1 The Definitive KVM (Kernel-based Virtual Machine) API Documentation
2 ===================================================================
7 The kvm API is a set of ioctls that are issued to control various aspects
8 of a virtual machine. The ioctls belong to three classes
10 - System ioctls: These query and set global attributes which affect the
11 whole kvm subsystem. In addition a system ioctl is used to create
14 - VM ioctls: These query and set attributes that affect an entire virtual
15 machine, for example memory layout. In addition a VM ioctl is used to
16 create virtual cpus (vcpus).
18 Only run VM ioctls from the same process (address space) that was used
21 - vcpu ioctls: These query and set attributes that control the operation
22 of a single virtual cpu.
24 Only run vcpu ioctls from the same thread that was used to create the
31 The kvm API is centered around file descriptors. An initial
32 open("/dev/kvm") obtains a handle to the kvm subsystem; this handle
33 can be used to issue system ioctls. A KVM_CREATE_VM ioctl on this
34 handle will create a VM file descriptor which can be used to issue VM
35 ioctls. A KVM_CREATE_VCPU ioctl on a VM fd will create a virtual cpu
36 and return a file descriptor pointing to it. Finally, ioctls on a vcpu
37 fd can be used to control the vcpu, including the important task of
38 actually running guest code.
40 In general file descriptors can be migrated among processes by means
41 of fork() and the SCM_RIGHTS facility of unix domain socket. These
42 kinds of tricks are explicitly not supported by kvm. While they will
43 not cause harm to the host, their actual behavior is not guaranteed by
44 the API. The only supported use is one virtual machine per process,
45 and one vcpu per thread.
51 As of Linux 2.6.22, the KVM ABI has been stabilized: no backward
52 incompatible change are allowed. However, there is an extension
53 facility that allows backward-compatible extensions to the API to be
56 The extension mechanism is not based on the Linux version number.
57 Instead, kvm defines extension identifiers and a facility to query
58 whether a particular extension identifier is available. If it is, a
59 set of ioctls is available for application use.
65 This section describes ioctls that can be used to control kvm guests.
66 For each ioctl, the following information is provided along with a
69 Capability: which KVM extension provides this ioctl. Can be 'basic',
70 which means that is will be provided by any kernel that supports
71 API version 12 (see section 4.1), or a KVM_CAP_xyz constant, which
72 means availability needs to be checked with KVM_CHECK_EXTENSION
75 Architectures: which instruction set architectures provide this ioctl.
76 x86 includes both i386 and x86_64.
78 Type: system, vm, or vcpu.
80 Parameters: what parameters are accepted by the ioctl.
82 Returns: the return value. General error numbers (EBADF, ENOMEM, EINVAL)
83 are not detailed, but errors with specific meanings are.
86 4.1 KVM_GET_API_VERSION
92 Returns: the constant KVM_API_VERSION (=12)
94 This identifies the API version as the stable kvm API. It is not
95 expected that this number will change. However, Linux 2.6.20 and
96 2.6.21 report earlier versions; these are not documented and not
97 supported. Applications should refuse to run if KVM_GET_API_VERSION
98 returns a value other than 12. If this check passes, all ioctls
99 described as 'basic' will be available.
107 Parameters: machine type identifier (KVM_VM_*)
108 Returns: a VM fd that can be used to control the new virtual machine.
110 The new VM has no virtual cpus and no memory. An mmap() of a VM fd
111 will access the virtual machine's physical address space; offset zero
112 corresponds to guest physical address zero. Use of mmap() on a VM fd
113 is discouraged if userspace memory allocation (KVM_CAP_USER_MEMORY) is
115 You most certainly want to use 0 as machine type.
117 In order to create user controlled virtual machines on S390, check
118 KVM_CAP_S390_UCONTROL and use the flag KVM_VM_S390_UCONTROL as
119 privileged user (CAP_SYS_ADMIN).
122 4.3 KVM_GET_MSR_INDEX_LIST
127 Parameters: struct kvm_msr_list (in/out)
128 Returns: 0 on success; -1 on error
130 E2BIG: the msr index list is to be to fit in the array specified by
133 struct kvm_msr_list {
134 __u32 nmsrs; /* number of msrs in entries */
138 This ioctl returns the guest msrs that are supported. The list varies
139 by kvm version and host processor, but does not change otherwise. The
140 user fills in the size of the indices array in nmsrs, and in return
141 kvm adjusts nmsrs to reflect the actual number of msrs and fills in
142 the indices array with their numbers.
144 Note: if kvm indicates supports MCE (KVM_CAP_MCE), then the MCE bank MSRs are
145 not returned in the MSR list, as different vcpus can have a different number
146 of banks, as set via the KVM_X86_SETUP_MCE ioctl.
149 4.4 KVM_CHECK_EXTENSION
154 Parameters: extension identifier (KVM_CAP_*)
155 Returns: 0 if unsupported; 1 (or some other positive integer) if supported
157 The API allows the application to query about extensions to the core
158 kvm API. Userspace passes an extension identifier (an integer) and
159 receives an integer that describes the extension availability.
160 Generally 0 means no and 1 means yes, but some extensions may report
161 additional information in the integer return value.
164 4.5 KVM_GET_VCPU_MMAP_SIZE
170 Returns: size of vcpu mmap area, in bytes
172 The KVM_RUN ioctl (cf.) communicates with userspace via a shared
173 memory region. This ioctl returns the size of that region. See the
174 KVM_RUN documentation for details.
177 4.6 KVM_SET_MEMORY_REGION
182 Parameters: struct kvm_memory_region (in)
183 Returns: 0 on success, -1 on error
185 This ioctl is obsolete and has been removed.
193 Parameters: vcpu id (apic id on x86)
194 Returns: vcpu fd on success, -1 on error
196 This API adds a vcpu to a virtual machine. The vcpu id is a small integer
197 in the range [0, max_vcpus).
199 The recommended max_vcpus value can be retrieved using the KVM_CAP_NR_VCPUS of
200 the KVM_CHECK_EXTENSION ioctl() at run-time.
201 The maximum possible value for max_vcpus can be retrieved using the
202 KVM_CAP_MAX_VCPUS of the KVM_CHECK_EXTENSION ioctl() at run-time.
204 If the KVM_CAP_NR_VCPUS does not exist, you should assume that max_vcpus is 4
206 If the KVM_CAP_MAX_VCPUS does not exist, you should assume that max_vcpus is
207 same as the value returned from KVM_CAP_NR_VCPUS.
209 On powerpc using book3s_hv mode, the vcpus are mapped onto virtual
210 threads in one or more virtual CPU cores. (This is because the
211 hardware requires all the hardware threads in a CPU core to be in the
212 same partition.) The KVM_CAP_PPC_SMT capability indicates the number
213 of vcpus per virtual core (vcore). The vcore id is obtained by
214 dividing the vcpu id by the number of vcpus per vcore. The vcpus in a
215 given vcore will always be in the same physical core as each other
216 (though that might be a different physical core from time to time).
217 Userspace can control the threading (SMT) mode of the guest by its
218 allocation of vcpu ids. For example, if userspace wants
219 single-threaded guest vcpus, it should make all vcpu ids be a multiple
220 of the number of vcpus per vcore.
222 For virtual cpus that have been created with S390 user controlled virtual
223 machines, the resulting vcpu fd can be memory mapped at page offset
224 KVM_S390_SIE_PAGE_OFFSET in order to obtain a memory map of the virtual
225 cpu's hardware control block.
228 4.8 KVM_GET_DIRTY_LOG (vm ioctl)
233 Parameters: struct kvm_dirty_log (in/out)
234 Returns: 0 on success, -1 on error
236 /* for KVM_GET_DIRTY_LOG */
237 struct kvm_dirty_log {
241 void __user *dirty_bitmap; /* one bit per page */
246 Given a memory slot, return a bitmap containing any pages dirtied
247 since the last call to this ioctl. Bit 0 is the first page in the
248 memory slot. Ensure the entire structure is cleared to avoid padding
252 4.9 KVM_SET_MEMORY_ALIAS
257 Parameters: struct kvm_memory_alias (in)
258 Returns: 0 (success), -1 (error)
260 This ioctl is obsolete and has been removed.
269 Returns: 0 on success, -1 on error
271 EINTR: an unmasked signal is pending
273 This ioctl is used to run a guest virtual cpu. While there are no
274 explicit parameters, there is an implicit parameter block that can be
275 obtained by mmap()ing the vcpu fd at offset 0, with the size given by
276 KVM_GET_VCPU_MMAP_SIZE. The parameter block is formatted as a 'struct
277 kvm_run' (see below).
283 Architectures: all except ARM, arm64
285 Parameters: struct kvm_regs (out)
286 Returns: 0 on success, -1 on error
288 Reads the general purpose registers from the vcpu.
292 /* out (KVM_GET_REGS) / in (KVM_SET_REGS) */
293 __u64 rax, rbx, rcx, rdx;
294 __u64 rsi, rdi, rsp, rbp;
295 __u64 r8, r9, r10, r11;
296 __u64 r12, r13, r14, r15;
304 Architectures: all except ARM, arm64
306 Parameters: struct kvm_regs (in)
307 Returns: 0 on success, -1 on error
309 Writes the general purpose registers into the vcpu.
311 See KVM_GET_REGS for the data structure.
317 Architectures: x86, ppc
319 Parameters: struct kvm_sregs (out)
320 Returns: 0 on success, -1 on error
322 Reads special registers from the vcpu.
326 struct kvm_segment cs, ds, es, fs, gs, ss;
327 struct kvm_segment tr, ldt;
328 struct kvm_dtable gdt, idt;
329 __u64 cr0, cr2, cr3, cr4, cr8;
332 __u64 interrupt_bitmap[(KVM_NR_INTERRUPTS + 63) / 64];
335 /* ppc -- see arch/powerpc/include/uapi/asm/kvm.h */
337 interrupt_bitmap is a bitmap of pending external interrupts. At most
338 one bit may be set. This interrupt has been acknowledged by the APIC
339 but not yet injected into the cpu core.
345 Architectures: x86, ppc
347 Parameters: struct kvm_sregs (in)
348 Returns: 0 on success, -1 on error
350 Writes special registers into the vcpu. See KVM_GET_SREGS for the
359 Parameters: struct kvm_translation (in/out)
360 Returns: 0 on success, -1 on error
362 Translates a virtual address according to the vcpu's current address
365 struct kvm_translation {
367 __u64 linear_address;
370 __u64 physical_address;
381 Architectures: x86, ppc
383 Parameters: struct kvm_interrupt (in)
384 Returns: 0 on success, -1 on error
386 Queues a hardware interrupt vector to be injected. This is only
387 useful if in-kernel local APIC or equivalent is not used.
389 /* for KVM_INTERRUPT */
390 struct kvm_interrupt {
397 Note 'irq' is an interrupt vector, not an interrupt pin or line.
401 Queues an external interrupt to be injected. This ioctl is overleaded
402 with 3 different irq values:
406 This injects an edge type external interrupt into the guest once it's ready
407 to receive interrupts. When injected, the interrupt is done.
409 b) KVM_INTERRUPT_UNSET
411 This unsets any pending interrupt.
413 Only available with KVM_CAP_PPC_UNSET_IRQ.
415 c) KVM_INTERRUPT_SET_LEVEL
417 This injects a level type external interrupt into the guest context. The
418 interrupt stays pending until a specific ioctl with KVM_INTERRUPT_UNSET
421 Only available with KVM_CAP_PPC_IRQ_LEVEL.
423 Note that any value for 'irq' other than the ones stated above is invalid
424 and incurs unexpected behavior.
435 Support for this has been removed. Use KVM_SET_GUEST_DEBUG instead.
443 Parameters: struct kvm_msrs (in/out)
444 Returns: 0 on success, -1 on error
446 Reads model-specific registers from the vcpu. Supported msr indices can
447 be obtained using KVM_GET_MSR_INDEX_LIST.
