userspaces which attempt to resolve memory faults so that they can retry8178KVM_RUN are encouraged to guard against repeatedly receiving the same8179error/annotated fault.8180 8181See KVM_EXIT_MEMORY_FAULT for more information.8182 81837.35 KVM_CAP_X86_APIC_BUS_CYCLES_NS8184-----------------------------------8185 8186:Architectures: x868187:Target: VM8188:Parameters: args[0] is the desired APIC bus clock rate, in nanoseconds8189:Returns: 0 on success, -EINVAL if args[0] contains an invalid value for the8190          frequency or if any vCPUs have been created, -ENXIO if a virtual8191          local APIC has not been created using KVM_CREATE_IRQCHIP.8192 8193This capability sets the VM's APIC bus clock frequency, used by KVM's in-kernel8194virtual APIC when emulating APIC timers.  KVM's default value can be retrieved8195by KVM_CHECK_EXTENSION.8196 8197Note: Userspace is responsible for correctly configuring CPUID 0x15, a.k.a. the8198core crystal clock frequency, if a non-zero CPUID 0x15 is exposed to the guest.8199 82007.36 KVM_CAP_X86_GUEST_MODE8201------------------------------8202 8203:Architectures: x868204:Returns: Informational only, -EINVAL on direct KVM_ENABLE_CAP.8205 8206The presence of this capability indicates that KVM_RUN will update the8207KVM_RUN_X86_GUEST_MODE bit in kvm_run.flags to indicate whether the8208vCPU was executing nested guest code when it exited.8209 8210KVM exits with the register state of either the L1 or L2 guest8211depending on which executed at the time of an exit. Userspace must8212take care to differentiate between these cases.8213 82148. Other capabilities.8215======================8216 8217This section lists capabilities that give information about other8218features of the KVM implementation.8219 82208.1 KVM_CAP_PPC_HWRNG8221---------------------8222 8223:Architectures: ppc8224 8225This capability, if KVM_CHECK_EXTENSION indicates that it is8226available, means that the kernel has an implementation of the8227H_RANDOM hypercall backed by a hardware random-number generator.8228If present, the kernel H_RANDOM handler can be enabled for guest use8229with the KVM_CAP_PPC_ENABLE_HCALL capability.8230 82318.2 KVM_CAP_HYPERV_SYNIC8232------------------------8233 8234:Architectures: x868235 8236This capability, if KVM_CHECK_EXTENSION indicates that it is8237available, means that the kernel has an implementation of the8238Hyper-V Synthetic interrupt controller(SynIC). Hyper-V SynIC is8239used to support Windows Hyper-V based guest paravirt drivers(VMBus).8240 8241In order to use SynIC, it has to be activated by setting this8242capability via KVM_ENABLE_CAP ioctl on the vcpu fd. Note that this8243will disable the use of APIC hardware virtualization even if supported8244by the CPU, as it's incompatible with SynIC auto-EOI behavior.8245 82468.3 KVM_CAP_PPC_MMU_RADIX8247-------------------------8248 8249:Architectures: ppc8250 8251This capability, if KVM_CHECK_EXTENSION indicates that it is8252available, means that the kernel can support guests using the8253radix MMU defined in Power ISA V3.00 (as implemented in the POWER98254processor).8255 82568.4 KVM_CAP_PPC_MMU_HASH_V38257---------------------------8258 8259:Architectures: ppc8260 8261This capability, if KVM_CHECK_EXTENSION indicates that it is8262available, means that the kernel can support guests using the8263hashed page table MMU defined in Power ISA V3.00 (as implemented in8264the POWER9 processor), including in-memory segment tables.8265 82668.5 KVM_CAP_MIPS_VZ8267-------------------8268 8269:Architectures: mips8270 8271This capability, if KVM_CHECK_EXTENSION on the main kvm handle indicates that8272it is available, means that full hardware assisted virtualization capabilities8273of the hardware are available for use through KVM. An appropriate8274KVM_VM_MIPS_* type must be passed to KVM_CREATE_VM to create a VM which8275utilises it.8276 8277If KVM_CHECK_EXTENSION on a kvm VM handle indicates that this capability is8278available, it means that the VM is using full hardware assisted virtualization8279capabilities of the hardware. This is useful to check after creating a VM with8280KVM_VM_MIPS_DEFAULT.8281 8282The value returned by KVM_CHECK_EXTENSION should be compared against known8283values (see below). All other values are reserved. This is to allow for the8284possibility of other hardware assisted virtualization implementations which8285may be incompatible with the MIPS VZ ASE.8286 8287==  ==========================================================================8288 0  The trap & emulate implementation is in use to run guest code in user8289    mode. Guest virtual memory segments are rearranged