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, KVM_CAP_S390_IRQCHIP (s390)
590 Architectures: x86, ia64, ARM, arm64, s390
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
599 created. On s390, a dummy irq routing table is created.
601 Note that on s390 the KVM_CAP_S390_IRQCHIP vm capability needs to be enabled
602 before KVM_CREATE_IRQCHIP can be used.
607 Capability: KVM_CAP_IRQCHIP
608 Architectures: x86, ia64, arm, arm64
610 Parameters: struct kvm_irq_level
611 Returns: 0 on success, -1 on error
613 Sets the level of a GSI input to the interrupt controller model in the kernel.
614 On some architectures it is required that an interrupt controller model has
615 been previously created with KVM_CREATE_IRQCHIP. Note that edge-triggered
616 interrupts require the level to be set to 1 and then back to 0.
618 On real hardware, interrupt pins can be active-low or active-high. This
619 does not matter for the level field of struct kvm_irq_level: 1 always
620 means active (asserted), 0 means inactive (deasserted).
622 x86 allows the operating system to program the interrupt polarity
623 (active-low/active-high) for level-triggered interrupts, and KVM used
624 to consider the polarity. However, due to bitrot in the handling of
625 active-low interrupts, the above convention is now valid on x86 too.
626 This is signaled by KVM_CAP_X86_IOAPIC_POLARITY_IGNORED. Userspace
627 should not present interrupts to the guest as active-low unless this
628 capability is present (or unless it is not using the in-kernel irqchip,
632 ARM/arm64 can signal an interrupt either at the CPU level, or at the
633 in-kernel irqchip (GIC), and for in-kernel irqchip can tell the GIC to
634 use PPIs designated for specific cpus. The irq field is interpreted
637 Â bits: | 31 ... 24 | 23 ... 16 | 15 ... 0 |
638 field: | irq_type | vcpu_index | irq_id |
640 The irq_type field has the following values:
641 - irq_type[0]: out-of-kernel GIC: irq_id 0 is IRQ, irq_id 1 is FIQ
642 - irq_type[1]: in-kernel GIC: SPI, irq_id between 32 and 1019 (incl.)
643 (the vcpu_index field is ignored)
644 - irq_type[2]: in-kernel GIC: PPI, irq_id between 16 and 31 (incl.)
646 (The irq_id field thus corresponds nicely to the IRQ ID in the ARM GIC specs)
648 In both cases, level is used to assert/deassert the line.
650 struct kvm_irq_level {
653 __s32 status; /* not used for KVM_IRQ_LEVEL */
655 __u32 level; /* 0 or 1 */
661 Capability: KVM_CAP_IRQCHIP
662 Architectures: x86, ia64
664 Parameters: struct kvm_irqchip (in/out)
665 Returns: 0 on success, -1 on error
667 Reads the state of a kernel interrupt controller created with
668 KVM_CREATE_IRQCHIP into a buffer provided by the caller.
671 __u32 chip_id; /* 0 = PIC1, 1 = PIC2, 2 = IOAPIC */
674 char dummy[512]; /* reserving space */
675 struct kvm_pic_state pic;
676 struct kvm_ioapic_state ioapic;
683 Capability: KVM_CAP_IRQCHIP
684 Architectures: x86, ia64
686 Parameters: struct kvm_irqchip (in)
687 Returns: 0 on success, -1 on error
689 Sets the state of a kernel interrupt controller created with
690 KVM_CREATE_IRQCHIP from a buffer provided by the caller.
693 __u32 chip_id; /* 0 = PIC1, 1 = PIC2, 2 = IOAPIC */
696 char dummy[512]; /* reserving space */
697 struct kvm_pic_state pic;
698 struct kvm_ioapic_state ioapic;
703 4.28 KVM_XEN_HVM_CONFIG
705 Capability: KVM_CAP_XEN_HVM
708 Parameters: struct kvm_xen_hvm_config (in)
709 Returns: 0 on success, -1 on error
711 Sets the MSR that the Xen HVM guest uses to initialize its hypercall
712 page, and provides the starting address and size of the hypercall
713 blobs in userspace. When the guest writes the MSR, kvm copies one
714 page of a blob (32- or 64-bit, depending on the vcpu mode) to guest
717 struct kvm_xen_hvm_config {
730 Capability: KVM_CAP_ADJUST_CLOCK
733 Parameters: struct kvm_clock_data (out)
734 Returns: 0 on success, -1 on error
736 Gets the current timestamp of kvmclock as seen by the current guest. In
737 conjunction with KVM_SET_CLOCK, it is used to ensure monotonicity on scenarios
740 struct kvm_clock_data {
741 __u64 clock; /* kvmclock current value */
749 Capability: KVM_CAP_ADJUST_CLOCK
752 Parameters: struct kvm_clock_data (in)
753 Returns: 0 on success, -1 on error
755 Sets the current timestamp of kvmclock to the value specified in its parameter.
756 In conjunction with KVM_GET_CLOCK, it is used to ensure monotonicity on scenarios
759 struct kvm_clock_data {
760 __u64 clock; /* kvmclock current value */
766 4.31 KVM_GET_VCPU_EVENTS
768 Capability: KVM_CAP_VCPU_EVENTS
769 Extended by: KVM_CAP_INTR_SHADOW
772 Parameters: struct kvm_vcpu_event (out)
773 Returns: 0 on success, -1 on error
775 Gets currently pending exceptions, interrupts, and NMIs as well as related
778 struct kvm_vcpu_events {
802 KVM_VCPUEVENT_VALID_SHADOW may be set in the flags field to signal that
803 interrupt.shadow contains a valid state. Otherwise, this field is undefined.
806 4.32 KVM_SET_VCPU_EVENTS
808 Capability: KVM_CAP_VCPU_EVENTS
809 Extended by: KVM_CAP_INTR_SHADOW
812 Parameters: struct kvm_vcpu_event (in)
813 Returns: 0 on success, -1 on error
815 Set pending exceptions, interrupts, and NMIs as well as related states of the
818 See KVM_GET_VCPU_EVENTS for the data structure.
820 Fields that may be modified asynchronously by running VCPUs can be excluded
821 from the update. These fields are nmi.pending and sipi_vector. Keep the
822 corresponding bits in the flags field cleared to suppress overwriting the
823 current in-kernel state. The bits are:
825 KVM_VCPUEVENT_VALID_NMI_PENDING - transfer nmi.pending to the kernel
826 KVM_VCPUEVENT_VALID_SIPI_VECTOR - transfer sipi_vector
828 If KVM_CAP_INTR_SHADOW is available, KVM_VCPUEVENT_VALID_SHADOW can be set in
829 the flags field to signal that interrupt.shadow contains a valid state and
830 shall be written into the VCPU.
833 4.33 KVM_GET_DEBUGREGS
835 Capability: KVM_CAP_DEBUGREGS
838 Parameters: struct kvm_debugregs (out)
839 Returns: 0 on success, -1 on error
841 Reads debug registers from the vcpu.
843 struct kvm_debugregs {
852 4.34 KVM_SET_DEBUGREGS
854 Capability: KVM_CAP_DEBUGREGS
857 Parameters: struct kvm_debugregs (in)
858 Returns: 0 on success, -1 on error
860 Writes debug registers into the vcpu.
862 See KVM_GET_DEBUGREGS for the data structure. The flags field is unused
863 yet and must be cleared on entry.
866 4.35 KVM_SET_USER_MEMORY_REGION
868 Capability: KVM_CAP_USER_MEM
871 Parameters: struct kvm_userspace_memory_region (in)
872 Returns: 0 on success, -1 on error
874 struct kvm_userspace_memory_region {
877 __u64 guest_phys_addr;
878 __u64 memory_size; /* bytes */
879 __u64 userspace_addr; /* start of the userspace allocated memory */
882 /* for kvm_memory_region::flags */
883 #define KVM_MEM_LOG_DIRTY_PAGES (1UL << 0)
884 #define KVM_MEM_READONLY (1UL << 1)
886 This ioctl allows the user to create or modify a guest physical memory
887 slot. When changing an existing slot, it may be moved in the guest
888 physical memory space, or its flags may be modified. It may not be
889 resized. Slots may not overlap in guest physical address space.
891 Memory for the region is taken starting at the address denoted by the
892 field userspace_addr, which must point at user addressable memory for
893 the entire memory slot size. Any object may back this memory, including
894 anonymous memory, ordinary files, and hugetlbfs.
896 It is recommended that the lower 21 bits of guest_phys_addr and userspace_addr
897 be identical. This allows large pages in the guest to be backed by large
900 The flags field supports two flags: KVM_MEM_LOG_DIRTY_PAGES and
901 KVM_MEM_READONLY. The former can be set to instruct KVM to keep track of
902 writes to memory within the slot. See KVM_GET_DIRTY_LOG ioctl to know how to
903 use it. The latter can be set, if KVM_CAP_READONLY_MEM capability allows it,
904 to make a new slot read-only. In this case, writes to this memory will be
905 posted to userspace as KVM_EXIT_MMIO exits.
907 When the KVM_CAP_SYNC_MMU capability is available, changes in the backing of
908 the memory region are automatically reflected into the guest. For example, an
909 mmap() that affects the region will be made visible immediately. Another
910 example is madvise(MADV_DROP).
912 It is recommended to use this API instead of the KVM_SET_MEMORY_REGION ioctl.
913 The KVM_SET_MEMORY_REGION does not allow fine grained control over memory
914 allocation and is deprecated.
917 4.36 KVM_SET_TSS_ADDR
919 Capability: KVM_CAP_SET_TSS_ADDR
922 Parameters: unsigned long tss_address (in)
923 Returns: 0 on success, -1 on error
925 This ioctl defines the physical address of a three-page region in the guest
926 physical address space. The region must be within the first 4GB of the
927 guest physical address space and must not conflict with any memory slot
928 or any mmio address. The guest may malfunction if it accesses this memory
931 This ioctl is required on Intel-based hosts. This is needed on Intel hardware
932 because of a quirk in the virtualization implementation (see the internals
933 documentation when it pops into existence).