450 __u32 nmsrs; /* number of msrs in entries */
453 struct kvm_msr_entry entries[0];
456 struct kvm_msr_entry {
462 Application code should set the 'nmsrs' member (which indicates the
463 size of the entries array) and the 'index' member of each array entry.
464 kvm will fill in the 'data' member.
472 Parameters: struct kvm_msrs (in)
473 Returns: 0 on success, -1 on error
475 Writes model-specific registers to the vcpu. See KVM_GET_MSRS for the
478 Application code should set the 'nmsrs' member (which indicates the
479 size of the entries array), and the 'index' and 'data' members of each
488 Parameters: struct kvm_cpuid (in)
489 Returns: 0 on success, -1 on error
491 Defines the vcpu responses to the cpuid instruction. Applications
492 should use the KVM_SET_CPUID2 ioctl if available.
495 struct kvm_cpuid_entry {
504 /* for KVM_SET_CPUID */
508 struct kvm_cpuid_entry entries[0];
512 4.21 KVM_SET_SIGNAL_MASK
517 Parameters: struct kvm_signal_mask (in)
518 Returns: 0 on success, -1 on error
520 Defines which signals are blocked during execution of KVM_RUN. This
521 signal mask temporarily overrides the threads signal mask. Any
522 unblocked signal received (except SIGKILL and SIGSTOP, which retain
523 their traditional behaviour) will cause KVM_RUN to return with -EINTR.
525 Note the signal will only be delivered if not blocked by the original
528 /* for KVM_SET_SIGNAL_MASK */
529 struct kvm_signal_mask {
540 Parameters: struct kvm_fpu (out)
541 Returns: 0 on success, -1 on error
543 Reads the floating point state from the vcpu.
545 /* for KVM_GET_FPU and KVM_SET_FPU */
550 __u8 ftwx; /* in fxsave format */
566 Parameters: struct kvm_fpu (in)
567 Returns: 0 on success, -1 on error
569 Writes the floating point state to the vcpu.
571 /* for KVM_GET_FPU and KVM_SET_FPU */
576 __u8 ftwx; /* in fxsave format */
587 4.24 KVM_CREATE_IRQCHIP
589 Capability: KVM_CAP_IRQCHIP
590 Architectures: x86, ia64, ARM, arm64
593 Returns: 0 on success, -1 on error
595 Creates an interrupt controller model in the kernel. On x86, creates a virtual
596 ioapic, a virtual PIC (two PICs, nested), and sets up future vcpus to have a
597 local APIC. IRQ routing for GSIs 0-15 is set to both PIC and IOAPIC; GSI 16-23
598 only go to the IOAPIC. On ia64, a IOSAPIC is created. On ARM/arm64, a GIC is
604 Capability: KVM_CAP_IRQCHIP
605 Architectures: x86, ia64, arm, arm64
607 Parameters: struct kvm_irq_level
608 Returns: 0 on success, -1 on error
610 Sets the level of a GSI input to the interrupt controller model in the kernel.
611 On some architectures it is required that an interrupt controller model has
612 been previously created with KVM_CREATE_IRQCHIP. Note that edge-triggered
613 interrupts require the level to be set to 1 and then back to 0.
615 On real hardware, interrupt pins can be active-low or active-high. This
616 does not matter for the level field of struct kvm_irq_level: 1 always
617 means active (asserted), 0 means inactive (deasserted).
619 x86 allows the operating system to program the interrupt polarity
620 (active-low/active-high) for level-triggered interrupts, and KVM used
621 to consider the polarity. However, due to bitrot in the handling of
622 active-low interrupts, the above convention is now valid on x86 too.
623 This is signaled by KVM_CAP_X86_IOAPIC_POLARITY_IGNORED. Userspace
624 should not present interrupts to the guest as active-low unless this
625 capability is present (or unless it is not using the in-kernel irqchip,
629 ARM/arm64 can signal an interrupt either at the CPU level, or at the
630 in-kernel irqchip (GIC), and for in-kernel irqchip can tell the GIC to
631 use PPIs designated for specific cpus. The irq field is interpreted
634 Â bits: | 31 ... 24 | 23 ... 16 | 15 ... 0 |
635 field: | irq_type | vcpu_index | irq_id |
637 The irq_type field has the following values:
638 - irq_type[0]: out-of-kernel GIC: irq_id 0 is IRQ, irq_id 1 is FIQ
639 - irq_type[1]: in-kernel GIC: SPI, irq_id between 32 and 1019 (incl.)
640 (the vcpu_index field is ignored)
641 - irq_type[2]: in-kernel GIC: PPI, irq_id between 16 and 31 (incl.)
643 (The irq_id field thus corresponds nicely to the IRQ ID in the ARM GIC specs)
645 In both cases, level is used to assert/deassert the line.
647 struct kvm_irq_level {
650 __s32 status; /* not used for KVM_IRQ_LEVEL */
652 __u32 level; /* 0 or 1 */
658 Capability: KVM_CAP_IRQCHIP
659 Architectures: x86, ia64
661 Parameters: struct kvm_irqchip (in/out)
662 Returns: 0 on success, -1 on error
664 Reads the state of a kernel interrupt controller created with
665 KVM_CREATE_IRQCHIP into a buffer provided by the caller.
668 __u32 chip_id; /* 0 = PIC1, 1 = PIC2, 2 = IOAPIC */
671 char dummy[512]; /* reserving space */
672 struct kvm_pic_state pic;
673 struct kvm_ioapic_state ioapic;
680 Capability: KVM_CAP_IRQCHIP
681 Architectures: x86, ia64
683 Parameters: struct kvm_irqchip (in)
684 Returns: 0 on success, -1 on error
686 Sets the state of a kernel interrupt controller created with
687 KVM_CREATE_IRQCHIP from a buffer provided by the caller.
690 __u32 chip_id; /* 0 = PIC1, 1 = PIC2, 2 = IOAPIC */
693 char dummy[512]; /* reserving space */
694 struct kvm_pic_state pic;
695 struct kvm_ioapic_state ioapic;
700 4.28 KVM_XEN_HVM_CONFIG
702 Capability: KVM_CAP_XEN_HVM
705 Parameters: struct kvm_xen_hvm_config (in)
706 Returns: 0 on success, -1 on error
708 Sets the MSR that the Xen HVM guest uses to initialize its hypercall
709 page, and provides the starting address and size of the hypercall
710 blobs in userspace. When the guest writes the MSR, kvm copies one
711 page of a blob (32- or 64-bit, depending on the vcpu mode) to guest
714 struct kvm_xen_hvm_config {
727 Capability: KVM_CAP_ADJUST_CLOCK
730 Parameters: struct kvm_clock_data (out)
731 Returns: 0 on success, -1 on error
733 Gets the current timestamp of kvmclock as seen by the current guest. In
734 conjunction with KVM_SET_CLOCK, it is used to ensure monotonicity on scenarios
737 struct kvm_clock_data {
738 __u64 clock; /* kvmclock current value */
746 Capability: KVM_CAP_ADJUST_CLOCK
749 Parameters: struct kvm_clock_data (in)
750 Returns: 0 on success, -1 on error
752 Sets the current timestamp of kvmclock to the value specified in its parameter.
753 In conjunction with KVM_GET_CLOCK, it is used to ensure monotonicity on scenarios
756 struct kvm_clock_data {
757 __u64 clock; /* kvmclock current value */
763 4.31 KVM_GET_VCPU_EVENTS
765 Capability: KVM_CAP_VCPU_EVENTS
766 Extended by: KVM_CAP_INTR_SHADOW
769 Parameters: struct kvm_vcpu_event (out)
770 Returns: 0 on success, -1 on error
772 Gets currently pending exceptions, interrupts, and NMIs as well as related
775 struct kvm_vcpu_events {
799 KVM_VCPUEVENT_VALID_SHADOW may be set in the flags field to signal that
800 interrupt.shadow contains a valid state. Otherwise, this field is undefined.
803 4.32 KVM_SET_VCPU_EVENTS
805 Capability: KVM_CAP_VCPU_EVENTS
806 Extended by: KVM_CAP_INTR_SHADOW
809 Parameters: struct kvm_vcpu_event (in)
810 Returns: 0 on success, -1 on error
812 Set pending exceptions, interrupts, and NMIs as well as related states of the
815 See KVM_GET_VCPU_EVENTS for the data structure.
817 Fields that may be modified asynchronously by running VCPUs can be excluded
818 from the update. These fields are nmi.pending and sipi_vector. Keep the
819 corresponding bits in the flags field cleared to suppress overwriting the
820 current in-kernel state. The bits are:
822 KVM_VCPUEVENT_VALID_NMI_PENDING - transfer nmi.pending to the kernel
823 KVM_VCPUEVENT_VALID_SIPI_VECTOR - transfer sipi_vector
825 If KVM_CAP_INTR_SHADOW is available, KVM_VCPUEVENT_VALID_SHADOW can be set in
826 the flags field to signal that interrupt.shadow contains a valid state and
827 shall be written into the VCPU.
830 4.33 KVM_GET_DEBUGREGS
832 Capability: KVM_CAP_DEBUGREGS
835 Parameters: struct kvm_debugregs (out)
836 Returns: 0 on success, -1 on error
838 Reads debug registers from the vcpu.
840 struct kvm_debugregs {
849 4.34 KVM_SET_DEBUGREGS
851 Capability: KVM_CAP_DEBUGREGS
854 Parameters: struct kvm_debugregs (in)
855 Returns: 0 on success, -1 on error
857 Writes debug registers into the vcpu.
859 See KVM_GET_DEBUGREGS for the data structure. The flags field is unused
860 yet and must be cleared on entry.
863 4.35 KVM_SET_USER_MEMORY_REGION
865 Capability: KVM_CAP_USER_MEM
868 Parameters: struct kvm_userspace_memory_region (in)
869 Returns: 0 on success, -1 on error
871 struct kvm_userspace_memory_region {
874 __u64 guest_phys_addr;
875 __u64 memory_size; /* bytes */
876 __u64 userspace_addr; /* start of the userspace allocated memory */
879 /* for kvm_memory_region::flags */
880 #define KVM_MEM_LOG_DIRTY_PAGES (1UL << 0)
881 #define KVM_MEM_READONLY (1UL << 1)
883 This ioctl allows the user to create or modify a guest physical memory
884 slot. When changing an existing slot, it may be moved in the guest
885 physical memory space, or its flags may be modified. It may not be
886 resized. Slots may not overlap in guest physical address space.
888 Memory for the region is taken starting at the address denoted by the
889 field userspace_addr, which must point at user addressable memory for
890 the entire memory slot size. Any object may back this memory, including
891 anonymous memory, ordinary files, and hugetlbfs.
893 It is recommended that the lower 21 bits of guest_phys_addr and userspace_addr
894 be identical. This allows large pages in the guest to be backed by large
897 The flags field supports two flags: KVM_MEM_LOG_DIRTY_PAGES and
898 KVM_MEM_READONLY. The former can be set to instruct KVM to keep track of
899 writes to memory within the slot. See KVM_GET_DIRTY_LOG ioctl to know how to
900 use it. The latter can be set, if KVM_CAP_READONLY_MEM capability allows it,
901 to make a new slot read-only. In this case, writes to this memory will be
902 posted to userspace as KVM_EXIT_MMIO exits.
904 When the KVM_CAP_SYNC_MMU capability is available, changes in the backing of
905 the memory region are automatically reflected into the guest. For example, an
906 mmap() that affects the region will be made visible immediately. Another
907 example is madvise(MADV_DROP).
909 It is recommended to use this API instead of the KVM_SET_MEMORY_REGION ioctl.
910 The KVM_SET_MEMORY_REGION does not allow fine grained control over memory
911 allocation and is deprecated.
