1 The Definitive KVM (Kernel-based Virtual Machine) API Documentation
2 ===================================================================
6 The kvm API is a set of ioctls that are issued to control various aspects
7 of a virtual machine. The ioctls belong to three classes
9 - System ioctls: These query and set global attributes which affect the
10 whole kvm subsystem. In addition a system ioctl is used to create
13 - VM ioctls: These query and set attributes that affect an entire virtual
14 machine, for example memory layout. In addition a VM ioctl is used to
15 create virtual cpus (vcpus).
17 Only run VM ioctls from the same process (address space) that was used
20 - vcpu ioctls: These query and set attributes that control the operation
21 of a single virtual cpu.
23 Only run vcpu ioctls from the same thread that was used to create the
28 The kvm API is centered around file descriptors. An initial
29 open("/dev/kvm") obtains a handle to the kvm subsystem; this handle
30 can be used to issue system ioctls. A KVM_CREATE_VM ioctl on this
31 handle will create a VM file descriptor which can be used to issue VM
32 ioctls. A KVM_CREATE_VCPU ioctl on a VM fd will create a virtual cpu
33 and return a file descriptor pointing to it. Finally, ioctls on a vcpu
34 fd can be used to control the vcpu, including the important task of
35 actually running guest code.
37 In general file descriptors can be migrated among processes by means
38 of fork() and the SCM_RIGHTS facility of unix domain socket. These
39 kinds of tricks are explicitly not supported by kvm. While they will
40 not cause harm to the host, their actual behavior is not guaranteed by
41 the API. The only supported use is one virtual machine per process,
42 and one vcpu per thread.
46 As of Linux 2.6.22, the KVM ABI has been stabilized: no backward
47 incompatible change are allowed. However, there is an extension
48 facility that allows backward-compatible extensions to the API to be
51 The extension mechanism is not based on on the Linux version number.
52 Instead, kvm defines extension identifiers and a facility to query
53 whether a particular extension identifier is available. If it is, a
54 set of ioctls is available for application use.
58 This section describes ioctls that can be used to control kvm guests.
59 For each ioctl, the following information is provided along with a
62 Capability: which KVM extension provides this ioctl. Can be 'basic',
63 which means that is will be provided by any kernel that supports
64 API version 12 (see section 4.1), or a KVM_CAP_xyz constant, which
65 means availability needs to be checked with KVM_CHECK_EXTENSION
68 Architectures: which instruction set architectures provide this ioctl.
69 x86 includes both i386 and x86_64.
71 Type: system, vm, or vcpu.
73 Parameters: what parameters are accepted by the ioctl.
75 Returns: the return value. General error numbers (EBADF, ENOMEM, EINVAL)
76 are not detailed, but errors with specific meanings are.
78 4.1 KVM_GET_API_VERSION
84 Returns: the constant KVM_API_VERSION (=12)
86 This identifies the API version as the stable kvm API. It is not
87 expected that this number will change. However, Linux 2.6.20 and
88 2.6.21 report earlier versions; these are not documented and not
89 supported. Applications should refuse to run if KVM_GET_API_VERSION
90 returns a value other than 12. If this check passes, all ioctls
91 described as 'basic' will be available.
99 Returns: a VM fd that can be used to control the new virtual machine.
101 The new VM has no virtual cpus and no memory. An mmap() of a VM fd
102 will access the virtual machine's physical address space; offset zero
103 corresponds to guest physical address zero. Use of mmap() on a VM fd
104 is discouraged if userspace memory allocation (KVM_CAP_USER_MEMORY) is
107 4.3 KVM_GET_MSR_INDEX_LIST
112 Parameters: struct kvm_msr_list (in/out)
113 Returns: 0 on success; -1 on error
115 E2BIG: the msr index list is to be to fit in the array specified by
118 struct kvm_msr_list {
119 __u32 nmsrs; /* number of msrs in entries */
123 This ioctl returns the guest msrs that are supported. The list varies
124 by kvm version and host processor, but does not change otherwise. The
125 user fills in the size of the indices array in nmsrs, and in return
126 kvm adjusts nmsrs to reflect the actual number of msrs and fills in
127 the indices array with their numbers.
129 Note: if kvm indicates supports MCE (KVM_CAP_MCE), then the MCE bank MSRs are
130 not returned in the MSR list, as different vcpus can have a different number
131 of banks, as set via the KVM_X86_SETUP_MCE ioctl.
133 4.4 KVM_CHECK_EXTENSION
138 Parameters: extension identifier (KVM_CAP_*)
139 Returns: 0 if unsupported; 1 (or some other positive integer) if supported
141 The API allows the application to query about extensions to the core
142 kvm API. Userspace passes an extension identifier (an integer) and
143 receives an integer that describes the extension availability.
144 Generally 0 means no and 1 means yes, but some extensions may report
145 additional information in the integer return value.