938 Capability: KVM_CAP_ENABLE_CAP, KVM_CAP_ENABLE_CAP_VM
939 Architectures: ppc, s390
940 Type: vcpu ioctl, vm ioctl (with KVM_CAP_ENABLE_CAP_VM)
941 Parameters: struct kvm_enable_cap (in)
942 Returns: 0 on success; -1 on error
944 +Not all extensions are enabled by default. Using this ioctl the application
945 can enable an extension, making it available to the guest.
947 On systems that do not support this ioctl, it always fails. On systems that
948 do support it, it only works for extensions that are supported for enablement.
950 To check if a capability can be enabled, the KVM_CHECK_EXTENSION ioctl should
953 struct kvm_enable_cap {
957 The capability that is supposed to get enabled.
961 A bitfield indicating future enhancements. Has to be 0 for now.
965 Arguments for enabling a feature. If a feature needs initial values to
966 function properly, this is the place to put them.
971 The vcpu ioctl should be used for vcpu-specific capabilities, the vm ioctl
972 for vm-wide capabilities.
974 4.38 KVM_GET_MP_STATE
976 Capability: KVM_CAP_MP_STATE
977 Architectures: x86, ia64
979 Parameters: struct kvm_mp_state (out)
980 Returns: 0 on success; -1 on error
982 struct kvm_mp_state {
986 Returns the vcpu's current "multiprocessing state" (though also valid on
987 uniprocessor guests).
991 - KVM_MP_STATE_RUNNABLE: the vcpu is currently running
992 - KVM_MP_STATE_UNINITIALIZED: the vcpu is an application processor (AP)
993 which has not yet received an INIT signal
994 - KVM_MP_STATE_INIT_RECEIVED: the vcpu has received an INIT signal, and is
996 - KVM_MP_STATE_HALTED: the vcpu has executed a HLT instruction and
997 is waiting for an interrupt
998 - KVM_MP_STATE_SIPI_RECEIVED: the vcpu has just received a SIPI (vector
999 accessible via KVM_GET_VCPU_EVENTS)
1001 This ioctl is only useful after KVM_CREATE_IRQCHIP. Without an in-kernel
1002 irqchip, the multiprocessing state must be maintained by userspace.
1005 4.39 KVM_SET_MP_STATE
1007 Capability: KVM_CAP_MP_STATE
1008 Architectures: x86, ia64
1010 Parameters: struct kvm_mp_state (in)
1011 Returns: 0 on success; -1 on error
1013 Sets the vcpu's current "multiprocessing state"; see KVM_GET_MP_STATE for
1016 This ioctl is only useful after KVM_CREATE_IRQCHIP. Without an in-kernel
1017 irqchip, the multiprocessing state must be maintained by userspace.
1020 4.40 KVM_SET_IDENTITY_MAP_ADDR
1022 Capability: KVM_CAP_SET_IDENTITY_MAP_ADDR
1025 Parameters: unsigned long identity (in)
1026 Returns: 0 on success, -1 on error
1028 This ioctl defines the physical address of a one-page region in the guest
1029 physical address space. The region must be within the first 4GB of the
1030 guest physical address space and must not conflict with any memory slot
1031 or any mmio address. The guest may malfunction if it accesses this memory
1034 This ioctl is required on Intel-based hosts. This is needed on Intel hardware
1035 because of a quirk in the virtualization implementation (see the internals
1036 documentation when it pops into existence).
1039 4.41 KVM_SET_BOOT_CPU_ID
1041 Capability: KVM_CAP_SET_BOOT_CPU_ID
1042 Architectures: x86, ia64
1044 Parameters: unsigned long vcpu_id
1045 Returns: 0 on success, -1 on error
1047 Define which vcpu is the Bootstrap Processor (BSP). Values are the same
1048 as the vcpu id in KVM_CREATE_VCPU. If this ioctl is not called, the default
1054 Capability: KVM_CAP_XSAVE
1057 Parameters: struct kvm_xsave (out)
1058 Returns: 0 on success, -1 on error
1064 This ioctl would copy current vcpu's xsave struct to the userspace.
1069 Capability: KVM_CAP_XSAVE
1072 Parameters: struct kvm_xsave (in)
1073 Returns: 0 on success, -1 on error
1079 This ioctl would copy userspace's xsave struct to the kernel.
1084 Capability: KVM_CAP_XCRS
1087 Parameters: struct kvm_xcrs (out)
1088 Returns: 0 on success, -1 on error
1099 struct kvm_xcr xcrs[KVM_MAX_XCRS];
1103 This ioctl would copy current vcpu's xcrs to the userspace.
1108 Capability: KVM_CAP_XCRS
1111 Parameters: struct kvm_xcrs (in)
1112 Returns: 0 on success, -1 on error
1123 struct kvm_xcr xcrs[KVM_MAX_XCRS];
1127 This ioctl would set vcpu's xcr to the value userspace specified.
1130 4.46 KVM_GET_SUPPORTED_CPUID
1132 Capability: KVM_CAP_EXT_CPUID
1135 Parameters: struct kvm_cpuid2 (in/out)
1136 Returns: 0 on success, -1 on error
1141 struct kvm_cpuid_entry2 entries[0];
1144 #define KVM_CPUID_FLAG_SIGNIFCANT_INDEX BIT(0)
1145 #define KVM_CPUID_FLAG_STATEFUL_FUNC BIT(1)
1146 #define KVM_CPUID_FLAG_STATE_READ_NEXT BIT(2)
1148 struct kvm_cpuid_entry2 {
1159 This ioctl returns x86 cpuid features which are supported by both the hardware
1160 and kvm. Userspace can use the information returned by this ioctl to
1161 construct cpuid information (for KVM_SET_CPUID2) that is consistent with
1162 hardware, kernel, and userspace capabilities, and with user requirements (for
1163 example, the user may wish to constrain cpuid to emulate older hardware,
1164 or for feature consistency across a cluster).
1166 Userspace invokes KVM_GET_SUPPORTED_CPUID by passing a kvm_cpuid2 structure
1167 with the 'nent' field indicating the number of entries in the variable-size
1168 array 'entries'. If the number of entries is too low to describe the cpu
1169 capabilities, an error (E2BIG) is returned. If the number is too high,
1170 the 'nent' field is adjusted and an error (ENOMEM) is returned. If the
1171 number is just right, the 'nent' field is adjusted to the number of valid
1172 entries in the 'entries' array, which is then filled.
1174 The entries returned are the host cpuid as returned by the cpuid instruction,
1175 with unknown or unsupported features masked out. Some features (for example,
1176 x2apic), may not be present in the host cpu, but are exposed by kvm if it can
1177 emulate them efficiently. The fields in each entry are defined as follows:
1179 function: the eax value used to obtain the entry
1180 index: the ecx value used to obtain the entry (for entries that are
1182 flags: an OR of zero or more of the following:
1183 KVM_CPUID_FLAG_SIGNIFCANT_INDEX:
1184 if the index field is valid
1185 KVM_CPUID_FLAG_STATEFUL_FUNC:
1186 if cpuid for this function returns different values for successive
1187 invocations; there will be several entries with the same function,
1188 all with this flag set
1189 KVM_CPUID_FLAG_STATE_READ_NEXT:
1190 for KVM_CPUID_FLAG_STATEFUL_FUNC entries, set if this entry is
1191 the first entry to be read by a cpu
1192 eax, ebx, ecx, edx: the values returned by the cpuid instruction for
1193 this function/index combination
1195 The TSC deadline timer feature (CPUID leaf 1, ecx[24]) is always returned
1196 as false, since the feature depends on KVM_CREATE_IRQCHIP for local APIC
1197 support. Instead it is reported via
1199 ioctl(KVM_CHECK_EXTENSION, KVM_CAP_TSC_DEADLINE_TIMER)
1201 if that returns true and you use KVM_CREATE_IRQCHIP, or if you emulate the
1202 feature in userspace, then you can enable the feature for KVM_SET_CPUID2.
1205 4.47 KVM_PPC_GET_PVINFO
1207 Capability: KVM_CAP_PPC_GET_PVINFO
1210 Parameters: struct kvm_ppc_pvinfo (out)
1211 Returns: 0 on success, !0 on error
1213 struct kvm_ppc_pvinfo {
1219 This ioctl fetches PV specific information that need to be passed to the guest
1220 using the device tree or other means from vm context.
1222 The hcall array defines 4 instructions that make up a hypercall.
1224 If any additional field gets added to this structure later on, a bit for that
1225 additional piece of information will be set in the flags bitmap.
1227 The flags bitmap is defined as:
1229 /* the host supports the ePAPR idle hcall
1230 #define KVM_PPC_PVINFO_FLAGS_EV_IDLE (1<<0)
1232 4.48 KVM_ASSIGN_PCI_DEVICE
1234 Capability: KVM_CAP_DEVICE_ASSIGNMENT
1235 Architectures: x86 ia64
1237 Parameters: struct kvm_assigned_pci_dev (in)
1238 Returns: 0 on success, -1 on error
1240 Assigns a host PCI device to the VM.
1242 struct kvm_assigned_pci_dev {
1243 __u32 assigned_dev_id;
1253 The PCI device is specified by the triple segnr, busnr, and devfn.
1254 Identification in succeeding service requests is done via assigned_dev_id. The
1255 following flags are specified:
1257 /* Depends on KVM_CAP_IOMMU */
1258 #define KVM_DEV_ASSIGN_ENABLE_IOMMU (1 << 0)
1259 /* The following two depend on KVM_CAP_PCI_2_3 */
1260 #define KVM_DEV_ASSIGN_PCI_2_3 (1 << 1)
1261 #define KVM_DEV_ASSIGN_MASK_INTX (1 << 2)
1263 If KVM_DEV_ASSIGN_PCI_2_3 is set, the kernel will manage legacy INTx interrupts
1264 via the PCI-2.3-compliant device-level mask, thus enable IRQ sharing with other
1265 assigned devices or host devices. KVM_DEV_ASSIGN_MASK_INTX specifies the
1266 guest's view on the INTx mask, see KVM_ASSIGN_SET_INTX_MASK for details.
1268 The KVM_DEV_ASSIGN_ENABLE_IOMMU flag is a mandatory option to ensure
1269 isolation of the device. Usages not specifying this flag are deprecated.
1271 Only PCI header type 0 devices with PCI BAR resources are supported by
1272 device assignment. The user requesting this ioctl must have read/write
1273 access to the PCI sysfs resource files associated with the device.