914 4.36 KVM_SET_TSS_ADDR
916 Capability: KVM_CAP_SET_TSS_ADDR
919 Parameters: unsigned long tss_address (in)
920 Returns: 0 on success, -1 on error
922 This ioctl defines the physical address of a three-page region in the guest
923 physical address space. The region must be within the first 4GB of the
924 guest physical address space and must not conflict with any memory slot
925 or any mmio address. The guest may malfunction if it accesses this memory
928 This ioctl is required on Intel-based hosts. This is needed on Intel hardware
929 because of a quirk in the virtualization implementation (see the internals
930 documentation when it pops into existence).
935 Capability: KVM_CAP_ENABLE_CAP, KVM_CAP_ENABLE_CAP_VM
936 Architectures: ppc, s390
937 Type: vcpu ioctl, vm ioctl (with KVM_CAP_ENABLE_CAP_VM)
938 Parameters: struct kvm_enable_cap (in)
939 Returns: 0 on success; -1 on error
941 +Not all extensions are enabled by default. Using this ioctl the application
942 can enable an extension, making it available to the guest.
944 On systems that do not support this ioctl, it always fails. On systems that
945 do support it, it only works for extensions that are supported for enablement.
947 To check if a capability can be enabled, the KVM_CHECK_EXTENSION ioctl should
950 struct kvm_enable_cap {
954 The capability that is supposed to get enabled.
958 A bitfield indicating future enhancements. Has to be 0 for now.
962 Arguments for enabling a feature. If a feature needs initial values to
963 function properly, this is the place to put them.
968 The vcpu ioctl should be used for vcpu-specific capabilities, the vm ioctl
969 for vm-wide capabilities.
971 4.38 KVM_GET_MP_STATE
973 Capability: KVM_CAP_MP_STATE
974 Architectures: x86, ia64
976 Parameters: struct kvm_mp_state (out)
977 Returns: 0 on success; -1 on error
979 struct kvm_mp_state {
983 Returns the vcpu's current "multiprocessing state" (though also valid on
984 uniprocessor guests).
988 - KVM_MP_STATE_RUNNABLE: the vcpu is currently running
989 - KVM_MP_STATE_UNINITIALIZED: the vcpu is an application processor (AP)
990 which has not yet received an INIT signal
991 - KVM_MP_STATE_INIT_RECEIVED: the vcpu has received an INIT signal, and is
993 - KVM_MP_STATE_HALTED: the vcpu has executed a HLT instruction and
994 is waiting for an interrupt
995 - KVM_MP_STATE_SIPI_RECEIVED: the vcpu has just received a SIPI (vector
996 accessible via KVM_GET_VCPU_EVENTS)
998 This ioctl is only useful after KVM_CREATE_IRQCHIP. Without an in-kernel
999 irqchip, the multiprocessing state must be maintained by userspace.
1002 4.39 KVM_SET_MP_STATE
1004 Capability: KVM_CAP_MP_STATE
1005 Architectures: x86, ia64
1007 Parameters: struct kvm_mp_state (in)
1008 Returns: 0 on success; -1 on error
1010 Sets the vcpu's current "multiprocessing state"; see KVM_GET_MP_STATE for
1013 This ioctl is only useful after KVM_CREATE_IRQCHIP. Without an in-kernel
1014 irqchip, the multiprocessing state must be maintained by userspace.
1017 4.40 KVM_SET_IDENTITY_MAP_ADDR
1019 Capability: KVM_CAP_SET_IDENTITY_MAP_ADDR
1022 Parameters: unsigned long identity (in)
1023 Returns: 0 on success, -1 on error
1025 This ioctl defines the physical address of a one-page region in the guest
1026 physical address space. The region must be within the first 4GB of the
1027 guest physical address space and must not conflict with any memory slot
1028 or any mmio address. The guest may malfunction if it accesses this memory
1031 This ioctl is required on Intel-based hosts. This is needed on Intel hardware
1032 because of a quirk in the virtualization implementation (see the internals
1033 documentation when it pops into existence).
1036 4.41 KVM_SET_BOOT_CPU_ID
1038 Capability: KVM_CAP_SET_BOOT_CPU_ID
1039 Architectures: x86, ia64
1041 Parameters: unsigned long vcpu_id
1042 Returns: 0 on success, -1 on error
1044 Define which vcpu is the Bootstrap Processor (BSP). Values are the same
1045 as the vcpu id in KVM_CREATE_VCPU. If this ioctl is not called, the default
1051 Capability: KVM_CAP_XSAVE
1054 Parameters: struct kvm_xsave (out)
1055 Returns: 0 on success, -1 on error
1061 This ioctl would copy current vcpu's xsave struct to the userspace.
1066 Capability: KVM_CAP_XSAVE
1069 Parameters: struct kvm_xsave (in)
1070 Returns: 0 on success, -1 on error
1076 This ioctl would copy userspace's xsave struct to the kernel.
1081 Capability: KVM_CAP_XCRS
1084 Parameters: struct kvm_xcrs (out)
1085 Returns: 0 on success, -1 on error
1096 struct kvm_xcr xcrs[KVM_MAX_XCRS];
1100 This ioctl would copy current vcpu's xcrs to the userspace.
1105 Capability: KVM_CAP_XCRS
1108 Parameters: struct kvm_xcrs (in)
1109 Returns: 0 on success, -1 on error
1120 struct kvm_xcr xcrs[KVM_MAX_XCRS];
1124 This ioctl would set vcpu's xcr to the value userspace specified.
1127 4.46 KVM_GET_SUPPORTED_CPUID
1129 Capability: KVM_CAP_EXT_CPUID
1132 Parameters: struct kvm_cpuid2 (in/out)
1133 Returns: 0 on success, -1 on error
1138 struct kvm_cpuid_entry2 entries[0];
1141 #define KVM_CPUID_FLAG_SIGNIFCANT_INDEX BIT(0)
1142 #define KVM_CPUID_FLAG_STATEFUL_FUNC BIT(1)
1143 #define KVM_CPUID_FLAG_STATE_READ_NEXT BIT(2)
1145 struct kvm_cpuid_entry2 {
1156 This ioctl returns x86 cpuid features which are supported by both the hardware
1157 and kvm. Userspace can use the information returned by this ioctl to
1158 construct cpuid information (for KVM_SET_CPUID2) that is consistent with
1159 hardware, kernel, and userspace capabilities, and with user requirements (for
1160 example, the user may wish to constrain cpuid to emulate older hardware,
1161 or for feature consistency across a cluster).
1163 Userspace invokes KVM_GET_SUPPORTED_CPUID by passing a kvm_cpuid2 structure
1164 with the 'nent' field indicating the number of entries in the variable-size
1165 array 'entries'. If the number of entries is too low to describe the cpu
1166 capabilities, an error (E2BIG) is returned. If the number is too high,
1167 the 'nent' field is adjusted and an error (ENOMEM) is returned. If the
1168 number is just right, the 'nent' field is adjusted to the number of valid
1169 entries in the 'entries' array, which is then filled.
1171 The entries returned are the host cpuid as returned by the cpuid instruction,
1172 with unknown or unsupported features masked out. Some features (for example,
1173 x2apic), may not be present in the host cpu, but are exposed by kvm if it can
1174 emulate them efficiently. The fields in each entry are defined as follows:
1176 function: the eax value used to obtain the entry
1177 index: the ecx value used to obtain the entry (for entries that are
1179 flags: an OR of zero or more of the following:
1180 KVM_CPUID_FLAG_SIGNIFCANT_INDEX:
1181 if the index field is valid
1182 KVM_CPUID_FLAG_STATEFUL_FUNC:
1183 if cpuid for this function returns different values for successive
1184 invocations; there will be several entries with the same function,
1185 all with this flag set
1186 KVM_CPUID_FLAG_STATE_READ_NEXT:
1187 for KVM_CPUID_FLAG_STATEFUL_FUNC entries, set if this entry is
1188 the first entry to be read by a cpu
1189 eax, ebx, ecx, edx: the values returned by the cpuid instruction for
1190 this function/index combination
1192 The TSC deadline timer feature (CPUID leaf 1, ecx[24]) is always returned
1193 as false, since the feature depends on KVM_CREATE_IRQCHIP for local APIC
1194 support. Instead it is reported via
1196 ioctl(KVM_CHECK_EXTENSION, KVM_CAP_TSC_DEADLINE_TIMER)
1198 if that returns true and you use KVM_CREATE_IRQCHIP, or if you emulate the
1199 feature in userspace, then you can enable the feature for KVM_SET_CPUID2.
1202 4.47 KVM_PPC_GET_PVINFO
1204 Capability: KVM_CAP_PPC_GET_PVINFO
1207 Parameters: struct kvm_ppc_pvinfo (out)
1208 Returns: 0 on success, !0 on error
1210 struct kvm_ppc_pvinfo {
1216 This ioctl fetches PV specific information that need to be passed to the guest
1217 using the device tree or other means from vm context.
1219 The hcall array defines 4 instructions that make up a hypercall.
1221 If any additional field gets added to this structure later on, a bit for that
1222 additional piece of information will be set in the flags bitmap.
1224 The flags bitmap is defined as:
1226 /* the host supports the ePAPR idle hcall
1227 #define KVM_PPC_PVINFO_FLAGS_EV_IDLE (1<<0)
1229 4.48 KVM_ASSIGN_PCI_DEVICE
1231 Capability: KVM_CAP_DEVICE_ASSIGNMENT
1232 Architectures: x86 ia64
1234 Parameters: struct kvm_assigned_pci_dev (in)
1235 Returns: 0 on success, -1 on error
1237 Assigns a host PCI device to the VM.
1239 struct kvm_assigned_pci_dev {
1240 __u32 assigned_dev_id;
1250 The PCI device is specified by the triple segnr, busnr, and devfn.
1251 Identification in succeeding service requests is done via assigned_dev_id. The
1252 following flags are specified:
1254 /* Depends on KVM_CAP_IOMMU */
1255 #define KVM_DEV_ASSIGN_ENABLE_IOMMU (1 << 0)
1256 /* The following two depend on KVM_CAP_PCI_2_3 */
1257 #define KVM_DEV_ASSIGN_PCI_2_3 (1 << 1)
1258 #define KVM_DEV_ASSIGN_MASK_INTX (1 << 2)
1260 If KVM_DEV_ASSIGN_PCI_2_3 is set, the kernel will manage legacy INTx interrupts
1261 via the PCI-2.3-compliant device-level mask, thus enable IRQ sharing with other
1262 assigned devices or host devices. KVM_DEV_ASSIGN_MASK_INTX specifies the
1263 guest's view on the INTx mask, see KVM_ASSIGN_SET_INTX_MASK for details.
1265 The KVM_DEV_ASSIGN_ENABLE_IOMMU flag is a mandatory option to ensure
1266 isolation of the device. Usages not specifying this flag are deprecated.
1268 Only PCI header type 0 devices with PCI BAR resources are supported by
1269 device assignment. The user requesting this ioctl must have read/write
1270 access to the PCI sysfs resource files associated with the device.
1273 4.49 KVM_DEASSIGN_PCI_DEVICE
1275 Capability: KVM_CAP_DEVICE_DEASSIGNMENT
1276 Architectures: x86 ia64
1278 Parameters: struct kvm_assigned_pci_dev (in)
1279 Returns: 0 on success, -1 on error
1281 Ends PCI device assignment, releasing all associated resources.
1283 See KVM_CAP_DEVICE_ASSIGNMENT for the data structure. Only assigned_dev_id is
1284 used in kvm_assigned_pci_dev to identify the device.