147 4.5 KVM_GET_VCPU_MMAP_SIZE
153 Returns: size of vcpu mmap area, in bytes
155 The KVM_RUN ioctl (cf.) communicates with userspace via a shared
156 memory region. This ioctl returns the size of that region. See the
157 KVM_RUN documentation for details.
159 4.6 KVM_SET_MEMORY_REGION
164 Parameters: struct kvm_memory_region (in)
165 Returns: 0 on success, -1 on error
167 This ioctl is obsolete and has been removed.
174 Parameters: vcpu id (apic id on x86)
175 Returns: vcpu fd on success, -1 on error
177 This API adds a vcpu to a virtual machine. The vcpu id is a small integer
178 in the range [0, max_vcpus). You can use KVM_CAP_NR_VCPUS of the
179 KVM_CHECK_EXTENSION ioctl() to determine the value for max_vcpus at run-time.
180 If the KVM_CAP_NR_VCPUS does not exist, you should assume that max_vcpus is 4
183 4.8 KVM_GET_DIRTY_LOG (vm ioctl)
188 Parameters: struct kvm_dirty_log (in/out)
189 Returns: 0 on success, -1 on error
191 /* for KVM_GET_DIRTY_LOG */
192 struct kvm_dirty_log {
196 void __user *dirty_bitmap; /* one bit per page */
201 Given a memory slot, return a bitmap containing any pages dirtied
202 since the last call to this ioctl. Bit 0 is the first page in the
203 memory slot. Ensure the entire structure is cleared to avoid padding
206 4.9 KVM_SET_MEMORY_ALIAS
211 Parameters: struct kvm_memory_alias (in)
212 Returns: 0 (success), -1 (error)
214 This ioctl is obsolete and has been removed.
222 Returns: 0 on success, -1 on error
224 EINTR: an unmasked signal is pending
226 This ioctl is used to run a guest virtual cpu. While there are no
227 explicit parameters, there is an implicit parameter block that can be
228 obtained by mmap()ing the vcpu fd at offset 0, with the size given by
229 KVM_GET_VCPU_MMAP_SIZE. The parameter block is formatted as a 'struct
230 kvm_run' (see below).
237 Parameters: struct kvm_regs (out)
238 Returns: 0 on success, -1 on error
240 Reads the general purpose registers from the vcpu.
244 /* out (KVM_GET_REGS) / in (KVM_SET_REGS) */
245 __u64 rax, rbx, rcx, rdx;
246 __u64 rsi, rdi, rsp, rbp;
247 __u64 r8, r9, r10, r11;
248 __u64 r12, r13, r14, r15;
257 Parameters: struct kvm_regs (in)
258 Returns: 0 on success, -1 on error
260 Writes the general purpose registers into the vcpu.
262 See KVM_GET_REGS for the data structure.
267 Architectures: x86, ppc
269 Parameters: struct kvm_sregs (out)
270 Returns: 0 on success, -1 on error
272 Reads special registers from the vcpu.
276 struct kvm_segment cs, ds, es, fs, gs, ss;
277 struct kvm_segment tr, ldt;
278 struct kvm_dtable gdt, idt;
279 __u64 cr0, cr2, cr3, cr4, cr8;
282 __u64 interrupt_bitmap[(KVM_NR_INTERRUPTS + 63) / 64];
285 /* ppc -- see arch/powerpc/include/asm/kvm.h */
287 interrupt_bitmap is a bitmap of pending external interrupts. At most
288 one bit may be set. This interrupt has been acknowledged by the APIC
289 but not yet injected into the cpu core.
294 Architectures: x86, ppc
296 Parameters: struct kvm_sregs (in)
297 Returns: 0 on success, -1 on error
299 Writes special registers into the vcpu. See KVM_GET_SREGS for the
307 Parameters: struct kvm_translation (in/out)
308 Returns: 0 on success, -1 on error
310 Translates a virtual address according to the vcpu's current address
313 struct kvm_translation {
315 __u64 linear_address;
318 __u64 physical_address;
328 Architectures: x86, ppc
330 Parameters: struct kvm_interrupt (in)
331 Returns: 0 on success, -1 on error
333 Queues a hardware interrupt vector to be injected. This is only
334 useful if in-kernel local APIC or equivalent is not used.
336 /* for KVM_INTERRUPT */
337 struct kvm_interrupt {
344 Note 'irq' is an interrupt vector, not an interrupt pin or line.
348 Queues an external interrupt to be injected. This ioctl is overleaded
349 with 3 different irq values:
353 This injects an edge type external interrupt into the guest once it's ready
354 to receive interrupts. When injected, the interrupt is done.
356 b) KVM_INTERRUPT_UNSET
358 This unsets any pending interrupt.
360 Only available with KVM_CAP_PPC_UNSET_IRQ.
362 c) KVM_INTERRUPT_SET_LEVEL
364 This injects a level type external interrupt into the guest context. The
365 interrupt stays pending until a specific ioctl with KVM_INTERRUPT_UNSET
368 Only available with KVM_CAP_PPC_IRQ_LEVEL.
370 Note that any value for 'irq' other than the ones stated above is invalid
371 and incurs unexpected behavior.