1276 4.49 KVM_DEASSIGN_PCI_DEVICE
1278 Capability: KVM_CAP_DEVICE_DEASSIGNMENT
1279 Architectures: x86 ia64
1281 Parameters: struct kvm_assigned_pci_dev (in)
1282 Returns: 0 on success, -1 on error
1284 Ends PCI device assignment, releasing all associated resources.
1286 See KVM_CAP_DEVICE_ASSIGNMENT for the data structure. Only assigned_dev_id is
1287 used in kvm_assigned_pci_dev to identify the device.
1290 4.50 KVM_ASSIGN_DEV_IRQ
1292 Capability: KVM_CAP_ASSIGN_DEV_IRQ
1293 Architectures: x86 ia64
1295 Parameters: struct kvm_assigned_irq (in)
1296 Returns: 0 on success, -1 on error
1298 Assigns an IRQ to a passed-through device.
1300 struct kvm_assigned_irq {
1301 __u32 assigned_dev_id;
1302 __u32 host_irq; /* ignored (legacy field) */
1310 The following flags are defined:
1312 #define KVM_DEV_IRQ_HOST_INTX (1 << 0)
1313 #define KVM_DEV_IRQ_HOST_MSI (1 << 1)
1314 #define KVM_DEV_IRQ_HOST_MSIX (1 << 2)
1316 #define KVM_DEV_IRQ_GUEST_INTX (1 << 8)
1317 #define KVM_DEV_IRQ_GUEST_MSI (1 << 9)
1318 #define KVM_DEV_IRQ_GUEST_MSIX (1 << 10)
1320 It is not valid to specify multiple types per host or guest IRQ. However, the
1321 IRQ type of host and guest can differ or can even be null.
1324 4.51 KVM_DEASSIGN_DEV_IRQ
1326 Capability: KVM_CAP_ASSIGN_DEV_IRQ
1327 Architectures: x86 ia64
1329 Parameters: struct kvm_assigned_irq (in)
1330 Returns: 0 on success, -1 on error
1332 Ends an IRQ assignment to a passed-through device.
1334 See KVM_ASSIGN_DEV_IRQ for the data structure. The target device is specified
1335 by assigned_dev_id, flags must correspond to the IRQ type specified on
1336 KVM_ASSIGN_DEV_IRQ. Partial deassignment of host or guest IRQ is allowed.
1339 4.52 KVM_SET_GSI_ROUTING
1341 Capability: KVM_CAP_IRQ_ROUTING
1342 Architectures: x86 ia64 s390
1344 Parameters: struct kvm_irq_routing (in)
1345 Returns: 0 on success, -1 on error
1347 Sets the GSI routing table entries, overwriting any previously set entries.
1349 struct kvm_irq_routing {
1352 struct kvm_irq_routing_entry entries[0];
1355 No flags are specified so far, the corresponding field must be set to zero.
1357 struct kvm_irq_routing_entry {
1363 struct kvm_irq_routing_irqchip irqchip;
1364 struct kvm_irq_routing_msi msi;
1365 struct kvm_irq_routing_s390_adapter adapter;
1370 /* gsi routing entry types */
1371 #define KVM_IRQ_ROUTING_IRQCHIP 1
1372 #define KVM_IRQ_ROUTING_MSI 2
1373 #define KVM_IRQ_ROUTING_S390_ADAPTER 3
1375 No flags are specified so far, the corresponding field must be set to zero.
1377 struct kvm_irq_routing_irqchip {
1382 struct kvm_irq_routing_msi {
1389 struct kvm_irq_routing_s390_adapter {
1393 __u32 summary_offset;
1398 4.53 KVM_ASSIGN_SET_MSIX_NR
1400 Capability: KVM_CAP_DEVICE_MSIX
1401 Architectures: x86 ia64
1403 Parameters: struct kvm_assigned_msix_nr (in)
1404 Returns: 0 on success, -1 on error
1406 Set the number of MSI-X interrupts for an assigned device. The number is
1407 reset again by terminating the MSI-X assignment of the device via
1408 KVM_DEASSIGN_DEV_IRQ. Calling this service more than once at any earlier
1411 struct kvm_assigned_msix_nr {
1412 __u32 assigned_dev_id;
1417 #define KVM_MAX_MSIX_PER_DEV 256
1420 4.54 KVM_ASSIGN_SET_MSIX_ENTRY
1422 Capability: KVM_CAP_DEVICE_MSIX
1423 Architectures: x86 ia64
1425 Parameters: struct kvm_assigned_msix_entry (in)
1426 Returns: 0 on success, -1 on error
1428 Specifies the routing of an MSI-X assigned device interrupt to a GSI. Setting
1429 the GSI vector to zero means disabling the interrupt.
1431 struct kvm_assigned_msix_entry {
1432 __u32 assigned_dev_id;
1434 __u16 entry; /* The index of entry in the MSI-X table */
1439 4.55 KVM_SET_TSC_KHZ
1441 Capability: KVM_CAP_TSC_CONTROL
1444 Parameters: virtual tsc_khz
1445 Returns: 0 on success, -1 on error
1447 Specifies the tsc frequency for the virtual machine. The unit of the
1451 4.56 KVM_GET_TSC_KHZ
1453 Capability: KVM_CAP_GET_TSC_KHZ
1457 Returns: virtual tsc-khz on success, negative value on error
1459 Returns the tsc frequency of the guest. The unit of the return value is
1460 KHz. If the host has unstable tsc this ioctl returns -EIO instead as an
1466 Capability: KVM_CAP_IRQCHIP
1469 Parameters: struct kvm_lapic_state (out)
1470 Returns: 0 on success, -1 on error
1472 #define KVM_APIC_REG_SIZE 0x400
1473 struct kvm_lapic_state {
1474 char regs[KVM_APIC_REG_SIZE];
1477 Reads the Local APIC registers and copies them into the input argument. The
1478 data format and layout are the same as documented in the architecture manual.
1483 Capability: KVM_CAP_IRQCHIP
1486 Parameters: struct kvm_lapic_state (in)
1487 Returns: 0 on success, -1 on error
1489 #define KVM_APIC_REG_SIZE 0x400
1490 struct kvm_lapic_state {
1491 char regs[KVM_APIC_REG_SIZE];
1494 Copies the input argument into the Local APIC registers. The data format
1495 and layout are the same as documented in the architecture manual.
1500 Capability: KVM_CAP_IOEVENTFD
1503 Parameters: struct kvm_ioeventfd (in)
1504 Returns: 0 on success, !0 on error
1506 This ioctl attaches or detaches an ioeventfd to a legal pio/mmio address
1507 within the guest. A guest write in the registered address will signal the
1508 provided event instead of triggering an exit.
1510 struct kvm_ioeventfd {
1512 __u64 addr; /* legal pio/mmio address */
1513 __u32 len; /* 1, 2, 4, or 8 bytes */
1519 For the special case of virtio-ccw devices on s390, the ioevent is matched
1520 to a subchannel/virtqueue tuple instead.
1522 The following flags are defined:
1524 #define KVM_IOEVENTFD_FLAG_DATAMATCH (1 << kvm_ioeventfd_flag_nr_datamatch)
1525 #define KVM_IOEVENTFD_FLAG_PIO (1 << kvm_ioeventfd_flag_nr_pio)
1526 #define KVM_IOEVENTFD_FLAG_DEASSIGN (1 << kvm_ioeventfd_flag_nr_deassign)
1527 #define KVM_IOEVENTFD_FLAG_VIRTIO_CCW_NOTIFY \
1528 (1 << kvm_ioeventfd_flag_nr_virtio_ccw_notify)
1530 If datamatch flag is set, the event will be signaled only if the written value
1531 to the registered address is equal to datamatch in struct kvm_ioeventfd.
1533 For virtio-ccw devices, addr contains the subchannel id and datamatch the
1539 Capability: KVM_CAP_SW_TLB
1542 Parameters: struct kvm_dirty_tlb (in)
1543 Returns: 0 on success, -1 on error
1545 struct kvm_dirty_tlb {
1550 This must be called whenever userspace has changed an entry in the shared
1551 TLB, prior to calling KVM_RUN on the associated vcpu.
1553 The "bitmap" field is the userspace address of an array. This array
1554 consists of a number of bits, equal to the total number of TLB entries as
1555 determined by the last successful call to KVM_CONFIG_TLB, rounded up to the
1556 nearest multiple of 64.
1558 Each bit corresponds to one TLB entry, ordered the same as in the shared TLB
1561 The array is little-endian: the bit 0 is the least significant bit of the
1562 first byte, bit 8 is the least significant bit of the second byte, etc.
1563 This avoids any complications with differing word sizes.
1565 The "num_dirty" field is a performance hint for KVM to determine whether it
1566 should skip processing the bitmap and just invalidate everything. It must
1567 be set to the number of set bits in the bitmap.
1570 4.61 KVM_ASSIGN_SET_INTX_MASK
1572 Capability: KVM_CAP_PCI_2_3
1575 Parameters: struct kvm_assigned_pci_dev (in)
1576 Returns: 0 on success, -1 on error
1578 Allows userspace to mask PCI INTx interrupts from the assigned device. The
1579 kernel will not deliver INTx interrupts to the guest between setting and
1580 clearing of KVM_ASSIGN_SET_INTX_MASK via this interface. This enables use of
1581 and emulation of PCI 2.3 INTx disable command register behavior.
1583 This may be used for both PCI 2.3 devices supporting INTx disable natively and
1584 older devices lacking this support. Userspace is responsible for emulating the
1585 read value of the INTx disable bit in the guest visible PCI command register.
1586 When modifying the INTx disable state, userspace should precede updating the
1587 physical device command register by calling this ioctl to inform the kernel of
1588 the new intended INTx mask state.