to fit the guest in the8290    user mode address space.8291 8292 1  The MIPS VZ ASE is in use, providing full hardware assisted8293    virtualization, including standard guest virtual memory segments.8294==  ==========================================================================8295 82968.6 KVM_CAP_MIPS_TE8297-------------------8298 8299:Architectures: mips8300 8301This capability, if KVM_CHECK_EXTENSION on the main kvm handle indicates that8302it is available, means that the trap & emulate implementation is available to8303run guest code in user mode, even if KVM_CAP_MIPS_VZ indicates that hardware8304assisted virtualisation is also available. KVM_VM_MIPS_TE (0) must be passed8305to KVM_CREATE_VM to create a VM which utilises it.8306 8307If KVM_CHECK_EXTENSION on a kvm VM handle indicates that this capability is8308available, it means that the VM is using trap & emulate.8309 83108.7 KVM_CAP_MIPS_64BIT8311----------------------8312 8313:Architectures: mips8314 8315This capability indicates the supported architecture type of the guest, i.e. the8316supported register and address width.8317 8318The values returned when this capability is checked by KVM_CHECK_EXTENSION on a8319kvm VM handle correspond roughly to the CP0_Config.AT register field, and should8320be checked specifically against known values (see below). All other values are8321reserved.8322 8323==  ========================================================================8324 0  MIPS32 or microMIPS32.8325    Both registers and addresses are 32-bits wide.8326    It will only be possible to run 32-bit guest code.8327 8328 1  MIPS64 or microMIPS64 with access only to 32-bit compatibility segments.8329    Registers are 64-bits wide, but addresses are 32-bits wide.8330    64-bit guest code may run but cannot access MIPS64 memory segments.8331    It will also be possible to run 32-bit guest code.8332 8333 2  MIPS64 or microMIPS64 with access to all address segments.8334    Both registers and addresses are 64-bits wide.8335    It will be possible to run 64-bit or 32-bit guest code.8336==  ========================================================================8337 83388.9 KVM_CAP_ARM_USER_IRQ8339------------------------8340 8341:Architectures: arm648342 8343This capability, if KVM_CHECK_EXTENSION indicates that it is available, means8344that if userspace creates a VM without an in-kernel interrupt controller, it8345will be notified of changes to the output level of in-kernel emulated devices,8346which can generate virtual interrupts, presented to the VM.8347For such VMs, on every return to userspace, the kernel8348updates the vcpu's run->s.regs.device_irq_level field to represent the actual8349output level of the device.8350 8351Whenever kvm detects a change in the device output level, kvm guarantees at8352least one return to userspace before running the VM.  This exit could either8353be a KVM_EXIT_INTR or any other exit event, like KVM_EXIT_MMIO. This way,8354userspace can always sample the device output level and re-compute the state of8355the userspace interrupt controller.  Userspace should always check the state8356of run->s.regs.device_irq_level on every kvm exit.8357The value in run->s.regs.device_irq_level can represent both level and edge8358triggered interrupt signals, depending on the device.  Edge triggered interrupt8359signals will exit to userspace with the bit in run->s.regs.device_irq_level8360set exactly once per edge signal.8361 8362The field run->s.regs.device_irq_level is available independent of8363run->kvm_valid_regs or run->kvm_dirty_regs bits.8364 8365If KVM_CAP_ARM_USER_IRQ is supported, the KVM_CHECK_EXTENSION ioctl returns a8366number larger than 0 indicating the version of this capability is implemented8367and thereby which bits in run->s.regs.device_irq_level can signal values.8368 8369Currently the following bits are defined for the device_irq_level bitmap::8370 8371  KVM_CAP_ARM_USER_IRQ >= 1:8372 8373    KVM_ARM_DEV_EL1_VTIMER -  EL1 virtual timer8374    KVM_ARM_DEV_EL1_PTIMER -  EL1 physical timer8375    KVM_ARM_DEV_PMU        -  ARM PMU overflow interrupt signal8376 8377Future versions of kvm may implement additional events. These will get8378indicated by returning a higher number from KVM_CHECK_EXTENSION and will be8379listed above.8380 83818.10 KVM_CAP_PPC_SMT_POSSIBLE8382-----------------------------8383 8384:Architectures: ppc8385 8386Querying this capability returns a bitmap indicating the possible8387virtual SMT modes that can be set using KVM_CAP_PPC_SMT.  