1287 4.50 KVM_ASSIGN_DEV_IRQ
1289 Capability: KVM_CAP_ASSIGN_DEV_IRQ
1290 Architectures: x86 ia64
1292 Parameters: struct kvm_assigned_irq (in)
1293 Returns: 0 on success, -1 on error
1295 Assigns an IRQ to a passed-through device.
1297 struct kvm_assigned_irq {
1298 __u32 assigned_dev_id;
1299 __u32 host_irq; /* ignored (legacy field) */
1307 The following flags are defined:
1309 #define KVM_DEV_IRQ_HOST_INTX (1 << 0)
1310 #define KVM_DEV_IRQ_HOST_MSI (1 << 1)
1311 #define KVM_DEV_IRQ_HOST_MSIX (1 << 2)
1313 #define KVM_DEV_IRQ_GUEST_INTX (1 << 8)
1314 #define KVM_DEV_IRQ_GUEST_MSI (1 << 9)
1315 #define KVM_DEV_IRQ_GUEST_MSIX (1 << 10)
1317 It is not valid to specify multiple types per host or guest IRQ. However, the
1318 IRQ type of host and guest can differ or can even be null.
1321 4.51 KVM_DEASSIGN_DEV_IRQ
1323 Capability: KVM_CAP_ASSIGN_DEV_IRQ
1324 Architectures: x86 ia64
1326 Parameters: struct kvm_assigned_irq (in)
1327 Returns: 0 on success, -1 on error
1329 Ends an IRQ assignment to a passed-through device.
1331 See KVM_ASSIGN_DEV_IRQ for the data structure. The target device is specified
1332 by assigned_dev_id, flags must correspond to the IRQ type specified on
1333 KVM_ASSIGN_DEV_IRQ. Partial deassignment of host or guest IRQ is allowed.
1336 4.52 KVM_SET_GSI_ROUTING
1338 Capability: KVM_CAP_IRQ_ROUTING
1339 Architectures: x86 ia64
1341 Parameters: struct kvm_irq_routing (in)
1342 Returns: 0 on success, -1 on error
1344 Sets the GSI routing table entries, overwriting any previously set entries.
1346 struct kvm_irq_routing {
1349 struct kvm_irq_routing_entry entries[0];
1352 No flags are specified so far, the corresponding field must be set to zero.
1354 struct kvm_irq_routing_entry {
1360 struct kvm_irq_routing_irqchip irqchip;
1361 struct kvm_irq_routing_msi msi;
1366 /* gsi routing entry types */
1367 #define KVM_IRQ_ROUTING_IRQCHIP 1
1368 #define KVM_IRQ_ROUTING_MSI 2
1370 No flags are specified so far, the corresponding field must be set to zero.
1372 struct kvm_irq_routing_irqchip {
1377 struct kvm_irq_routing_msi {
1385 4.53 KVM_ASSIGN_SET_MSIX_NR
1387 Capability: KVM_CAP_DEVICE_MSIX
1388 Architectures: x86 ia64
1390 Parameters: struct kvm_assigned_msix_nr (in)
1391 Returns: 0 on success, -1 on error
1393 Set the number of MSI-X interrupts for an assigned device. The number is
1394 reset again by terminating the MSI-X assignment of the device via
1395 KVM_DEASSIGN_DEV_IRQ. Calling this service more than once at any earlier
1398 struct kvm_assigned_msix_nr {
1399 __u32 assigned_dev_id;
1404 #define KVM_MAX_MSIX_PER_DEV 256
1407 4.54 KVM_ASSIGN_SET_MSIX_ENTRY
1409 Capability: KVM_CAP_DEVICE_MSIX
1410 Architectures: x86 ia64
1412 Parameters: struct kvm_assigned_msix_entry (in)
1413 Returns: 0 on success, -1 on error
1415 Specifies the routing of an MSI-X assigned device interrupt to a GSI. Setting
1416 the GSI vector to zero means disabling the interrupt.
1418 struct kvm_assigned_msix_entry {
1419 __u32 assigned_dev_id;
1421 __u16 entry; /* The index of entry in the MSI-X table */
1426 4.55 KVM_SET_TSC_KHZ
1428 Capability: KVM_CAP_TSC_CONTROL
1431 Parameters: virtual tsc_khz
1432 Returns: 0 on success, -1 on error
1434 Specifies the tsc frequency for the virtual machine. The unit of the
1438 4.56 KVM_GET_TSC_KHZ
1440 Capability: KVM_CAP_GET_TSC_KHZ
1444 Returns: virtual tsc-khz on success, negative value on error
1446 Returns the tsc frequency of the guest. The unit of the return value is
1447 KHz. If the host has unstable tsc this ioctl returns -EIO instead as an
1453 Capability: KVM_CAP_IRQCHIP
1456 Parameters: struct kvm_lapic_state (out)
1457 Returns: 0 on success, -1 on error
1459 #define KVM_APIC_REG_SIZE 0x400
1460 struct kvm_lapic_state {
1461 char regs[KVM_APIC_REG_SIZE];
1464 Reads the Local APIC registers and copies them into the input argument. The
1465 data format and layout are the same as documented in the architecture manual.
1470 Capability: KVM_CAP_IRQCHIP
1473 Parameters: struct kvm_lapic_state (in)
1474 Returns: 0 on success, -1 on error
1476 #define KVM_APIC_REG_SIZE 0x400
1477 struct kvm_lapic_state {
1478 char regs[KVM_APIC_REG_SIZE];
1481 Copies the input argument into the the Local APIC registers. The data format
1482 and layout are the same as documented in the architecture manual.
1487 Capability: KVM_CAP_IOEVENTFD
1490 Parameters: struct kvm_ioeventfd (in)
1491 Returns: 0 on success, !0 on error
1493 This ioctl attaches or detaches an ioeventfd to a legal pio/mmio address
1494 within the guest. A guest write in the registered address will signal the
1495 provided event instead of triggering an exit.
1497 struct kvm_ioeventfd {
1499 __u64 addr; /* legal pio/mmio address */
1500 __u32 len; /* 1, 2, 4, or 8 bytes */
1506 For the special case of virtio-ccw devices on s390, the ioevent is matched
1507 to a subchannel/virtqueue tuple instead.
1509 The following flags are defined:
1511 #define KVM_IOEVENTFD_FLAG_DATAMATCH (1 << kvm_ioeventfd_flag_nr_datamatch)
1512 #define KVM_IOEVENTFD_FLAG_PIO (1 << kvm_ioeventfd_flag_nr_pio)
1513 #define KVM_IOEVENTFD_FLAG_DEASSIGN (1 << kvm_ioeventfd_flag_nr_deassign)
1514 #define KVM_IOEVENTFD_FLAG_VIRTIO_CCW_NOTIFY \
1515 (1 << kvm_ioeventfd_flag_nr_virtio_ccw_notify)
1517 If datamatch flag is set, the event will be signaled only if the written value
1518 to the registered address is equal to datamatch in struct kvm_ioeventfd.
1520 For virtio-ccw devices, addr contains the subchannel id and datamatch the
1526 Capability: KVM_CAP_SW_TLB
1529 Parameters: struct kvm_dirty_tlb (in)
1530 Returns: 0 on success, -1 on error
1532 struct kvm_dirty_tlb {
1537 This must be called whenever userspace has changed an entry in the shared
1538 TLB, prior to calling KVM_RUN on the associated vcpu.
1540 The "bitmap" field is the userspace address of an array. This array
1541 consists of a number of bits, equal to the total number of TLB entries as
1542 determined by the last successful call to KVM_CONFIG_TLB, rounded up to the
1543 nearest multiple of 64.
1545 Each bit corresponds to one TLB entry, ordered the same as in the shared TLB
1548 The array is little-endian: the bit 0 is the least significant bit of the
1549 first byte, bit 8 is the least significant bit of the second byte, etc.
1550 This avoids any complications with differing word sizes.
1552 The "num_dirty" field is a performance hint for KVM to determine whether it
1553 should skip processing the bitmap and just invalidate everything. It must
1554 be set to the number of set bits in the bitmap.
1557 4.61 KVM_ASSIGN_SET_INTX_MASK
1559 Capability: KVM_CAP_PCI_2_3
1562 Parameters: struct kvm_assigned_pci_dev (in)
1563 Returns: 0 on success, -1 on error
1565 Allows userspace to mask PCI INTx interrupts from the assigned device. The
1566 kernel will not deliver INTx interrupts to the guest between setting and
1567 clearing of KVM_ASSIGN_SET_INTX_MASK via this interface. This enables use of
1568 and emulation of PCI 2.3 INTx disable command register behavior.
1570 This may be used for both PCI 2.3 devices supporting INTx disable natively and
1571 older devices lacking this support. Userspace is responsible for emulating the
1572 read value of the INTx disable bit in the guest visible PCI command register.
1573 When modifying the INTx disable state, userspace should precede updating the
1574 physical device command register by calling this ioctl to inform the kernel of
1575 the new intended INTx mask state.
1577 Note that the kernel uses the device INTx disable bit to internally manage the
1578 device interrupt state for PCI 2.3 devices. Reads of this register may
1579 therefore not match the expected value. Writes should always use the guest
1580 intended INTx disable value rather than attempting to read-copy-update the
1581 current physical device state. Races between user and kernel updates to the
1582 INTx disable bit are handled lazily in the kernel. It's possible the device
1583 may generate unintended interrupts, but they will not be injected into the
1586 See KVM_ASSIGN_DEV_IRQ for the data structure. The target device is specified
1587 by assigned_dev_id. In the flags field, only KVM_DEV_ASSIGN_MASK_INTX is
1591 4.62 KVM_CREATE_SPAPR_TCE
1593 Capability: KVM_CAP_SPAPR_TCE
1594 Architectures: powerpc
1596 Parameters: struct kvm_create_spapr_tce (in)
1597 Returns: file descriptor for manipulating the created TCE table
1599 This creates a virtual TCE (translation control entry) table, which
1600 is an IOMMU for PAPR-style virtual I/O. It is used to translate
1601 logical addresses used in virtual I/O into guest physical addresses,
1602 and provides a scatter/gather capability for PAPR virtual I/O.
1604 /* for KVM_CAP_SPAPR_TCE */
1605 struct kvm_create_spapr_tce {
1610 The liobn field gives the logical IO bus number for which to create a
1611 TCE table. The window_size field specifies the size of the DMA window
1612 which this TCE table will translate - the table will contain one 64
1613 bit TCE entry for every 4kiB of the DMA window.
1615 When the guest issues an H_PUT_TCE hcall on a liobn for which a TCE
1616 table has been created using this ioctl(), the kernel will handle it
1617 in real mode, updating the TCE table. H_PUT_TCE calls for other
1618 liobns will cause a vm exit and must be handled by userspace.
1620 The return value is a file descriptor which can be passed to mmap(2)
1621 to map the created TCE table into userspace. This lets userspace read
1622 the entries written by kernel-handled H_PUT_TCE calls, and also lets
1623 userspace update the TCE table directly which is useful in some
1627 4.63 KVM_ALLOCATE_RMA
1629 Capability: KVM_CAP_PPC_RMA
1630 Architectures: powerpc
1632 Parameters: struct kvm_allocate_rma (out)
1633 Returns: file descriptor for mapping the allocated RMA
1635 This allocates a Real Mode Area (RMA) from the pool allocated at boot
1636 time by the kernel. An RMA is a physically-contiguous, aligned region
1637 of memory used on older POWER processors to provide the memory which
1638 will be accessed by real-mode (MMU off) accesses in a KVM guest.
1639 POWER processors support a set of sizes for the RMA that usually
1640 includes 64MB, 128MB, 256MB and some larger powers of two.
1642 /* for KVM_ALLOCATE_RMA */
1643 struct kvm_allocate_rma {
1647 The return value is a file descriptor which can be passed to mmap(2)
1648 to map the allocated RMA into userspace. The mapped area can then be
1649 passed to the KVM_SET_USER_MEMORY_REGION ioctl to establish it as the
1650 RMA for a virtual machine. The size of the RMA in bytes (which is
1651 fixed at host kernel boot time) is returned in the rma_size field of
1652 the argument structure.
1654 The KVM_CAP_PPC_RMA capability is 1 or 2 if the KVM_ALLOCATE_RMA ioctl
1655 is supported; 2 if the processor requires all virtual machines to have
1656 an RMA, or 1 if the processor can use an RMA but doesn't require it,
1657 because it supports the Virtual RMA (VRMA) facility.