381 Support for this has been removed. Use KVM_SET_GUEST_DEBUG instead.
388 Parameters: struct kvm_msrs (in/out)
389 Returns: 0 on success, -1 on error
391 Reads model-specific registers from the vcpu. Supported msr indices can
392 be obtained using KVM_GET_MSR_INDEX_LIST.
395 __u32 nmsrs; /* number of msrs in entries */
398 struct kvm_msr_entry entries[0];
401 struct kvm_msr_entry {
407 Application code should set the 'nmsrs' member (which indicates the
408 size of the entries array) and the 'index' member of each array entry.
409 kvm will fill in the 'data' member.
416 Parameters: struct kvm_msrs (in)
417 Returns: 0 on success, -1 on error
419 Writes model-specific registers to the vcpu. See KVM_GET_MSRS for the
422 Application code should set the 'nmsrs' member (which indicates the
423 size of the entries array), and the 'index' and 'data' members of each
431 Parameters: struct kvm_cpuid (in)
432 Returns: 0 on success, -1 on error
434 Defines the vcpu responses to the cpuid instruction. Applications
435 should use the KVM_SET_CPUID2 ioctl if available.
438 struct kvm_cpuid_entry {
447 /* for KVM_SET_CPUID */
451 struct kvm_cpuid_entry entries[0];
454 4.21 KVM_SET_SIGNAL_MASK
459 Parameters: struct kvm_signal_mask (in)
460 Returns: 0 on success, -1 on error
462 Defines which signals are blocked during execution of KVM_RUN. This
463 signal mask temporarily overrides the threads signal mask. Any
464 unblocked signal received (except SIGKILL and SIGSTOP, which retain
465 their traditional behaviour) will cause KVM_RUN to return with -EINTR.
467 Note the signal will only be delivered if not blocked by the original
470 /* for KVM_SET_SIGNAL_MASK */
471 struct kvm_signal_mask {
481 Parameters: struct kvm_fpu (out)
482 Returns: 0 on success, -1 on error
484 Reads the floating point state from the vcpu.
486 /* for KVM_GET_FPU and KVM_SET_FPU */
491 __u8 ftwx; /* in fxsave format */
506 Parameters: struct kvm_fpu (in)
507 Returns: 0 on success, -1 on error
509 Writes the floating point state to the vcpu.
511 /* for KVM_GET_FPU and KVM_SET_FPU */
516 __u8 ftwx; /* in fxsave format */
526 4.24 KVM_CREATE_IRQCHIP
528 Capability: KVM_CAP_IRQCHIP
529 Architectures: x86, ia64
532 Returns: 0 on success, -1 on error
534 Creates an interrupt controller model in the kernel. On x86, creates a virtual
535 ioapic, a virtual PIC (two PICs, nested), and sets up future vcpus to have a
536 local APIC. IRQ routing for GSIs 0-15 is set to both PIC and IOAPIC; GSI 16-23
537 only go to the IOAPIC. On ia64, a IOSAPIC is created.
541 Capability: KVM_CAP_IRQCHIP
542 Architectures: x86, ia64
544 Parameters: struct kvm_irq_level
545 Returns: 0 on success, -1 on error
547 Sets the level of a GSI input to the interrupt controller model in the kernel.
548 Requires that an interrupt controller model has been previously created with
549 KVM_CREATE_IRQCHIP. Note that edge-triggered interrupts require the level
550 to be set to 1 and then back to 0.
552 struct kvm_irq_level {
555 __s32 status; /* not used for KVM_IRQ_LEVEL */
557 __u32 level; /* 0 or 1 */
562 Capability: KVM_CAP_IRQCHIP
563 Architectures: x86, ia64
565 Parameters: struct kvm_irqchip (in/out)
566 Returns: 0 on success, -1 on error
568 Reads the state of a kernel interrupt controller created with
569 KVM_CREATE_IRQCHIP into a buffer provided by the caller.
572 __u32 chip_id; /* 0 = PIC1, 1 = PIC2, 2 = IOAPIC */
575 char dummy[512]; /* reserving space */
576 struct kvm_pic_state pic;
577 struct kvm_ioapic_state ioapic;