1590 Note that the kernel uses the device INTx disable bit to internally manage the
1591 device interrupt state for PCI 2.3 devices. Reads of this register may
1592 therefore not match the expected value. Writes should always use the guest
1593 intended INTx disable value rather than attempting to read-copy-update the
1594 current physical device state. Races between user and kernel updates to the
1595 INTx disable bit are handled lazily in the kernel. It's possible the device
1596 may generate unintended interrupts, but they will not be injected into the
1599 See KVM_ASSIGN_DEV_IRQ for the data structure. The target device is specified
1600 by assigned_dev_id. In the flags field, only KVM_DEV_ASSIGN_MASK_INTX is
1604 4.62 KVM_CREATE_SPAPR_TCE
1606 Capability: KVM_CAP_SPAPR_TCE
1607 Architectures: powerpc
1609 Parameters: struct kvm_create_spapr_tce (in)
1610 Returns: file descriptor for manipulating the created TCE table
1612 This creates a virtual TCE (translation control entry) table, which
1613 is an IOMMU for PAPR-style virtual I/O. It is used to translate
1614 logical addresses used in virtual I/O into guest physical addresses,
1615 and provides a scatter/gather capability for PAPR virtual I/O.
1617 /* for KVM_CAP_SPAPR_TCE */
1618 struct kvm_create_spapr_tce {
1623 The liobn field gives the logical IO bus number for which to create a
1624 TCE table. The window_size field specifies the size of the DMA window
1625 which this TCE table will translate - the table will contain one 64
1626 bit TCE entry for every 4kiB of the DMA window.
1628 When the guest issues an H_PUT_TCE hcall on a liobn for which a TCE
1629 table has been created using this ioctl(), the kernel will handle it
1630 in real mode, updating the TCE table. H_PUT_TCE calls for other
1631 liobns will cause a vm exit and must be handled by userspace.
1633 The return value is a file descriptor which can be passed to mmap(2)
1634 to map the created TCE table into userspace. This lets userspace read
1635 the entries written by kernel-handled H_PUT_TCE calls, and also lets
1636 userspace update the TCE table directly which is useful in some
1640 4.63 KVM_ALLOCATE_RMA
1642 Capability: KVM_CAP_PPC_RMA
1643 Architectures: powerpc
1645 Parameters: struct kvm_allocate_rma (out)
1646 Returns: file descriptor for mapping the allocated RMA
1648 This allocates a Real Mode Area (RMA) from the pool allocated at boot
1649 time by the kernel. An RMA is a physically-contiguous, aligned region
1650 of memory used on older POWER processors to provide the memory which
1651 will be accessed by real-mode (MMU off) accesses in a KVM guest.
1652 POWER processors support a set of sizes for the RMA that usually
1653 includes 64MB, 128MB, 256MB and some larger powers of two.
1655 /* for KVM_ALLOCATE_RMA */
1656 struct kvm_allocate_rma {
1660 The return value is a file descriptor which can be passed to mmap(2)
1661 to map the allocated RMA into userspace. The mapped area can then be
1662 passed to the KVM_SET_USER_MEMORY_REGION ioctl to establish it as the
1663 RMA for a virtual machine. The size of the RMA in bytes (which is
1664 fixed at host kernel boot time) is returned in the rma_size field of
1665 the argument structure.
1667 The KVM_CAP_PPC_RMA capability is 1 or 2 if the KVM_ALLOCATE_RMA ioctl
1668 is supported; 2 if the processor requires all virtual machines to have
1669 an RMA, or 1 if the processor can use an RMA but doesn't require it,
1670 because it supports the Virtual RMA (VRMA) facility.
1675 Capability: KVM_CAP_USER_NMI
1679 Returns: 0 on success, -1 on error
1681 Queues an NMI on the thread's vcpu. Note this is well defined only
1682 when KVM_CREATE_IRQCHIP has not been called, since this is an interface
1683 between the virtual cpu core and virtual local APIC. After KVM_CREATE_IRQCHIP
1684 has been called, this interface is completely emulated within the kernel.
1686 To use this to emulate the LINT1 input with KVM_CREATE_IRQCHIP, use the
1687 following algorithm:
1690 - read the local APIC's state (KVM_GET_LAPIC)
1691 - check whether changing LINT1 will queue an NMI (see the LVT entry for LINT1)
1692 - if so, issue KVM_NMI
1695 Some guests configure the LINT1 NMI input to cause a panic, aiding in
1699 4.65 KVM_S390_UCAS_MAP
1701 Capability: KVM_CAP_S390_UCONTROL
1704 Parameters: struct kvm_s390_ucas_mapping (in)
1705 Returns: 0 in case of success
1707 The parameter is defined like this:
1708 struct kvm_s390_ucas_mapping {
1714 This ioctl maps the memory at "user_addr" with the length "length" to
1715 the vcpu's address space starting at "vcpu_addr". All parameters need to
1716 be aligned by 1 megabyte.
1719 4.66 KVM_S390_UCAS_UNMAP
1721 Capability: KVM_CAP_S390_UCONTROL
1724 Parameters: struct kvm_s390_ucas_mapping (in)
1725 Returns: 0 in case of success
1727 The parameter is defined like this:
1728 struct kvm_s390_ucas_mapping {
1734 This ioctl unmaps the memory in the vcpu's address space starting at
1735 "vcpu_addr" with the length "length". The field "user_addr" is ignored.
1736 All parameters need to be aligned by 1 megabyte.
1739 4.67 KVM_S390_VCPU_FAULT
1741 Capability: KVM_CAP_S390_UCONTROL
1744 Parameters: vcpu absolute address (in)
1745 Returns: 0 in case of success
1747 This call creates a page table entry on the virtual cpu's address space
1748 (for user controlled virtual machines) or the virtual machine's address
1749 space (for regular virtual machines). This only works for minor faults,
1750 thus it's recommended to access subject memory page via the user page
1751 table upfront. This is useful to handle validity intercepts for user
1752 controlled virtual machines to fault in the virtual cpu's lowcore pages
1753 prior to calling the KVM_RUN ioctl.
1756 4.68 KVM_SET_ONE_REG
1758 Capability: KVM_CAP_ONE_REG
1761 Parameters: struct kvm_one_reg (in)
1762 Returns: 0 on success, negative value on failure
1764 struct kvm_one_reg {
1769 Using this ioctl, a single vcpu register can be set to a specific value
1770 defined by user space with the passed in struct kvm_one_reg, where id
1771 refers to the register identifier as described below and addr is a pointer
1772 to a variable with the respective size. There can be architecture agnostic
1773 and architecture specific registers. Each have their own range of operation
1774 and their own constants and width. To keep track of the implemented
1775 registers, find a list below:
1777 Arch | Register | Width (bits)
1779 PPC | KVM_REG_PPC_HIOR | 64
1780 PPC | KVM_REG_PPC_IAC1 | 64
1781 PPC | KVM_REG_PPC_IAC2 | 64
1782 PPC | KVM_REG_PPC_IAC3 | 64
1783 PPC | KVM_REG_PPC_IAC4 | 64
1784 PPC | KVM_REG_PPC_DAC1 | 64
1785 PPC | KVM_REG_PPC_DAC2 | 64
1786 PPC | KVM_REG_PPC_DABR | 64
1787 PPC | KVM_REG_PPC_DSCR | 64
1788 PPC | KVM_REG_PPC_PURR | 64
1789 PPC | KVM_REG_PPC_SPURR | 64
1790 PPC | KVM_REG_PPC_DAR | 64
1791 PPC | KVM_REG_PPC_DSISR | 32
1792 PPC | KVM_REG_PPC_AMR | 64
1793 PPC | KVM_REG_PPC_UAMOR | 64
1794 PPC | KVM_REG_PPC_MMCR0 | 64
1795 PPC | KVM_REG_PPC_MMCR1 | 64
1796 PPC | KVM_REG_PPC_MMCRA | 64
1797 PPC | KVM_REG_PPC_PMC1 | 32
1798 PPC | KVM_REG_PPC_PMC2 | 32
1799 PPC | KVM_REG_PPC_PMC3 | 32
1800 PPC | KVM_REG_PPC_PMC4 | 32
1801 PPC | KVM_REG_PPC_PMC5 | 32
1802 PPC | KVM_REG_PPC_PMC6 | 32
1803 PPC | KVM_REG_PPC_PMC7 | 32
1804 PPC | KVM_REG_PPC_PMC8 | 32
1805 PPC | KVM_REG_PPC_FPR0 | 64
1807 PPC | KVM_REG_PPC_FPR31 | 64
1808 PPC | KVM_REG_PPC_VR0 | 128
1810 PPC | KVM_REG_PPC_VR31 | 128
1811 PPC | KVM_REG_PPC_VSR0 | 128
1813 PPC | KVM_REG_PPC_VSR31 | 128
1814 PPC | KVM_REG_PPC_FPSCR | 64
1815 PPC | KVM_REG_PPC_VSCR | 32
1816 PPC | KVM_REG_PPC_VPA_ADDR | 64
1817 PPC | KVM_REG_PPC_VPA_SLB | 128
1818 PPC | KVM_REG_PPC_VPA_DTL | 128
1819 PPC | KVM_REG_PPC_EPCR | 32
1820 PPC | KVM_REG_PPC_EPR | 32
1821 PPC | KVM_REG_PPC_TCR | 32
1822 PPC | KVM_REG_PPC_TSR | 32
1823 PPC | KVM_REG_PPC_OR_TSR | 32
1824 PPC | KVM_REG_PPC_CLEAR_TSR | 32
1825 PPC | KVM_REG_PPC_MAS0 | 32
1826 PPC | KVM_REG_PPC_MAS1 | 32
1827 PPC | KVM_REG_PPC_MAS2 | 64
1828 PPC | KVM_REG_PPC_MAS7_3 | 64
1829 PPC | KVM_REG_PPC_MAS4 | 32
1830 PPC | KVM_REG_PPC_MAS6 | 32
1831 PPC | KVM_REG_PPC_MMUCFG | 32
1832 PPC | KVM_REG_PPC_TLB0CFG | 32
1833 PPC | KVM_REG_PPC_TLB1CFG | 32
1834 PPC | KVM_REG_PPC_TLB2CFG | 32
1835 PPC | KVM_REG_PPC_TLB3CFG | 32
1836 PPC | KVM_REG_PPC_TLB0PS | 32
1837 PPC | KVM_REG_PPC_TLB1PS | 32