If bit N8388(counting from the right) is set, then a virtual SMT mode of 2^N is8389available.8390 83918.11 KVM_CAP_HYPERV_SYNIC28392--------------------------8393 8394:Architectures: x868395 8396This capability enables a newer version of Hyper-V Synthetic interrupt8397controller (SynIC).  The only difference with KVM_CAP_HYPERV_SYNIC is that KVM8398doesn't clear SynIC message and event flags pages when they are enabled by8399writing to the respective MSRs.8400 84018.12 KVM_CAP_HYPERV_VP_INDEX8402----------------------------8403 8404:Architectures: x868405 8406This capability indicates that userspace can load HV_X64_MSR_VP_INDEX msr.  Its8407value is used to denote the target vcpu for a SynIC interrupt.  For8408compatibility, KVM initializes this msr to KVM's internal vcpu index.  When this8409capability is absent, userspace can still query this msr's value.8410 84118.13 KVM_CAP_S390_AIS_MIGRATION8412-------------------------------8413 8414:Architectures: s3908415:Parameters: none8416 8417This capability indicates if the flic device will be able to get/set the8418AIS states for migration via the KVM_DEV_FLIC_AISM_ALL attribute and allows8419to discover this without having to create a flic device.8420 84218.14 KVM_CAP_S390_PSW8422---------------------8423 8424:Architectures: s3908425 8426This capability indicates that the PSW is exposed via the kvm_run structure.8427 84288.15 KVM_CAP_S390_GMAP8429----------------------8430 8431:Architectures: s3908432 8433This capability indicates that the user space memory used as guest mapping can8434be anywhere in the user memory address space, as long as the memory slots are8435aligned and sized to a segment (1MB) boundary.8436 84378.16 KVM_CAP_S390_COW8438---------------------8439 8440:Architectures: s3908441 8442This capability indicates that the user space memory used as guest mapping can8443use copy-on-write semantics as well as dirty pages tracking via read-only page8444tables.8445 84468.17 KVM_CAP_S390_BPB8447---------------------8448 8449:Architectures: s3908450 8451This capability indicates that kvm will implement the interfaces to handle8452reset, migration and nested KVM for branch prediction blocking. The stfle8453facility 82 should not be provided to the guest without this capability.8454 84558.18 KVM_CAP_HYPERV_TLBFLUSH8456----------------------------8457 8458:Architectures: x868459 8460This capability indicates that KVM supports paravirtualized Hyper-V TLB Flush8461hypercalls:8462HvFlushVirtualAddressSpace, HvFlushVirtualAddressSpaceEx,8463HvFlushVirtualAddressList, HvFlushVirtualAddressListEx.8464 84658.19 KVM_CAP_ARM_INJECT_SERROR_ESR8466----------------------------------8467 8468:Architectures: arm648469 8470This capability indicates that userspace can specify (via the8471KVM_SET_VCPU_EVENTS ioctl) the syndrome value reported to the guest when it8472takes a virtual SError interrupt exception.8473If KVM advertises this capability, userspace can only specify the ISS field for8474the ESR syndrome. Other parts of the ESR, such as the EC are generated by the8475CPU when the exception is taken. If this virtual SError is taken to EL1 using8476AArch64, this value will be reported in the ISS field of ESR_ELx.8477 8478See KVM_CAP_VCPU_EVENTS for more details.8479 84808.20 KVM_CAP_HYPERV_SEND_IPI8481----------------------------8482 8483:Architectures: x868484 8485This capability indicates that KVM supports paravirtualized Hyper-V IPI send8486hypercalls:8487HvCallSendSyntheticClusterIpi, HvCallSendSyntheticClusterIpiEx.8488 84898.21 KVM_CAP_HYPERV_DIRECT_TLBFLUSH8490-----------------------------------8491 8492:Architectures: x868493 8494This capability indicates that KVM running on top of Hyper-V hypervisor8495enables Direct TLB flush for its guests meaning that TLB flush8496hypercalls are handled by Level 0 hypervisor (Hyper-V) bypassing KVM.8497Due to the different ABI for hypercall parameters between Hyper-V and8498KVM, enabling this capability effectively disables all hypercall8499handling by KVM (as some KVM hypercall may be mistakenly treated as TLB8500flush hypercalls by Hyper-V) so userspace should disable KVM identification8501in CPUID and only exposes Hyper-V identification. In this case, guest8502thinks it's running on Hyper-V and only use Hyper-V hypercalls.8503 85048.22 KVM_CAP_S390_VCPU_RESETS8505-----------------------------8506 8507:Architectures: s3908508 8509This capability indicates that the KVM_S390_NORMAL_RESET and8510KVM_S390_CLEAR_RESET ioctls are available.8511 85128.23 KVM_CAP_S390_PROTECTED8513---------------------------8514 8515:Architectures: s3908516 8517This capability indicates that the Ultravisor has been initialized and8518KVM can therefore start protected VMs.8519This capability governs the KVM_S390_PV_COMMAND ioctl and the8520KVM_MP_STATE_LOAD MP_STATE. KVM_SET_MP_STATE can fail for protected8521guests when the state change is invalid.8522 85238.24 KVM_CAP_STEAL_TIME8524-----------------------8525 8526:Architectures: arm64, x868527 8528This capability indicates that KVM supports steal time accounting.8529When steal time accounting is supported it may be enabled with8530architecture-specific interfaces.  