583 Capability: KVM_CAP_IRQCHIP
584 Architectures: x86, ia64
586 Parameters: struct kvm_irqchip (in)
587 Returns: 0 on success, -1 on error
589 Sets the state of a kernel interrupt controller created with
590 KVM_CREATE_IRQCHIP from a buffer provided by the caller.
593 __u32 chip_id; /* 0 = PIC1, 1 = PIC2, 2 = IOAPIC */
596 char dummy[512]; /* reserving space */
597 struct kvm_pic_state pic;
598 struct kvm_ioapic_state ioapic;
602 4.28 KVM_XEN_HVM_CONFIG
604 Capability: KVM_CAP_XEN_HVM
607 Parameters: struct kvm_xen_hvm_config (in)
608 Returns: 0 on success, -1 on error
610 Sets the MSR that the Xen HVM guest uses to initialize its hypercall
611 page, and provides the starting address and size of the hypercall
612 blobs in userspace. When the guest writes the MSR, kvm copies one
613 page of a blob (32- or 64-bit, depending on the vcpu mode) to guest
616 struct kvm_xen_hvm_config {
628 Capability: KVM_CAP_ADJUST_CLOCK
631 Parameters: struct kvm_clock_data (out)
632 Returns: 0 on success, -1 on error
634 Gets the current timestamp of kvmclock as seen by the current guest. In
635 conjunction with KVM_SET_CLOCK, it is used to ensure monotonicity on scenarios
638 struct kvm_clock_data {
639 __u64 clock; /* kvmclock current value */
646 Capability: KVM_CAP_ADJUST_CLOCK
649 Parameters: struct kvm_clock_data (in)
650 Returns: 0 on success, -1 on error
652 Sets the current timestamp of kvmclock to the value specified in its parameter.
653 In conjunction with KVM_GET_CLOCK, it is used to ensure monotonicity on scenarios
656 struct kvm_clock_data {
657 __u64 clock; /* kvmclock current value */
662 4.31 KVM_GET_VCPU_EVENTS
664 Capability: KVM_CAP_VCPU_EVENTS
665 Extended by: KVM_CAP_INTR_SHADOW
668 Parameters: struct kvm_vcpu_event (out)
669 Returns: 0 on success, -1 on error
671 Gets currently pending exceptions, interrupts, and NMIs as well as related
674 struct kvm_vcpu_events {
698 KVM_VCPUEVENT_VALID_SHADOW may be set in the flags field to signal that
699 interrupt.shadow contains a valid state. Otherwise, this field is undefined.
701 4.32 KVM_SET_VCPU_EVENTS
703 Capability: KVM_CAP_VCPU_EVENTS
704 Extended by: KVM_CAP_INTR_SHADOW
707 Parameters: struct kvm_vcpu_event (in)
708 Returns: 0 on success, -1 on error
710 Set pending exceptions, interrupts, and NMIs as well as related states of the
713 See KVM_GET_VCPU_EVENTS for the data structure.
715 Fields that may be modified asynchronously by running VCPUs can be excluded
716 from the update. These fields are nmi.pending and sipi_vector. Keep the
717 corresponding bits in the flags field cleared to suppress overwriting the
718 current in-kernel state. The bits are:
720 KVM_VCPUEVENT_VALID_NMI_PENDING - transfer nmi.pending to the kernel
721 KVM_VCPUEVENT_VALID_SIPI_VECTOR - transfer sipi_vector
723 If KVM_CAP_INTR_SHADOW is available, KVM_VCPUEVENT_VALID_SHADOW can be set in
724 the flags field to signal that interrupt.shadow contains a valid state and
725 shall be written into the VCPU.
727 4.33 KVM_GET_DEBUGREGS
729 Capability: KVM_CAP_DEBUGREGS
732 Parameters: struct kvm_debugregs (out)
733 Returns: 0 on success, -1 on error
735 Reads debug registers from the vcpu.
737 struct kvm_debugregs {
745 4.34 KVM_SET_DEBUGREGS
747 Capability: KVM_CAP_DEBUGREGS
750 Parameters: struct kvm_debugregs (in)
751 Returns: 0 on success, -1 on error
753 Writes debug registers into the vcpu.
755 See KVM_GET_DEBUGREGS for the data structure. The flags field is unused
756 yet and must be cleared on entry.
758 4.35 KVM_SET_USER_MEMORY_REGION
760 Capability: KVM_CAP_USER_MEM
763 Parameters: struct kvm_userspace_memory_region (in)
764 Returns: 0 on success, -1 on error
766 struct kvm_userspace_memory_region {
769 __u64 guest_phys_addr;
770 __u64 memory_size; /* bytes */
771 __u64 userspace_addr; /* start of the userspace allocated memory */
774 /* for kvm_memory_region::flags */
775 #define KVM_MEM_LOG_DIRTY_PAGES 1UL
777 This ioctl allows the user to create or modify a guest physical memory
778 slot. When changing an existing slot, it may be moved in the guest
779 physical memory space, or its flags may be modified. It may not be
780 resized. Slots may not overlap in guest physical address space.
782 Memory for the region is taken starting at the address denoted by the
783 field userspace_addr, which must point at user addressable memory for
784 the entire memory slot size. Any object may back this memory, including
785 anonymous memory, ordinary files, and hugetlbfs.