1838 PPC | KVM_REG_PPC_TLB2PS | 32
1839 PPC | KVM_REG_PPC_TLB3PS | 32
1840 PPC | KVM_REG_PPC_EPTCFG | 32
1841 PPC | KVM_REG_PPC_ICP_STATE | 64
1842 PPC | KVM_REG_PPC_TB_OFFSET | 64
1843 PPC | KVM_REG_PPC_SPMC1 | 32
1844 PPC | KVM_REG_PPC_SPMC2 | 32
1845 PPC | KVM_REG_PPC_IAMR | 64
1846 PPC | KVM_REG_PPC_TFHAR | 64
1847 PPC | KVM_REG_PPC_TFIAR | 64
1848 PPC | KVM_REG_PPC_TEXASR | 64
1849 PPC | KVM_REG_PPC_FSCR | 64
1850 PPC | KVM_REG_PPC_PSPB | 32
1851 PPC | KVM_REG_PPC_EBBHR | 64
1852 PPC | KVM_REG_PPC_EBBRR | 64
1853 PPC | KVM_REG_PPC_BESCR | 64
1854 PPC | KVM_REG_PPC_TAR | 64
1855 PPC | KVM_REG_PPC_DPDES | 64
1856 PPC | KVM_REG_PPC_DAWR | 64
1857 PPC | KVM_REG_PPC_DAWRX | 64
1858 PPC | KVM_REG_PPC_CIABR | 64
1859 PPC | KVM_REG_PPC_IC | 64
1860 PPC | KVM_REG_PPC_VTB | 64
1861 PPC | KVM_REG_PPC_CSIGR | 64
1862 PPC | KVM_REG_PPC_TACR | 64
1863 PPC | KVM_REG_PPC_TCSCR | 64
1864 PPC | KVM_REG_PPC_PID | 64
1865 PPC | KVM_REG_PPC_ACOP | 64
1866 PPC | KVM_REG_PPC_VRSAVE | 32
1867 PPC | KVM_REG_PPC_LPCR | 64
1868 PPC | KVM_REG_PPC_PPR | 64
1869 PPC | KVM_REG_PPC_ARCH_COMPAT 32
1870 PPC | KVM_REG_PPC_DABRX | 32
1871 PPC | KVM_REG_PPC_TM_GPR0 | 64
1873 PPC | KVM_REG_PPC_TM_GPR31 | 64
1874 PPC | KVM_REG_PPC_TM_VSR0 | 128
1876 PPC | KVM_REG_PPC_TM_VSR63 | 128
1877 PPC | KVM_REG_PPC_TM_CR | 64
1878 PPC | KVM_REG_PPC_TM_LR | 64
1879 PPC | KVM_REG_PPC_TM_CTR | 64
1880 PPC | KVM_REG_PPC_TM_FPSCR | 64
1881 PPC | KVM_REG_PPC_TM_AMR | 64
1882 PPC | KVM_REG_PPC_TM_PPR | 64
1883 PPC | KVM_REG_PPC_TM_VRSAVE | 64
1884 PPC | KVM_REG_PPC_TM_VSCR | 32
1885 PPC | KVM_REG_PPC_TM_DSCR | 64
1886 PPC | KVM_REG_PPC_TM_TAR | 64
1888 ARM registers are mapped using the lower 32 bits. The upper 16 of that
1889 is the register group type, or coprocessor number:
1891 ARM core registers have the following id bit patterns:
1892 0x4020 0000 0010 <index into the kvm_regs struct:16>
1894 ARM 32-bit CP15 registers have the following id bit patterns:
1895 0x4020 0000 000F <zero:1> <crn:4> <crm:4> <opc1:4> <opc2:3>
1897 ARM 64-bit CP15 registers have the following id bit patterns:
1898 0x4030 0000 000F <zero:1> <zero:4> <crm:4> <opc1:4> <zero:3>
1900 ARM CCSIDR registers are demultiplexed by CSSELR value:
1901 0x4020 0000 0011 00 <csselr:8>
1903 ARM 32-bit VFP control registers have the following id bit patterns:
1904 0x4020 0000 0012 1 <regno:12>
1906 ARM 64-bit FP registers have the following id bit patterns:
1907 0x4030 0000 0012 0 <regno:12>
1910 arm64 registers are mapped using the lower 32 bits. The upper 16 of
1911 that is the register group type, or coprocessor number:
1913 arm64 core/FP-SIMD registers have the following id bit patterns. Note
1914 that the size of the access is variable, as the kvm_regs structure
1915 contains elements ranging from 32 to 128 bits. The index is a 32bit
1916 value in the kvm_regs structure seen as a 32bit array.
1917 0x60x0 0000 0010 <index into the kvm_regs struct:16>
1919 arm64 CCSIDR registers are demultiplexed by CSSELR value:
1920 0x6020 0000 0011 00 <csselr:8>
1922 arm64 system registers have the following id bit patterns:
1923 0x6030 0000 0013 <op0:2> <op1:3> <crn:4> <crm:4> <op2:3>
1925 4.69 KVM_GET_ONE_REG
1927 Capability: KVM_CAP_ONE_REG
1930 Parameters: struct kvm_one_reg (in and out)
1931 Returns: 0 on success, negative value on failure
1933 This ioctl allows to receive the value of a single register implemented
1934 in a vcpu. The register to read is indicated by the "id" field of the
1935 kvm_one_reg struct passed in. On success, the register value can be found
1936 at the memory location pointed to by "addr".
1938 The list of registers accessible using this interface is identical to the
1942 4.70 KVM_KVMCLOCK_CTRL
1944 Capability: KVM_CAP_KVMCLOCK_CTRL
1945 Architectures: Any that implement pvclocks (currently x86 only)
1948 Returns: 0 on success, -1 on error
1950 This signals to the host kernel that the specified guest is being paused by
1951 userspace. The host will set a flag in the pvclock structure that is checked
1952 from the soft lockup watchdog. The flag is part of the pvclock structure that
1953 is shared between guest and host, specifically the second bit of the flags
1954 field of the pvclock_vcpu_time_info structure. It will be set exclusively by
1955 the host and read/cleared exclusively by the guest. The guest operation of
1956 checking and clearing the flag must an atomic operation so
1957 load-link/store-conditional, or equivalent must be used. There are two cases
1958 where the guest will clear the flag: when the soft lockup watchdog timer resets
1959 itself or when a soft lockup is detected. This ioctl can be called any time
1960 after pausing the vcpu, but before it is resumed.
1965 Capability: KVM_CAP_SIGNAL_MSI
1968 Parameters: struct kvm_msi (in)
1969 Returns: >0 on delivery, 0 if guest blocked the MSI, and -1 on error
1971 Directly inject a MSI message. Only valid with in-kernel irqchip that handles
1982 No flags are defined so far. The corresponding field must be 0.
1985 4.71 KVM_CREATE_PIT2
1987 Capability: KVM_CAP_PIT2
1990 Parameters: struct kvm_pit_config (in)
1991 Returns: 0 on success, -1 on error
1993 Creates an in-kernel device model for the i8254 PIT. This call is only valid
1994 after enabling in-kernel irqchip support via KVM_CREATE_IRQCHIP. The following
1995 parameters have to be passed:
1997 struct kvm_pit_config {
2004 #define KVM_PIT_SPEAKER_DUMMY 1 /* emulate speaker port stub */
2006 PIT timer interrupts may use a per-VM kernel thread for injection. If it
2007 exists, this thread will have a name of the following pattern:
2009 kvm-pit/<owner-process-pid>
2011 When running a guest with elevated priorities, the scheduling parameters of
2012 this thread may have to be adjusted accordingly.
2014 This IOCTL replaces the obsolete KVM_CREATE_PIT.
2019 Capability: KVM_CAP_PIT_STATE2
2022 Parameters: struct kvm_pit_state2 (out)
2023 Returns: 0 on success, -1 on error
2025 Retrieves the state of the in-kernel PIT model. Only valid after
2026 KVM_CREATE_PIT2. The state is returned in the following structure:
2028 struct kvm_pit_state2 {
2029 struct kvm_pit_channel_state channels[3];
2036 /* disable PIT in HPET legacy mode */
2037 #define KVM_PIT_FLAGS_HPET_LEGACY 0x00000001
2039 This IOCTL replaces the obsolete KVM_GET_PIT.
2044 Capability: KVM_CAP_PIT_STATE2
2047 Parameters: struct kvm_pit_state2 (in)
2048 Returns: 0 on success, -1 on error
2050 Sets the state of the in-kernel PIT model. Only valid after KVM_CREATE_PIT2.
2051 See KVM_GET_PIT2 for details on struct kvm_pit_state2.
2053 This IOCTL replaces the obsolete KVM_SET_PIT.
2056 4.74 KVM_PPC_GET_SMMU_INFO
2058 Capability: KVM_CAP_PPC_GET_SMMU_INFO
2059 Architectures: powerpc
2062 Returns: 0 on success, -1 on error
2064 This populates and returns a structure describing the features of
2065 the "Server" class MMU emulation supported by KVM.
2066 This can in turn be used by userspace to generate the appropriate
2067 device-tree properties for the guest operating system.
2069 The structure contains some global informations, followed by an
2070 array of supported segment page sizes:
2072 struct kvm_ppc_smmu_info {
2076 struct kvm_ppc_one_seg_page_size sps[KVM_PPC_PAGE_SIZES_MAX_SZ];
2079 The supported flags are:
2081 - KVM_PPC_PAGE_SIZES_REAL:
2082 When that flag is set, guest page sizes must "fit" the backing
2083 store page sizes. When not set, any page size in the list can
2084 be used regardless of how they are backed by userspace.
2086 - KVM_PPC_1T_SEGMENTS
2087 The emulated MMU supports 1T segments in addition to the
2090 The "slb_size" field indicates how many SLB entries are supported
2092 The "sps" array contains 8 entries indicating the supported base
2093 page sizes for a segment in increasing order. Each entry is defined
2096 struct kvm_ppc_one_seg_page_size {
2097 __u32 page_shift; /* Base page shift of segment (or 0) */
2098 __u32 slb_enc; /* SLB encoding for BookS */
2099 struct kvm_ppc_one_page_size enc[KVM_PPC_PAGE_SIZES_MAX_SZ];
2102 An entry with a "page_shift" of 0 is unused. Because the array is
2103 organized in increasing order, a lookup can stop when encoutering
2106 The "slb_enc" field provides the encoding to use in the SLB for the
2107 page size. The bits are in positions such as the value can directly
2108 be OR'ed into the "vsid" argument of the slbmte instruction.