1662 Capability: KVM_CAP_USER_NMI
1666 Returns: 0 on success, -1 on error
1668 Queues an NMI on the thread's vcpu. Note this is well defined only
1669 when KVM_CREATE_IRQCHIP has not been called, since this is an interface
1670 between the virtual cpu core and virtual local APIC. After KVM_CREATE_IRQCHIP
1671 has been called, this interface is completely emulated within the kernel.
1673 To use this to emulate the LINT1 input with KVM_CREATE_IRQCHIP, use the
1674 following algorithm:
1677 - read the local APIC's state (KVM_GET_LAPIC)
1678 - check whether changing LINT1 will queue an NMI (see the LVT entry for LINT1)
1679 - if so, issue KVM_NMI
1682 Some guests configure the LINT1 NMI input to cause a panic, aiding in
1686 4.65 KVM_S390_UCAS_MAP
1688 Capability: KVM_CAP_S390_UCONTROL
1691 Parameters: struct kvm_s390_ucas_mapping (in)
1692 Returns: 0 in case of success
1694 The parameter is defined like this:
1695 struct kvm_s390_ucas_mapping {
1701 This ioctl maps the memory at "user_addr" with the length "length" to
1702 the vcpu's address space starting at "vcpu_addr". All parameters need to
1703 be aligned by 1 megabyte.
1706 4.66 KVM_S390_UCAS_UNMAP
1708 Capability: KVM_CAP_S390_UCONTROL
1711 Parameters: struct kvm_s390_ucas_mapping (in)
1712 Returns: 0 in case of success
1714 The parameter is defined like this:
1715 struct kvm_s390_ucas_mapping {
1721 This ioctl unmaps the memory in the vcpu's address space starting at
1722 "vcpu_addr" with the length "length". The field "user_addr" is ignored.
1723 All parameters need to be aligned by 1 megabyte.
1726 4.67 KVM_S390_VCPU_FAULT
1728 Capability: KVM_CAP_S390_UCONTROL
1731 Parameters: vcpu absolute address (in)
1732 Returns: 0 in case of success
1734 This call creates a page table entry on the virtual cpu's address space
1735 (for user controlled virtual machines) or the virtual machine's address
1736 space (for regular virtual machines). This only works for minor faults,
1737 thus it's recommended to access subject memory page via the user page
1738 table upfront. This is useful to handle validity intercepts for user
1739 controlled virtual machines to fault in the virtual cpu's lowcore pages
1740 prior to calling the KVM_RUN ioctl.
1743 4.68 KVM_SET_ONE_REG
1745 Capability: KVM_CAP_ONE_REG
1748 Parameters: struct kvm_one_reg (in)
1749 Returns: 0 on success, negative value on failure
1751 struct kvm_one_reg {
1756 Using this ioctl, a single vcpu register can be set to a specific value
1757 defined by user space with the passed in struct kvm_one_reg, where id
1758 refers to the register identifier as described below and addr is a pointer
1759 to a variable with the respective size. There can be architecture agnostic
1760 and architecture specific registers. Each have their own range of operation
1761 and their own constants and width. To keep track of the implemented
1762 registers, find a list below:
1764 Arch | Register | Width (bits)
1766 PPC | KVM_REG_PPC_HIOR | 64
1767 PPC | KVM_REG_PPC_IAC1 | 64
1768 PPC | KVM_REG_PPC_IAC2 | 64
1769 PPC | KVM_REG_PPC_IAC3 | 64
1770 PPC | KVM_REG_PPC_IAC4 | 64
1771 PPC | KVM_REG_PPC_DAC1 | 64
1772 PPC | KVM_REG_PPC_DAC2 | 64
1773 PPC | KVM_REG_PPC_DABR | 64
1774 PPC | KVM_REG_PPC_DSCR | 64
1775 PPC | KVM_REG_PPC_PURR | 64
1776 PPC | KVM_REG_PPC_SPURR | 64
1777 PPC | KVM_REG_PPC_DAR | 64
1778 PPC | KVM_REG_PPC_DSISR | 32
1779 PPC | KVM_REG_PPC_AMR | 64
1780 PPC | KVM_REG_PPC_UAMOR | 64
1781 PPC | KVM_REG_PPC_MMCR0 | 64
1782 PPC | KVM_REG_PPC_MMCR1 | 64
1783 PPC | KVM_REG_PPC_MMCRA | 64
1784 PPC | KVM_REG_PPC_PMC1 | 32
1785 PPC | KVM_REG_PPC_PMC2 | 32
1786 PPC | KVM_REG_PPC_PMC3 | 32
1787 PPC | KVM_REG_PPC_PMC4 | 32
1788 PPC | KVM_REG_PPC_PMC5 | 32
1789 PPC | KVM_REG_PPC_PMC6 | 32
1790 PPC | KVM_REG_PPC_PMC7 | 32
1791 PPC | KVM_REG_PPC_PMC8 | 32
1792 PPC | KVM_REG_PPC_FPR0 | 64
1794 PPC | KVM_REG_PPC_FPR31 | 64
1795 PPC | KVM_REG_PPC_VR0 | 128
1797 PPC | KVM_REG_PPC_VR31 | 128
1798 PPC | KVM_REG_PPC_VSR0 | 128
1800 PPC | KVM_REG_PPC_VSR31 | 128
1801 PPC | KVM_REG_PPC_FPSCR | 64
1802 PPC | KVM_REG_PPC_VSCR | 32
1803 PPC | KVM_REG_PPC_VPA_ADDR | 64
1804 PPC | KVM_REG_PPC_VPA_SLB | 128
1805 PPC | KVM_REG_PPC_VPA_DTL | 128
1806 PPC | KVM_REG_PPC_EPCR | 32
1807 PPC | KVM_REG_PPC_EPR | 32
1808 PPC | KVM_REG_PPC_TCR | 32
1809 PPC | KVM_REG_PPC_TSR | 32
1810 PPC | KVM_REG_PPC_OR_TSR | 32
1811 PPC | KVM_REG_PPC_CLEAR_TSR | 32
1812 PPC | KVM_REG_PPC_MAS0 | 32
1813 PPC | KVM_REG_PPC_MAS1 | 32
1814 PPC | KVM_REG_PPC_MAS2 | 64
1815 PPC | KVM_REG_PPC_MAS7_3 | 64
1816 PPC | KVM_REG_PPC_MAS4 | 32
1817 PPC | KVM_REG_PPC_MAS6 | 32
1818 PPC | KVM_REG_PPC_MMUCFG | 32
1819 PPC | KVM_REG_PPC_TLB0CFG | 32
1820 PPC | KVM_REG_PPC_TLB1CFG | 32
1821 PPC | KVM_REG_PPC_TLB2CFG | 32
1822 PPC | KVM_REG_PPC_TLB3CFG | 32
1823 PPC | KVM_REG_PPC_TLB0PS | 32
1824 PPC | KVM_REG_PPC_TLB1PS | 32
1825 PPC | KVM_REG_PPC_TLB2PS | 32
1826 PPC | KVM_REG_PPC_TLB3PS | 32
1827 PPC | KVM_REG_PPC_EPTCFG | 32
1828 PPC | KVM_REG_PPC_ICP_STATE | 64
1829 PPC | KVM_REG_PPC_TB_OFFSET | 64
1830 PPC | KVM_REG_PPC_SPMC1 | 32
1831 PPC | KVM_REG_PPC_SPMC2 | 32
1832 PPC | KVM_REG_PPC_IAMR | 64
1833 PPC | KVM_REG_PPC_TFHAR | 64
1834 PPC | KVM_REG_PPC_TFIAR | 64
1835 PPC | KVM_REG_PPC_TEXASR | 64
1836 PPC | KVM_REG_PPC_FSCR | 64
1837 PPC | KVM_REG_PPC_PSPB | 32
1838 PPC | KVM_REG_PPC_EBBHR | 64
1839 PPC | KVM_REG_PPC_EBBRR | 64
1840 PPC | KVM_REG_PPC_BESCR | 64
1841 PPC | KVM_REG_PPC_TAR | 64
1842 PPC | KVM_REG_PPC_DPDES | 64
1843 PPC | KVM_REG_PPC_DAWR | 64
1844 PPC | KVM_REG_PPC_DAWRX | 64
1845 PPC | KVM_REG_PPC_CIABR | 64
1846 PPC | KVM_REG_PPC_IC | 64
1847 PPC | KVM_REG_PPC_VTB | 64
1848 PPC | KVM_REG_PPC_CSIGR | 64
1849 PPC | KVM_REG_PPC_TACR | 64
1850 PPC | KVM_REG_PPC_TCSCR | 64
1851 PPC | KVM_REG_PPC_PID | 64
1852 PPC | KVM_REG_PPC_ACOP | 64
1853 PPC | KVM_REG_PPC_VRSAVE | 32
1854 PPC | KVM_REG_PPC_LPCR | 64
1855 PPC | KVM_REG_PPC_PPR | 64
1856 PPC | KVM_REG_PPC_ARCH_COMPAT 32
1857 PPC | KVM_REG_PPC_DABRX | 32
1858 PPC | KVM_REG_PPC_TM_GPR0 | 64
1860 PPC | KVM_REG_PPC_TM_GPR31 | 64
1861 PPC | KVM_REG_PPC_TM_VSR0 | 128
1863 PPC | KVM_REG_PPC_TM_VSR63 | 128
1864 PPC | KVM_REG_PPC_TM_CR | 64
1865 PPC | KVM_REG_PPC_TM_LR | 64
1866 PPC | KVM_REG_PPC_TM_CTR | 64
1867 PPC | KVM_REG_PPC_TM_FPSCR | 64
1868 PPC | KVM_REG_PPC_TM_AMR | 64
1869 PPC | KVM_REG_PPC_TM_PPR | 64
1870 PPC | KVM_REG_PPC_TM_VRSAVE | 64
1871 PPC | KVM_REG_PPC_TM_VSCR | 32
1872 PPC | KVM_REG_PPC_TM_DSCR | 64
1873 PPC | KVM_REG_PPC_TM_TAR | 64
1875 ARM registers are mapped using the lower 32 bits. The upper 16 of that
1876 is the register group type, or coprocessor number:
1878 ARM core registers have the following id bit patterns:
1879 0x4020 0000 0010 <index into the kvm_regs struct:16>
1881 ARM 32-bit CP15 registers have the following id bit patterns:
1882 0x4020 0000 000F <zero:1> <crn:4> <crm:4> <opc1:4> <opc2:3>
1884 ARM 64-bit CP15 registers have the following id bit patterns:
1885 0x4030 0000 000F <zero:1> <zero:4> <crm:4> <opc1:4> <zero:3>
1887 ARM CCSIDR registers are demultiplexed by CSSELR value:
1888 0x4020 0000 0011 00 <csselr:8>
1890 ARM 32-bit VFP control registers have the following id bit patterns:
1891 0x4020 0000 0012 1 <regno:12>
1893 ARM 64-bit FP registers have the following id bit patterns:
1894 0x4030 0000 0012 0 <regno:12>
1897 arm64 registers are mapped using the lower 32 bits. The upper 16 of
1898 that is the register group type, or coprocessor number:
1900 arm64 core/FP-SIMD registers have the following id bit patterns. Note
1901 that the size of the access is variable, as the kvm_regs structure
1902 contains elements ranging from 32 to 128 bits. The index is a 32bit
1903 value in the kvm_regs structure seen as a 32bit array.
1904 0x60x0 0000 0010 <index into the kvm_regs struct:16>
1906 arm64 CCSIDR registers are demultiplexed by CSSELR value:
1907 0x6020 0000 0011 00 <csselr:8>
1909 arm64 system registers have the following id bit patterns:
1910 0x6030 0000 0013 <op0:2> <op1:3> <crn:4> <crm:4> <op2:3>
1912 4.69 KVM_GET_ONE_REG
1914 Capability: KVM_CAP_ONE_REG
1917 Parameters: struct kvm_one_reg (in and out)
1918 Returns: 0 on success, negative value on failure
1920 This ioctl allows to receive the value of a single register implemented
1921 in a vcpu. The register to read is indicated by the "id" field of the
1922 kvm_one_reg struct passed in. On success, the register value can be found
1923 at the memory location pointed to by "addr".