This capability and the architecture-8531specific interfaces must be consistent, i.e. if one says the feature8532is supported, than the other should as well and vice versa.  For arm648533see Documentation/virt/kvm/devices/vcpu.rst "KVM_ARM_VCPU_PVTIME_CTRL".8534For x86 see Documentation/virt/kvm/x86/msr.rst "MSR_KVM_STEAL_TIME".8535 85368.25 KVM_CAP_S390_DIAG3188537-------------------------8538 8539:Architectures: s3908540 8541This capability enables a guest to set information about its control program8542(i.e. guest kernel type and version). The information is helpful during8543system/firmware service events, providing additional data about the guest8544environments running on the machine.8545 8546The information is associated with the DIAGNOSE 0x318 instruction, which sets8547an 8-byte value consisting of a one-byte Control Program Name Code (CPNC) and8548a 7-byte Control Program Version Code (CPVC). The CPNC determines what8549environment the control program is running in (e.g. Linux, z/VM...), and the8550CPVC is used for information specific to OS (e.g. Linux version, Linux8551distribution...)8552 8553If this capability is available, then the CPNC and CPVC can be synchronized8554between KVM and userspace via the sync regs mechanism (KVM_SYNC_DIAG318).8555 85568.26 KVM_CAP_X86_USER_SPACE_MSR8557-------------------------------8558 8559:Architectures: x868560 8561This capability indicates that KVM supports deflection of MSR reads and8562writes to user space. It can be enabled on a VM level. If enabled, MSR8563accesses that would usually trigger a #GP by KVM into the guest will8564instead get bounced to user space through the KVM_EXIT_X86_RDMSR and8565KVM_EXIT_X86_WRMSR exit notifications.8566 85678.27 KVM_CAP_X86_MSR_FILTER8568---------------------------8569 8570:Architectures: x868571 8572This capability indicates that KVM supports that accesses to user defined MSRs8573may be rejected. With this capability exposed, KVM exports new VM ioctl8574KVM_X86_SET_MSR_FILTER which user space can call to specify bitmaps of MSR8575ranges that KVM should deny access to.8576 8577In combination with KVM_CAP_X86_USER_SPACE_MSR, this allows user space to8578trap and emulate MSRs that are outside of the scope of KVM as well as8579limit the attack surface on KVM's MSR emulation code.8580 85818.28 KVM_CAP_ENFORCE_PV_FEATURE_CPUID8582-------------------------------------8583 8584Architectures: x868585 8586When enabled, KVM will disable paravirtual features provided to the8587guest according to the bits in the KVM_CPUID_FEATURES CPUID leaf8588(0x40000001). Otherwise, a guest may use the paravirtual features8589regardless of what has actually been exposed through the CPUID leaf.8590 85918.29 KVM_CAP_DIRTY_LOG_RING/KVM_CAP_DIRTY_LOG_RING_ACQ_REL8592----------------------------------------------------------8593 8594:Architectures: x86, arm648595:Parameters: args[0] - size of the dirty log ring8596 8597KVM is capable of tracking dirty memory using ring buffers that are8598mmapped into userspace; there is one dirty ring per vcpu.8599 8600The dirty ring is available to userspace as an array of8601``struct kvm_dirty_gfn``.  Each dirty entry is defined as::8602 8603  struct kvm_dirty_gfn {8604          __u32 flags;8605          __u32 slot; /* as_id | slot_id */8606          __u64 offset;8607  };8608 8609The following values are defined for the flags field to define the8610current state of the entry::8611 8612  #define KVM_DIRTY_GFN_F_DIRTY           BIT(0)8613  #define KVM_DIRTY_GFN_F_RESET           BIT(1)8614  #define KVM_DIRTY_GFN_F_MASK            0x38615 8616Userspace should call KVM_ENABLE_CAP ioctl right after KVM_CREATE_VM8617ioctl to enable this capability for the new guest and set the size of8618the rings.  Enabling the capability is only allowed before creating any8619vCPU, and the size of the ring must be a power of two.  The larger the8620ring buffer, the less likely the ring is full and the VM is forced to8621exit to userspace. The optimal size depends on the workload, but it is8622recommended that it be at least 64 KiB (4096 entries).8623 8624Just like for dirty page bitmaps, the buffer tracks writes to8625all user memory regions for which the KVM_MEM_LOG_DIRTY_PAGES flag was8626set in KVM_SET_USER_MEMORY_REGION.  