2110 The "enc" array is a list which for each of those segment base page
2111 size provides the list of supported actual page sizes (which can be
2112 only larger or equal to the base page size), along with the
2113 corresponding encoding in the hash PTE. Similarly, the array is
2114 8 entries sorted by increasing sizes and an entry with a "0" shift
2115 is an empty entry and a terminator:
2117 struct kvm_ppc_one_page_size {
2118 __u32 page_shift; /* Page shift (or 0) */
2119 __u32 pte_enc; /* Encoding in the HPTE (>>12) */
2122 The "pte_enc" field provides a value that can OR'ed into the hash
2123 PTE's RPN field (ie, it needs to be shifted left by 12 to OR it
2124 into the hash PTE second double word).
2128 Capability: KVM_CAP_IRQFD
2131 Parameters: struct kvm_irqfd (in)
2132 Returns: 0 on success, -1 on error
2134 Allows setting an eventfd to directly trigger a guest interrupt.
2135 kvm_irqfd.fd specifies the file descriptor to use as the eventfd and
2136 kvm_irqfd.gsi specifies the irqchip pin toggled by this event. When
2137 an event is triggered on the eventfd, an interrupt is injected into
2138 the guest using the specified gsi pin. The irqfd is removed using
2139 the KVM_IRQFD_FLAG_DEASSIGN flag, specifying both kvm_irqfd.fd
2142 With KVM_CAP_IRQFD_RESAMPLE, KVM_IRQFD supports a de-assert and notify
2143 mechanism allowing emulation of level-triggered, irqfd-based
2144 interrupts. When KVM_IRQFD_FLAG_RESAMPLE is set the user must pass an
2145 additional eventfd in the kvm_irqfd.resamplefd field. When operating
2146 in resample mode, posting of an interrupt through kvm_irq.fd asserts
2147 the specified gsi in the irqchip. When the irqchip is resampled, such
2148 as from an EOI, the gsi is de-asserted and the user is notified via
2149 kvm_irqfd.resamplefd. It is the user's responsibility to re-queue
2150 the interrupt if the device making use of it still requires service.
2151 Note that closing the resamplefd is not sufficient to disable the
2152 irqfd. The KVM_IRQFD_FLAG_RESAMPLE is only necessary on assignment
2153 and need not be specified with KVM_IRQFD_FLAG_DEASSIGN.
2155 4.76 KVM_PPC_ALLOCATE_HTAB
2157 Capability: KVM_CAP_PPC_ALLOC_HTAB
2158 Architectures: powerpc
2160 Parameters: Pointer to u32 containing hash table order (in/out)
2161 Returns: 0 on success, -1 on error
2163 This requests the host kernel to allocate an MMU hash table for a
2164 guest using the PAPR paravirtualization interface. This only does
2165 anything if the kernel is configured to use the Book 3S HV style of
2166 virtualization. Otherwise the capability doesn't exist and the ioctl
2167 returns an ENOTTY error. The rest of this description assumes Book 3S
2170 There must be no vcpus running when this ioctl is called; if there
2171 are, it will do nothing and return an EBUSY error.
2173 The parameter is a pointer to a 32-bit unsigned integer variable
2174 containing the order (log base 2) of the desired size of the hash
2175 table, which must be between 18 and 46. On successful return from the
2176 ioctl, it will have been updated with the order of the hash table that
2179 If no hash table has been allocated when any vcpu is asked to run
2180 (with the KVM_RUN ioctl), the host kernel will allocate a
2181 default-sized hash table (16 MB).
2183 If this ioctl is called when a hash table has already been allocated,
2184 the kernel will clear out the existing hash table (zero all HPTEs) and
2185 return the hash table order in the parameter. (If the guest is using
2186 the virtualized real-mode area (VRMA) facility, the kernel will
2187 re-create the VMRA HPTEs on the next KVM_RUN of any vcpu.)
2189 4.77 KVM_S390_INTERRUPT
2193 Type: vm ioctl, vcpu ioctl
2194 Parameters: struct kvm_s390_interrupt (in)
2195 Returns: 0 on success, -1 on error
2197 Allows to inject an interrupt to the guest. Interrupts can be floating
2198 (vm ioctl) or per cpu (vcpu ioctl), depending on the interrupt type.
2200 Interrupt parameters are passed via kvm_s390_interrupt:
2202 struct kvm_s390_interrupt {
2208 type can be one of the following:
2210 KVM_S390_SIGP_STOP (vcpu) - sigp restart
2211 KVM_S390_PROGRAM_INT (vcpu) - program check; code in parm
2212 KVM_S390_SIGP_SET_PREFIX (vcpu) - sigp set prefix; prefix address in parm
2213 KVM_S390_RESTART (vcpu) - restart
2214 KVM_S390_INT_VIRTIO (vm) - virtio external interrupt; external interrupt
2215 parameters in parm and parm64
2216 KVM_S390_INT_SERVICE (vm) - sclp external interrupt; sclp parameter in parm
2217 KVM_S390_INT_EMERGENCY (vcpu) - sigp emergency; source cpu in parm
2218 KVM_S390_INT_EXTERNAL_CALL (vcpu) - sigp external call; source cpu in parm
2219 KVM_S390_INT_IO(ai,cssid,ssid,schid) (vm) - compound value to indicate an
2220 I/O interrupt (ai - adapter interrupt; cssid,ssid,schid - subchannel);
2221 I/O interruption parameters in parm (subchannel) and parm64 (intparm,
2222 interruption subclass)
2223 KVM_S390_MCHK (vm, vcpu) - machine check interrupt; cr 14 bits in parm,
2224 machine check interrupt code in parm64 (note that
2225 machine checks needing further payload are not
2226 supported by this ioctl)
2228 Note that the vcpu ioctl is asynchronous to vcpu execution.
2230 4.78 KVM_PPC_GET_HTAB_FD
2232 Capability: KVM_CAP_PPC_HTAB_FD
2233 Architectures: powerpc
2235 Parameters: Pointer to struct kvm_get_htab_fd (in)
2236 Returns: file descriptor number (>= 0) on success, -1 on error
2238 This returns a file descriptor that can be used either to read out the
2239 entries in the guest's hashed page table (HPT), or to write entries to
2240 initialize the HPT. The returned fd can only be written to if the
2241 KVM_GET_HTAB_WRITE bit is set in the flags field of the argument, and
2242 can only be read if that bit is clear. The argument struct looks like
2245 /* For KVM_PPC_GET_HTAB_FD */
2246 struct kvm_get_htab_fd {
2252 /* Values for kvm_get_htab_fd.flags */
2253 #define KVM_GET_HTAB_BOLTED_ONLY ((__u64)0x1)
2254 #define KVM_GET_HTAB_WRITE ((__u64)0x2)
2256 The `start_index' field gives the index in the HPT of the entry at
2257 which to start reading. It is ignored when writing.
2259 Reads on the fd will initially supply information about all
2260 "interesting" HPT entries. Interesting entries are those with the
2261 bolted bit set, if the KVM_GET_HTAB_BOLTED_ONLY bit is set, otherwise
2262 all entries. When the end of the HPT is reached, the read() will
2263 return. If read() is called again on the fd, it will start again from
2264 the beginning of the HPT, but will only return HPT entries that have
2265 changed since they were last read.
2267 Data read or written is structured as a header (8 bytes) followed by a
2268 series of valid HPT entries (16 bytes) each. The header indicates how
2269 many valid HPT entries there are and how many invalid entries follow
2270 the valid entries. The invalid entries are not represented explicitly
2271 in the stream. The header format is:
2273 struct kvm_get_htab_header {
2279 Writes to the fd create HPT entries starting at the index given in the
2280 header; first `n_valid' valid entries with contents from the data
2281 written, then `n_invalid' invalid entries, invalidating any previously
2282 valid entries found.
2284 4.79 KVM_CREATE_DEVICE
2286 Capability: KVM_CAP_DEVICE_CTRL
2288 Parameters: struct kvm_create_device (in/out)
2289 Returns: 0 on success, -1 on error
2291 ENODEV: The device type is unknown or unsupported
2292 EEXIST: Device already created, and this type of device may not
2293 be instantiated multiple times
2295 Other error conditions may be defined by individual device types or
2296 have their standard meanings.
2298 Creates an emulated device in the kernel. The file descriptor returned
2299 in fd can be used with KVM_SET/GET/HAS_DEVICE_ATTR.
2301 If the KVM_CREATE_DEVICE_TEST flag is set, only test whether the
2302 device type is supported (not necessarily whether it can be created
2305 Individual devices should not define flags. Attributes should be used
2306 for specifying any behavior that is not implied by the device type
2309 struct kvm_create_device {
2310 __u32 type; /* in: KVM_DEV_TYPE_xxx */
2311 __u32 fd; /* out: device handle */
2312 __u32 flags; /* in: KVM_CREATE_DEVICE_xxx */
2315 4.80 KVM_SET_DEVICE_ATTR/KVM_GET_DEVICE_ATTR
2317 Capability: KVM_CAP_DEVICE_CTRL
2319 Parameters: struct kvm_device_attr
2320 Returns: 0 on success, -1 on error
2322 ENXIO: The group or attribute is unknown/unsupported for this device
2323 EPERM: The attribute cannot (currently) be accessed this way
2324 (e.g. read-only attribute, or attribute that only makes
2325 sense when the device is in a different state)
2327 Other error conditions may be defined by individual device types.
2329 Gets/sets a specified piece of device configuration and/or state. The
2330 semantics are device-specific. See individual device documentation in
2331 the "devices" directory. As with ONE_REG, the size of the data
2332 transferred is defined by the particular attribute.
2334 struct kvm_device_attr {
2335 __u32 flags; /* no flags currently defined */
2336 __u32 group; /* device-defined */
2337 __u64 attr; /* group-defined */
2338 __u64 addr; /* userspace address of attr data */
2341 4.81 KVM_HAS_DEVICE_ATTR
2343 Capability: KVM_CAP_DEVICE_CTRL
2345 Parameters: struct kvm_device_attr
2346 Returns: 0 on success, -1 on error
2348 ENXIO: The group or attribute is unknown/unsupported for this device
2350 Tests whether a device supports a particular attribute. A successful
2351 return indicates the attribute is implemented. It does not necessarily
2352 indicate that the attribute can be read or written in the device's
2353 current state. "addr" is ignored.