1925 The list of registers accessible using this interface is identical to the
1929 4.70 KVM_KVMCLOCK_CTRL
1931 Capability: KVM_CAP_KVMCLOCK_CTRL
1932 Architectures: Any that implement pvclocks (currently x86 only)
1935 Returns: 0 on success, -1 on error
1937 This signals to the host kernel that the specified guest is being paused by
1938 userspace. The host will set a flag in the pvclock structure that is checked
1939 from the soft lockup watchdog. The flag is part of the pvclock structure that
1940 is shared between guest and host, specifically the second bit of the flags
1941 field of the pvclock_vcpu_time_info structure. It will be set exclusively by
1942 the host and read/cleared exclusively by the guest. The guest operation of
1943 checking and clearing the flag must an atomic operation so
1944 load-link/store-conditional, or equivalent must be used. There are two cases
1945 where the guest will clear the flag: when the soft lockup watchdog timer resets
1946 itself or when a soft lockup is detected. This ioctl can be called any time
1947 after pausing the vcpu, but before it is resumed.
1952 Capability: KVM_CAP_SIGNAL_MSI
1955 Parameters: struct kvm_msi (in)
1956 Returns: >0 on delivery, 0 if guest blocked the MSI, and -1 on error
1958 Directly inject a MSI message. Only valid with in-kernel irqchip that handles
1969 No flags are defined so far. The corresponding field must be 0.
1972 4.71 KVM_CREATE_PIT2
1974 Capability: KVM_CAP_PIT2
1977 Parameters: struct kvm_pit_config (in)
1978 Returns: 0 on success, -1 on error
1980 Creates an in-kernel device model for the i8254 PIT. This call is only valid
1981 after enabling in-kernel irqchip support via KVM_CREATE_IRQCHIP. The following
1982 parameters have to be passed:
1984 struct kvm_pit_config {
1991 #define KVM_PIT_SPEAKER_DUMMY 1 /* emulate speaker port stub */
1993 PIT timer interrupts may use a per-VM kernel thread for injection. If it
1994 exists, this thread will have a name of the following pattern:
1996 kvm-pit/<owner-process-pid>
1998 When running a guest with elevated priorities, the scheduling parameters of
1999 this thread may have to be adjusted accordingly.
2001 This IOCTL replaces the obsolete KVM_CREATE_PIT.
2006 Capability: KVM_CAP_PIT_STATE2
2009 Parameters: struct kvm_pit_state2 (out)
2010 Returns: 0 on success, -1 on error
2012 Retrieves the state of the in-kernel PIT model. Only valid after
2013 KVM_CREATE_PIT2. The state is returned in the following structure:
2015 struct kvm_pit_state2 {
2016 struct kvm_pit_channel_state channels[3];
2023 /* disable PIT in HPET legacy mode */
2024 #define KVM_PIT_FLAGS_HPET_LEGACY 0x00000001
2026 This IOCTL replaces the obsolete KVM_GET_PIT.
2031 Capability: KVM_CAP_PIT_STATE2
2034 Parameters: struct kvm_pit_state2 (in)
2035 Returns: 0 on success, -1 on error
2037 Sets the state of the in-kernel PIT model. Only valid after KVM_CREATE_PIT2.
2038 See KVM_GET_PIT2 for details on struct kvm_pit_state2.
2040 This IOCTL replaces the obsolete KVM_SET_PIT.
2043 4.74 KVM_PPC_GET_SMMU_INFO
2045 Capability: KVM_CAP_PPC_GET_SMMU_INFO
2046 Architectures: powerpc
2049 Returns: 0 on success, -1 on error
2051 This populates and returns a structure describing the features of
2052 the "Server" class MMU emulation supported by KVM.
2053 This can in turn be used by userspace to generate the appropriate
2054 device-tree properties for the guest operating system.
2056 The structure contains some global informations, followed by an
2057 array of supported segment page sizes:
2059 struct kvm_ppc_smmu_info {
2063 struct kvm_ppc_one_seg_page_size sps[KVM_PPC_PAGE_SIZES_MAX_SZ];
2066 The supported flags are:
2068 - KVM_PPC_PAGE_SIZES_REAL:
2069 When that flag is set, guest page sizes must "fit" the backing
2070 store page sizes. When not set, any page size in the list can
2071 be used regardless of how they are backed by userspace.
2073 - KVM_PPC_1T_SEGMENTS
2074 The emulated MMU supports 1T segments in addition to the
2077 The "slb_size" field indicates how many SLB entries are supported
2079 The "sps" array contains 8 entries indicating the supported base
2080 page sizes for a segment in increasing order. Each entry is defined
2083 struct kvm_ppc_one_seg_page_size {
2084 __u32 page_shift; /* Base page shift of segment (or 0) */
2085 __u32 slb_enc; /* SLB encoding for BookS */
2086 struct kvm_ppc_one_page_size enc[KVM_PPC_PAGE_SIZES_MAX_SZ];
2089 An entry with a "page_shift" of 0 is unused. Because the array is
2090 organized in increasing order, a lookup can stop when encoutering
2093 The "slb_enc" field provides the encoding to use in the SLB for the
2094 page size. The bits are in positions such as the value can directly
2095 be OR'ed into the "vsid" argument of the slbmte instruction.
2097 The "enc" array is a list which for each of those segment base page
2098 size provides the list of supported actual page sizes (which can be
2099 only larger or equal to the base page size), along with the
2100 corresponding encoding in the hash PTE. Similarly, the array is
2101 8 entries sorted by increasing sizes and an entry with a "0" shift
2102 is an empty entry and a terminator:
2104 struct kvm_ppc_one_page_size {
2105 __u32 page_shift; /* Page shift (or 0) */
2106 __u32 pte_enc; /* Encoding in the HPTE (>>12) */
2109 The "pte_enc" field provides a value that can OR'ed into the hash
2110 PTE's RPN field (ie, it needs to be shifted left by 12 to OR it
2111 into the hash PTE second double word).
2115 Capability: KVM_CAP_IRQFD
2118 Parameters: struct kvm_irqfd (in)
2119 Returns: 0 on success, -1 on error
2121 Allows setting an eventfd to directly trigger a guest interrupt.
2122 kvm_irqfd.fd specifies the file descriptor to use as the eventfd and
2123 kvm_irqfd.gsi specifies the irqchip pin toggled by this event. When
2124 an event is triggered on the eventfd, an interrupt is injected into
2125 the guest using the specified gsi pin. The irqfd is removed using
2126 the KVM_IRQFD_FLAG_DEASSIGN flag, specifying both kvm_irqfd.fd
2129 With KVM_CAP_IRQFD_RESAMPLE, KVM_IRQFD supports a de-assert and notify
2130 mechanism allowing emulation of level-triggered, irqfd-based
2131 interrupts. When KVM_IRQFD_FLAG_RESAMPLE is set the user must pass an
2132 additional eventfd in the kvm_irqfd.resamplefd field. When operating
2133 in resample mode, posting of an interrupt through kvm_irq.fd asserts
2134 the specified gsi in the irqchip. When the irqchip is resampled, such
2135 as from an EOI, the gsi is de-asserted and the user is notified via
2136 kvm_irqfd.resamplefd. It is the user's responsibility to re-queue
2137 the interrupt if the device making use of it still requires service.
2138 Note that closing the resamplefd is not sufficient to disable the
2139 irqfd. The KVM_IRQFD_FLAG_RESAMPLE is only necessary on assignment
2140 and need not be specified with KVM_IRQFD_FLAG_DEASSIGN.
2142 4.76 KVM_PPC_ALLOCATE_HTAB
2144 Capability: KVM_CAP_PPC_ALLOC_HTAB
2145 Architectures: powerpc
2147 Parameters: Pointer to u32 containing hash table order (in/out)
2148 Returns: 0 on success, -1 on error
2150 This requests the host kernel to allocate an MMU hash table for a
2151 guest using the PAPR paravirtualization interface. This only does
2152 anything if the kernel is configured to use the Book 3S HV style of
2153 virtualization. Otherwise the capability doesn't exist and the ioctl
2154 returns an ENOTTY error. The rest of this description assumes Book 3S
2157 There must be no vcpus running when this ioctl is called; if there
2158 are, it will do nothing and return an EBUSY error.
2160 The parameter is a pointer to a 32-bit unsigned integer variable
2161 containing the order (log base 2) of the desired size of the hash
2162 table, which must be between 18 and 46. On successful return from the
2163 ioctl, it will have been updated with the order of the hash table that
2166 If no hash table has been allocated when any vcpu is asked to run
2167 (with the KVM_RUN ioctl), the host kernel will allocate a
2168 default-sized hash table (16 MB).
2170 If this ioctl is called when a hash table has already been allocated,
2171 the kernel will clear out the existing hash table (zero all HPTEs) and
2172 return the hash table order in the parameter. (If the guest is using
2173 the virtualized real-mode area (VRMA) facility, the kernel will
2174 re-create the VMRA HPTEs on the next KVM_RUN of any vcpu.)
2176 4.77 KVM_S390_INTERRUPT
2180 Type: vm ioctl, vcpu ioctl
2181 Parameters: struct kvm_s390_interrupt (in)
2182 Returns: 0 on success, -1 on error
2184 Allows to inject an interrupt to the guest. Interrupts can be floating
2185 (vm ioctl) or per cpu (vcpu ioctl), depending on the interrupt type.
2187 Interrupt parameters are passed via kvm_s390_interrupt:
2189 struct kvm_s390_interrupt {
2195 type can be one of the following:
2197 KVM_S390_SIGP_STOP (vcpu) - sigp restart
2198 KVM_S390_PROGRAM_INT (vcpu) - program check; code in parm
2199 KVM_S390_SIGP_SET_PREFIX (vcpu) - sigp set prefix; prefix address in parm
2200 KVM_S390_RESTART (vcpu) - restart
2201 KVM_S390_INT_VIRTIO (vm) - virtio external interrupt; external interrupt
2202 parameters in parm and parm64
2203 KVM_S390_INT_SERVICE (vm) - sclp external interrupt; sclp parameter in parm
2204 KVM_S390_INT_EMERGENCY (vcpu) - sigp emergency; source cpu in parm
2205 KVM_S390_INT_EXTERNAL_CALL (vcpu) - sigp external call; source cpu in parm
2206 KVM_S390_INT_IO(ai,cssid,ssid,schid) (vm) - compound value to indicate an
2207 I/O interrupt (ai - adapter interrupt; cssid,ssid,schid - subchannel);
2208 I/O interruption parameters in parm (subchannel) and parm64 (intparm,
2209 interruption subclass)
2210 KVM_S390_MCHK (vm, vcpu) - machine check interrupt; cr 14 bits in parm,
2211 machine check interrupt code in parm64 (note that
2212 machine checks needing further payload are not
2213 supported by this ioctl)
2215 Note that the vcpu ioctl is asynchronous to vcpu execution.
2217 4.78 KVM_PPC_GET_HTAB_FD
2219 Capability: KVM_CAP_PPC_HTAB_FD
2220 Architectures: powerpc
2222 Parameters: Pointer to struct kvm_get_htab_fd (in)
2223 Returns: file descriptor number (>= 0) on success, -1 on error
2225 This returns a file descriptor that can be used either to read out the
2226 entries in the guest's hashed page table (HPT), or to write entries to
2227 initialize the HPT. The returned fd can only be written to if the
2228 KVM_GET_HTAB_WRITE bit is set in the flags field of the argument, and
2229 can only be read if that bit is clear. The argument struct looks like
2232 /* For KVM_PPC_GET_HTAB_FD */
2233 struct kvm_get_htab_fd {
2239 /* Values for kvm_get_htab_fd.flags */
2240 #define KVM_GET_HTAB_BOLTED_ONLY ((__u64)0x1)
2241 #define KVM_GET_HTAB_WRITE ((__u64)0x2)
2243 The `start_index' field gives the index in the HPT of the entry at
2244 which to start reading. It is ignored when writing.