787 It is recommended that the lower 21 bits of guest_phys_addr and userspace_addr
788 be identical. This allows large pages in the guest to be backed by large
791 The flags field supports just one flag, KVM_MEM_LOG_DIRTY_PAGES, which
792 instructs kvm to keep track of writes to memory within the slot. See
793 the KVM_GET_DIRTY_LOG ioctl.
795 When the KVM_CAP_SYNC_MMU capability, changes in the backing of the memory
796 region are automatically reflected into the guest. For example, an mmap()
797 that affects the region will be made visible immediately. Another example
798 is madvise(MADV_DROP).
800 It is recommended to use this API instead of the KVM_SET_MEMORY_REGION ioctl.
801 The KVM_SET_MEMORY_REGION does not allow fine grained control over memory
802 allocation and is deprecated.
804 4.36 KVM_SET_TSS_ADDR
806 Capability: KVM_CAP_SET_TSS_ADDR
809 Parameters: unsigned long tss_address (in)
810 Returns: 0 on success, -1 on error
812 This ioctl defines the physical address of a three-page region in the guest
813 physical address space. The region must be within the first 4GB of the
814 guest physical address space and must not conflict with any memory slot
815 or any mmio address. The guest may malfunction if it accesses this memory
818 This ioctl is required on Intel-based hosts. This is needed on Intel hardware
819 because of a quirk in the virtualization implementation (see the internals
820 documentation when it pops into existence).
824 Capability: KVM_CAP_ENABLE_CAP
827 Parameters: struct kvm_enable_cap (in)
828 Returns: 0 on success; -1 on error
830 +Not all extensions are enabled by default. Using this ioctl the application
831 can enable an extension, making it available to the guest.
833 On systems that do not support this ioctl, it always fails. On systems that
834 do support it, it only works for extensions that are supported for enablement.
836 To check if a capability can be enabled, the KVM_CHECK_EXTENSION ioctl should
839 struct kvm_enable_cap {
843 The capability that is supposed to get enabled.
847 A bitfield indicating future enhancements. Has to be 0 for now.
851 Arguments for enabling a feature. If a feature needs initial values to
852 function properly, this is the place to put them.
857 4.38 KVM_GET_MP_STATE
859 Capability: KVM_CAP_MP_STATE
860 Architectures: x86, ia64
862 Parameters: struct kvm_mp_state (out)
863 Returns: 0 on success; -1 on error
865 struct kvm_mp_state {
869 Returns the vcpu's current "multiprocessing state" (though also valid on
870 uniprocessor guests).
874 - KVM_MP_STATE_RUNNABLE: the vcpu is currently running
875 - KVM_MP_STATE_UNINITIALIZED: the vcpu is an application processor (AP)
876 which has not yet received an INIT signal
877 - KVM_MP_STATE_INIT_RECEIVED: the vcpu has received an INIT signal, and is
879 - KVM_MP_STATE_HALTED: the vcpu has executed a HLT instruction and
880 is waiting for an interrupt
881 - KVM_MP_STATE_SIPI_RECEIVED: the vcpu has just received a SIPI (vector
882 accessible via KVM_GET_VCPU_EVENTS)
884 This ioctl is only useful after KVM_CREATE_IRQCHIP. Without an in-kernel
885 irqchip, the multiprocessing state must be maintained by userspace.
887 4.39 KVM_SET_MP_STATE
889 Capability: KVM_CAP_MP_STATE
890 Architectures: x86, ia64
892 Parameters: struct kvm_mp_state (in)
893 Returns: 0 on success; -1 on error
895 Sets the vcpu's current "multiprocessing state"; see KVM_GET_MP_STATE for
898 This ioctl is only useful after KVM_CREATE_IRQCHIP. Without an in-kernel
899 irqchip, the multiprocessing state must be maintained by userspace.
901 4.40 KVM_SET_IDENTITY_MAP_ADDR
903 Capability: KVM_CAP_SET_IDENTITY_MAP_ADDR
906 Parameters: unsigned long identity (in)
907 Returns: 0 on success, -1 on error
909 This ioctl defines the physical address of a one-page region in the guest
910 physical address space. The region must be within the first 4GB of the
911 guest physical address space and must not conflict with any memory slot
912 or any mmio address. The guest may malfunction if it accesses this memory
915 This ioctl is required on Intel-based hosts. This is needed on Intel hardware
916 because of a quirk in the virtualization implementation (see the internals
917 documentation when it pops into existence).