2355 4.82 KVM_ARM_VCPU_INIT
2358 Architectures: arm, arm64
2360 Parameters: struct kvm_vcpu_init (in)
2361 Returns: 0 on success; -1 on error
2363 Â EINVAL: Â Â Â the target is unknown, or the combination of features is invalid.
2364 Â ENOENT: Â Â Â a features bit specified is unknown.
2366 This tells KVM what type of CPU to present to the guest, and what
2367 optional features it should have. Â This will cause a reset of the cpu
2368 registers to their initial values. Â If this is not called, KVM_RUN will
2369 return ENOEXEC for that vcpu.
2371 Note that because some registers reflect machine topology, all vcpus
2372 should be created before this ioctl is invoked.
2375 - KVM_ARM_VCPU_POWER_OFF: Starts the CPU in a power-off state.
2376 Depends on KVM_CAP_ARM_PSCI.
2377 - KVM_ARM_VCPU_EL1_32BIT: Starts the CPU in a 32bit mode.
2378 Depends on KVM_CAP_ARM_EL1_32BIT (arm64 only).
2381 4.83 KVM_ARM_PREFERRED_TARGET
2384 Architectures: arm, arm64
2386 Parameters: struct struct kvm_vcpu_init (out)
2387 Returns: 0 on success; -1 on error
2389 ENODEV: no preferred target available for the host
2391 This queries KVM for preferred CPU target type which can be emulated
2392 by KVM on underlying host.
2394 The ioctl returns struct kvm_vcpu_init instance containing information
2395 about preferred CPU target type and recommended features for it. The
2396 kvm_vcpu_init->features bitmap returned will have feature bits set if
2397 the preferred target recommends setting these features, but this is
2400 The information returned by this ioctl can be used to prepare an instance
2401 of struct kvm_vcpu_init for KVM_ARM_VCPU_INIT ioctl which will result in
2402 in VCPU matching underlying host.
2405 4.84 KVM_GET_REG_LIST
2408 Architectures: arm, arm64
2410 Parameters: struct kvm_reg_list (in/out)
2411 Returns: 0 on success; -1 on error
2413 Â E2BIG: Â Â Â Â the reg index list is too big to fit in the array specified by
2414 Â Â Â Â Â Â Â Â Â Â Â Â the user (the number required will be written into n).
2416 struct kvm_reg_list {
2417 __u64 n; /* number of registers in reg[] */
2421 This ioctl returns the guest registers that are supported for the
2422 KVM_GET_ONE_REG/KVM_SET_ONE_REG calls.
2425 4.85 KVM_ARM_SET_DEVICE_ADDR (deprecated)
2427 Capability: KVM_CAP_ARM_SET_DEVICE_ADDR
2428 Architectures: arm, arm64
2430 Parameters: struct kvm_arm_device_address (in)
2431 Returns: 0 on success, -1 on error
2433 ENODEV: The device id is unknown
2434 ENXIO: Device not supported on current system
2435 EEXIST: Address already set
2436 E2BIG: Address outside guest physical address space
2437 EBUSY: Address overlaps with other device range
2439 struct kvm_arm_device_addr {
2444 Specify a device address in the guest's physical address space where guests
2445 can access emulated or directly exposed devices, which the host kernel needs
2446 to know about. The id field is an architecture specific identifier for a
2449 ARM/arm64 divides the id field into two parts, a device id and an
2450 address type id specific to the individual device.
2452 Â bits: | 63 ... 32 | 31 ... 16 | 15 ... 0 |
2453 field: | 0x00000000 | device id | addr type id |
2455 ARM/arm64 currently only require this when using the in-kernel GIC
2456 support for the hardware VGIC features, using KVM_ARM_DEVICE_VGIC_V2
2457 as the device id. When setting the base address for the guest's
2458 mapping of the VGIC virtual CPU and distributor interface, the ioctl
2459 must be called after calling KVM_CREATE_IRQCHIP, but before calling
2460 KVM_RUN on any of the VCPUs. Calling this ioctl twice for any of the
2461 base addresses will return -EEXIST.
2463 Note, this IOCTL is deprecated and the more flexible SET/GET_DEVICE_ATTR API
2464 should be used instead.
2467 4.86 KVM_PPC_RTAS_DEFINE_TOKEN
2469 Capability: KVM_CAP_PPC_RTAS
2472 Parameters: struct kvm_rtas_token_args
2473 Returns: 0 on success, -1 on error
2475 Defines a token value for a RTAS (Run Time Abstraction Services)
2476 service in order to allow it to be handled in the kernel. The
2477 argument struct gives the name of the service, which must be the name
2478 of a service that has a kernel-side implementation. If the token
2479 value is non-zero, it will be associated with that service, and
2480 subsequent RTAS calls by the guest specifying that token will be
2481 handled by the kernel. If the token value is 0, then any token
2482 associated with the service will be forgotten, and subsequent RTAS
2483 calls by the guest for that service will be passed to userspace to be
2487 5. The kvm_run structure
2488 ------------------------
2490 Application code obtains a pointer to the kvm_run structure by
2491 mmap()ing a vcpu fd. From that point, application code can control
2492 execution by changing fields in kvm_run prior to calling the KVM_RUN
2493 ioctl, and obtain information about the reason KVM_RUN returned by
2494 looking up structure members.
2498 __u8 request_interrupt_window;
2500 Request that KVM_RUN return when it becomes possible to inject external
2501 interrupts into the guest. Useful in conjunction with KVM_INTERRUPT.
2508 When KVM_RUN has returned successfully (return value 0), this informs
2509 application code why KVM_RUN has returned. Allowable values for this
2510 field are detailed below.
2512 __u8 ready_for_interrupt_injection;
2514 If request_interrupt_window has been specified, this field indicates
2515 an interrupt can be injected now with KVM_INTERRUPT.
2519 The value of the current interrupt flag. Only valid if in-kernel
2520 local APIC is not used.
2524 /* in (pre_kvm_run), out (post_kvm_run) */
2527 The value of the cr8 register. Only valid if in-kernel local APIC is
2528 not used. Both input and output.
2532 The value of the APIC BASE msr. Only valid if in-kernel local
2533 APIC is not used. Both input and output.
2536 /* KVM_EXIT_UNKNOWN */
2538 __u64 hardware_exit_reason;
2541 If exit_reason is KVM_EXIT_UNKNOWN, the vcpu has exited due to unknown
2542 reasons. Further architecture-specific information is available in
2543 hardware_exit_reason.
2545 /* KVM_EXIT_FAIL_ENTRY */
2547 __u64 hardware_entry_failure_reason;
2550 If exit_reason is KVM_EXIT_FAIL_ENTRY, the vcpu could not be run due
2551 to unknown reasons. Further architecture-specific information is
2552 available in hardware_entry_failure_reason.
2554 /* KVM_EXIT_EXCEPTION */
2564 #define KVM_EXIT_IO_IN 0
2565 #define KVM_EXIT_IO_OUT 1
2567 __u8 size; /* bytes */
2570 __u64 data_offset; /* relative to kvm_run start */
2573 If exit_reason is KVM_EXIT_IO, then the vcpu has
2574 executed a port I/O instruction which could not be satisfied by kvm.
2575 data_offset describes where the data is located (KVM_EXIT_IO_OUT) or
2576 where kvm expects application code to place the data for the next
2577 KVM_RUN invocation (KVM_EXIT_IO_IN). Data format is a packed array.
2580 struct kvm_debug_exit_arch arch;
2593 If exit_reason is KVM_EXIT_MMIO, then the vcpu has
2594 executed a memory-mapped I/O instruction which could not be satisfied
2595 by kvm. The 'data' member contains the written data if 'is_write' is
2596 true, and should be filled by application code otherwise.
2598 The 'data' member contains, in its first 'len' bytes, the value as it would
2599 appear if the VCPU performed a load or store of the appropriate width directly
2602 NOTE: For KVM_EXIT_IO, KVM_EXIT_MMIO, KVM_EXIT_OSI, KVM_EXIT_DCR,
2603 KVM_EXIT_PAPR and KVM_EXIT_EPR the corresponding
2604 operations are complete (and guest state is consistent) only after userspace
2605 has re-entered the kernel with KVM_RUN. The kernel side will first finish
2606 incomplete operations and then check for pending signals. Userspace
2607 can re-enter the guest with an unmasked signal pending to complete
2610 /* KVM_EXIT_HYPERCALL */
2619 Unused. This was once used for 'hypercall to userspace'. To implement
2620 such functionality, use KVM_EXIT_IO (x86) or KVM_EXIT_MMIO (all except s390).
2621 Note KVM_EXIT_IO is significantly faster than KVM_EXIT_MMIO.
2623 /* KVM_EXIT_TPR_ACCESS */
2630 To be documented (KVM_TPR_ACCESS_REPORTING).
2632 /* KVM_EXIT_S390_SIEIC */
2635 __u64 mask; /* psw upper half */
2636 __u64 addr; /* psw lower half */
2643 /* KVM_EXIT_S390_RESET */
2644 #define KVM_S390_RESET_POR 1
2645 #define KVM_S390_RESET_CLEAR 2
2646 #define KVM_S390_RESET_SUBSYSTEM 4
2647 #define KVM_S390_RESET_CPU_INIT 8
2648 #define KVM_S390_RESET_IPL 16
2649 __u64 s390_reset_flags;
2653 /* KVM_EXIT_S390_UCONTROL */
2655 __u64 trans_exc_code;
2659 s390 specific. A page fault has occurred for a user controlled virtual
2660 machine (KVM_VM_S390_UNCONTROL) on it's host page table that cannot be
2661 resolved by the kernel.
2662 The program code and the translation exception code that were placed
2663 in the cpu's lowcore are presented here as defined by the z Architecture
2664 Principles of Operation Book in the Chapter for Dynamic Address Translation
2681 MOL uses a special hypercall interface it calls 'OSI'. To enable it, we catch
2682 hypercalls and exit with this exit struct that contains all the guest gprs.