2246 Reads on the fd will initially supply information about all
2247 "interesting" HPT entries. Interesting entries are those with the
2248 bolted bit set, if the KVM_GET_HTAB_BOLTED_ONLY bit is set, otherwise
2249 all entries. When the end of the HPT is reached, the read() will
2250 return. If read() is called again on the fd, it will start again from
2251 the beginning of the HPT, but will only return HPT entries that have
2252 changed since they were last read.
2254 Data read or written is structured as a header (8 bytes) followed by a
2255 series of valid HPT entries (16 bytes) each. The header indicates how
2256 many valid HPT entries there are and how many invalid entries follow
2257 the valid entries. The invalid entries are not represented explicitly
2258 in the stream. The header format is:
2260 struct kvm_get_htab_header {
2266 Writes to the fd create HPT entries starting at the index given in the
2267 header; first `n_valid' valid entries with contents from the data
2268 written, then `n_invalid' invalid entries, invalidating any previously
2269 valid entries found.
2271 4.79 KVM_CREATE_DEVICE
2273 Capability: KVM_CAP_DEVICE_CTRL
2275 Parameters: struct kvm_create_device (in/out)
2276 Returns: 0 on success, -1 on error
2278 ENODEV: The device type is unknown or unsupported
2279 EEXIST: Device already created, and this type of device may not
2280 be instantiated multiple times
2282 Other error conditions may be defined by individual device types or
2283 have their standard meanings.
2285 Creates an emulated device in the kernel. The file descriptor returned
2286 in fd can be used with KVM_SET/GET/HAS_DEVICE_ATTR.
2288 If the KVM_CREATE_DEVICE_TEST flag is set, only test whether the
2289 device type is supported (not necessarily whether it can be created
2292 Individual devices should not define flags. Attributes should be used
2293 for specifying any behavior that is not implied by the device type
2296 struct kvm_create_device {
2297 __u32 type; /* in: KVM_DEV_TYPE_xxx */
2298 __u32 fd; /* out: device handle */
2299 __u32 flags; /* in: KVM_CREATE_DEVICE_xxx */
2302 4.80 KVM_SET_DEVICE_ATTR/KVM_GET_DEVICE_ATTR
2304 Capability: KVM_CAP_DEVICE_CTRL
2306 Parameters: struct kvm_device_attr
2307 Returns: 0 on success, -1 on error
2309 ENXIO: The group or attribute is unknown/unsupported for this device
2310 EPERM: The attribute cannot (currently) be accessed this way
2311 (e.g. read-only attribute, or attribute that only makes
2312 sense when the device is in a different state)
2314 Other error conditions may be defined by individual device types.
2316 Gets/sets a specified piece of device configuration and/or state. The
2317 semantics are device-specific. See individual device documentation in
2318 the "devices" directory. As with ONE_REG, the size of the data
2319 transferred is defined by the particular attribute.
2321 struct kvm_device_attr {
2322 __u32 flags; /* no flags currently defined */
2323 __u32 group; /* device-defined */
2324 __u64 attr; /* group-defined */
2325 __u64 addr; /* userspace address of attr data */
2328 4.81 KVM_HAS_DEVICE_ATTR
2330 Capability: KVM_CAP_DEVICE_CTRL
2332 Parameters: struct kvm_device_attr
2333 Returns: 0 on success, -1 on error
2335 ENXIO: The group or attribute is unknown/unsupported for this device
2337 Tests whether a device supports a particular attribute. A successful
2338 return indicates the attribute is implemented. It does not necessarily
2339 indicate that the attribute can be read or written in the device's
2340 current state. "addr" is ignored.
2342 4.82 KVM_ARM_VCPU_INIT
2345 Architectures: arm, arm64
2347 Parameters: struct kvm_vcpu_init (in)
2348 Returns: 0 on success; -1 on error
2350 Â EINVAL: Â Â Â the target is unknown, or the combination of features is invalid.
2351 Â ENOENT: Â Â Â a features bit specified is unknown.
2353 This tells KVM what type of CPU to present to the guest, and what
2354 optional features it should have. Â This will cause a reset of the cpu
2355 registers to their initial values. Â If this is not called, KVM_RUN will
2356 return ENOEXEC for that vcpu.
2358 Note that because some registers reflect machine topology, all vcpus
2359 should be created before this ioctl is invoked.
2362 - KVM_ARM_VCPU_POWER_OFF: Starts the CPU in a power-off state.
2363 Depends on KVM_CAP_ARM_PSCI.
2364 - KVM_ARM_VCPU_EL1_32BIT: Starts the CPU in a 32bit mode.
2365 Depends on KVM_CAP_ARM_EL1_32BIT (arm64 only).
2368 4.83 KVM_ARM_PREFERRED_TARGET
2371 Architectures: arm, arm64
2373 Parameters: struct struct kvm_vcpu_init (out)
2374 Returns: 0 on success; -1 on error
2376 ENODEV: no preferred target available for the host
2378 This queries KVM for preferred CPU target type which can be emulated
2379 by KVM on underlying host.
2381 The ioctl returns struct kvm_vcpu_init instance containing information
2382 about preferred CPU target type and recommended features for it. The
2383 kvm_vcpu_init->features bitmap returned will have feature bits set if
2384 the preferred target recommends setting these features, but this is
2387 The information returned by this ioctl can be used to prepare an instance
2388 of struct kvm_vcpu_init for KVM_ARM_VCPU_INIT ioctl which will result in
2389 in VCPU matching underlying host.
2392 4.84 KVM_GET_REG_LIST
2395 Architectures: arm, arm64
2397 Parameters: struct kvm_reg_list (in/out)
2398 Returns: 0 on success; -1 on error
2400 Â E2BIG: Â Â Â Â the reg index list is too big to fit in the array specified by
2401 Â Â Â Â Â Â Â Â Â Â Â Â the user (the number required will be written into n).
2403 struct kvm_reg_list {
2404 __u64 n; /* number of registers in reg[] */
2408 This ioctl returns the guest registers that are supported for the
2409 KVM_GET_ONE_REG/KVM_SET_ONE_REG calls.
2412 4.85 KVM_ARM_SET_DEVICE_ADDR (deprecated)
2414 Capability: KVM_CAP_ARM_SET_DEVICE_ADDR
2415 Architectures: arm, arm64
2417 Parameters: struct kvm_arm_device_address (in)
2418 Returns: 0 on success, -1 on error
2420 ENODEV: The device id is unknown
2421 ENXIO: Device not supported on current system
2422 EEXIST: Address already set
2423 E2BIG: Address outside guest physical address space
2424 EBUSY: Address overlaps with other device range
2426 struct kvm_arm_device_addr {
2431 Specify a device address in the guest's physical address space where guests
2432 can access emulated or directly exposed devices, which the host kernel needs
2433 to know about. The id field is an architecture specific identifier for a
2436 ARM/arm64 divides the id field into two parts, a device id and an
2437 address type id specific to the individual device.
2439 Â bits: | 63 ... 32 | 31 ... 16 | 15 ... 0 |
2440 field: | 0x00000000 | device id | addr type id |
2442 ARM/arm64 currently only require this when using the in-kernel GIC
2443 support for the hardware VGIC features, using KVM_ARM_DEVICE_VGIC_V2
2444 as the device id. When setting the base address for the guest's
2445 mapping of the VGIC virtual CPU and distributor interface, the ioctl
2446 must be called after calling KVM_CREATE_IRQCHIP, but before calling
2447 KVM_RUN on any of the VCPUs. Calling this ioctl twice for any of the
2448 base addresses will return -EEXIST.
2450 Note, this IOCTL is deprecated and the more flexible SET/GET_DEVICE_ATTR API
2451 should be used instead.
2454 4.86 KVM_PPC_RTAS_DEFINE_TOKEN
2456 Capability: KVM_CAP_PPC_RTAS
2459 Parameters: struct kvm_rtas_token_args
2460 Returns: 0 on success, -1 on error
2462 Defines a token value for a RTAS (Run Time Abstraction Services)
2463 service in order to allow it to be handled in the kernel. The
2464 argument struct gives the name of the service, which must be the name
2465 of a service that has a kernel-side implementation. If the token
2466 value is non-zero, it will be associated with that service, and
2467 subsequent RTAS calls by the guest specifying that token will be
2468 handled by the kernel. If the token value is 0, then any token
2469 associated with the service will be forgotten, and subsequent RTAS
2470 calls by the guest for that service will be passed to userspace to be
2474 5. The kvm_run structure
2475 ------------------------
2477 Application code obtains a pointer to the kvm_run structure by
2478 mmap()ing a vcpu fd. From that point, application code can control
2479 execution by changing fields in kvm_run prior to calling the KVM_RUN
2480 ioctl, and obtain information about the reason KVM_RUN returned by
2481 looking up structure members.
2485 __u8 request_interrupt_window;
2487 Request that KVM_RUN return when it becomes possible to inject external
2488 interrupts into the guest. Useful in conjunction with KVM_INTERRUPT.
2495 When KVM_RUN has returned successfully (return value 0), this informs
2496 application code why KVM_RUN has returned. Allowable values for this
2497 field are detailed below.
2499 __u8 ready_for_interrupt_injection;
2501 If request_interrupt_window has been specified, this field indicates
2502 an interrupt can be injected now with KVM_INTERRUPT.
2506 The value of the current interrupt flag. Only valid if in-kernel
2507 local APIC is not used.
2511 /* in (pre_kvm_run), out (post_kvm_run) */
2514 The value of the cr8 register. Only valid if in-kernel local APIC is
2515 not used. Both input and output.
2519 The value of the APIC BASE msr. Only valid if in-kernel local
2520 APIC is not used. Both input and output.
2523 /* KVM_EXIT_UNKNOWN */
2525 __u64 hardware_exit_reason;
2528 If exit_reason is KVM_EXIT_UNKNOWN, the vcpu has exited due to unknown
2529 reasons. Further architecture-specific information is available in
2530 hardware_exit_reason.
2532 /* KVM_EXIT_FAIL_ENTRY */
2534 __u64 hardware_entry_failure_reason;
2537 If exit_reason is KVM_EXIT_FAIL_ENTRY, the vcpu could not be run due
2538 to unknown reasons. Further architecture-specific information is
2539 available in hardware_entry_failure_reason.
2541 /* KVM_EXIT_EXCEPTION */
2551 #define KVM_EXIT_IO_IN 0
2552 #define KVM_EXIT_IO_OUT 1
2554 __u8 size; /* bytes */
2557 __u64 data_offset; /* relative to kvm_run start */
2560 If exit_reason is KVM_EXIT_IO, then the vcpu has
2561 executed a port I/O instruction which could not be satisfied by kvm.
2562 data_offset describes where the data is located (KVM_EXIT_IO_OUT) or
2563 where kvm expects application code to place the data for the next
2564 KVM_RUN invocation (KVM_EXIT_IO_IN). Data format is a packed array.
2567 struct kvm_debug_exit_arch arch;
2580 If exit_reason is KVM_EXIT_MMIO, then the vcpu has
2581 executed a memory-mapped I/O instruction which could not be satisfied
2582 by kvm. The 'data' member contains the written data if 'is_write' is
2583 true, and should be filled by application code otherwise.
2585 NOTE: For KVM_EXIT_IO, KVM_EXIT_MMIO, KVM_EXIT_OSI, KVM_EXIT_DCR,
2586 KVM_EXIT_PAPR and KVM_EXIT_EPR the corresponding
2587 operations are complete (and guest state is consistent) only after userspace
2588 has re-entered the kernel with KVM_RUN. The kernel side will first finish
2589 incomplete operations and then check for pending signals. Userspace
2590 can re-enter the guest with an unmasked signal pending to complete
2593 /* KVM_EXIT_HYPERCALL */
2602 Unused. This was once used for 'hypercall to userspace'. To implement
2603 such functionality, use KVM_EXIT_IO (x86) or KVM_EXIT_MMIO (all except s390).