919 4.41 KVM_SET_BOOT_CPU_ID
921 Capability: KVM_CAP_SET_BOOT_CPU_ID
922 Architectures: x86, ia64
924 Parameters: unsigned long vcpu_id
925 Returns: 0 on success, -1 on error
927 Define which vcpu is the Bootstrap Processor (BSP). Values are the same
928 as the vcpu id in KVM_CREATE_VCPU. If this ioctl is not called, the default
933 Capability: KVM_CAP_XSAVE
936 Parameters: struct kvm_xsave (out)
937 Returns: 0 on success, -1 on error
943 This ioctl would copy current vcpu's xsave struct to the userspace.
947 Capability: KVM_CAP_XSAVE
950 Parameters: struct kvm_xsave (in)
951 Returns: 0 on success, -1 on error
957 This ioctl would copy userspace's xsave struct to the kernel.
961 Capability: KVM_CAP_XCRS
964 Parameters: struct kvm_xcrs (out)
965 Returns: 0 on success, -1 on error
976 struct kvm_xcr xcrs[KVM_MAX_XCRS];
980 This ioctl would copy current vcpu's xcrs to the userspace.
984 Capability: KVM_CAP_XCRS
987 Parameters: struct kvm_xcrs (in)
988 Returns: 0 on success, -1 on error
999 struct kvm_xcr xcrs[KVM_MAX_XCRS];
1003 This ioctl would set vcpu's xcr to the value userspace specified.
1005 4.46 KVM_GET_SUPPORTED_CPUID
1007 Capability: KVM_CAP_EXT_CPUID
1010 Parameters: struct kvm_cpuid2 (in/out)
1011 Returns: 0 on success, -1 on error
1016 struct kvm_cpuid_entry2 entries[0];
1019 #define KVM_CPUID_FLAG_SIGNIFCANT_INDEX 1
1020 #define KVM_CPUID_FLAG_STATEFUL_FUNC 2
1021 #define KVM_CPUID_FLAG_STATE_READ_NEXT 4
1023 struct kvm_cpuid_entry2 {
1034 This ioctl returns x86 cpuid features which are supported by both the hardware
1035 and kvm. Userspace can use the information returned by this ioctl to
1036 construct cpuid information (for KVM_SET_CPUID2) that is consistent with
1037 hardware, kernel, and userspace capabilities, and with user requirements (for
1038 example, the user may wish to constrain cpuid to emulate older hardware,
1039 or for feature consistency across a cluster).
1041 Userspace invokes KVM_GET_SUPPORTED_CPUID by passing a kvm_cpuid2 structure
1042 with the 'nent' field indicating the number of entries in the variable-size
1043 array 'entries'. If the number of entries is too low to describe the cpu
1044 capabilities, an error (E2BIG) is returned. If the number is too high,
1045 the 'nent' field is adjusted and an error (ENOMEM) is returned. If the
1046 number is just right, the 'nent' field is adjusted to the number of valid
1047 entries in the 'entries' array, which is then filled.
1049 The entries returned are the host cpuid as returned by the cpuid instruction,
1050 with unknown or unsupported features masked out. Some features (for example,
1051 x2apic), may not be present in the host cpu, but are exposed by kvm if it can
1052 emulate them efficiently. The fields in each entry are defined as follows:
1054 function: the eax value used to obtain the entry
1055 index: the ecx value used to obtain the entry (for entries that are
1057 flags: an OR of zero or more of the following:
1058 KVM_CPUID_FLAG_SIGNIFCANT_INDEX:
1059 if the index field is valid
1060 KVM_CPUID_FLAG_STATEFUL_FUNC:
1061 if cpuid for this function returns different values for successive
1062 invocations; there will be several entries with the same function,
1063 all with this flag set
1064 KVM_CPUID_FLAG_STATE_READ_NEXT:
1065 for KVM_CPUID_FLAG_STATEFUL_FUNC entries, set if this entry is
1066 the first entry to be read by a cpu
1067 eax, ebx, ecx, edx: the values returned by the cpuid instruction for
1068 this function/index combination
1070 4.47 KVM_PPC_GET_PVINFO
1072 Capability: KVM_CAP_PPC_GET_PVINFO
1075 Parameters: struct kvm_ppc_pvinfo (out)
1076 Returns: 0 on success, !0 on error
1078 struct kvm_ppc_pvinfo {
1084 This ioctl fetches PV specific information that need to be passed to the guest
1085 using the device tree or other means from vm context.
1087 For now the only implemented piece of information distributed here is an array
1088 of 4 instructions that make up a hypercall.
1090 If any additional field gets added to this structure later on, a bit for that
1091 additional piece of information will be set in the flags bitmap.
1093 4.48 KVM_ASSIGN_PCI_DEVICE
1095 Capability: KVM_CAP_DEVICE_ASSIGNMENT
1096 Architectures: x86 ia64
1098 Parameters: struct kvm_assigned_pci_dev (in)
1099 Returns: 0 on success, -1 on error
1101 Assigns a host PCI device to the VM.
1103 struct kvm_assigned_pci_dev {
1104 __u32 assigned_dev_id;
1114 The PCI device is specified by the triple segnr, busnr, and devfn.
1115 Identification in succeeding service requests is done via assigned_dev_id. The
1116 following flags are specified:
1118 /* Depends on KVM_CAP_IOMMU */
1119 #define KVM_DEV_ASSIGN_ENABLE_IOMMU (1 << 0)
1121 4.49 KVM_DEASSIGN_PCI_DEVICE
1123 Capability: KVM_CAP_DEVICE_DEASSIGNMENT
1124 Architectures: x86 ia64
1126 Parameters: struct kvm_assigned_pci_dev (in)
1127 Returns: 0 on success, -1 on error
1129 Ends PCI device assignment, releasing all associated resources.
1131 See KVM_CAP_DEVICE_ASSIGNMENT for the data structure. Only assigned_dev_id is
1132 used in kvm_assigned_pci_dev to identify the device.