2684 If exit_reason is KVM_EXIT_OSI, then the vcpu has triggered such a hypercall.
2685 Userspace can now handle the hypercall and when it's done modify the gprs as
2686 necessary. Upon guest entry all guest GPRs will then be replaced by the values
2689 /* KVM_EXIT_PAPR_HCALL */
2696 This is used on 64-bit PowerPC when emulating a pSeries partition,
2697 e.g. with the 'pseries' machine type in qemu. It occurs when the
2698 guest does a hypercall using the 'sc 1' instruction. The 'nr' field
2699 contains the hypercall number (from the guest R3), and 'args' contains
2700 the arguments (from the guest R4 - R12). Userspace should put the
2701 return code in 'ret' and any extra returned values in args[].
2702 The possible hypercalls are defined in the Power Architecture Platform
2703 Requirements (PAPR) document available from www.power.org (free
2704 developer registration required to access it).
2706 /* KVM_EXIT_S390_TSCH */
2708 __u16 subchannel_id;
2709 __u16 subchannel_nr;
2716 s390 specific. This exit occurs when KVM_CAP_S390_CSS_SUPPORT has been enabled
2717 and TEST SUBCHANNEL was intercepted. If dequeued is set, a pending I/O
2718 interrupt for the target subchannel has been dequeued and subchannel_id,
2719 subchannel_nr, io_int_parm and io_int_word contain the parameters for that
2720 interrupt. ipb is needed for instruction parameter decoding.
2727 On FSL BookE PowerPC chips, the interrupt controller has a fast patch
2728 interrupt acknowledge path to the core. When the core successfully
2729 delivers an interrupt, it automatically populates the EPR register with
2730 the interrupt vector number and acknowledges the interrupt inside
2731 the interrupt controller.
2733 In case the interrupt controller lives in user space, we need to do
2734 the interrupt acknowledge cycle through it to fetch the next to be
2735 delivered interrupt vector using this exit.
2737 It gets triggered whenever both KVM_CAP_PPC_EPR are enabled and an
2738 external interrupt has just been delivered into the guest. User space
2739 should put the acknowledged interrupt vector into the 'epr' field.
2741 /* Fix the size of the union. */
2746 * shared registers between kvm and userspace.
2747 * kvm_valid_regs specifies the register classes set by the host
2748 * kvm_dirty_regs specified the register classes dirtied by userspace
2749 * struct kvm_sync_regs is architecture specific, as well as the
2750 * bits for kvm_valid_regs and kvm_dirty_regs
2752 __u64 kvm_valid_regs;
2753 __u64 kvm_dirty_regs;
2755 struct kvm_sync_regs regs;
2759 If KVM_CAP_SYNC_REGS is defined, these fields allow userspace to access
2760 certain guest registers without having to call SET/GET_*REGS. Thus we can
2761 avoid some system call overhead if userspace has to handle the exit.
2762 Userspace can query the validity of the structure by checking
2763 kvm_valid_regs for specific bits. These bits are architecture specific
2764 and usually define the validity of a groups of registers. (e.g. one bit
2765 for general purpose registers)
2770 4.81 KVM_GET_EMULATED_CPUID
2772 Capability: KVM_CAP_EXT_EMUL_CPUID
2775 Parameters: struct kvm_cpuid2 (in/out)
2776 Returns: 0 on success, -1 on error
2781 struct kvm_cpuid_entry2 entries[0];
2784 The member 'flags' is used for passing flags from userspace.
2786 #define KVM_CPUID_FLAG_SIGNIFCANT_INDEX BIT(0)
2787 #define KVM_CPUID_FLAG_STATEFUL_FUNC BIT(1)
2788 #define KVM_CPUID_FLAG_STATE_READ_NEXT BIT(2)
2790 struct kvm_cpuid_entry2 {
2801 This ioctl returns x86 cpuid features which are emulated by
2802 kvm.Userspace can use the information returned by this ioctl to query
2803 which features are emulated by kvm instead of being present natively.
2805 Userspace invokes KVM_GET_EMULATED_CPUID by passing a kvm_cpuid2
2806 structure with the 'nent' field indicating the number of entries in
2807 the variable-size array 'entries'. If the number of entries is too low
2808 to describe the cpu capabilities, an error (E2BIG) is returned. If the
2809 number is too high, the 'nent' field is adjusted and an error (ENOMEM)
2810 is returned. If the number is just right, the 'nent' field is adjusted
2811 to the number of valid entries in the 'entries' array, which is then
2814 The entries returned are the set CPUID bits of the respective features
2815 which kvm emulates, as returned by the CPUID instruction, with unknown
2816 or unsupported feature bits cleared.
2818 Features like x2apic, for example, may not be present in the host cpu
2819 but are exposed by kvm in KVM_GET_SUPPORTED_CPUID because they can be
2820 emulated efficiently and thus not included here.
2822 The fields in each entry are defined as follows:
2824 function: the eax value used to obtain the entry
2825 index: the ecx value used to obtain the entry (for entries that are
2827 flags: an OR of zero or more of the following:
2828 KVM_CPUID_FLAG_SIGNIFCANT_INDEX:
2829 if the index field is valid
2830 KVM_CPUID_FLAG_STATEFUL_FUNC:
2831 if cpuid for this function returns different values for successive
2832 invocations; there will be several entries with the same function,
2833 all with this flag set
2834 KVM_CPUID_FLAG_STATE_READ_NEXT:
2835 for KVM_CPUID_FLAG_STATEFUL_FUNC entries, set if this entry is
2836 the first entry to be read by a cpu
2837 eax, ebx, ecx, edx: the values returned by the cpuid instruction for
2838 this function/index combination
2841 6. Capabilities that can be enabled
2842 -----------------------------------
2844 There are certain capabilities that change the behavior of the virtual CPU when
2845 enabled. To enable them, please see section 4.37. Below you can find a list of
2846 capabilities and what their effect on the vCPU is when enabling them.
2848 The following information is provided along with the description:
2850 Architectures: which instruction set architectures provide this ioctl.
2851 x86 includes both i386 and x86_64.
2853 Parameters: what parameters are accepted by the capability.
2855 Returns: the return value. General error numbers (EBADF, ENOMEM, EINVAL)
2856 are not detailed, but errors with specific meanings are.
2863 Returns: 0 on success; -1 on error
2865 This capability enables interception of OSI hypercalls that otherwise would
2866 be treated as normal system calls to be injected into the guest. OSI hypercalls
2867 were invented by Mac-on-Linux to have a standardized communication mechanism
2868 between the guest and the host.
2870 When this capability is enabled, KVM_EXIT_OSI can occur.
2873 6.2 KVM_CAP_PPC_PAPR
2877 Returns: 0 on success; -1 on error
2879 This capability enables interception of PAPR hypercalls. PAPR hypercalls are
2880 done using the hypercall instruction "sc 1".
2882 It also sets the guest privilege level to "supervisor" mode. Usually the guest
2883 runs in "hypervisor" privilege mode with a few missing features.
2885 In addition to the above, it changes the semantics of SDR1. In this mode, the
2886 HTAB address part of SDR1 contains an HVA instead of a GPA, as PAPR keeps the
2887 HTAB invisible to the guest.
2889 When this capability is enabled, KVM_EXIT_PAPR_HCALL can occur.
2895 Parameters: args[0] is the address of a struct kvm_config_tlb
2896 Returns: 0 on success; -1 on error
2898 struct kvm_config_tlb {
2905 Configures the virtual CPU's TLB array, establishing a shared memory area
2906 between userspace and KVM. The "params" and "array" fields are userspace
2907 addresses of mmu-type-specific data structures. The "array_len" field is an
2908 safety mechanism, and should be set to the size in bytes of the memory that
2909 userspace has reserved for the array. It must be at least the size dictated
2910 by "mmu_type" and "params".
2912 While KVM_RUN is active, the shared region is under control of KVM. Its
2913 contents are undefined, and any modification by userspace results in
2914 boundedly undefined behavior.
2916 On return from KVM_RUN, the shared region will reflect the current state of
2917 the guest's TLB. If userspace makes any changes, it must call KVM_DIRTY_TLB
2918 to tell KVM which entries have been changed, prior to calling KVM_RUN again
2921 For mmu types KVM_MMU_FSL_BOOKE_NOHV and KVM_MMU_FSL_BOOKE_HV:
2922 - The "params" field is of type "struct kvm_book3e_206_tlb_params".
2923 - The "array" field points to an array of type "struct
2924 kvm_book3e_206_tlb_entry".
2925 - The array consists of all entries in the first TLB, followed by all
2926 entries in the second TLB.
2927 - Within a TLB, entries are ordered first by increasing set number. Within a
2928 set, entries are ordered by way (increasing ESEL).
2929 - The hash for determining set number in TLB0 is: (MAS2 >> 12) & (num_sets - 1)
2930 where "num_sets" is the tlb_sizes[] value divided by the tlb_ways[] value.
2931 - The tsize field of mas1 shall be set to 4K on TLB0, even though the
2932 hardware ignores this value for TLB0.
2934 6.4 KVM_CAP_S390_CSS_SUPPORT
2938 Returns: 0 on success; -1 on error
2940 This capability enables support for handling of channel I/O instructions.
2942 TEST PENDING INTERRUPTION and the interrupt portion of TEST SUBCHANNEL are
2943 handled in-kernel, while the other I/O instructions are passed to userspace.
2945 When this capability is enabled, KVM_EXIT_S390_TSCH will occur on TEST
2946 SUBCHANNEL intercepts.
2951 Parameters: args[0] defines whether the proxy facility is active
2952 Returns: 0 on success; -1 on error
2954 This capability enables or disables the delivery of interrupts through the
2955 external proxy facility.
2957 When enabled (args[0] != 0), every time the guest gets an external interrupt
2958 delivered, it automatically exits into user space with a KVM_EXIT_EPR exit
2959 to receive the topmost interrupt vector.
2961 When disabled (args[0] == 0), behavior is as if this facility is unsupported.
2963 When this capability is enabled, KVM_EXIT_EPR can occur.
2965 6.6 KVM_CAP_IRQ_MPIC
2968 Parameters: args[0] is the MPIC device fd
2969 args[1] is the MPIC CPU number for this vcpu
2971 This capability connects the vcpu to an in-kernel MPIC device.
2973 6.7 KVM_CAP_IRQ_XICS
2976 Parameters: args[0] is the XICS device fd
2977 args[1] is the XICS CPU number (server ID) for this vcpu
2979 This capability connects the vcpu to an in-kernel XICS device.