2604 Note KVM_EXIT_IO is significantly faster than KVM_EXIT_MMIO.
2606 /* KVM_EXIT_TPR_ACCESS */
2613 To be documented (KVM_TPR_ACCESS_REPORTING).
2615 /* KVM_EXIT_S390_SIEIC */
2618 __u64 mask; /* psw upper half */
2619 __u64 addr; /* psw lower half */
2626 /* KVM_EXIT_S390_RESET */
2627 #define KVM_S390_RESET_POR 1
2628 #define KVM_S390_RESET_CLEAR 2
2629 #define KVM_S390_RESET_SUBSYSTEM 4
2630 #define KVM_S390_RESET_CPU_INIT 8
2631 #define KVM_S390_RESET_IPL 16
2632 __u64 s390_reset_flags;
2636 /* KVM_EXIT_S390_UCONTROL */
2638 __u64 trans_exc_code;
2642 s390 specific. A page fault has occurred for a user controlled virtual
2643 machine (KVM_VM_S390_UNCONTROL) on it's host page table that cannot be
2644 resolved by the kernel.
2645 The program code and the translation exception code that were placed
2646 in the cpu's lowcore are presented here as defined by the z Architecture
2647 Principles of Operation Book in the Chapter for Dynamic Address Translation
2664 MOL uses a special hypercall interface it calls 'OSI'. To enable it, we catch
2665 hypercalls and exit with this exit struct that contains all the guest gprs.
2667 If exit_reason is KVM_EXIT_OSI, then the vcpu has triggered such a hypercall.
2668 Userspace can now handle the hypercall and when it's done modify the gprs as
2669 necessary. Upon guest entry all guest GPRs will then be replaced by the values
2672 /* KVM_EXIT_PAPR_HCALL */
2679 This is used on 64-bit PowerPC when emulating a pSeries partition,
2680 e.g. with the 'pseries' machine type in qemu. It occurs when the
2681 guest does a hypercall using the 'sc 1' instruction. The 'nr' field
2682 contains the hypercall number (from the guest R3), and 'args' contains
2683 the arguments (from the guest R4 - R12). Userspace should put the
2684 return code in 'ret' and any extra returned values in args[].
2685 The possible hypercalls are defined in the Power Architecture Platform
2686 Requirements (PAPR) document available from www.power.org (free
2687 developer registration required to access it).
2689 /* KVM_EXIT_S390_TSCH */
2691 __u16 subchannel_id;
2692 __u16 subchannel_nr;
2699 s390 specific. This exit occurs when KVM_CAP_S390_CSS_SUPPORT has been enabled
2700 and TEST SUBCHANNEL was intercepted. If dequeued is set, a pending I/O
2701 interrupt for the target subchannel has been dequeued and subchannel_id,
2702 subchannel_nr, io_int_parm and io_int_word contain the parameters for that
2703 interrupt. ipb is needed for instruction parameter decoding.
2710 On FSL BookE PowerPC chips, the interrupt controller has a fast patch
2711 interrupt acknowledge path to the core. When the core successfully
2712 delivers an interrupt, it automatically populates the EPR register with
2713 the interrupt vector number and acknowledges the interrupt inside
2714 the interrupt controller.
2716 In case the interrupt controller lives in user space, we need to do
2717 the interrupt acknowledge cycle through it to fetch the next to be
2718 delivered interrupt vector using this exit.
2720 It gets triggered whenever both KVM_CAP_PPC_EPR are enabled and an
2721 external interrupt has just been delivered into the guest. User space
2722 should put the acknowledged interrupt vector into the 'epr' field.
2724 /* Fix the size of the union. */
2729 * shared registers between kvm and userspace.
2730 * kvm_valid_regs specifies the register classes set by the host
2731 * kvm_dirty_regs specified the register classes dirtied by userspace
2732 * struct kvm_sync_regs is architecture specific, as well as the
2733 * bits for kvm_valid_regs and kvm_dirty_regs
2735 __u64 kvm_valid_regs;
2736 __u64 kvm_dirty_regs;
2738 struct kvm_sync_regs regs;
2742 If KVM_CAP_SYNC_REGS is defined, these fields allow userspace to access
2743 certain guest registers without having to call SET/GET_*REGS. Thus we can
2744 avoid some system call overhead if userspace has to handle the exit.
2745 Userspace can query the validity of the structure by checking
2746 kvm_valid_regs for specific bits. These bits are architecture specific
2747 and usually define the validity of a groups of registers. (e.g. one bit
2748 for general purpose registers)
2753 4.81 KVM_GET_EMULATED_CPUID
2755 Capability: KVM_CAP_EXT_EMUL_CPUID
2758 Parameters: struct kvm_cpuid2 (in/out)
2759 Returns: 0 on success, -1 on error
2764 struct kvm_cpuid_entry2 entries[0];
2767 The member 'flags' is used for passing flags from userspace.
2769 #define KVM_CPUID_FLAG_SIGNIFCANT_INDEX BIT(0)
2770 #define KVM_CPUID_FLAG_STATEFUL_FUNC BIT(1)
2771 #define KVM_CPUID_FLAG_STATE_READ_NEXT BIT(2)
2773 struct kvm_cpuid_entry2 {
2784 This ioctl returns x86 cpuid features which are emulated by
2785 kvm.Userspace can use the information returned by this ioctl to query
2786 which features are emulated by kvm instead of being present natively.
2788 Userspace invokes KVM_GET_EMULATED_CPUID by passing a kvm_cpuid2
2789 structure with the 'nent' field indicating the number of entries in
2790 the variable-size array 'entries'. If the number of entries is too low
2791 to describe the cpu capabilities, an error (E2BIG) is returned. If the
2792 number is too high, the 'nent' field is adjusted and an error (ENOMEM)
2793 is returned. If the number is just right, the 'nent' field is adjusted
2794 to the number of valid entries in the 'entries' array, which is then
2797 The entries returned are the set CPUID bits of the respective features
2798 which kvm emulates, as returned by the CPUID instruction, with unknown
2799 or unsupported feature bits cleared.
2801 Features like x2apic, for example, may not be present in the host cpu
2802 but are exposed by kvm in KVM_GET_SUPPORTED_CPUID because they can be
2803 emulated efficiently and thus not included here.
2805 The fields in each entry are defined as follows:
2807 function: the eax value used to obtain the entry
2808 index: the ecx value used to obtain the entry (for entries that are
2810 flags: an OR of zero or more of the following:
2811 KVM_CPUID_FLAG_SIGNIFCANT_INDEX:
2812 if the index field is valid
2813 KVM_CPUID_FLAG_STATEFUL_FUNC:
2814 if cpuid for this function returns different values for successive
2815 invocations; there will be several entries with the same function,
2816 all with this flag set
2817 KVM_CPUID_FLAG_STATE_READ_NEXT:
2818 for KVM_CPUID_FLAG_STATEFUL_FUNC entries, set if this entry is
2819 the first entry to be read by a cpu
2820 eax, ebx, ecx, edx: the values returned by the cpuid instruction for
2821 this function/index combination
2824 6. Capabilities that can be enabled
2825 -----------------------------------
2827 There are certain capabilities that change the behavior of the virtual CPU when
2828 enabled. To enable them, please see section 4.37. Below you can find a list of
2829 capabilities and what their effect on the vCPU is when enabling them.
2831 The following information is provided along with the description:
2833 Architectures: which instruction set architectures provide this ioctl.
2834 x86 includes both i386 and x86_64.
2836 Parameters: what parameters are accepted by the capability.
2838 Returns: the return value. General error numbers (EBADF, ENOMEM, EINVAL)
2839 are not detailed, but errors with specific meanings are.
2846 Returns: 0 on success; -1 on error
2848 This capability enables interception of OSI hypercalls that otherwise would
2849 be treated as normal system calls to be injected into the guest. OSI hypercalls
2850 were invented by Mac-on-Linux to have a standardized communication mechanism
2851 between the guest and the host.
2853 When this capability is enabled, KVM_EXIT_OSI can occur.
2856 6.2 KVM_CAP_PPC_PAPR
2860 Returns: 0 on success; -1 on error
2862 This capability enables interception of PAPR hypercalls. PAPR hypercalls are
2863 done using the hypercall instruction "sc 1".
2865 It also sets the guest privilege level to "supervisor" mode. Usually the guest
2866 runs in "hypervisor" privilege mode with a few missing features.
2868 In addition to the above, it changes the semantics of SDR1. In this mode, the
2869 HTAB address part of SDR1 contains an HVA instead of a GPA, as PAPR keeps the
2870 HTAB invisible to the guest.
2872 When this capability is enabled, KVM_EXIT_PAPR_HCALL can occur.
2878 Parameters: args[0] is the address of a struct kvm_config_tlb
2879 Returns: 0 on success; -1 on error
2881 struct kvm_config_tlb {
2888 Configures the virtual CPU's TLB array, establishing a shared memory area
2889 between userspace and KVM. The "params" and "array" fields are userspace
2890 addresses of mmu-type-specific data structures. The "array_len" field is an
2891 safety mechanism, and should be set to the size in bytes of the memory that
2892 userspace has reserved for the array. It must be at least the size dictated
2893 by "mmu_type" and "params".
2895 While KVM_RUN is active, the shared region is under control of KVM. Its
2896 contents are undefined, and any modification by userspace results in
2897 boundedly undefined behavior.
2899 On return from KVM_RUN, the shared region will reflect the current state of
2900 the guest's TLB. If userspace makes any changes, it must call KVM_DIRTY_TLB
2901 to tell KVM which entries have been changed, prior to calling KVM_RUN again
2904 For mmu types KVM_MMU_FSL_BOOKE_NOHV and KVM_MMU_FSL_BOOKE_HV:
2905 - The "params" field is of type "struct kvm_book3e_206_tlb_params".
2906 - The "array" field points to an array of type "struct
2907 kvm_book3e_206_tlb_entry".
2908 - The array consists of all entries in the first TLB, followed by all
2909 entries in the second TLB.
2910 - Within a TLB, entries are ordered first by increasing set number. Within a
2911 set, entries are ordered by way (increasing ESEL).
2912 - The hash for determining set number in TLB0 is: (MAS2 >> 12) & (num_sets - 1)
2913 where "num_sets" is the tlb_sizes[] value divided by the tlb_ways[] value.
2914 - The tsize field of mas1 shall be set to 4K on TLB0, even though the
2915 hardware ignores this value for TLB0.
2917 6.4 KVM_CAP_S390_CSS_SUPPORT
2921 Returns: 0 on success; -1 on error
2923 This capability enables support for handling of channel I/O instructions.
2925 TEST PENDING INTERRUPTION and the interrupt portion of TEST SUBCHANNEL are
2926 handled in-kernel, while the other I/O instructions are passed to userspace.
2928 When this capability is enabled, KVM_EXIT_S390_TSCH will occur on TEST
2929 SUBCHANNEL intercepts.
2934 Parameters: args[0] defines whether the proxy facility is active
2935 Returns: 0 on success; -1 on error
2937 This capability enables or disables the delivery of interrupts through the
2938 external proxy facility.
2940 When enabled (args[0] != 0), every time the guest gets an external interrupt
2941 delivered, it automatically exits into user space with a KVM_EXIT_EPR exit
2942 to receive the topmost interrupt vector.
2944 When disabled (args[0] == 0), behavior is as if this facility is unsupported.
2946 When this capability is enabled, KVM_EXIT_EPR can occur.
2948 6.6 KVM_CAP_IRQ_MPIC
2951 Parameters: args[0] is the MPIC device fd
2952 args[1] is the MPIC CPU number for this vcpu
2954 This capability connects the vcpu to an in-kernel MPIC device.
2956 6.7 KVM_CAP_IRQ_XICS
2959 Parameters: args[0] is the XICS device fd
2960 args[1] is the XICS CPU number (server ID) for this vcpu
2962 This capability connects the vcpu to an in-kernel XICS device.