Once a memory region is registered8627with the flag set, userspace can start harvesting dirty pages from the8628ring buffer.8629 8630An entry in the ring buffer can be unused (flag bits ``00``),8631dirty (flag bits ``01``) or harvested (flag bits ``1X``).  The8632state machine for the entry is as follows::8633 8634          dirtied         harvested        reset8635     00 -----------> 01 -------------> 1X -------+8636      ^                                          |8637      |                                          |8638      +------------------------------------------+8639 8640To harvest the dirty pages, userspace accesses the mmapped ring buffer8641to read the dirty GFNs.  If the flags has the DIRTY bit set (at this stage8642the RESET bit must be cleared), then it means this GFN is a dirty GFN.8643The userspace should harvest this GFN and mark the flags from state8644``01b`` to ``1Xb`` (bit 0 will be ignored by KVM, but bit 1 must be set8645to show that this GFN is harvested and waiting for a reset), and move8646on to the next GFN.  The userspace should continue to do this until the8647flags of a GFN have the DIRTY bit cleared, meaning that it has harvested8648all the dirty GFNs that were available.8649 8650Note that on weakly ordered architectures, userspace accesses to the8651ring buffer (and more specifically the 'flags' field) must be ordered,8652using load-acquire/store-release accessors when available, or any8653other memory barrier that will ensure this ordering.8654 8655It's not necessary for userspace to harvest the all dirty GFNs at once.8656However it must collect the dirty GFNs in sequence, i.e., the userspace8657program cannot skip one dirty GFN to collect the one next to it.8658 8659After processing one or more entries in the ring buffer, userspace8660calls the VM ioctl KVM_RESET_DIRTY_RINGS to notify the kernel about8661it, so that the kernel will reprotect those collected GFNs.8662Therefore, the ioctl must be called *before* reading the content of8663the dirty pages.8664 8665The dirty ring can get full.  When it happens, the KVM_RUN of the8666vcpu will return with exit reason KVM_EXIT_DIRTY_LOG_FULL.8667 8668The dirty ring interface has a major difference comparing to the8669KVM_GET_DIRTY_LOG interface in that, when reading the dirty ring from8670userspace, it's still possible that the kernel has not yet flushed the8671processor's dirty page buffers into the kernel buffer (with dirty bitmaps, the8672flushing is done by the KVM_GET_DIRTY_LOG ioctl).  To achieve that, one8673needs to kick the vcpu out of KVM_RUN using a signal.  The resulting8674vmexit ensures that all dirty GFNs are flushed to the dirty rings.8675 8676NOTE: KVM_CAP_DIRTY_LOG_RING_ACQ_REL is the only capability that8677should be exposed by weakly ordered architecture, in order to indicate8678the additional memory ordering requirements imposed on userspace when8679reading the state of an entry and mutating it from DIRTY to HARVESTED.8680Architecture with TSO-like ordering (such as x86) are allowed to8681expose both KVM_CAP_DIRTY_LOG_RING and KVM_CAP_DIRTY_LOG_RING_ACQ_REL8682to userspace.8683 8684After enabling the dirty rings, the userspace needs to detect the8685capability of KVM_CAP_DIRTY_LOG_RING_WITH_BITMAP to see whether the8686ring structures can be backed by per-slot bitmaps. With this capability8687advertised, it means the architecture can dirty guest pages without8688vcpu/ring context, so that some of the dirty information will still be8689maintained in the bitmap structure. KVM_CAP_DIRTY_LOG_RING_WITH_BITMAP8690can't be enabled if the capability of KVM_CAP_DIRTY_LOG_RING_ACQ_REL8691hasn't been enabled, or any memslot has been existing.8692 8693Note that the bitmap here is only a backup of the ring structure. The8694use of the ring and bitmap combination is only beneficial if there is8695only a very small amount of memory that is dirtied out of vcpu/ring8696context. Otherwise, the stand-alone per-slot bitmap mechanism needs to8697be considered.8698 8699To collect dirty bits in the backup bitmap, userspace can use the same8700KVM_GET_DIRTY_LOG ioctl. KVM_CLEAR_DIRTY_LOG isn't needed as long as all8701the generation of the dirty bits is done in a single pass. Collecting8702the dirty bitmap should be the very last thing that the VMM does before8703considering the state as complete. VMM needs to ensure that the dirty8704state is final and avoid missing dirty pages from another ioctl ordered8705after the bitmap collection.8706 8707NOTE: Multiple examples of using the backup bitmap: (1) save vgic/its8708tables through command KVM_DEV_ARM_{VGIC_GRP_CTRL, ITS_SAVE_TABLES} on8709KVM device "kvm-arm-vgic-its". (2) restore vgic/its tables through8710command KVM_DEV_ARM_{VGIC_GRP_CTRL, ITS_RESTORE_TABLES} on KVM device8711"kvm-arm-vgic-its". VGICv3 LPI pending status is restored. (3) save8712vgic3 pending table through KVM_DEV_ARM_VGIC_{GRP_CTRL, SAVE_PENDING_TABLES}8713command on KVM device "kvm-arm-vgic-v3".8714 87158.30 KVM_CAP_XEN_HVM8716--------------------8717 8718:Architectures: x868719 8720This capability indicates the features that Xen supports for hosting Xen8721PVHVM guests. Valid flags are::8722 8723  #define KVM_XEN_HVM_CONFIG_HYPERCALL_MSR		(1 << 0)8724  #define KVM_XEN_HVM_CONFIG_INTERCEPT_HCALL		(1 << 1)8725  #define KVM_XEN_HVM_CONFIG_SHARED_INFO		(1 << 2)8726  #define KVM_XEN_HVM_CONFIG_RUNSTATE			(1 << 3)8727  #define KVM_XEN_HVM_CONFIG_EVTCHN_2LEVEL		(1 << 4)8728  #define KVM_XEN_HVM_CONFIG_EVTCHN_SEND		(1 << 5)8729  #define KVM_XEN_HVM_CONFIG_RUNSTATE_UPDATE_FLAG	(1 << 6)8730  #define KVM_XEN_HVM_CONFIG_PVCLOCK_TSC_UNSTABLE	(1 << 7)8731 8732The KVM_XEN_HVM_CONFIG_HYPERCALL_MSR flag indicates that the KVM_XEN_HVM_CONFIG8733ioctl is available, for the guest to set its hypercall page.8734 8735If KVM_XEN_HVM_CONFIG_INTERCEPT_HCALL is also set, the same flag may also be8736provided in the flags to KVM_XEN_HVM_CONFIG, without providing hypercall page8737contents, to request that KVM generate hypercall page content automatically8738and also enable interception of guest hypercalls with KVM_EXIT_XEN.8739 8740The KVM_XEN_HVM_CONFIG_SHARED_INFO flag indicates the availability of the8741KVM_XEN_HVM_SET_ATTR, KVM_XEN_HVM_GET_ATTR, KVM_XEN_VCPU_SET_ATTR and8742KVM_XEN_VCPU_GET_ATTR ioctls, as well as the delivery of exception vectors8743for event channel upcalls when the evtchn_upcall_pending field of a vcpu's8744vcpu_info is set.8745 8746The KVM_XEN_HVM_CONFIG_RUNSTATE flag indicates that the runstate-related8747features KVM_XEN_VCPU_ATTR_TYPE_RUNSTATE_ADDR/_CURRENT/_DATA/_ADJUST are8748supported by the KVM_XEN_VCPU_SET_ATTR/KVM_XEN_VCPU_GET_ATTR ioctls.8749 8750The KVM_XEN_HVM_CONFIG_EVTCHN_2LEVEL flag indicates that IRQ routing entries8751of the type KVM_IRQ_ROUTING_XEN_EVTCHN are supported, with the priority8752field set to indicate 2 level event channel delivery.8753 8754The KVM_XEN_HVM_CONFIG_EVTCHN_SEND flag indicates that KVM supports8755injecting event channel events directly into the guest with the8756KVM_XEN_HVM_EVTCHN_SEND ioctl. It also indicates support for the8757KVM_XEN_ATTR_TYPE_EVTCHN/XEN_VERSION HVM attributes and the8758KVM_XEN_VCPU_ATTR_TYPE_VCPU_ID/TIMER/UPCALL_VECTOR vCPU attributes.8759related to event channel delivery, timers, and the XENVER_version8760interception.8761 8762The KVM_XEN_HVM_CONFIG_RUNSTATE_UPDATE_FLAG flag indicates that KVM supports8763the KVM_XEN_ATTR_TYPE_RUNSTATE_UPDATE_FLAG attribute in the KVM_XEN_SET_ATTR8764and KVM_XEN_GET_ATTR ioctls. This controls whether KVM will set the8765XEN_RUNSTATE_UPDATE flag in guest memory mapped vcpu_runstate_info during8766updates of the runstate information. Note that versions of KVM which support8767the RUNSTATE feature above, but not the RUNSTATE_UPDATE_FLAG feature, will8768always set the XEN_RUNSTATE_UPDATE flag when updating the guest structure,8769which is perhaps counterintuitive. When this flag is advertised, KVM will8770behave more correctly, not using the XEN_RUNSTATE_UPDATE flag until/unless8771specifically enabled (by the guest making the hypercall, causing the VMM8772to enable the KVM_XEN_ATTR_TYPE_RUNSTATE_UPDATE_FLAG attribute).8773 8774The KVM_XEN_HVM_CONFIG_PVCLOCK_TSC_UNSTABLE flag indicates that KVM supports8775clearing the PVCLOCK_TSC_STABLE_BIT flag in Xen pvclock sources. This will be8776done when the KVM_CAP_XEN_HVM ioctl sets the8777KVM_XEN_HVM_CONFIG_PVCLOCK_TSC_UNSTABLE flag.8778 87798.31 KVM_CAP_PPC_MULTITCE8780-------------------------8781 8782:Capability: KVM_CAP_PPC_MULTITCE8783:Architectures: ppc8784:Type: vm8785 8786This capability means the kernel is capable of handling hypercalls8787H_PUT_TCE_INDIRECT and H_STUFF_TCE without passing those into the user8788space. This significantly accelerates DMA operations for PPC KVM guests.8789User space should expect that its handlers for these hypercalls8790are not going to be called if user space previously registered LIOBN8791in KVM (via KVM_CREATE_SPAPR_TCE or similar calls).8792 8793In order to enable H_PUT_TCE_INDIRECT and H_STUFF_TCE use in the guest,8794user space might have to advertise it for the guest. For example,8795IBM pSeries (sPAPR) guest starts using them if "hcall-multi-tce" is8796present in the "ibm,hypertas-functions" device-tree property.8797 8798The hypercalls mentioned above may or may not be