1134 4.50 KVM_ASSIGN_DEV_IRQ
1136 Capability: KVM_CAP_ASSIGN_DEV_IRQ
1137 Architectures: x86 ia64
1139 Parameters: struct kvm_assigned_irq (in)
1140 Returns: 0 on success, -1 on error
1142 Assigns an IRQ to a passed-through device.
1144 struct kvm_assigned_irq {
1145 __u32 assigned_dev_id;
1159 The following flags are defined:
1161 #define KVM_DEV_IRQ_HOST_INTX (1 << 0)
1162 #define KVM_DEV_IRQ_HOST_MSI (1 << 1)
1163 #define KVM_DEV_IRQ_HOST_MSIX (1 << 2)
1165 #define KVM_DEV_IRQ_GUEST_INTX (1 << 8)
1166 #define KVM_DEV_IRQ_GUEST_MSI (1 << 9)
1167 #define KVM_DEV_IRQ_GUEST_MSIX (1 << 10)
1169 It is not valid to specify multiple types per host or guest IRQ. However, the
1170 IRQ type of host and guest can differ or can even be null.
1172 4.51 KVM_DEASSIGN_DEV_IRQ
1174 Capability: KVM_CAP_ASSIGN_DEV_IRQ
1175 Architectures: x86 ia64
1177 Parameters: struct kvm_assigned_irq (in)
1178 Returns: 0 on success, -1 on error
1180 Ends an IRQ assignment to a passed-through device.
1182 See KVM_ASSIGN_DEV_IRQ for the data structure. The target device is specified
1183 by assigned_dev_id, flags must correspond to the IRQ type specified on
1184 KVM_ASSIGN_DEV_IRQ. Partial deassignment of host or guest IRQ is allowed.
1186 4.52 KVM_SET_GSI_ROUTING
1188 Capability: KVM_CAP_IRQ_ROUTING
1189 Architectures: x86 ia64
1191 Parameters: struct kvm_irq_routing (in)
1192 Returns: 0 on success, -1 on error
1194 Sets the GSI routing table entries, overwriting any previously set entries.
1196 struct kvm_irq_routing {
1199 struct kvm_irq_routing_entry entries[0];
1202 No flags are specified so far, the corresponding field must be set to zero.
1204 struct kvm_irq_routing_entry {
1210 struct kvm_irq_routing_irqchip irqchip;
1211 struct kvm_irq_routing_msi msi;
1216 /* gsi routing entry types */
1217 #define KVM_IRQ_ROUTING_IRQCHIP 1
1218 #define KVM_IRQ_ROUTING_MSI 2
1220 No flags are specified so far, the corresponding field must be set to zero.
1222 struct kvm_irq_routing_irqchip {
1227 struct kvm_irq_routing_msi {
1234 4.53 KVM_ASSIGN_SET_MSIX_NR
1236 Capability: KVM_CAP_DEVICE_MSIX
1237 Architectures: x86 ia64
1239 Parameters: struct kvm_assigned_msix_nr (in)
1240 Returns: 0 on success, -1 on error
1242 Set the number of MSI-X interrupts for an assigned device. This service can
1243 only be called once in the lifetime of an assigned device.
1245 struct kvm_assigned_msix_nr {
1246 __u32 assigned_dev_id;
1251 #define KVM_MAX_MSIX_PER_DEV 256
1253 4.54 KVM_ASSIGN_SET_MSIX_ENTRY
1255 Capability: KVM_CAP_DEVICE_MSIX
1256 Architectures: x86 ia64
1258 Parameters: struct kvm_assigned_msix_entry (in)
1259 Returns: 0 on success, -1 on error
1261 Specifies the routing of an MSI-X assigned device interrupt to a GSI. Setting
1262 the GSI vector to zero means disabling the interrupt.
1264 struct kvm_assigned_msix_entry {
1265 __u32 assigned_dev_id;
1267 __u16 entry; /* The index of entry in the MSI-X table */
1271 4.54 KVM_SET_TSC_KHZ
1273 Capability: KVM_CAP_TSC_CONTROL
1276 Parameters: virtual tsc_khz
1277 Returns: 0 on success, -1 on error
1279 Specifies the tsc frequency for the virtual machine. The unit of the
1282 4.55 KVM_GET_TSC_KHZ
1284 Capability: KVM_CAP_GET_TSC_KHZ
1288 Returns: virtual tsc-khz on success, negative value on error
1290 Returns the tsc frequency of the guest. The unit of the return value is
1291 KHz. If the host has unstable tsc this ioctl returns -EIO instead as an
1294 5. The kvm_run structure
1296 Application code obtains a pointer to the kvm_run structure by
1297 mmap()ing a vcpu fd. From that point, application code can control
1298 execution by changing fields in kvm_run prior to calling the KVM_RUN
1299 ioctl, and obtain information about the reason KVM_RUN returned by
1300 looking up structure members.