processed successfully8799in the kernel based fast path. If they can not be handled by the kernel,8800they will get passed on to user space. So user space still has to have8801an implementation for these despite the in kernel acceleration.8802 8803This capability is always enabled.8804 88058.32 KVM_CAP_PTP_KVM8806--------------------8807 8808:Architectures: arm648809 8810This capability indicates that the KVM virtual PTP service is8811supported in the host. A VMM can check whether the service is8812available to the guest on migration.8813 88148.33 KVM_CAP_HYPERV_ENFORCE_CPUID8815---------------------------------8816 8817Architectures: x868818 8819When enabled, KVM will disable emulated Hyper-V features provided to the8820guest according to the bits Hyper-V CPUID feature leaves. Otherwise, all8821currently implemented Hyper-V features are provided unconditionally when8822Hyper-V identification is set in the HYPERV_CPUID_INTERFACE (0x40000001)8823leaf.8824 88258.34 KVM_CAP_EXIT_HYPERCALL8826---------------------------8827 8828:Capability: KVM_CAP_EXIT_HYPERCALL8829:Architectures: x868830:Type: vm8831 8832This capability, if enabled, will cause KVM to exit to userspace8833with KVM_EXIT_HYPERCALL exit reason to process some hypercalls.8834 8835Calling KVM_CHECK_EXTENSION for this capability will return a bitmask8836of hypercalls that can be configured to exit to userspace.8837Right now, the only such hypercall is KVM_HC_MAP_GPA_RANGE.8838 8839The argument to KVM_ENABLE_CAP is also a bitmask, and must be a subset8840of the result of KVM_CHECK_EXTENSION.  KVM will forward to userspace8841the hypercalls whose corresponding bit is in the argument, and return8842ENOSYS for the others.8843 88448.35 KVM_CAP_PMU_CAPABILITY8845---------------------------8846 8847:Capability: KVM_CAP_PMU_CAPABILITY8848:Architectures: x868849:Type: vm8850:Parameters: arg[0] is bitmask of PMU virtualization capabilities.8851:Returns: 0 on success, -EINVAL when arg[0] contains invalid bits8852 8853This capability alters PMU virtualization in KVM.8854 8855Calling KVM_CHECK_EXTENSION for this capability returns a bitmask of8856PMU virtualization capabilities that can be adjusted on a VM.8857 8858The argument to KVM_ENABLE_CAP is also a bitmask and selects specific8859PMU virtualization capabilities to be applied to the VM.  This can8860only be invoked on a VM prior to the creation of VCPUs.8861 8862At this time, KVM_PMU_CAP_DISABLE is the only capability.  Setting8863this capability will disable PMU virtualization for that VM.  Usermode8864should adjust CPUID leaf 0xA to reflect that the PMU is disabled.8865 88668.36 KVM_CAP_ARM_SYSTEM_SUSPEND8867-------------------------------8868 8869:Capability: KVM_CAP_ARM_SYSTEM_SUSPEND8870:Architectures: arm648871:Type: vm8872 8873When enabled, KVM will exit to userspace with KVM_EXIT_SYSTEM_EVENT of8874type KVM_SYSTEM_EVENT_SUSPEND to process the guest suspend request.8875 88768.37 KVM_CAP_S390_PROTECTED_DUMP8877--------------------------------8878 8879:Capability: KVM_CAP_S390_PROTECTED_DUMP8880:Architectures: s3908881:Type: vm8882 8883This capability indicates that KVM and the Ultravisor support dumping8884PV guests. The `KVM_PV_DUMP` command is available for the8885`KVM_S390_PV_COMMAND` ioctl and the `KVM_PV_INFO` command provides8886dump related UV data. Also the vcpu ioctl `KVM_S390_PV_CPU_COMMAND` is8887available and supports the `KVM_PV_DUMP_CPU` subcommand.8888 88898.38 KVM_CAP_VM_DISABLE_NX_HUGE_PAGES8890-------------------------------------8891 8892:Capability: KVM_CAP_VM_DISABLE_NX_HUGE_PAGES8893:Architectures: x868894:Type: vm8895:Parameters: arg[0] must be 0.8896:Returns: 0 on success, -EPERM if the userspace process does not8897          have CAP_SYS_BOOT, -EINVAL if args[0] is not 0 or any vCPUs have been8898          created.8899 8900This capability disables the NX huge pages mitigation for iTLB MULTIHIT.8901 8902The capability has no effect if the nx_huge_pages module parameter is not set.8903 8904This capability may only be set before any vCPUs are created.8905 89068.39 KVM_CAP_S390_CPU_TOPOLOGY8907------------------------------8908 8909:Capability: KVM_CAP_S390_CPU_TOPOLOGY8910:Architectures: s3908911:Type: vm8912 8913This capability indicates that KVM will provide the S390 CPU Topology8914facility which consist of the interpretation of the PTF instruction for8915the function code 2 along with interception and forwarding of both the8916PTF instruction with function codes 0 or 1 and the STSI(15,1,x)8917instruction to the userland hypervisor.8918 8919The stfle facility 11, CPU Topology facility, should not be indicated8920to the guest without this capability.8921 