1304 __u8 request_interrupt_window;
1306 Request that KVM_RUN return when it becomes possible to inject external
1307 interrupts into the guest. Useful in conjunction with KVM_INTERRUPT.
1314 When KVM_RUN has returned successfully (return value 0), this informs
1315 application code why KVM_RUN has returned. Allowable values for this
1316 field are detailed below.
1318 __u8 ready_for_interrupt_injection;
1320 If request_interrupt_window has been specified, this field indicates
1321 an interrupt can be injected now with KVM_INTERRUPT.
1325 The value of the current interrupt flag. Only valid if in-kernel
1326 local APIC is not used.
1330 /* in (pre_kvm_run), out (post_kvm_run) */
1333 The value of the cr8 register. Only valid if in-kernel local APIC is
1334 not used. Both input and output.
1338 The value of the APIC BASE msr. Only valid if in-kernel local
1339 APIC is not used. Both input and output.
1342 /* KVM_EXIT_UNKNOWN */
1344 __u64 hardware_exit_reason;
1347 If exit_reason is KVM_EXIT_UNKNOWN, the vcpu has exited due to unknown
1348 reasons. Further architecture-specific information is available in
1349 hardware_exit_reason.
1351 /* KVM_EXIT_FAIL_ENTRY */
1353 __u64 hardware_entry_failure_reason;
1356 If exit_reason is KVM_EXIT_FAIL_ENTRY, the vcpu could not be run due
1357 to unknown reasons. Further architecture-specific information is
1358 available in hardware_entry_failure_reason.
1360 /* KVM_EXIT_EXCEPTION */
1370 #define KVM_EXIT_IO_IN 0
1371 #define KVM_EXIT_IO_OUT 1
1373 __u8 size; /* bytes */
1376 __u64 data_offset; /* relative to kvm_run start */
1379 If exit_reason is KVM_EXIT_IO, then the vcpu has
1380 executed a port I/O instruction which could not be satisfied by kvm.
1381 data_offset describes where the data is located (KVM_EXIT_IO_OUT) or
1382 where kvm expects application code to place the data for the next
1383 KVM_RUN invocation (KVM_EXIT_IO_IN). Data format is a packed array.
1386 struct kvm_debug_exit_arch arch;
1399 If exit_reason is KVM_EXIT_MMIO, then the vcpu has
1400 executed a memory-mapped I/O instruction which could not be satisfied
1401 by kvm. The 'data' member contains the written data if 'is_write' is
1402 true, and should be filled by application code otherwise.
1404 NOTE: For KVM_EXIT_IO, KVM_EXIT_MMIO and KVM_EXIT_OSI, the corresponding
1405 operations are complete (and guest state is consistent) only after userspace
1406 has re-entered the kernel with KVM_RUN. The kernel side will first finish
1407 incomplete operations and then check for pending signals. Userspace
1408 can re-enter the guest with an unmasked signal pending to complete
1411 /* KVM_EXIT_HYPERCALL */
1420 Unused. This was once used for 'hypercall to userspace'. To implement
1421 such functionality, use KVM_EXIT_IO (x86) or KVM_EXIT_MMIO (all except s390).
1422 Note KVM_EXIT_IO is significantly faster than KVM_EXIT_MMIO.
1424 /* KVM_EXIT_TPR_ACCESS */
1431 To be documented (KVM_TPR_ACCESS_REPORTING).
1433 /* KVM_EXIT_S390_SIEIC */
1436 __u64 mask; /* psw upper half */
1437 __u64 addr; /* psw lower half */
1444 /* KVM_EXIT_S390_RESET */
1445 #define KVM_S390_RESET_POR 1
1446 #define KVM_S390_RESET_CLEAR 2
1447 #define KVM_S390_RESET_SUBSYSTEM 4
1448 #define KVM_S390_RESET_CPU_INIT 8
1449 #define KVM_S390_RESET_IPL 16
1450 __u64 s390_reset_flags;
1468 MOL uses a special hypercall interface it calls 'OSI'. To enable it, we catch
1469 hypercalls and exit with this exit struct that contains all the guest gprs.
1471 If exit_reason is KVM_EXIT_OSI, then the vcpu has triggered such a hypercall.
1472 Userspace can now handle the hypercall and when it's done modify the gprs as
1473 necessary. Upon guest entry all guest GPRs will then be replaced by the values
1476 /* Fix the size of the union. */