8922When this capability is present, KVM provides a new attribute group8923on vm fd, KVM_S390_VM_CPU_TOPOLOGY.8924This new attribute allows to get, set or clear the Modified Change8925Topology Report (MTCR) bit of the SCA through the kvm_device_attr8926structure.8927 8928When getting the Modified Change Topology Report value, the attr->addr8929must point to a byte where the value will be stored or retrieved from.8930 89318.40 KVM_CAP_ARM_EAGER_SPLIT_CHUNK_SIZE8932---------------------------------------8933 8934:Capability: KVM_CAP_ARM_EAGER_SPLIT_CHUNK_SIZE8935:Architectures: arm648936:Type: vm8937:Parameters: arg[0] is the new split chunk size.8938:Returns: 0 on success, -EINVAL if any memslot was already created.8939 8940This capability sets the chunk size used in Eager Page Splitting.8941 8942Eager Page Splitting improves the performance of dirty-logging (used8943in live migrations) when guest memory is backed by huge-pages.  It8944avoids splitting huge-pages (into PAGE_SIZE pages) on fault, by doing8945it eagerly when enabling dirty logging (with the8946KVM_MEM_LOG_DIRTY_PAGES flag for a memory region), or when using8947KVM_CLEAR_DIRTY_LOG.8948 8949The chunk size specifies how many pages to break at a time, using a8950single allocation for each chunk. Bigger the chunk size, more pages8951need to be allocated ahead of time.8952 8953The chunk size needs to be a valid block size. The list of acceptable8954block sizes is exposed in KVM_CAP_ARM_SUPPORTED_BLOCK_SIZES as a895564-bit bitmap (each bit describing a block size). The default value is89560, to disable the eager page splitting.8957 89588.41 KVM_CAP_VM_TYPES8959---------------------8960 8961:Capability: KVM_CAP_MEMORY_ATTRIBUTES8962:Architectures: x868963:Type: system ioctl8964 8965This capability returns a bitmap of support VM types.  The 1-setting of bit @n8966means the VM type with value @n is supported.  Possible values of @n are::8967 8968  #define KVM_X86_DEFAULT_VM	08969  #define KVM_X86_SW_PROTECTED_VM	18970  #define KVM_X86_SEV_VM	28971  #define KVM_X86_SEV_ES_VM	38972 8973Note, KVM_X86_SW_PROTECTED_VM is currently only for development and testing.8974Do not use KVM_X86_SW_PROTECTED_VM for "real" VMs, and especially not in8975production.  The behavior and effective ABI for software-protected VMs is8976unstable.8977 89789. Known KVM API problems8979=========================8980 8981In some cases, KVM's API has some inconsistencies or common pitfalls8982that userspace need to be aware of.  This section details some of8983these issues.8984 8985Most of them are architecture specific, so the section is split by8986architecture.8987 89889.1. x868989--------8990 8991``KVM_GET_SUPPORTED_CPUID`` issues8992^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^8993 8994In general, ``KVM_GET_SUPPORTED_CPUID`` is designed so that it is possible8995to take its result and pass it directly to ``KVM_SET_CPUID2``.  This section8996documents some cases in which that requires some care.8997 8998Local APIC features8999~~~~~~~~~~~~~~~~~~~9000 9001CPU[EAX=1]:ECX[21] (X2APIC) is reported by ``KVM_GET_SUPPORTED_CPUID``,9002but it can only be enabled if ``KVM_CREATE_IRQCHIP`` or9003``KVM_ENABLE_CAP(KVM_CAP_IRQCHIP_SPLIT)`` are used to enable in-kernel emulation of9004the local APIC.9005 9006The same is true for the ``KVM_FEATURE_PV_UNHALT`` paravirtualized feature.9007 9008CPU[EAX=1]:ECX[24] (TSC_DEADLINE) is not reported by ``KVM_GET_SUPPORTED_CPUID``.9009It can be enabled if ``KVM_CAP_TSC_DEADLINE_TIMER`` is present and the kernel9010has enabled in-kernel emulation of the local APIC.9011 9012CPU topology9013~~~~~~~~~~~~9014 9015Several CPUID values include topology information for the host CPU:90160x0b and 0x1f for Intel systems, 0x8000001e for AMD systems.  Different9017versions of KVM return different values for this information and userspace9018should not rely on it.  Currently they return all zeroes.9019 9020If userspace wishes to set up a guest topology, it should be careful that9021the values of these three leaves differ for each CPU.  In particular,9022the APIC ID is found in EDX for all subleaves of 0x0b and 0x1f, and in EAX9023for 0x8000001e; the latter also encodes the core id and node id in bits90247:0 of EBX and ECX respectively.9025 9026Obsolete ioctls and capabilities9027^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^9028 9029KVM_CAP_DISABLE_QUIRKS does not let userspace know which quirks are actually9030available.  Use ``KVM_CHECK_EXTENSION(KVM_CAP_DISABLE_QUIRKS2)`` instead if9031available.9032 9033Ordering of KVM_GET_*/KVM_SET_* ioctls9034^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^9035 9036TBD9037